CINXE.COM

<!doctype html><html data-n-head-ssr lang="en" data-n-head="%7B%22lang%22:%7B%22ssr%22:%22en%22%7D%7D"><head ><meta data-n-head="ssr" charset="utf-8"><meta data-n-head="ssr" name="viewport" content="width=device-width, initial-scale=1"><meta data-n-head="ssr" name="msapplication-TileColor" content="#247CB3"><meta data-n-head="ssr" name="msapplication-TileImage" content="https://asset.jmir.pub/assets/static/images/mstile-144x144.png"><meta data-n-head="ssr" name="description" content="Background: Natural language processing (NLP), a branch of artificial intelligence that analyzes unstructured language, is being increasingly used in health care. However, the extent to which NLP has been formally studied in neurological disorders remains unclear. Objective: We sought to characterize studies that applied NLP to the diagnosis, prediction, or treatment of common neurological disorders. Methods: This review followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) standards. The search was conducted using MEDLINE and Embase on May 11, 2022. Studies of NLP use in migraine, Parkinson disease, Alzheimer disease, stroke and transient ischemic attack, epilepsy, or multiple sclerosis were included. We excluded conference abstracts, review papers, as well as studies involving heterogeneous clinical populations or indirect clinical uses of NLP. Study characteristics were extracted and analyzed using descriptive statistics. We did not aggregate measurements of performance in our review due to the high variability in study outcomes, which is the main limitation of the study. Results: In total, 916 studies were identified, of which 41 (4.5%) met all eligibility criteria and were included in the final review. Of the 41 included studies, the most frequently represented disorders were stroke and transient ischemic attack (n=20, 49%), followed by epilepsy (n=10, 24%), Alzheimer disease (n=6, 15%), and multiple sclerosis (n=5, 12%). We found no studies of NLP use in migraine or Parkinson disease that met our eligibility criteria. The main objective of NLP was diagnosis (n=20, 49%), followed by disease phenotyping (n=17, 41%), prognostication (n=9, 22%), and treatment (n=4, 10%). In total, 18 (44%) studies used only machine learning approaches, 6 (15%) used only rule-based methods, and 17 (41%) used both. Conclusions: We found that NLP was most commonly applied for diagnosis, implying a potential role for NLP in augmenting diagnostic accuracy in settings with limited access to neurological expertise. We also found several gaps in neurological NLP research, with few to no studies addressing certain disorders, which may suggest additional areas of inquiry. Trial Registration: Prospective Register of Systematic Reviews (PROSPERO) CRD42021228703; https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=228703 "><meta data-n-head="ssr" name="keywords" content="artificial intelligence; machine learning; neurology; stroke; parkinson; deep learning; cardiovascular; multiple sclerosis; natural language processing; treatment; epilepsy; scoping review; parkinson disease; prediction; diagnosis; migraine; cerebrovascular disease; neurological; neurological disorder; unstructured; transient ischemic attack; nlp; text; headache disorders"><meta data-n-head="ssr" name="DC.Title" content="Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review"><meta data-n-head="ssr" name="DC.Subject" content="artificial intelligence; machine learning; neurology; stroke; parkinson; deep learning; cardiovascular; multiple sclerosis; natural language processing; treatment; epilepsy; scoping review; parkinson disease; prediction; diagnosis; migraine; cerebrovascular disease; neurological; neurological disorder; unstructured; transient ischemic attack; nlp; text; headache disorders"><meta data-n-head="ssr" name="DC.Description" content="Background: Natural language processing (NLP), a branch of artificial intelligence that analyzes unstructured language, is being increasingly used in health care. However, the extent to which NLP has been formally studied in neurological disorders remains unclear. Objective: We sought to characterize studies that applied NLP to the diagnosis, prediction, or treatment of common neurological disorders. Methods: This review followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) standards. The search was conducted using MEDLINE and Embase on May 11, 2022. Studies of NLP use in migraine, Parkinson disease, Alzheimer disease, stroke and transient ischemic attack, epilepsy, or multiple sclerosis were included. We excluded conference abstracts, review papers, as well as studies involving heterogeneous clinical populations or indirect clinical uses of NLP. Study characteristics were extracted and analyzed using descriptive statistics. We did not aggregate measurements of performance in our review due to the high variability in study outcomes, which is the main limitation of the study. Results: In total, 916 studies were identified, of which 41 (4.5%) met all eligibility criteria and were included in the final review. Of the 41 included studies, the most frequently represented disorders were stroke and transient ischemic attack (n=20, 49%), followed by epilepsy (n=10, 24%), Alzheimer disease (n=6, 15%), and multiple sclerosis (n=5, 12%). We found no studies of NLP use in migraine or Parkinson disease that met our eligibility criteria. The main objective of NLP was diagnosis (n=20, 49%), followed by disease phenotyping (n=17, 41%), prognostication (n=9, 22%), and treatment (n=4, 10%). In total, 18 (44%) studies used only machine learning approaches, 6 (15%) used only rule-based methods, and 17 (41%) used both. Conclusions: We found that NLP was most commonly applied for diagnosis, implying a potential role for NLP in augmenting diagnostic accuracy in settings with limited access to neurological expertise. We also found several gaps in neurological NLP research, with few to no studies addressing certain disorders, which may suggest additional areas of inquiry. Trial Registration: Prospective Register of Systematic Reviews (PROSPERO) CRD42021228703; https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=228703 "><meta data-n-head="ssr" name="DC.Publisher" content="JMIR Neurotechnology"><meta data-n-head="ssr" name="DC.Publisher.Address" content="JMIR Publications // 130 Queens Quay East, Unit 1100 // Toronto, ON, M5A 0P6"><meta data-n-head="ssr" name="DC.Date" scheme="ISO8601" content="2024-05-22"><meta data-n-head="ssr" name="DC.Type" content="Text.Serial.Journal"><meta data-n-head="ssr" name="DC.Format" scheme="IMT" content="text/xml"><meta data-n-head="ssr" name="DC.Identifier" content="doi:10.2196/51822"><meta data-n-head="ssr" name="DC.Language" scheme="ISO639-1" content="EN"><meta data-n-head="ssr" name="DC.Relation" content="World"><meta data-n-head="ssr" name="DC.Source" content="JMIR Neurotech 2024;3:e51822 https://neuro.jmir.org/2024/1/e51822"><meta data-n-head="ssr" name="DC.Rights" content=""><meta data-n-head="ssr" property="og:title" content="Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review"><meta data-n-head="ssr" property="og:type" content="article"><meta data-n-head="ssr" property="og:url" content="https://neuro.jmir.org/2024/1/e51822"><meta data-n-head="ssr" property="og:image" content="https://asset.jmir.pub/assets/92d91f88ddb727d190767f7247b60cd4.png"><meta data-n-head="ssr" property="og:site_name" content="JMIR Neurotechnology"><meta data-n-head="ssr" name="twitter:card" content="summary_large_image"><meta data-n-head="ssr" name="twitter:site" content="@jmirpub"><meta data-n-head="ssr" name="twitter:title" content="Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review"><meta data-n-head="ssr" name="twitter:description" content="Background: Natural language processing (NLP), a branch of artificial intelligence that analyzes unstructured language, is being increasingly used in health care. However, the extent to which NLP has been formally studied in neurological disorders remains unclear. Objective: We sought to characterize studies that applied NLP to the diagnosis, prediction, or treatment of common neurological disorders. Methods: This review followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) standards. The search was conducted using MEDLINE and Embase on May 11, 2022. Studies of NLP use in migraine, Parkinson disease, Alzheimer disease, stroke and transient ischemic attack, epilepsy, or multiple sclerosis were included. We excluded conference abstracts, review papers, as well as studies involving heterogeneous clinical populations or indirect clinical uses of NLP. Study characteristics were extracted and analyzed using descriptive statistics. We did not aggregate measurements of performance in our review due to the high variability in study outcomes, which is the main limitation of the study. Results: In total, 916 studies were identified, of which 41 (4.5%) met all eligibility criteria and were included in the final review. Of the 41 included studies, the most frequently represented disorders were stroke and transient ischemic attack (n=20, 49%), followed by epilepsy (n=10, 24%), Alzheimer disease (n=6, 15%), and multiple sclerosis (n=5, 12%). We found no studies of NLP use in migraine or Parkinson disease that met our eligibility criteria. The main objective of NLP was diagnosis (n=20, 49%), followed by disease phenotyping (n=17, 41%), prognostication (n=9, 22%), and treatment (n=4, 10%). In total, 18 (44%) studies used only machine learning approaches, 6 (15%) used only rule-based methods, and 17 (41%) used both. Conclusions: We found that NLP was most commonly applied for diagnosis, implying a potential role for NLP in augmenting diagnostic accuracy in settings with limited access to neurological expertise. We also found several gaps in neurological NLP research, with few to no studies addressing certain disorders, which may suggest additional areas of inquiry. Trial Registration: Prospective Register of Systematic Reviews (PROSPERO) CRD42021228703; https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=228703 "><meta data-n-head="ssr" name="twitter:image" content="https://asset.jmir.pub/assets/92d91f88ddb727d190767f7247b60cd4.png"><meta data-n-head="ssr" name="citation_title" content="Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review"><meta data-n-head="ssr" name="citation_journal_title" content="JMIR Neurotechnology"><meta data-n-head="ssr" name="citation_publisher" content="JMIR Publications Inc., Toronto, Canada"><meta data-n-head="ssr" name="citation_doi" content="10.2196/51822"><meta data-n-head="ssr" name="citation_issue" content="1"><meta data-n-head="ssr" name="citation_volume" content="3"><meta data-n-head="ssr" name="citation_firstpage" content="e51822"><meta data-n-head="ssr" name="citation_date" content="2024-05-22"><meta data-n-head="ssr" name="citation_abstract_html_url" content="https://neuro.jmir.org/2024/1/e51822"><meta data-n-head="ssr" name="citation_abstract_pdf_url" content="https://neuro.jmir.org/2024/1/e51822/PDF"><meta data-n-head="ssr" name="DC.Creator" content="Ilana"><meta data-n-head="ssr" name="DC.Contributor" content="Ilana Lefkovitz"><meta data-n-head="ssr" name="DC.Contributor" content="Samantha Walsh"><meta data-n-head="ssr" name="DC.Contributor" content="Leah J Blank"><meta data-n-head="ssr" name="DC.Contributor" content="Nathalie Jetté"><meta data-n-head="ssr" name="DC.Contributor" content="Benjamin R Kummer"><meta data-n-head="ssr" name="citation_authors" content="Ilana Lefkovitz"><meta data-n-head="ssr" name="citation_authors" content="Samantha Walsh"><meta data-n-head="ssr" name="citation_authors" content="Leah J Blank"><meta data-n-head="ssr" name="citation_authors" content="Nathalie Jetté"><meta data-n-head="ssr" name="citation_authors" content="Benjamin R Kummer"><title>JMIR Neurotechnology - Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review</title><link data-n-head="ssr" rel="apple-touch-icon" sizes="57x57" href="https://asset.jmir.pub/assets/static/images/apple-touch-icon-57x57.png"><link data-n-head="ssr" rel="apple-touch-icon" sizes="114x114" href="https://asset.jmir.pub/assets/static/images/apple-touch-icon-114x114.png"><link data-n-head="ssr" rel="apple-touch-icon" sizes="72x72" href="https://asset.jmir.pub/assets/static/images/apple-touch-icon-72x72.png"><link data-n-head="ssr" rel="apple-touch-icon" sizes="144x144" href="https://asset.jmir.pub/assets/static/images/apple-touch-icon-144x144.png"><link data-n-head="ssr" rel="apple-touch-icon" sizes="60x60" href="https://asset.jmir.pub/assets/static/images/apple-touch-icon-60x60.png"><link data-n-head="ssr" rel="apple-touch-icon" sizes="120x120" href="https://asset.jmir.pub/assets/static/images/apple-touch-icon-120x120.png"><link data-n-head="ssr" rel="apple-touch-icon" sizes="76x76" href="https://asset.jmir.pub/assets/static/images/apple-touch-icon-76x76.png"><link data-n-head="ssr" rel="apple-touch-icon" sizes="152x152" href="https://asset.jmir.pub/assets/static/images/apple-touch-icon-152x152.png"><link data-n-head="ssr" rel="icon" type="image/png" href="https://asset.jmir.pub/assets/static/images/favicon-196x196.png" sizes="196x196"><link data-n-head="ssr" rel="icon" type="image/png" href="https://asset.jmir.pub/assets/static/images/favicon-160x160.png" sizes="160x160"><link data-n-head="ssr" rel="icon" type="image/png" href="https://asset.jmir.pub/assets/static/images/favicon-96x96.png" sizes="96x96"><link data-n-head="ssr" rel="icon" type="image/png" href="https://asset.jmir.pub/assets/static/images/favicon-16x16.png" sizes="16x16"><link data-n-head="ssr" rel="icon" type="image/png" href="https://asset.jmir.pub/assets/static/images/favicon-32x32.png" sizes="32x32"><link data-n-head="ssr" rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/twitter-bootstrap/4.3.1/css/bootstrap-grid.min.css" defer><link data-n-head="ssr" rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.5.2/css/all.min.css" defer><link data-n-head="ssr" rel="stylesheet" href="https://fonts.googleapis.com/css?family=Roboto:100,100i,300,300i,400,400i,500,500i,700,700i,900,900i&display=swap" defer><link data-n-head="ssr" rel="schema.DC" href="http://purl.org/dc/elements/1.1/"><script data-n-head="ssr" type="text/javascript" id="hs-script-loader" src="//js.hs-scripts.com/19668141.js"></script><script data-n-head="ssr" data-hid="twitter-script" type="text/javascript" charset="utf-8"> !function(e,t,n,s,u,a){e.twq||(s=e.twq=function(){s.exe?s.exe.apply(s,arguments):s.queue.push(arguments);},s.version='1.1',s.queue=[],u=t.createElement(n),u.async=!0,u.src='https://static.ads-twitter.com/uwt.js',a=t.getElementsByTagName(n)[0],a.parentNode.insertBefore(u,a))}(window,document,'script');twq('config','o7i83'); </script><script data-n-head="ssr" src="https://js.trendmd.com/trendmd-ns.min.js" defer data-trendmdconfig="{"element": "#trendmd-suggestions"}"></script><link rel="preload" href="/_nuxt/d762be0.js" as="script"><link rel="preload" href="/_nuxt/114bf28.js" as="script"><link rel="preload" href="/_nuxt/557a776.js" as="script"><link rel="preload" href="/_nuxt/f65c9d0.js" as="script"><link rel="preload" href="/_nuxt/b7cb769.js" as="script"><link rel="preload" href="/_nuxt/65075b6.js" as="script"><link rel="preload" href="/_nuxt/8499a81.js" as="script"><link rel="preload" href="/_nuxt/60620ca.js" as="script"><link rel="preload" href="/_nuxt/7bab8a9.js" as="script"><link rel="preload" href="/_nuxt/6c336c8.js" as="script"><link rel="preload" href="/_nuxt/e63f738.js" as="script"><link rel="preload" href="/_nuxt/2f3783d.js" as="script"><link rel="preload" href="/_nuxt/8679c54.js" as="script"><link rel="preload" href="/_nuxt/84ba7a4.js" as="script"><link rel="preload" href="/_nuxt/2042386.js" as="script"><link rel="preload" href="/_nuxt/7aa873e.js" as="script"><link rel="preload" href="/_nuxt/4f5e72f.js" as="script"><link rel="preload" href="/_nuxt/512387a.js" as="script"><link rel="preload" href="/_nuxt/69df2f4.js" as="script"><link rel="preload" href="/_nuxt/c8abcfd.js" as="script"><link rel="preload" href="/_nuxt/ef88041.js" as="script"><link rel="preload" href="/_nuxt/4e67ec6.js" as="script"><link rel="preload" href="/_nuxt/edec9c1.js" as="script"><link rel="preload" href="/_nuxt/a0160a8.js" as="script"><link rel="preload" href="/_nuxt/88d69a9.js" as="script"><link rel="preload" href="/_nuxt/456ce06.js" as="script"><link rel="preload" href="/_nuxt/3049f10.js" as="script"><link rel="preload" href="/_nuxt/fec3470.js" as="script"><link rel="preload" href="/_nuxt/48052c9.js" as="script"><link rel="preload" href="/_nuxt/f76b897.js" as="script"><link rel="preload" href="/_nuxt/f2fb177.js" as="script"><link rel="preload" href="/_nuxt/fdd6c0c.js" as="script"><link rel="preload" href="/_nuxt/51c3064.js" as="script"><link rel="preload" href="/_nuxt/67892e3.js" as="script"><link rel="preload" href="/_nuxt/6c2c095.js" as="script"><link rel="preload" href="/_nuxt/9a11cd2.js" as="script"><link rel="preload" href="/_nuxt/e3a1965.js" as="script"><link rel="preload" href="/_nuxt/c14e3a6.js" as="script"><link rel="preload" href="/_nuxt/ceb5b90.js" as="script"><link rel="preload" href="/_nuxt/2715511.js" as="script"><link rel="preload" href="/_nuxt/7819e44.js" as="script"><link rel="preload" href="/_nuxt/8d00881.js" as="script"><link rel="preload" href="/_nuxt/6171ac4.js" as="script"><link rel="preload" href="/_nuxt/9a8e2c4.js" as="script"><link rel="preload" href="/_nuxt/0f313a0.js" as="script"><link rel="preload" href="/_nuxt/90db445.js" as="script"><link rel="preload" href="/_nuxt/ee7c71d.js" as="script"><link rel="preload" href="/_nuxt/0c77ea6.js" as="script"><link rel="preload" href="/_nuxt/c4453ca.js" as="script"><link rel="preload" href="/_nuxt/f3339c8.js" as="script"><link rel="preload" href="/_nuxt/76715db.js" as="script"><style data-vue-ssr-id="7975b328:0 c46114bc:0 b8d2539e:0 f8852dd4:0 f2a96380:0 19d22970:0 e58746f2:0 6f03c485:0 aca923aa:0 79d6150e:0 cd6e1cc0:0 31a45c1a:0 d8137ec4:0">.mt-0{margin-top:0!important}.mt-5{margin-top:5px!important}.mt-10{margin-top:10px!important}.mt-15{margin-top:15px!important}.mt-20{margin-top:20px!important}.mt-25{margin-top:25px!important}.mt-30{margin-top:30px!important}.mt-35{margin-top:35px!important}.mt-40{margin-top:40px!important}.mt-45{margin-top:45px!important}.mt-50{margin-top:50px!important}.mt-55{margin-top:55px!important}.mt-60{margin-top:60px!important}.mt-65{margin-top:65px!important}.mt-70{margin-top:70px!important}.mt-75{margin-top:75px!important}.mt-80{margin-top:80px!important}.mt-85{margin-top:85px!important}.mt-90{margin-top:90px!important}.mt-95{margin-top:95px!important}.mt-100{margin-top:100px!important}.mb-0{margin-bottom:0!important}.mb-5{margin-bottom:5px!important}.mb-10{margin-bottom:10px!important}.mb-15{margin-bottom:15px!important}.mb-20{margin-bottom:20px!important}.mb-25{margin-bottom:25px!important}.mb-30{margin-bottom:30px!important}.mb-35{margin-bottom:35px!important}.mb-40{margin-bottom:40px!important}.mb-45{margin-bottom:45px!important}.mb-50{margin-bottom:50px!important}.mb-55{margin-bottom:55px!important}.mb-60{margin-bottom:60px!important}.mb-65{margin-bottom:65px!important}.mb-70{margin-bottom:70px!important}.mb-75{margin-bottom:75px!important}.mb-80{margin-bottom:80px!important}.mb-85{margin-bottom:85px!important}.mb-90{margin-bottom:90px!important}.mb-95{margin-bottom:95px!important}.mb-100{margin-bottom:100px!important}.ml-0{margin-left:0!important}.ml-5{margin-left:5px!important}.ml-10{margin-left:10px!important}.ml-15{margin-left:15px!important}.ml-20{margin-left:20px!important}.ml-25{margin-left:25px!important}.ml-30{margin-left:30px!important}.ml-35{margin-left:35px!important}.ml-40{margin-left:40px!important}.ml-45{margin-left:45px!important}.ml-50{margin-left:50px!important}.ml-55{margin-left:55px!important}.ml-60{margin-left:60px!important}.ml-65{margin-left:65px!important}.ml-70{margin-left:70px!important}.ml-75{margin-left:75px!important}.ml-80{margin-left:80px!important}.ml-85{margin-left:85px!important}.ml-90{margin-left:90px!important}.ml-95{margin-left:95px!important}.ml-100{margin-left:100px!important}.mr-0{margin-right:0!important}.mr-5{margin-right:5px!important}.mr-10{margin-right:10px!important}.mr-15{margin-right:15px!important}.mr-20{margin-right:20px!important}.mr-25{margin-right:25px!important}.mr-30{margin-right:30px!important}.mr-35{margin-right:35px!important}.mr-40{margin-right:40px!important}.mr-45{margin-right:45px!important}.mr-50{margin-right:50px!important}.mr-55{margin-right:55px!important}.mr-60{margin-right:60px!important}.mr-65{margin-right:65px!important}.mr-70{margin-right:70px!important}.mr-75{margin-right:75px!important}.mr-80{margin-right:80px!important}.mr-85{margin-right:85px!important}.mr-90{margin-right:90px!important}.mr-95{margin-right:95px!important}.mr-100{margin-right:100px!important}.pt-0{padding-top:0!important}.pt-5{padding-top:5px!important}.pt-10{padding-top:10px!important}.pt-15{padding-top:15px!important}.pt-20{padding-top:20px!important}.pt-25{padding-top:25px!important}.pt-30{padding-top:30px!important}.pt-35{padding-top:35px!important}.pt-40{padding-top:40px!important}.pt-45{padding-top:45px!important}.pt-50{padding-top:50px!important}.pt-55{padding-top:55px!important}.pt-60{padding-top:60px!important}.pt-65{padding-top:65px!important}.pt-70{padding-top:70px!important}.pt-75{padding-top:75px!important}.pt-80{padding-top:80px!important}.pt-85{padding-top:85px!important}.pt-90{padding-top:90px!important}.pt-95{padding-top:95px!important}.pt-100{padding-top:100px!important}.pb-0{padding-bottom:0!important}.pb-5{padding-bottom:5px!important}.pb-10{padding-bottom:10px!important}.pb-15{padding-bottom:15px!important}.pb-20{padding-bottom:20px!important}.pb-25{padding-bottom:25px!important}.pb-30{padding-bottom:30px!important}.pb-35{padding-bottom:35px!important}.pb-40{padding-bottom:40px!important}.pb-45{padding-bottom:45px!important}.pb-50{padding-bottom:50px!important}.pb-55{padding-bottom:55px!important}.pb-60{padding-bottom:60px!important}.pb-65{padding-bottom:65px!important}.pb-70{padding-bottom:70px!important}.pb-75{padding-bottom:75px!important}.pb-80{padding-bottom:80px!important}.pb-85{padding-bottom:85px!important}.pb-90{padding-bottom:90px!important}.pb-95{padding-bottom:95px!important}.pb-100{padding-bottom:100px!important}.pl-0{padding-left:0!important}.pl-5{padding-left:5px!important}.pl-10{padding-left:10px!important}.pl-15{padding-left:15px!important}.pl-20{padding-left:20px!important}.pl-25{padding-left:25px!important}.pl-30{padding-left:30px!important}.pl-35{padding-left:35px!important}.pl-40{padding-left:40px!important}.pl-45{padding-left:45px!important}.pl-50{padding-left:50px!important}.pl-55{padding-left:55px!important}.pl-60{padding-left:60px!important}.pl-65{padding-left:65px!important}.pl-70{padding-left:70px!important}.pl-75{padding-left:75px!important}.pl-80{padding-left:80px!important}.pl-85{padding-left:85px!important}.pl-90{padding-left:90px!important}.pl-95{padding-left:95px!important}.pl-100{padding-left:100px!important}.pr-0{padding-right:0!important}.pr-5{padding-right:5px!important}.pr-10{padding-right:10px!important}.pr-15{padding-right:15px!important}.pr-20{padding-right:20px!important}.pr-25{padding-right:25px!important}.pr-30{padding-right:30px!important}.pr-35{padding-right:35px!important}.pr-40{padding-right:40px!important}.pr-45{padding-right:45px!important}.pr-50{padding-right:50px!important}.pr-55{padding-right:55px!important}.pr-60{padding-right:60px!important}.pr-65{padding-right:65px!important}.pr-70{padding-right:70px!important}.pr-75{padding-right:75px!important}.pr-80{padding-right:80px!important}.pr-85{padding-right:85px!important}.pr-90{padding-right:90px!important}.pr-95{padding-right:95px!important}.pr-100{padding-right:100px!important}input,select,textarea{border-radius:3px!important;font-size:1.6rem!important}textarea{font-family:Roboto-Regular,Roboto}input[type=date],input[type=email],input[type=hidden],input[type=number],input[type=search],input[type=tel],input[type=text],input[type=url]{border:1px solid rgba(26,37,76,.396);height:3.2rem;padding:0 10px}input[type=date]:focus,input[type=email]:focus,input[type=hidden]:focus,input[type=number]:focus,input[type=search]:focus,input[type=tel]:focus,input[type=text]:focus,input[type=url]:focus{background-color:rgba(48,136,223,.071);border:1px solid #1e70c2;outline:none}input::-moz-placeholder{opacity:.4}input::placeholder{opacity:.4}.v-select .vs__selected{margin:2px}.v-select .vs__dropdown-toggle{border:1px solid rgba(26,37,76,.396)!important;padding:0!important}.v-select input{border:1px solid transparent!important;margin:0}.v-select ul.vs__dropdown-menu{paddinng-top:0!important}.v-select.vs--single .vs__selected{margin:0 3px 0 5px}textarea{border:1px solid rgba(26,37,76,.396);padding:5px 10px;width:100%}textarea:focus{background-color:rgba(48,136,223,.071);border:1px solid #1e70c2;outline:none}select{border:1px solid rgba(26,37,76,.396);cursor:pointer;height:3.2rem}select:focus{background-color:rgba(48,136,223,.071);border:1px solid #1e70c2;outline:none}input::-webkit-input-placeholder,textarea::-webkit-input-placeholder{font-size:1.2rem}select::-webkit-input-placeholder{font-size:1.2rem}::-moz-selection{background-color:#1e70c2;color:#fff}::selection{background-color:#1e70c2;color:#fff}a{color:#1e70c2;outline:none;-webkit-text-decoration:none;text-decoration:none}a:focus,a:hover{-webkit-text-decoration:underline;text-decoration:underline}a:focus{font-weight:700}button:focus{outline:none}.deactive{cursor:not-allowed;pointer-events:none}.element-wrapper{margin-bottom:7rem}.page-heading{background-color:#f1f3f5;border-bottom:1px solid hsla(0,0%,44%,.161);margin-bottom:60px;padding:30px 0}.page-heading h1{margin:0}.link{color:#2078cf;outline:none;-webkit-text-decoration:none;text-decoration:none}.link:focus,.link:hover{color:#2078cf;-webkit-text-decoration:underline;text-decoration:underline}.title-link{color:#1a254c;outline:none;-webkit-text-decoration:none;text-decoration:none}.title-link:focus,.title-link:hover{color:#2078cf;-webkit-text-decoration:underline;text-decoration:underline}.h1,.h2,.h3,.h4,.h5,.h6,h2,h3,h4,h5,h6,pwdh1{font-weight:700}.h1,h1{font-size:4rem;line-height:5rem}.h2,h2{font-size:3.2rem;line-height:4rem}.h3,h3{font-size:2.6rem;line-height:3.4rem}.h4,h4{font-size:1.8rem;line-height:2.4rem}.h5,h5{font-size:1.6rem;line-height:2.2rem}.h6,h6{font-size:1.4rem}.h6,h6,p{line-height:2rem}small{line-height:1.8rem}.h1{font-size:4rem!important;line-height:5rem!important}.h2{font-size:3.2rem!important;line-height:4rem!important}.h3{font-size:2.6rem!important;line-height:3.4rem!important}.h4{font-size:1.8rem!important;line-height:2.4rem!important}.h5{font-size:1.6rem!important;line-height:2.2rem!important}.h6{font-size:1.4rem!important;line-height:2rem!important}input.disabled,select.disabled,textarea.disabled{background:hsla(0,0%,82%,.29);cursor:not-allowed}button.disabled{cursor:not-allowed;opacity:.5;pointer-events:none}strong{font-weight:700!important}.disabled-section{cursor:not-allowed;opacity:.5}.fa,.fas{font-weight:900}.errors{color:red;display:block}.screen-readers-only{height:1px;left:-10000px;overflow:hidden;position:absolute;top:auto;width:1px}input[type=text].input-error{border:1px solid red!important;border-radius:3px}.input-error{border:1px solid red!important;border-radius:3px}.popper{max-width:400px;padding:10px;text-align:justify}.vue-notification{margin:20px 20px 0 0}.vue-notification.toast-success{background:#4caf50;border-left:5px solid #1a254c}.vue-dropzone{border:2px dashed #e5e5e5}.vue-dropzone .icon{display:block;font-size:25px;margin-bottom:10px}.required:before{color:red;content:"*"}.grey-heading-underline{border-bottom:2px solid #c8cad4}.green-heading-underline{border-bottom:2px solid #367c3a}.green-underline{background:#367c3a;content:"";height:3px;margin:0 auto 40px;width:100px}.separator{color:#000;margin:0 10px;opacity:.4}.list-style-none{list-style-type:none}.list-style-none li{margin-bottom:10px}.width-100{width:100%!important}.width-fit-content{width:-moz-fit-content;width:fit-content}.break-word{word-break:break-word}.text-center{text-align:center}.d-inline-block{display:inline-block}.d-flex{display:flex}.d-block{display:block}.flex-direction-column{flex-direction:column}.justify-content-space-between{justify-content:space-between}.align-items-center{align-items:center}.align-items-baseline{align-items:baseline}.fs-10{font-size:1rem;line-height:1.6rem}.fs-12{font-size:1.2rem;line-height:1.8rem}.fs-14{font-size:1.4rem;line-height:2rem}.fs-16{font-size:1.6rem;line-height:2.4rem}.fs-18{font-size:1.8rem;line-height:2.6rem}.fs-20{font-size:2rem;line-height:3rem}.fs-italic{font-style:italic}.fw-bold{font-weight:700}.color-blue{color:#1e70c2}.color-green{color:#367c3a}.color-red{color:#b30000}.ql-toolbar{background-color:#f8f9fa}.btn{cursor:pointer;opacity:1;text-align:center;transition:.3s}.btn:focus,.btn:hover{font-weight:400;opacity:.9;-webkit-text-decoration:none;text-decoration:none}.btn:focus{outline:2px solid #f69038!important;outline-offset:6px!important}.btn-disabled{cursor:not-allowed;opacity:.6;pointer-events:none}.btn-small{font-size:1.2rem;padding:5px 10px}.btn-medium{font-size:1.4rem;padding:10px 20px}.btn-large{font-size:1.8rem;padding:20px 40px}.btn-blue{background-color:#1e70c2;border:1px solid #1e70c2;color:#fff}.btn-blue:active{background-color:#2b7bca;border:1px solid #2b7bca}.btn-green{background-color:#367c3a;border:1px solid #367c3a;color:#fff}.btn-green:active{background-color:#3b9d3f;border:1px solid #3b9d3f}.btn-grey{background-color:#f1f3f5;border:1px solid #dcdee0;color:#1a254c}.btn-grey:active{background-color:#dcdee0}.btn-red{background-color:#b30000;border:1px solid #b30000;color:#fff}.btn-red:active{background-color:#ba302d;border:1px solid #ba302d}.btn-blue-pill{background-color:inherit;border:none;border-radius:20px;color:#1e70c2;cursor:pointer;font-size:1.4rem;padding:5px 10px;transition:.3s}.btn-blue-pill:focus,.btn-blue-pill:hover{background-color:rgba(48,136,223,.161);-webkit-text-decoration:none;text-decoration:none}.btn-blue-pill:active{background-color:rgba(48,136,223,.29)}.btn-green-pill{background-color:inherit;border:none;border-radius:20px;color:#367c3a;cursor:pointer;font-size:1.4rem;padding:5px 10px;transition:.3s}.btn-green-pill:focus,.btn-green-pill:hover{background-color:rgba(76,175,80,.188);-webkit-text-decoration:none;text-decoration:none}.btn-green-pill:active{background-color:rgba(76,175,80,.29)}.btn-grey-pill{background-color:inherit;border:none;border-radius:20px;color:#1a254c;cursor:pointer;font-size:1.4rem;padding:5px 10px;transition:.3s}.btn-grey-pill:focus,.btn-grey-pill:hover{background-color:#dcdee0;-webkit-text-decoration:none;text-decoration:none}.btn-grey-pill:active{background-color:#cccdce}.btn-red-pill{background-color:inherit;border:none;border-radius:20px;color:#fa2a24;cursor:pointer;font-size:1.4rem;padding:5px 10px;transition:.3s}.btn-red-pill:focus,.btn-red-pill:hover{background-color:rgba(255,0,0,.122);-webkit-text-decoration:none;text-decoration:none}.btn-red-pill:active{background-color:rgba(255,0,0,.22)}.sm-icons a{border-radius:50%;color:#fff;cursor:pointer;display:inline-block;font-size:14px;height:25px;line-height:25px;padding:0;text-align:center;-webkit-text-decoration:none;text-decoration:none;transition:.3s;width:25px}.sm-icons a:hover{opacity:.9}.sm-icons a:focus{opacity:.9;outline:none;transform:scale(1.2)}.sm-icons .bluesky{background-color:#0085ff}.sm-icons .bluesky:before{content:"";font-family:"Font Awesome 6 Brands"}.sm-icons .twitter{background-color:#000}.sm-icons .twitter:before{content:"";font-family:"Font Awesome 5 Brands";font-weight:900}.sm-icons .facebook{background-color:#3b5a98}.sm-icons .facebook:before{content:"";font-family:"Font Awesome 5 Brands";font-weight:900}.sm-icons .linkedin{background-color:#0077b5}.sm-icons .linkedin:before{content:"";font-family:"Font Awesome 5 Brands";font-weight:900}.sm-icons .youtube{background-color:red}.sm-icons .youtube:before{content:"";font-family:"Font Awesome 5 Brands";font-weight:900}.sm-icons .instagram{background:radial-gradient(circle at 30% 107%,#fdf497 0,#fdf497 5%,#fd5949 45%,#d6249f 60%,#285aeb 90%)}.sm-icons .instagram:before{content:"";font-family:"Font Awesome 5 Brands";font-weight:900}.sm-icons .email{background-color:#cc2126}.sm-icons .email:before{content:"";font-family:"Font Awesome 5 Free";font-weight:900}.sm-icons .rss{background-color:#ee802f}.sm-icons .rss:before{content:"";font-family:"Font Awesome 5 Free";font-weight:900}.full-width-card-wrapper .full-width-card{border:1px solid #ced1dc;display:flex;margin:10px 0 20px}@media screen and (max-width:61.9375em){.full-width-card-wrapper .full-width-card{flex-wrap:wrap;justify-content:space-between}}.full-width-card-wrapper .full-width-card-img{height:auto;position:relative;width:250px}@media screen and (max-width:61.9375em){.full-width-card-wrapper .full-width-card-img{flex-basis:50%}}@media screen and (max-width:47.9375em){.full-width-card-wrapper .full-width-card-img{width:100%}}.full-width-card-wrapper .full-width-card-img img{border:1px solid #ced1dc;height:auto;width:100%}.full-width-card-wrapper .full-width-card-img-info{background-color:#1e70c2;border-radius:3px 0 0 0;bottom:4px;cursor:pointer;outline:none;padding:11px;position:absolute;right:1px}.full-width-card-wrapper .full-width-card-img-info .icon{color:#fff;font-size:1.8rem;transition:all .3s ease}.full-width-card-wrapper .full-width-card-img-info:hover{padding-bottom:10px}.full-width-card-wrapper .full-width-card-img-info:hover .icon{font-size:2rem}.full-width-card-wrapper .full-width-card-img-info:focus .icon{font-size:2.4rem}.full-width-card-wrapper .full-width-card-info{flex-grow:1;padding:15px 15px 15px 0}@media screen and (max-width:61.9375em){.full-width-card-wrapper .full-width-card-info{flex-basis:50%}}@media screen and (max-width:47.9375em){.full-width-card-wrapper .full-width-card-info{flex-basis:100%;padding:15px}}.full-width-card-wrapper .full-width-card-info-title{margin-top:0}.full-width-card-wrapper .full-width-card-highlight:after,.full-width-card-wrapper .full-width-card-highlight:before{color:#5d6581;content:"..."}.full-width-card-wrapper .full-width-card-info-download-links a{margin-right:9px}@media screen and (max-width:61.9375em){.full-width-card-wrapper .full-width-card-info-button button{margin:10px 0 0;width:100%}}.full-width-card-wrapper .full-width-card-info-group-buttons{display:flex;margin-top:14px}@media screen and (max-width:61.9375em){.full-width-card-wrapper .full-width-card-info-group-buttons a:first-child{display:block;margin-bottom:14px;margin-left:0!important;margin-right:0!important}.full-width-card-wrapper .full-width-card-info-group-buttons a:last-child{display:block;margin-left:0!important;margin-right:0!important}.full-width-card-wrapper .full-width-card-info-group-buttons{display:block}}.full-width-card-wrapper .full-width-card-info-group-buttons button{margin-right:10px}@media screen and (max-width:61.9375em){.full-width-card-wrapper .full-width-card-info-group-buttons button{margin:10px 0 0;width:100%}}.full-width-card-wrapper .full-width-card-info-date-published{margin-bottom:20px}.full-width-card-wrapper .full-width-card-altmetric{align-self:center;margin-left:auto;padding-right:15px}@media screen and (max-width:61.9375em){.full-width-card-wrapper .full-width-card-altmetric{margin:10px auto 15px;padding:0}}.full-width-card-wrapper .full-width-card-altmetric img{width:auto}.img-cont{padding:15px}@media screen and (max-width:47.9375em){.img-cont{width:100%}}.cards{justify-content:space-between;margin-bottom:14px}.cards,.cards .card{display:flex;flex-wrap:wrap}.cards .card{flex-basis:31%;flex-direction:column;flex-grow:0;flex-shrink:0;transition:.3s}@media screen and (max-width:61.9375em){.cards .card{flex-basis:48%}}@media screen and (max-width:47.9375em){.cards .card{flex-basis:100%}}.cards .card:hover .card-header .card-img img{filter:brightness(1);transform:scale(1.1);transition:all .3s ease}.cards .card-header,.cards .card-img{border-radius:10px;border-radius:3px;overflow:hidden}.cards .card-img{height:200px;position:relative}.cards .card-img img{filter:brightness(.6);height:auto;transition:all .3s ease;width:100%}@media screen and (max-width:61.9375em){.cards .card-img{height:180px}}@media screen and (max-width:47.9375em){.cards .card-img{height:200px}}.cards .card-img-info{background-color:#1e70c2;border-radius:3px 0 0 0;cursor:pointer;outline:none;padding:11px;position:absolute;right:0;top:160px}@media screen and (max-width:61.9375em){.cards .card-img-info{top:140px}}@media screen and (max-width:47.9375em){.cards .card-img-info{top:160px}}.cards .card-img-info .icon{color:#fff;font-size:1.8rem;transition:all .3s ease}.cards .card-img-info:hover .icon{font-size:2rem}.cards .card-img-info:focus .icon{font-size:2.4rem}.cards .card-body{flex-grow:1;flex-shrink:0}.cards .card-body,.cards .card-title{display:flex;flex-direction:column}.cards .card-title a{color:#1a254c;outline:none;transition:all .3s ease}.cards .card-title a:focus,.cards .card-title a:hover{color:#1e70c2;-webkit-text-decoration:underline;text-decoration:underline;transition:all .3s ease}.cards .card-info p{max-height:125px;overflow:scroll}.cards .card-info p a{color:#1e70c2;cursor:pointer}.cards .card-info p a:hover{-webkit-text-decoration:underline;text-decoration:underline}.cards .card-years{display:flex;flex-wrap:wrap}.cards .card-years a{color:#1e70c2;cursor:pointer;margin:0 10px 5px 0}.cards .card-years a:focus,.cards .card-years a:hover{outline:none;-webkit-text-decoration:underline;text-decoration:underline}.cards .card-date-social{color:#5d6581;display:flex;flex-wrap:wrap}.v--modal-overlay{background:rgba(17,26,55,.7);box-sizing:border-box;height:100vh;left:0;opacity:1;position:fixed;top:0;width:100%;z-index:99999}.v--modal-overlay .v--modal-background-click{min-height:100%;padding-bottom:10px;width:100%}.v--modal-overlay .v--modal-background-click .v--modal-top-right{display:block;position:absolute;right:0;top:0}.v--modal-overlay .v--modal-background-click .v--modal-box{box-sizing:border-box;position:relative}.v--modal-overlay .v--modal-background-click .v--modal{border-radius:3px;box-shadow:0 20px 60px -2px rgba(27,33,58,.4);text-align:left}.modal-window{background-color:#fff;height:100%;overflow:hidden;position:relative}.modal-window-header{background:#f3f3f5;border-bottom:1px solid #bab4b4;display:flex;flex-wrap:wrap;justify-content:space-between;padding:10px 20px}.modal-window-title{margin:0}.modal-window-close{background-color:#f3f3f5;border:none;color:gray;cursor:pointer;font-size:2rem;transition:.3s}.modal-window-close:hover{color:#000}.modal-window-body{height:-webkit-fill-available;max-height:400px;overflow-y:auto;padding:20px}.modal-window-footer{background-color:#f8f9fa;border-top:1px solid #bab4b4}.modal-window-footer div{float:right;padding:10px 20px} .cookieControl__Modal-enter-active,.cookieControl__Modal-leave-active{transition:opacity .25s}.cookieControl__Modal-enter,.cookieControl__Modal-leave-to{opacity:0}.cookieControl__Bar--center{left:50%;top:50%;transform:translate(-50%,-50%)}.cookieControl__Bar--bottom-full-enter-active,.cookieControl__Bar--bottom-full-leave-active,.cookieControl__Bar--bottom-left-enter-active,.cookieControl__Bar--bottom-left-leave-active,.cookieControl__Bar--bottom-right-enter-active,.cookieControl__Bar--bottom-right-leave-active,.cookieControl__Bar--center-enter-active,.cookieControl__Bar--center-leave-active,.cookieControl__Bar--top-full-enter-active,.cookieControl__Bar--top-full-leave-active,.cookieControl__Bar--top-left-enter-active,.cookieControl__Bar--top-left-leave-active,.cookieControl__Bar--top-right-enter-active,.cookieControl__Bar--top-right-leave-active{transition:transform .25s}.cookieControl__Bar--top-full-enter,.cookieControl__Bar--top-full-leave-to,.cookieControl__Bar--top-left-enter,.cookieControl__Bar--top-left-leave-to,.cookieControl__Bar--top-right-enter,.cookieControl__Bar--top-right-leave-to{transform:translateY(-100%)}.cookieControl__Bar--bottom-full-enter,.cookieControl__Bar--bottom-full-leave-to,.cookieControl__Bar--bottom-left-enter,.cookieControl__Bar--bottom-left-leave-to,.cookieControl__Bar--bottom-right-enter,.cookieControl__Bar--bottom-right-leave-to{transform:translateY(100%)}.cookieControl__Bar--center-enter,.cookieControl__Bar--center-leave-to{transform:translate(-50%,-50%) scale(.95)}.cookieControl{position:relative;z-index:100000}.cookieControl button{backface-visibility:hidden;border:0;cursor:pointer;font-size:16px;outline:0;padding:12px 20px;transition:background-color .2s,color .2s}.cookieControl__Bar{background-color:var(--cookie-control-barBackground);font-family:Arial,"Helvetica Neue",Helvetica,sans-serif;position:fixed}.cookieControl__Bar h3,.cookieControl__Bar p{color:var(--cookie-control-barTextColor);max-width:900px}.cookieControl__Bar h3{font-size:20px;margin:0}.cookieControl__Bar p{font-size:16px;margin:5px 0 0}.cookieControl__Bar button{background-color:var(--cookie-control-barButtonBackground);color:var(--cookie-control-barButtonColor)}.cookieControl__Bar button:hover{background-color:var(--cookie-control-barButtonHoverBackground);color:var(--cookie-control-barButtonHoverColor)}.cookieControl__Bar button+button{margin-left:10px}.cookieControl__BarContainer{align-items:flex-end;display:flex;justify-content:space-between;padding:20px}.cookieControl__Bar--bottom-full,.cookieControl__Bar--top-full{left:0;right:0}.cookieControl__Bar--top-full{top:0}.cookieControl__Bar--bottom-full{bottom:0}.cookieControl__Bar--bottom-left p,.cookieControl__Bar--bottom-right p,.cookieControl__Bar--center p,.cookieControl__Bar--top-left p,.cookieControl__Bar--top-right p{max-width:400px}.cookieControl__Bar--bottom-left .cookieControl__BarContainer,.cookieControl__Bar--bottom-right .cookieControl__BarContainer,.cookieControl__Bar--center .cookieControl__BarContainer,.cookieControl__Bar--top-left .cookieControl__BarContainer,.cookieControl__Bar--top-right .cookieControl__BarContainer{flex-direction:column}.cookieControl__Bar--bottom-left .cookieControl__BarButtons,.cookieControl__Bar--bottom-right .cookieControl__BarButtons,.cookieControl__Bar--center .cookieControl__BarButtons,.cookieControl__Bar--top-left .cookieControl__BarButtons,.cookieControl__Bar--top-right .cookieControl__BarButtons{margin-top:20px}.cookieControl__Bar--top-left,.cookieControl__Bar--top-right{top:20px}.cookieControl__Bar--bottom-left,.cookieControl__Bar--bottom-right{bottom:20px}.cookieControl__Bar--bottom-left,.cookieControl__Bar--top-left{left:20px}.cookieControl__Bar--bottom-right,.cookieControl__Bar--top-right{right:20px}.cookieControl__BarButtons{display:flex}.cookieControl__Modal{bottom:0;font-size:0;left:0;position:fixed;right:0;text-align:center;top:0;z-index:1}.cookieControl__Modal:before{content:"";display:inline-block;min-height:100vh;vertical-align:middle}.cookieControl__Modal:after{background-color:var(--cookie-control-modalOverlay);bottom:0;content:"";left:0;opacity:var(--cookie-control-modalOverlayOpacity);position:absolute;right:0;top:0;z-index:-1}.cookieControl__Modal>div{font-size:medium;padding-top:80px}.cookieControl__Modal button{background-color:var(--cookie-control-modalButtonBackground);color:var(--cookie-control-modalButtonColor)}.cookieControl__Modal button:hover{background-color:var(--cookie-control-modalButtonHoverBackground);color:var(--cookie-control-modalButtonHoverColor)}.cookieControl__ModalContent{background-color:var(--cookie-control-modalBackground);display:inline-block;max-height:80vh;max-width:550px;overflow-y:scroll;padding:40px;position:relative;text-align:left;vertical-align:middle;width:100%}.cookieControl__ModalContent,.cookieControl__ModalContent :not(button){color:var(--cookie-control-modalTextColor)}.cookieControl__ModalContent h3{font-size:24px;margin:50px 0 25px}.cookieControl__ModalContent h3:first-of-type{margin-top:0}.cookieControl__ModalContent ul{font-size:16px;list-style-type:none;padding:0}.cookieControl__ModalContent ul ul{padding:5px 56px 0}.cookieControl__ModalContent ul ul li+li{margin-top:5px}.cookieControl__ModalContent li{align-items:center}.cookieControl__ModalContent li+li{margin-top:20px}.cookieControl__ModalContent input{display:none}.cookieControl__ModalContent input:checked+label{background-color:var(--cookie-control-checkboxActiveBackground)}.cookieControl__ModalContent input:checked+label:before{background-color:var(--cookie-control-checkboxActiveCircleBackground);transform:translate3d(100%,-50%,0)}.cookieControl__ModalContent input:checked:disabled+label{background-color:var(--cookie-control-checkboxDisabledBackground)}.cookieControl__ModalContent input:checked:disabled+label:before{background-color:var(--cookie-control-checkboxDisabledCircleBackground)}.cookieControl__ModalContent label{backface-visibility:hidden;background-color:var(--cookie-control-checkboxInactiveBackground);border-radius:20px;display:block;font-size:0;margin-right:20px;min-height:20px;min-width:36px;position:relative;transition:background-color .2s}.cookieControl__ModalContent label:before{background-color:var(--cookie-control-checkboxInactiveCircleBackground);border-radius:50%;content:"";height:15px;left:3px;position:absolute;top:50%;transform:translate3d(0,-50%,0);transition:transform .2s;width:15px}.cookieControl__ModalInputWrapper{align-items:flex-start;display:flex}.cookieControl__ModalCookieName{font-weight:700;text-transform:uppercase}.cookieControl__ModalCookieName span{font-weight:400;text-transform:none}.cookieControl__ModalClose{position:absolute;right:20px;top:20px}.cookieControl__ModalButtons{align-items:flex-start;display:flex;margin-top:80px}.cookieControl__ModalButtons button+button{margin-left:20px}.cookieControl__ModalUnsaved{bottom:40px;color:var(--cookie-control-modalUnsavedColor);font-size:14px;left:50%;margin:0;position:absolute;transform:translateX(-50%)}.cookieControl__BlockedIframe{border:2px solid #ddd;padding:20px}.cookieControl__BlockedIframe a,.cookieControl__BlockedIframe p{font-family:Arial,"Helvetica Neue",Helvetica,sans-serif}@media screen and (max-width:768px){.cookieControl__Bar{flex-direction:column;left:0;right:0}.cookieControl__Bar h3,.cookieControl__Bar p{max-width:100%}.cookieControl__Bar--top-full,.cookieControl__Bar--top-left,.cookieControl__Bar--top-right{top:0}.cookieControl__Bar--bottom-full,.cookieControl__Bar--bottom-left,.cookieControl__Bar--bottom-right{bottom:0}.cookieControl__ModalContent{bottom:0;left:0;max-height:100%;max-width:none;padding:80px 20px 20px;position:absolute;right:0;top:0}.cookieControl__BarButtons{flex-direction:column;justify-content:center;margin-top:20px;width:100%}.cookieControl__BarButtons button{width:100%}.cookieControl__BarButtons button+button{margin:10px 0 0}.cookieControl__BarContainer,.cookieControl__ModalButtons{flex-direction:column}.cookieControl__ModalButtons button{width:100%}.cookieControl__ModalButtons button+button{margin:10px 0 0}}.cookieControl__ControlButton{backface-visibility:hidden;background:var(--cookie-control-controlButtonBackground);border:0;border-radius:50%;bottom:20px;box-shadow:0 0 10px rgba(0,0,0,.3);cursor:pointer;height:40px;min-height:40px;min-width:40px;outline:0;position:fixed;right:20px;transition:background-color .2s;width:40px}.cookieControl__ControlButton svg{backface-visibility:hidden;color:var(--cookie-control-controlButtonIconColor);left:50%;max-height:24px;max-width:24px;min-height:24px;min-width:24px;position:absolute;top:50%;transform:translate(-50%,-50%);transition:color .2s}.cookieControl__ControlButton:hover{background-color:var(--cookie-control-controlButtonHoverBackground)}.cookieControl__ControlButton:hover svg{color:var(--cookie-control-controlButtonIconHoverColor)} .vue-modal-resizer{bottom:0;cursor:se-resize;height:12px;overflow:hidden;right:0;width:12px;z-index:9999999}.vue-modal-resizer,.vue-modal-resizer:after{background:transparent;display:block;position:absolute}.vue-modal-resizer:after{border-bottom:10px solid #ddd;border-left:10px solid transparent;content:"";height:0;left:0;top:0;width:0}.vue-modal-resizer.clicked:after{border-bottom:10px solid #369be9}.v--modal-block-scroll{overflow:hidden;width:100vw}.v--modal-overlay{background:rgba(0,0,0,.2);box-sizing:border-box;height:100vh;left:0;opacity:1;position:fixed;top:0;width:100%;z-index:999}.v--modal-overlay.scrollable{-webkit-overflow-scrolling:touch;height:100%;min-height:100vh;overflow-y:auto}.v--modal-overlay .v--modal-background-click{height:auto;min-height:100%;width:100%}.v--modal-overlay .v--modal-box{box-sizing:border-box;overflow:hidden;position:relative}.v--modal-overlay.scrollable .v--modal-box{margin-bottom:2px}.v--modal{background-color:#fff;border-radius:3px;box-shadow:0 20px 60px -2px rgba(27,33,58,.4);padding:0;text-align:left}.v--modal.v--modal-fullscreen{height:100vh;left:0;margin:0;top:0;width:100vw}.v--modal-top-right{display:block;position:absolute;right:0;top:0}.overlay-fade-enter-active,.overlay-fade-leave-active{transition:all .2s}.overlay-fade-enter,.overlay-fade-leave-active{opacity:0}.nice-modal-fade-enter-active,.nice-modal-fade-leave-active{transition:all .4s}.nice-modal-fade-enter,.nice-modal-fade-leave-active{opacity:0;-webkit-transform:translateY(-20px);transform:translateY(-20px)}.vue-dialog div{box-sizing:border-box}.vue-dialog .dialog-flex{height:100%;width:100%}.vue-dialog .dialog-content{flex:1 0 auto;font-size:14px;padding:15px;width:100%}.vue-dialog .dialog-c-title{font-weight:600;padding-bottom:15px}.vue-dialog .vue-dialog-buttons{border-top:1px solid #eee;display:flex;flex:0 1 auto;width:100%}.vue-dialog .vue-dialog-buttons-none{padding-bottom:15px;width:100%}.vue-dialog-button{background:transparent;border:0;box-sizing:border-box;color:inherit;cursor:pointer;font-size:12px!important;height:40px;line-height:40px;font:inherit;margin:0;outline:none;padding:0}.vue-dialog-button:hover{background:rgba(0,0,0,.01)}.vue-dialog-button:active{background:rgba(0,0,0,.025)}.vue-dialog-button:not(:first-of-type){border-left:1px solid #eee} .nuxt-progress{background-color:#3088df;height:2px;left:0;opacity:1;position:fixed;right:0;top:0;transition:width .1s,opacity .4s;width:0;z-index:999999}.nuxt-progress.nuxt-progress-notransition{transition:none}.nuxt-progress-failed{background-color:red} .mt-0{margin-top:0!important}.mt-5{margin-top:5px!important}.mt-10{margin-top:10px!important}.mt-15{margin-top:15px!important}.mt-20{margin-top:20px!important}.mt-25{margin-top:25px!important}.mt-30{margin-top:30px!important}.mt-35{margin-top:35px!important}.mt-40{margin-top:40px!important}.mt-45{margin-top:45px!important}.mt-50{margin-top:50px!important}.mt-55{margin-top:55px!important}.mt-60{margin-top:60px!important}.mt-65{margin-top:65px!important}.mt-70{margin-top:70px!important}.mt-75{margin-top:75px!important}.mt-80{margin-top:80px!important}.mt-85{margin-top:85px!important}.mt-90{margin-top:90px!important}.mt-95{margin-top:95px!important}.mt-100{margin-top:100px!important}.mb-0{margin-bottom:0!important}.mb-5{margin-bottom:5px!important}.mb-10{margin-bottom:10px!important}.mb-15{margin-bottom:15px!important}.mb-20{margin-bottom:20px!important}.mb-25{margin-bottom:25px!important}.mb-30{margin-bottom:30px!important}.mb-35{margin-bottom:35px!important}.mb-40{margin-bottom:40px!important}.mb-45{margin-bottom:45px!important}.mb-50{margin-bottom:50px!important}.mb-55{margin-bottom:55px!important}.mb-60{margin-bottom:60px!important}.mb-65{margin-bottom:65px!important}.mb-70{margin-bottom:70px!important}.mb-75{margin-bottom:75px!important}.mb-80{margin-bottom:80px!important}.mb-85{margin-bottom:85px!important}.mb-90{margin-bottom:90px!important}.mb-95{margin-bottom:95px!important}.mb-100{margin-bottom:100px!important}.ml-0{margin-left:0!important}.ml-5{margin-left:5px!important}.ml-10{margin-left:10px!important}.ml-15{margin-left:15px!important}.ml-20{margin-left:20px!important}.ml-25{margin-left:25px!important}.ml-30{margin-left:30px!important}.ml-35{margin-left:35px!important}.ml-40{margin-left:40px!important}.ml-45{margin-left:45px!important}.ml-50{margin-left:50px!important}.ml-55{margin-left:55px!important}.ml-60{margin-left:60px!important}.ml-65{margin-left:65px!important}.ml-70{margin-left:70px!important}.ml-75{margin-left:75px!important}.ml-80{margin-left:80px!important}.ml-85{margin-left:85px!important}.ml-90{margin-left:90px!important}.ml-95{margin-left:95px!important}.ml-100{margin-left:100px!important}.mr-0{margin-right:0!important}.mr-5{margin-right:5px!important}.mr-10{margin-right:10px!important}.mr-15{margin-right:15px!important}.mr-20{margin-right:20px!important}.mr-25{margin-right:25px!important}.mr-30{margin-right:30px!important}.mr-35{margin-right:35px!important}.mr-40{margin-right:40px!important}.mr-45{margin-right:45px!important}.mr-50{margin-right:50px!important}.mr-55{margin-right:55px!important}.mr-60{margin-right:60px!important}.mr-65{margin-right:65px!important}.mr-70{margin-right:70px!important}.mr-75{margin-right:75px!important}.mr-80{margin-right:80px!important}.mr-85{margin-right:85px!important}.mr-90{margin-right:90px!important}.mr-95{margin-right:95px!important}.mr-100{margin-right:100px!important}.pt-0{padding-top:0!important}.pt-5{padding-top:5px!important}.pt-10{padding-top:10px!important}.pt-15{padding-top:15px!important}.pt-20{padding-top:20px!important}.pt-25{padding-top:25px!important}.pt-30{padding-top:30px!important}.pt-35{padding-top:35px!important}.pt-40{padding-top:40px!important}.pt-45{padding-top:45px!important}.pt-50{padding-top:50px!important}.pt-55{padding-top:55px!important}.pt-60{padding-top:60px!important}.pt-65{padding-top:65px!important}.pt-70{padding-top:70px!important}.pt-75{padding-top:75px!important}.pt-80{padding-top:80px!important}.pt-85{padding-top:85px!important}.pt-90{padding-top:90px!important}.pt-95{padding-top:95px!important}.pt-100{padding-top:100px!important}.pb-0{padding-bottom:0!important}.pb-5{padding-bottom:5px!important}.pb-10{padding-bottom:10px!important}.pb-15{padding-bottom:15px!important}.pb-20{padding-bottom:20px!important}.pb-25{padding-bottom:25px!important}.pb-30{padding-bottom:30px!important}.pb-35{padding-bottom:35px!important}.pb-40{padding-bottom:40px!important}.pb-45{padding-bottom:45px!important}.pb-50{padding-bottom:50px!important}.pb-55{padding-bottom:55px!important}.pb-60{padding-bottom:60px!important}.pb-65{padding-bottom:65px!important}.pb-70{padding-bottom:70px!important}.pb-75{padding-bottom:75px!important}.pb-80{padding-bottom:80px!important}.pb-85{padding-bottom:85px!important}.pb-90{padding-bottom:90px!important}.pb-95{padding-bottom:95px!important}.pb-100{padding-bottom:100px!important}.pl-0{padding-left:0!important}.pl-5{padding-left:5px!important}.pl-10{padding-left:10px!important}.pl-15{padding-left:15px!important}.pl-20{padding-left:20px!important}.pl-25{padding-left:25px!important}.pl-30{padding-left:30px!important}.pl-35{padding-left:35px!important}.pl-40{padding-left:40px!important}.pl-45{padding-left:45px!important}.pl-50{padding-left:50px!important}.pl-55{padding-left:55px!important}.pl-60{padding-left:60px!important}.pl-65{padding-left:65px!important}.pl-70{padding-left:70px!important}.pl-75{padding-left:75px!important}.pl-80{padding-left:80px!important}.pl-85{padding-left:85px!important}.pl-90{padding-left:90px!important}.pl-95{padding-left:95px!important}.pl-100{padding-left:100px!important}.pr-0{padding-right:0!important}.pr-5{padding-right:5px!important}.pr-10{padding-right:10px!important}.pr-15{padding-right:15px!important}.pr-20{padding-right:20px!important}.pr-25{padding-right:25px!important}.pr-30{padding-right:30px!important}.pr-35{padding-right:35px!important}.pr-40{padding-right:40px!important}.pr-45{padding-right:45px!important}.pr-50{padding-right:50px!important}.pr-55{padding-right:55px!important}.pr-60{padding-right:60px!important}.pr-65{padding-right:65px!important}.pr-70{padding-right:70px!important}.pr-75{padding-right:75px!important}.pr-80{padding-right:80px!important}.pr-85{padding-right:85px!important}.pr-90{padding-right:90px!important}.pr-95{padding-right:95px!important}.pr-100{padding-right:100px!important}*,:after,:before{box-sizing:inherit}html{font-size:62.5%}body{background-color:#fff;box-sizing:border-box;color:#1a254c;font-family:"Roboto",sans-serif;font-size:1.4rem;font-weight:400;margin:0;padding:0}#jmir-html{position:relative}#skip-link a{font-size:1.6rem;font-weight:700;height:1px;left:-10000px;margin:10px 0 10px 10px;overflow:hidden;padding:10px;position:absolute;-webkit-text-decoration:underline;text-decoration:underline;top:auto;width:1px}#skip-link a:focus{border:2px solid #f69038;display:inline-block;height:auto;position:static;width:auto}#main-layout-container{display:flex;flex-direction:column;height:100vh;justify-content:space-between}.cookieControl{z-index:999999999}.cookieControl__Bar{border-radius:0;box-shadow:0 0 40px -10px rgba(0,0,0,.75)}.cookieControl__BarButtons button,.cookieControl__ModalButtons button{cursor:pointer;font-size:14px;opacity:1;padding:10px 20px;text-align:center;transition:.3s}.cookieControl__BarButtons button:focus,.cookieControl__BarButtons button:hover,.cookieControl__ModalButtons button:focus,.cookieControl__ModalButtons button:hover{font-weight:400;opacity:.9;-webkit-text-decoration:none;text-decoration:none}.cookieControl__BarButtons button:focus,.cookieControl__ModalButtons button:focus{outline:2px solid #f69038!important;outline-offset:6px!important}.cookieControl__BarButtons button:first-child{background-color:#f1f3f5;border:1px solid #dcdee0;color:#1a254c;display:none}.cookieControl__BarButtons button:first-child:active{background-color:#dcdee0}.cookieControl__BarButtons button:last-child{background-color:#367c3a;border:1px solid #367c3a;color:#fff}.cookieControl__BarButtons button:last-child:active{background-color:#3b9d3f;border:1px solid #3b9d3f}.cookieControl__ModalButtons button:first-child{background-color:#367c3a;border:1px solid #367c3a;color:#fff}.cookieControl__ModalButtons button:first-child:active{background-color:#3b9d3f;border:1px solid #3b9d3f}.cookieControl__ModalButtons button:last-child{background-color:#b30000;border:1px solid #b30000;color:#fff}.cookieControl__ModalButtons button:last-child:active{background-color:#ba302d;border:1px solid #ba302d}.cookieControl__Modal>div{padding-top:50px}.cookieControl__ModalContent{background-color:#f3f3f5;padding-left:0;padding-right:0;position:relative}.cookieControl__ModalContent div:first-child{background-color:#fff;border-top:1px solid #bab4b4;padding-left:20px;padding-right:20px}.cookieControl__ModalClose{top:5px!important}.cookieControl__ModalContent ul li div{border-top:none!important}.cookieControl__ModalContent ul li ul li{font-size:14px;font-style:italic;margin-left:21px}.cookieControl__ModalContent ul:last-child{padding-bottom:20px}.cookieControl__ModalContent h2{font-size:18px;margin:0;position:absolute;top:5px!important}.cookieControl__ModalContent h3{padding-top:20px}.cookieControl__ModalButtons{bottom:10px;margin-left:auto;position:absolute;right:20px;width:-moz-fit-content;width:fit-content}.cookieControl__ModalClose{background-color:transparent!important;color:gray!important}.cookieControl__ModalClose:hover{color:#000!important}.cookieControl__ModalCookieName{text-transform:none!important}.toasted-container.top-right{right:2%!important;top:2%!important}._hj-3ZiaL__MinimizedWidgetBottom__container{bottom:.5%!important;flex-direction:row!important;justify-content:flex-end;width:84%!important}span a.a-select-membership{font-size:1.2rem;padding:5px 10px;-webkit-text-decoration:none!important;text-decoration:none!important} .mt-0[data-v-6301a668]{margin-top:0!important}.mt-5[data-v-6301a668]{margin-top:5px!important}.mt-10[data-v-6301a668]{margin-top:10px!important}.mt-15[data-v-6301a668]{margin-top:15px!important}.mt-20[data-v-6301a668]{margin-top:20px!important}.mt-25[data-v-6301a668]{margin-top:25px!important}.mt-30[data-v-6301a668]{margin-top:30px!important}.mt-35[data-v-6301a668]{margin-top:35px!important}.mt-40[data-v-6301a668]{margin-top:40px!important}.mt-45[data-v-6301a668]{margin-top:45px!important}.mt-50[data-v-6301a668]{margin-top:50px!important}.mt-55[data-v-6301a668]{margin-top:55px!important}.mt-60[data-v-6301a668]{margin-top:60px!important}.mt-65[data-v-6301a668]{margin-top:65px!important}.mt-70[data-v-6301a668]{margin-top:70px!important}.mt-75[data-v-6301a668]{margin-top:75px!important}.mt-80[data-v-6301a668]{margin-top:80px!important}.mt-85[data-v-6301a668]{margin-top:85px!important}.mt-90[data-v-6301a668]{margin-top:90px!important}.mt-95[data-v-6301a668]{margin-top:95px!important}.mt-100[data-v-6301a668]{margin-top:100px!important}.mb-0[data-v-6301a668]{margin-bottom:0!important}.mb-5[data-v-6301a668]{margin-bottom:5px!important}.mb-10[data-v-6301a668]{margin-bottom:10px!important}.mb-15[data-v-6301a668]{margin-bottom:15px!important}.mb-20[data-v-6301a668]{margin-bottom:20px!important}.mb-25[data-v-6301a668]{margin-bottom:25px!important}.mb-30[data-v-6301a668]{margin-bottom:30px!important}.mb-35[data-v-6301a668]{margin-bottom:35px!important}.mb-40[data-v-6301a668]{margin-bottom:40px!important}.mb-45[data-v-6301a668]{margin-bottom:45px!important}.mb-50[data-v-6301a668]{margin-bottom:50px!important}.mb-55[data-v-6301a668]{margin-bottom:55px!important}.mb-60[data-v-6301a668]{margin-bottom:60px!important}.mb-65[data-v-6301a668]{margin-bottom:65px!important}.mb-70[data-v-6301a668]{margin-bottom:70px!important}.mb-75[data-v-6301a668]{margin-bottom:75px!important}.mb-80[data-v-6301a668]{margin-bottom:80px!important}.mb-85[data-v-6301a668]{margin-bottom:85px!important}.mb-90[data-v-6301a668]{margin-bottom:90px!important}.mb-95[data-v-6301a668]{margin-bottom:95px!important}.mb-100[data-v-6301a668]{margin-bottom:100px!important}.ml-0[data-v-6301a668]{margin-left:0!important}.ml-5[data-v-6301a668]{margin-left:5px!important}.ml-10[data-v-6301a668]{margin-left:10px!important}.ml-15[data-v-6301a668]{margin-left:15px!important}.ml-20[data-v-6301a668]{margin-left:20px!important}.ml-25[data-v-6301a668]{margin-left:25px!important}.ml-30[data-v-6301a668]{margin-left:30px!important}.ml-35[data-v-6301a668]{margin-left:35px!important}.ml-40[data-v-6301a668]{margin-left:40px!important}.ml-45[data-v-6301a668]{margin-left:45px!important}.ml-50[data-v-6301a668]{margin-left:50px!important}.ml-55[data-v-6301a668]{margin-left:55px!important}.ml-60[data-v-6301a668]{margin-left:60px!important}.ml-65[data-v-6301a668]{margin-left:65px!important}.ml-70[data-v-6301a668]{margin-left:70px!important}.ml-75[data-v-6301a668]{margin-left:75px!important}.ml-80[data-v-6301a668]{margin-left:80px!important}.ml-85[data-v-6301a668]{margin-left:85px!important}.ml-90[data-v-6301a668]{margin-left:90px!important}.ml-95[data-v-6301a668]{margin-left:95px!important}.ml-100[data-v-6301a668]{margin-left:100px!important}.mr-0[data-v-6301a668]{margin-right:0!important}.mr-5[data-v-6301a668]{margin-right:5px!important}.mr-10[data-v-6301a668]{margin-right:10px!important}.mr-15[data-v-6301a668]{margin-right:15px!important}.mr-20[data-v-6301a668]{margin-right:20px!important}.mr-25[data-v-6301a668]{margin-right:25px!important}.mr-30[data-v-6301a668]{margin-right:30px!important}.mr-35[data-v-6301a668]{margin-right:35px!important}.mr-40[data-v-6301a668]{margin-right:40px!important}.mr-45[data-v-6301a668]{margin-right:45px!important}.mr-50[data-v-6301a668]{margin-right:50px!important}.mr-55[data-v-6301a668]{margin-right:55px!important}.mr-60[data-v-6301a668]{margin-right:60px!important}.mr-65[data-v-6301a668]{margin-right:65px!important}.mr-70[data-v-6301a668]{margin-right:70px!important}.mr-75[data-v-6301a668]{margin-right:75px!important}.mr-80[data-v-6301a668]{margin-right:80px!important}.mr-85[data-v-6301a668]{margin-right:85px!important}.mr-90[data-v-6301a668]{margin-right:90px!important}.mr-95[data-v-6301a668]{margin-right:95px!important}.mr-100[data-v-6301a668]{margin-right:100px!important}.pt-0[data-v-6301a668]{padding-top:0!important}.pt-5[data-v-6301a668]{padding-top:5px!important}.pt-10[data-v-6301a668]{padding-top:10px!important}.pt-15[data-v-6301a668]{padding-top:15px!important}.pt-20[data-v-6301a668]{padding-top:20px!important}.pt-25[data-v-6301a668]{padding-top:25px!important}.pt-30[data-v-6301a668]{padding-top:30px!important}.pt-35[data-v-6301a668]{padding-top:35px!important}.pt-40[data-v-6301a668]{padding-top:40px!important}.pt-45[data-v-6301a668]{padding-top:45px!important}.pt-50[data-v-6301a668]{padding-top:50px!important}.pt-55[data-v-6301a668]{padding-top:55px!important}.pt-60[data-v-6301a668]{padding-top:60px!important}.pt-65[data-v-6301a668]{padding-top:65px!important}.pt-70[data-v-6301a668]{padding-top:70px!important}.pt-75[data-v-6301a668]{padding-top:75px!important}.pt-80[data-v-6301a668]{padding-top:80px!important}.pt-85[data-v-6301a668]{padding-top:85px!important}.pt-90[data-v-6301a668]{padding-top:90px!important}.pt-95[data-v-6301a668]{padding-top:95px!important}.pt-100[data-v-6301a668]{padding-top:100px!important}.pb-0[data-v-6301a668]{padding-bottom:0!important}.pb-5[data-v-6301a668]{padding-bottom:5px!important}.pb-10[data-v-6301a668]{padding-bottom:10px!important}.pb-15[data-v-6301a668]{padding-bottom:15px!important}.pb-20[data-v-6301a668]{padding-bottom:20px!important}.pb-25[data-v-6301a668]{padding-bottom:25px!important}.pb-30[data-v-6301a668]{padding-bottom:30px!important}.pb-35[data-v-6301a668]{padding-bottom:35px!important}.pb-40[data-v-6301a668]{padding-bottom:40px!important}.pb-45[data-v-6301a668]{padding-bottom:45px!important}.pb-50[data-v-6301a668]{padding-bottom:50px!important}.pb-55[data-v-6301a668]{padding-bottom:55px!important}.pb-60[data-v-6301a668]{padding-bottom:60px!important}.pb-65[data-v-6301a668]{padding-bottom:65px!important}.pb-70[data-v-6301a668]{padding-bottom:70px!important}.pb-75[data-v-6301a668]{padding-bottom:75px!important}.pb-80[data-v-6301a668]{padding-bottom:80px!important}.pb-85[data-v-6301a668]{padding-bottom:85px!important}.pb-90[data-v-6301a668]{padding-bottom:90px!important}.pb-95[data-v-6301a668]{padding-bottom:95px!important}.pb-100[data-v-6301a668]{padding-bottom:100px!important}.pl-0[data-v-6301a668]{padding-left:0!important}.pl-5[data-v-6301a668]{padding-left:5px!important}.pl-10[data-v-6301a668]{padding-left:10px!important}.pl-15[data-v-6301a668]{padding-left:15px!important}.pl-20[data-v-6301a668]{padding-left:20px!important}.pl-25[data-v-6301a668]{padding-left:25px!important}.pl-30[data-v-6301a668]{padding-left:30px!important}.pl-35[data-v-6301a668]{padding-left:35px!important}.pl-40[data-v-6301a668]{padding-left:40px!important}.pl-45[data-v-6301a668]{padding-left:45px!important}.pl-50[data-v-6301a668]{padding-left:50px!important}.pl-55[data-v-6301a668]{padding-left:55px!important}.pl-60[data-v-6301a668]{padding-left:60px!important}.pl-65[data-v-6301a668]{padding-left:65px!important}.pl-70[data-v-6301a668]{padding-left:70px!important}.pl-75[data-v-6301a668]{padding-left:75px!important}.pl-80[data-v-6301a668]{padding-left:80px!important}.pl-85[data-v-6301a668]{padding-left:85px!important}.pl-90[data-v-6301a668]{padding-left:90px!important}.pl-95[data-v-6301a668]{padding-left:95px!important}.pl-100[data-v-6301a668]{padding-left:100px!important}.pr-0[data-v-6301a668]{padding-right:0!important}.pr-5[data-v-6301a668]{padding-right:5px!important}.pr-10[data-v-6301a668]{padding-right:10px!important}.pr-15[data-v-6301a668]{padding-right:15px!important}.pr-20[data-v-6301a668]{padding-right:20px!important}.pr-25[data-v-6301a668]{padding-right:25px!important}.pr-30[data-v-6301a668]{padding-right:30px!important}.pr-35[data-v-6301a668]{padding-right:35px!important}.pr-40[data-v-6301a668]{padding-right:40px!important}.pr-45[data-v-6301a668]{padding-right:45px!important}.pr-50[data-v-6301a668]{padding-right:50px!important}.pr-55[data-v-6301a668]{padding-right:55px!important}.pr-60[data-v-6301a668]{padding-right:60px!important}.pr-65[data-v-6301a668]{padding-right:65px!important}.pr-70[data-v-6301a668]{padding-right:70px!important}.pr-75[data-v-6301a668]{padding-right:75px!important}.pr-80[data-v-6301a668]{padding-right:80px!important}.pr-85[data-v-6301a668]{padding-right:85px!important}.pr-90[data-v-6301a668]{padding-right:90px!important}.pr-95[data-v-6301a668]{padding-right:95px!important}.pr-100[data-v-6301a668]{padding-right:100px!important}.top-hero-banner[data-v-6301a668]{background-color:#e5f2fe;padding:0 10vw 0 5vw;position:relative}.top-hero-banner__divider[data-v-6301a668]{border-top:1px solid #1a254c}.top-hero-banner__close[data-v-6301a668]{color:#1a254c;cursor:pointer;font-size:2rem;position:absolute;right:10vw;transition:.3s}.top-hero-banner__close[data-v-6301a668]:hover{color:#2e4185}.top-hero-banner__info[data-v-6301a668]{display:flex}.top-hero-banner__info .icon[data-v-6301a668]{font-size:2rem;margin-top:3px}.top-hero-banner__actions[data-v-6301a668]{margin-left:2.9rem}.top-hero-banner__actions a[data-v-6301a668]{display:inline-block;margin-right:2rem}@media screen and (max-width:31.25em){.top-hero-banner__actions a[data-v-6301a668]{padding:5px 0}}.top-hero-banner__link[data-v-6301a668]{color:#1a254c;display:inline-block;font-size:1.6rem;font-weight:700;margin-bottom:5px;margin-right:4rem;-webkit-text-decoration:none;text-decoration:none}.top-hero-banner__link[data-v-6301a668]:focus,.top-hero-banner__link[data-v-6301a668]:hover{-webkit-text-decoration:underline;text-decoration:underline}@media screen and (max-width:47.9375em){.top-hero-banner__link[data-v-6301a668]{display:block;margin-right:0}} .mt-0[data-v-49c694ee]{margin-top:0!important}.mt-5[data-v-49c694ee]{margin-top:5px!important}.mt-10[data-v-49c694ee]{margin-top:10px!important}.mt-15[data-v-49c694ee]{margin-top:15px!important}.mt-20[data-v-49c694ee]{margin-top:20px!important}.mt-25[data-v-49c694ee]{margin-top:25px!important}.mt-30[data-v-49c694ee]{margin-top:30px!important}.mt-35[data-v-49c694ee]{margin-top:35px!important}.mt-40[data-v-49c694ee]{margin-top:40px!important}.mt-45[data-v-49c694ee]{margin-top:45px!important}.mt-50[data-v-49c694ee]{margin-top:50px!important}.mt-55[data-v-49c694ee]{margin-top:55px!important}.mt-60[data-v-49c694ee]{margin-top:60px!important}.mt-65[data-v-49c694ee]{margin-top:65px!important}.mt-70[data-v-49c694ee]{margin-top:70px!important}.mt-75[data-v-49c694ee]{margin-top:75px!important}.mt-80[data-v-49c694ee]{margin-top:80px!important}.mt-85[data-v-49c694ee]{margin-top:85px!important}.mt-90[data-v-49c694ee]{margin-top:90px!important}.mt-95[data-v-49c694ee]{margin-top:95px!important}.mt-100[data-v-49c694ee]{margin-top:100px!important}.mb-0[data-v-49c694ee]{margin-bottom:0!important}.mb-5[data-v-49c694ee]{margin-bottom:5px!important}.mb-10[data-v-49c694ee]{margin-bottom:10px!important}.mb-15[data-v-49c694ee]{margin-bottom:15px!important}.mb-20[data-v-49c694ee]{margin-bottom:20px!important}.mb-25[data-v-49c694ee]{margin-bottom:25px!important}.mb-30[data-v-49c694ee]{margin-bottom:30px!important}.mb-35[data-v-49c694ee]{margin-bottom:35px!important}.mb-40[data-v-49c694ee]{margin-bottom:40px!important}.mb-45[data-v-49c694ee]{margin-bottom:45px!important}.mb-50[data-v-49c694ee]{margin-bottom:50px!important}.mb-55[data-v-49c694ee]{margin-bottom:55px!important}.mb-60[data-v-49c694ee]{margin-bottom:60px!important}.mb-65[data-v-49c694ee]{margin-bottom:65px!important}.mb-70[data-v-49c694ee]{margin-bottom:70px!important}.mb-75[data-v-49c694ee]{margin-bottom:75px!important}.mb-80[data-v-49c694ee]{margin-bottom:80px!important}.mb-85[data-v-49c694ee]{margin-bottom:85px!important}.mb-90[data-v-49c694ee]{margin-bottom:90px!important}.mb-95[data-v-49c694ee]{margin-bottom:95px!important}.mb-100[data-v-49c694ee]{margin-bottom:100px!important}.ml-0[data-v-49c694ee]{margin-left:0!important}.ml-5[data-v-49c694ee]{margin-left:5px!important}.ml-10[data-v-49c694ee]{margin-left:10px!important}.ml-15[data-v-49c694ee]{margin-left:15px!important}.ml-20[data-v-49c694ee]{margin-left:20px!important}.ml-25[data-v-49c694ee]{margin-left:25px!important}.ml-30[data-v-49c694ee]{margin-left:30px!important}.ml-35[data-v-49c694ee]{margin-left:35px!important}.ml-40[data-v-49c694ee]{margin-left:40px!important}.ml-45[data-v-49c694ee]{margin-left:45px!important}.ml-50[data-v-49c694ee]{margin-left:50px!important}.ml-55[data-v-49c694ee]{margin-left:55px!important}.ml-60[data-v-49c694ee]{margin-left:60px!important}.ml-65[data-v-49c694ee]{margin-left:65px!important}.ml-70[data-v-49c694ee]{margin-left:70px!important}.ml-75[data-v-49c694ee]{margin-left:75px!important}.ml-80[data-v-49c694ee]{margin-left:80px!important}.ml-85[data-v-49c694ee]{margin-left:85px!important}.ml-90[data-v-49c694ee]{margin-left:90px!important}.ml-95[data-v-49c694ee]{margin-left:95px!important}.ml-100[data-v-49c694ee]{margin-left:100px!important}.mr-0[data-v-49c694ee]{margin-right:0!important}.mr-5[data-v-49c694ee]{margin-right:5px!important}.mr-10[data-v-49c694ee]{margin-right:10px!important}.mr-15[data-v-49c694ee]{margin-right:15px!important}.mr-20[data-v-49c694ee]{margin-right:20px!important}.mr-25[data-v-49c694ee]{margin-right:25px!important}.mr-30[data-v-49c694ee]{margin-right:30px!important}.mr-35[data-v-49c694ee]{margin-right:35px!important}.mr-40[data-v-49c694ee]{margin-right:40px!important}.mr-45[data-v-49c694ee]{margin-right:45px!important}.mr-50[data-v-49c694ee]{margin-right:50px!important}.mr-55[data-v-49c694ee]{margin-right:55px!important}.mr-60[data-v-49c694ee]{margin-right:60px!important}.mr-65[data-v-49c694ee]{margin-right:65px!important}.mr-70[data-v-49c694ee]{margin-right:70px!important}.mr-75[data-v-49c694ee]{margin-right:75px!important}.mr-80[data-v-49c694ee]{margin-right:80px!important}.mr-85[data-v-49c694ee]{margin-right:85px!important}.mr-90[data-v-49c694ee]{margin-right:90px!important}.mr-95[data-v-49c694ee]{margin-right:95px!important}.mr-100[data-v-49c694ee]{margin-right:100px!important}.pt-0[data-v-49c694ee]{padding-top:0!important}.pt-5[data-v-49c694ee]{padding-top:5px!important}.pt-10[data-v-49c694ee]{padding-top:10px!important}.pt-15[data-v-49c694ee]{padding-top:15px!important}.pt-20[data-v-49c694ee]{padding-top:20px!important}.pt-25[data-v-49c694ee]{padding-top:25px!important}.pt-30[data-v-49c694ee]{padding-top:30px!important}.pt-35[data-v-49c694ee]{padding-top:35px!important}.pt-40[data-v-49c694ee]{padding-top:40px!important}.pt-45[data-v-49c694ee]{padding-top:45px!important}.pt-50[data-v-49c694ee]{padding-top:50px!important}.pt-55[data-v-49c694ee]{padding-top:55px!important}.pt-60[data-v-49c694ee]{padding-top:60px!important}.pt-65[data-v-49c694ee]{padding-top:65px!important}.pt-70[data-v-49c694ee]{padding-top:70px!important}.pt-75[data-v-49c694ee]{padding-top:75px!important}.pt-80[data-v-49c694ee]{padding-top:80px!important}.pt-85[data-v-49c694ee]{padding-top:85px!important}.pt-90[data-v-49c694ee]{padding-top:90px!important}.pt-95[data-v-49c694ee]{padding-top:95px!important}.pt-100[data-v-49c694ee]{padding-top:100px!important}.pb-0[data-v-49c694ee]{padding-bottom:0!important}.pb-5[data-v-49c694ee]{padding-bottom:5px!important}.pb-10[data-v-49c694ee]{padding-bottom:10px!important}.pb-15[data-v-49c694ee]{padding-bottom:15px!important}.pb-20[data-v-49c694ee]{padding-bottom:20px!important}.pb-25[data-v-49c694ee]{padding-bottom:25px!important}.pb-30[data-v-49c694ee]{padding-bottom:30px!important}.pb-35[data-v-49c694ee]{padding-bottom:35px!important}.pb-40[data-v-49c694ee]{padding-bottom:40px!important}.pb-45[data-v-49c694ee]{padding-bottom:45px!important}.pb-50[data-v-49c694ee]{padding-bottom:50px!important}.pb-55[data-v-49c694ee]{padding-bottom:55px!important}.pb-60[data-v-49c694ee]{padding-bottom:60px!important}.pb-65[data-v-49c694ee]{padding-bottom:65px!important}.pb-70[data-v-49c694ee]{padding-bottom:70px!important}.pb-75[data-v-49c694ee]{padding-bottom:75px!important}.pb-80[data-v-49c694ee]{padding-bottom:80px!important}.pb-85[data-v-49c694ee]{padding-bottom:85px!important}.pb-90[data-v-49c694ee]{padding-bottom:90px!important}.pb-95[data-v-49c694ee]{padding-bottom:95px!important}.pb-100[data-v-49c694ee]{padding-bottom:100px!important}.pl-0[data-v-49c694ee]{padding-left:0!important}.pl-5[data-v-49c694ee]{padding-left:5px!important}.pl-10[data-v-49c694ee]{padding-left:10px!important}.pl-15[data-v-49c694ee]{padding-left:15px!important}.pl-20[data-v-49c694ee]{padding-left:20px!important}.pl-25[data-v-49c694ee]{padding-left:25px!important}.pl-30[data-v-49c694ee]{padding-left:30px!important}.pl-35[data-v-49c694ee]{padding-left:35px!important}.pl-40[data-v-49c694ee]{padding-left:40px!important}.pl-45[data-v-49c694ee]{padding-left:45px!important}.pl-50[data-v-49c694ee]{padding-left:50px!important}.pl-55[data-v-49c694ee]{padding-left:55px!important}.pl-60[data-v-49c694ee]{padding-left:60px!important}.pl-65[data-v-49c694ee]{padding-left:65px!important}.pl-70[data-v-49c694ee]{padding-left:70px!important}.pl-75[data-v-49c694ee]{padding-left:75px!important}.pl-80[data-v-49c694ee]{padding-left:80px!important}.pl-85[data-v-49c694ee]{padding-left:85px!important}.pl-90[data-v-49c694ee]{padding-left:90px!important}.pl-95[data-v-49c694ee]{padding-left:95px!important}.pl-100[data-v-49c694ee]{padding-left:100px!important}.pr-0[data-v-49c694ee]{padding-right:0!important}.pr-5[data-v-49c694ee]{padding-right:5px!important}.pr-10[data-v-49c694ee]{padding-right:10px!important}.pr-15[data-v-49c694ee]{padding-right:15px!important}.pr-20[data-v-49c694ee]{padding-right:20px!important}.pr-25[data-v-49c694ee]{padding-right:25px!important}.pr-30[data-v-49c694ee]{padding-right:30px!important}.pr-35[data-v-49c694ee]{padding-right:35px!important}.pr-40[data-v-49c694ee]{padding-right:40px!important}.pr-45[data-v-49c694ee]{padding-right:45px!important}.pr-50[data-v-49c694ee]{padding-right:50px!important}.pr-55[data-v-49c694ee]{padding-right:55px!important}.pr-60[data-v-49c694ee]{padding-right:60px!important}.pr-65[data-v-49c694ee]{padding-right:65px!important}.pr-70[data-v-49c694ee]{padding-right:70px!important}.pr-75[data-v-49c694ee]{padding-right:75px!important}.pr-80[data-v-49c694ee]{padding-right:80px!important}.pr-85[data-v-49c694ee]{padding-right:85px!important}.pr-90[data-v-49c694ee]{padding-right:90px!important}.pr-95[data-v-49c694ee]{padding-right:95px!important}.pr-100[data-v-49c694ee]{padding-right:100px!important}.universal-access-btn[data-v-49c694ee]{background-color:#fff;border:2px solid transparent;border-radius:100px;color:#1e70c2;cursor:pointer;display:block;font-size:4rem;position:fixed;right:2px;top:2px;z-index:200}.universal-access-btn[data-v-49c694ee]:focus{border:2px solid #f69038;outline:none}.accessibility *[data-v-49c694ee]{line-height:inherit!important}.accessibility[data-v-49c694ee] :not(.icon){font-weight:400!important}.accessibility[data-v-49c694ee]{background-color:#fff;border:2px solid #f69038;border-radius:20px 20px 20px 20px;box-shadow:0 0 40px -10px rgba(0,0,0,.75);padding:30px 30px 10px;position:fixed;right:0;top:0;z-index:1000000000000000}@media screen and (max-width:33.1875em){.accessibility[data-v-49c694ee]{width:100%}}.accessibility__wrapper[data-v-49c694ee]{position:relative}@media screen and (max-width:59.6875em){.accessibility__wrapper[data-v-49c694ee]{flex-direction:column}}.accessibility__wrapper button[data-v-49c694ee]{border:2px solid #dcdee0}.accessibility__wrapper button[data-v-49c694ee]:focus{border:2px solid #f69038}.accessibility__wrapper h3[data-v-49c694ee]{font-size:26px!important}.accessibility__wrapper h3 .icon[data-v-49c694ee]{color:#1e70c2}.accessibility__wrapper h4[data-v-49c694ee]{font-size:18px!important}.accessibility__settings[data-v-49c694ee]{max-height:600px;overflow-x:hidden;overflow-y:auto;padding-right:20px}.accessibility__settings[data-v-49c694ee]::-webkit-scrollbar{-webkit-appearance:none}.accessibility__settings[data-v-49c694ee]::-webkit-scrollbar-track{border-radius:8px;box-shadow:inset 0 0 5px #dcdee0}.accessibility__settings[data-v-49c694ee]::-webkit-scrollbar:vertical{width:3px}.accessibility__settings[data-v-49c694ee]::-webkit-scrollbar-thumb{background-color:#1e70c2;border-radius:1px}.accessibility__settings button[data-v-49c694ee]{border-radius:10px;width:150px}@media screen and (max-width:45.9375em){.accessibility__settings button[data-v-49c694ee]{width:100%}}@media screen and (max-width:59.6875em){.accessibility__settings[data-v-49c694ee]{max-height:200px;padding-right:0;width:100%}}.accessibility__close-top[data-v-49c694ee]{background-color:transparent;border:2px solid #fff!important;color:gray;cursor:pointer;font-size:20px;position:absolute;right:-27px;top:-27px;transition:.3s;width:auto!important}.accessibility__close-top[data-v-49c694ee]:hover{color:#000}.accessibility__close-top[data-v-49c694ee]:focus{border:2px solid #f69038!important;color:#000}.accessibility__link[data-v-49c694ee]{font-size:14px;font-weight:400}.accessibility__link[data-v-49c694ee],.accessibility__section[data-v-49c694ee]{text-align:left!important}.accessibility__colour button[data-v-49c694ee],.accessibility__content button[data-v-49c694ee],.accessibility__font button[data-v-49c694ee]{text-align:center}.accessibility__colour button span[data-v-49c694ee],.accessibility__content button span[data-v-49c694ee],.accessibility__font button span[data-v-49c694ee]{display:block;font-size:20px}.accessibility__font-menu[data-v-49c694ee]{background-color:#f0f3f5;border-radius:20px;padding:20px}.accessibility__font-menu button[data-v-49c694ee]{text-align:center;width:-moz-fit-content;width:fit-content}.accessibility__font-menu button span[data-v-49c694ee]{display:block;font-size:20px}.accessibility__font-menu p[data-v-49c694ee]{display:inline-block;font-size:14px}.accessibility__content[data-v-49c694ee]{margin-bottom:20px}.accessibility__footer[data-v-49c694ee]{float:right;margin-top:20px}.accessibility__reset-settings[data-v-49c694ee]{border:2px solid #dcdee0}.accessibility__reset-settings[data-v-49c694ee]:focus{border:2px solid #f69038}.accessibility__close-bottom[data-v-49c694ee]{border:2px solid #dcdee0}.accessibility__close-bottom[data-v-49c694ee]:focus{border:2px solid #f69038}.accessibility .btn[data-v-49c694ee]{outline:none!important}.accessibility .deactive[data-v-49c694ee]{color:#000;cursor:not-allowed;opacity:.5;pointer-events:none}.accessibility .acc-active[data-v-49c694ee]{background-color:#1e70c2!important;border:1px solid #1e70c2!important;color:#fff!important}.accessibility button[data-v-49c694ee]{font-size:14px!important} .mt-0[data-v-575455fb]{margin-top:0!important}.mt-5[data-v-575455fb]{margin-top:5px!important}.mt-10[data-v-575455fb]{margin-top:10px!important}.mt-15[data-v-575455fb]{margin-top:15px!important}.mt-20[data-v-575455fb]{margin-top:20px!important}.mt-25[data-v-575455fb]{margin-top:25px!important}.mt-30[data-v-575455fb]{margin-top:30px!important}.mt-35[data-v-575455fb]{margin-top:35px!important}.mt-40[data-v-575455fb]{margin-top:40px!important}.mt-45[data-v-575455fb]{margin-top:45px!important}.mt-50[data-v-575455fb]{margin-top:50px!important}.mt-55[data-v-575455fb]{margin-top:55px!important}.mt-60[data-v-575455fb]{margin-top:60px!important}.mt-65[data-v-575455fb]{margin-top:65px!important}.mt-70[data-v-575455fb]{margin-top:70px!important}.mt-75[data-v-575455fb]{margin-top:75px!important}.mt-80[data-v-575455fb]{margin-top:80px!important}.mt-85[data-v-575455fb]{margin-top:85px!important}.mt-90[data-v-575455fb]{margin-top:90px!important}.mt-95[data-v-575455fb]{margin-top:95px!important}.mt-100[data-v-575455fb]{margin-top:100px!important}.mb-0[data-v-575455fb]{margin-bottom:0!important}.mb-5[data-v-575455fb]{margin-bottom:5px!important}.mb-10[data-v-575455fb]{margin-bottom:10px!important}.mb-15[data-v-575455fb]{margin-bottom:15px!important}.mb-20[data-v-575455fb]{margin-bottom:20px!important}.mb-25[data-v-575455fb]{margin-bottom:25px!important}.mb-30[data-v-575455fb]{margin-bottom:30px!important}.mb-35[data-v-575455fb]{margin-bottom:35px!important}.mb-40[data-v-575455fb]{margin-bottom:40px!important}.mb-45[data-v-575455fb]{margin-bottom:45px!important}.mb-50[data-v-575455fb]{margin-bottom:50px!important}.mb-55[data-v-575455fb]{margin-bottom:55px!important}.mb-60[data-v-575455fb]{margin-bottom:60px!important}.mb-65[data-v-575455fb]{margin-bottom:65px!important}.mb-70[data-v-575455fb]{margin-bottom:70px!important}.mb-75[data-v-575455fb]{margin-bottom:75px!important}.mb-80[data-v-575455fb]{margin-bottom:80px!important}.mb-85[data-v-575455fb]{margin-bottom:85px!important}.mb-90[data-v-575455fb]{margin-bottom:90px!important}.mb-95[data-v-575455fb]{margin-bottom:95px!important}.mb-100[data-v-575455fb]{margin-bottom:100px!important}.ml-0[data-v-575455fb]{margin-left:0!important}.ml-5[data-v-575455fb]{margin-left:5px!important}.ml-10[data-v-575455fb]{margin-left:10px!important}.ml-15[data-v-575455fb]{margin-left:15px!important}.ml-20[data-v-575455fb]{margin-left:20px!important}.ml-25[data-v-575455fb]{margin-left:25px!important}.ml-30[data-v-575455fb]{margin-left:30px!important}.ml-35[data-v-575455fb]{margin-left:35px!important}.ml-40[data-v-575455fb]{margin-left:40px!important}.ml-45[data-v-575455fb]{margin-left:45px!important}.ml-50[data-v-575455fb]{margin-left:50px!important}.ml-55[data-v-575455fb]{margin-left:55px!important}.ml-60[data-v-575455fb]{margin-left:60px!important}.ml-65[data-v-575455fb]{margin-left:65px!important}.ml-70[data-v-575455fb]{margin-left:70px!important}.ml-75[data-v-575455fb]{margin-left:75px!important}.ml-80[data-v-575455fb]{margin-left:80px!important}.ml-85[data-v-575455fb]{margin-left:85px!important}.ml-90[data-v-575455fb]{margin-left:90px!important}.ml-95[data-v-575455fb]{margin-left:95px!important}.ml-100[data-v-575455fb]{margin-left:100px!important}.mr-0[data-v-575455fb]{margin-right:0!important}.mr-5[data-v-575455fb]{margin-right:5px!important}.mr-10[data-v-575455fb]{margin-right:10px!important}.mr-15[data-v-575455fb]{margin-right:15px!important}.mr-20[data-v-575455fb]{margin-right:20px!important}.mr-25[data-v-575455fb]{margin-right:25px!important}.mr-30[data-v-575455fb]{margin-right:30px!important}.mr-35[data-v-575455fb]{margin-right:35px!important}.mr-40[data-v-575455fb]{margin-right:40px!important}.mr-45[data-v-575455fb]{margin-right:45px!important}.mr-50[data-v-575455fb]{margin-right:50px!important}.mr-55[data-v-575455fb]{margin-right:55px!important}.mr-60[data-v-575455fb]{margin-right:60px!important}.mr-65[data-v-575455fb]{margin-right:65px!important}.mr-70[data-v-575455fb]{margin-right:70px!important}.mr-75[data-v-575455fb]{margin-right:75px!important}.mr-80[data-v-575455fb]{margin-right:80px!important}.mr-85[data-v-575455fb]{margin-right:85px!important}.mr-90[data-v-575455fb]{margin-right:90px!important}.mr-95[data-v-575455fb]{margin-right:95px!important}.mr-100[data-v-575455fb]{margin-right:100px!important}.pt-0[data-v-575455fb]{padding-top:0!important}.pt-5[data-v-575455fb]{padding-top:5px!important}.pt-10[data-v-575455fb]{padding-top:10px!important}.pt-15[data-v-575455fb]{padding-top:15px!important}.pt-20[data-v-575455fb]{padding-top:20px!important}.pt-25[data-v-575455fb]{padding-top:25px!important}.pt-30[data-v-575455fb]{padding-top:30px!important}.pt-35[data-v-575455fb]{padding-top:35px!important}.pt-40[data-v-575455fb]{padding-top:40px!important}.pt-45[data-v-575455fb]{padding-top:45px!important}.pt-50[data-v-575455fb]{padding-top:50px!important}.pt-55[data-v-575455fb]{padding-top:55px!important}.pt-60[data-v-575455fb]{padding-top:60px!important}.pt-65[data-v-575455fb]{padding-top:65px!important}.pt-70[data-v-575455fb]{padding-top:70px!important}.pt-75[data-v-575455fb]{padding-top:75px!important}.pt-80[data-v-575455fb]{padding-top:80px!important}.pt-85[data-v-575455fb]{padding-top:85px!important}.pt-90[data-v-575455fb]{padding-top:90px!important}.pt-95[data-v-575455fb]{padding-top:95px!important}.pt-100[data-v-575455fb]{padding-top:100px!important}.pb-0[data-v-575455fb]{padding-bottom:0!important}.pb-5[data-v-575455fb]{padding-bottom:5px!important}.pb-10[data-v-575455fb]{padding-bottom:10px!important}.pb-15[data-v-575455fb]{padding-bottom:15px!important}.pb-20[data-v-575455fb]{padding-bottom:20px!important}.pb-25[data-v-575455fb]{padding-bottom:25px!important}.pb-30[data-v-575455fb]{padding-bottom:30px!important}.pb-35[data-v-575455fb]{padding-bottom:35px!important}.pb-40[data-v-575455fb]{padding-bottom:40px!important}.pb-45[data-v-575455fb]{padding-bottom:45px!important}.pb-50[data-v-575455fb]{padding-bottom:50px!important}.pb-55[data-v-575455fb]{padding-bottom:55px!important}.pb-60[data-v-575455fb]{padding-bottom:60px!important}.pb-65[data-v-575455fb]{padding-bottom:65px!important}.pb-70[data-v-575455fb]{padding-bottom:70px!important}.pb-75[data-v-575455fb]{padding-bottom:75px!important}.pb-80[data-v-575455fb]{padding-bottom:80px!important}.pb-85[data-v-575455fb]{padding-bottom:85px!important}.pb-90[data-v-575455fb]{padding-bottom:90px!important}.pb-95[data-v-575455fb]{padding-bottom:95px!important}.pb-100[data-v-575455fb]{padding-bottom:100px!important}.pl-0[data-v-575455fb]{padding-left:0!important}.pl-5[data-v-575455fb]{padding-left:5px!important}.pl-10[data-v-575455fb]{padding-left:10px!important}.pl-15[data-v-575455fb]{padding-left:15px!important}.pl-20[data-v-575455fb]{padding-left:20px!important}.pl-25[data-v-575455fb]{padding-left:25px!important}.pl-30[data-v-575455fb]{padding-left:30px!important}.pl-35[data-v-575455fb]{padding-left:35px!important}.pl-40[data-v-575455fb]{padding-left:40px!important}.pl-45[data-v-575455fb]{padding-left:45px!important}.pl-50[data-v-575455fb]{padding-left:50px!important}.pl-55[data-v-575455fb]{padding-left:55px!important}.pl-60[data-v-575455fb]{padding-left:60px!important}.pl-65[data-v-575455fb]{padding-left:65px!important}.pl-70[data-v-575455fb]{padding-left:70px!important}.pl-75[data-v-575455fb]{padding-left:75px!important}.pl-80[data-v-575455fb]{padding-left:80px!important}.pl-85[data-v-575455fb]{padding-left:85px!important}.pl-90[data-v-575455fb]{padding-left:90px!important}.pl-95[data-v-575455fb]{padding-left:95px!important}.pl-100[data-v-575455fb]{padding-left:100px!important}.pr-0[data-v-575455fb]{padding-right:0!important}.pr-5[data-v-575455fb]{padding-right:5px!important}.pr-10[data-v-575455fb]{padding-right:10px!important}.pr-15[data-v-575455fb]{padding-right:15px!important}.pr-20[data-v-575455fb]{padding-right:20px!important}.pr-25[data-v-575455fb]{padding-right:25px!important}.pr-30[data-v-575455fb]{padding-right:30px!important}.pr-35[data-v-575455fb]{padding-right:35px!important}.pr-40[data-v-575455fb]{padding-right:40px!important}.pr-45[data-v-575455fb]{padding-right:45px!important}.pr-50[data-v-575455fb]{padding-right:50px!important}.pr-55[data-v-575455fb]{padding-right:55px!important}.pr-60[data-v-575455fb]{padding-right:60px!important}.pr-65[data-v-575455fb]{padding-right:65px!important}.pr-70[data-v-575455fb]{padding-right:70px!important}.pr-75[data-v-575455fb]{padding-right:75px!important}.pr-80[data-v-575455fb]{padding-right:80px!important}.pr-85[data-v-575455fb]{padding-right:85px!important}.pr-90[data-v-575455fb]{padding-right:90px!important}.pr-95[data-v-575455fb]{padding-right:95px!important}.pr-100[data-v-575455fb]{padding-right:100px!important}.top-nav[data-v-575455fb]{background-color:#fff}@media screen and (max-width:64.0625em){.top-nav[data-v-575455fb]{margin-top:35px}}@media screen and (max-width:61.9375em){.top-nav .container[data-v-575455fb]{max-width:none}}.top-nav .corporate[data-v-575455fb]{align-items:center;display:flex;flex-wrap:wrap;justify-content:space-between}@media screen and (max-width:47.9375em){.top-nav .corporate[data-v-575455fb]{justify-content:flex-start}}.top-nav .corporate__logo[data-v-575455fb]{order:1}.top-nav .corporate__logo a[data-v-575455fb]:focus{display:block;font-weight:400!important;outline:2px solid #f69038!important;outline-offset:6px!important}@media screen and (max-width:47.9375em){.top-nav .corporate__logo[data-v-575455fb]{margin-right:auto}}.top-nav .corporate__mobile-search[data-v-575455fb]{display:none}@media screen and (max-width:47.9375em){.top-nav .corporate__mobile-search[data-v-575455fb]{display:block;order:2}}.top-nav .corporate__mobile-search-btn[data-v-575455fb]{background-color:#fff;border:none;margin-right:15px;padding:0}@media screen and (max-width:28.125em){.top-nav .corporate__mobile-search-btn[data-v-575455fb]{display:block}}.top-nav .corporate__mobile-search-btn .icon[data-v-575455fb]{font-size:2.5rem}.top-nav .corporate__mobile-search-invisible[data-v-575455fb]{margin-top:3px}.top-nav .corporate__mobile-search-visible[data-v-575455fb]{position:relative}.top-nav .corporate__mobile-search-visible .icon.fas.fa-search[data-v-575455fb]{margin-top:3px}.top-nav .corporate__mobile-search-visible .icon.fas.fa-slash[data-v-575455fb]{left:-5px;position:absolute;top:0;transform:rotate(90deg)}.top-nav .corporate__mobile-menu[data-v-575455fb]{display:none}@media screen and (max-width:59.625em){.top-nav .corporate__mobile-menu[data-v-575455fb]{display:block;order:3}}.top-nav .corporate__mobile-menu button[data-v-575455fb]{background-color:#fff;border:none;padding:0}.top-nav .corporate__mobile-menu button .icon[data-v-575455fb]{font-size:2.5rem}.top-nav .corporate__search[data-v-575455fb]{align-items:strech;background:#fff;border-radius:3px;display:flex;flex:1;margin:0 25px;order:2;position:relative;z-index:99999}@media screen and (max-width:47.9375em){.top-nav .corporate__search[data-v-575455fb]{flex-basis:100%;margin:5px 0 10px;order:4}}.top-nav .corporate__search-form[data-v-575455fb]{display:flex;flex:1;position:relative}.top-nav .corporate__search-select[data-v-575455fb]{-moz-appearance:none;appearance:none;-webkit-appearance:none;background-color:#fff;background-image:linear-gradient(45deg,transparent 50%,#5d6581 0),linear-gradient(135deg,#5d6581 50%,transparent 0);background-position:calc(100% - 8px) calc(1em - 1px),calc(100% - 3px) calc(1em - 1px),100% 0;background-repeat:no-repeat;background-size:5px 5px,5px 5px,2.5em 2.5em;border-radius:3px 0 0 3px!important;-webkit-border-radius:0;border-right:1px solid transparent;padding:0 20px 0 5px}.top-nav .corporate__search-select[data-v-575455fb]:focus{background-color:rgba(48,136,223,.071);border:1px solid #3088df;outline:none}.top-nav .corporate__search-input[data-v-575455fb]{font-size:1.2rem;margin-left:auto;transition:.3s;width:100%}.top-nav .corporate__search-input[data-v-575455fb]:focus{outline:none}.top-nav .corporate__search-box[data-v-575455fb]{background-color:#fff;border:1px solid #b1b4bf;border-radius:3px;box-shadow:0 4px 10px 0 rgba(0,0,0,.2);display:block;position:absolute;top:32px;width:130%;z-index:99999}@media screen and (max-width:59.625em){.top-nav .corporate__search-box[data-v-575455fb]{width:100%}}.top-nav .corporate__search-main-link[data-v-575455fb]{color:#1a254c;display:block;font-size:1.6rem;padding:7px 10px;-webkit-text-decoration:none;text-decoration:none;transition:all .3s;width:100%;word-break:break-word}.top-nav .corporate__search-main-link[data-v-575455fb]:hover{background-color:#e3edfe;text-indent:3px}.top-nav .corporate__search-results[data-v-575455fb]{background-color:#fff;border-top:1px solid #1a254c;list-style:none;margin:0;max-height:410px;overflow-y:scroll;padding:0}.top-nav .corporate__search-heading-search[data-v-575455fb]{color:gray;font-size:1.2rem;font-weight:700;margin-bottom:0;padding:0 10px}.top-nav .corporate__search-list[data-v-575455fb]{display:flex;padding:7px 10px;transition:all .3s}.top-nav .corporate__search-list[data-v-575455fb]:hover{background-color:#e8ecee}.top-nav .corporate__search-list .icon[data-v-575455fb]{color:#1a254c;font-size:1rem;margin-right:5px;margin-top:2px}.top-nav .corporate__search-link[data-v-575455fb]{color:#1a254c;display:inline-block}.top-nav .corporate__search-link[data-v-575455fb],.top-nav .corporate__search-related-link[data-v-575455fb]{cursor:pointer;overflow:hidden;text-overflow:ellipsis;white-space:nowrap;width:-moz-fit-content;width:fit-content}.top-nav .corporate__search-related-link[data-v-575455fb]{color:#3088df;font-size:1.2rem;line-height:17px;margin-left:5px}.top-nav .corporate__search-related-link[data-v-575455fb]:hover{-webkit-text-decoration:underline;text-decoration:underline}.top-nav .corporate__search-btn[data-v-575455fb]{border-radius:0 3px 3px 0;margin:0;padding:6px 25px}.top-nav .corporate__search-btn[data-v-575455fb]:focus{outline:2px solid #f69038!important;outline-offset:6px!important}.top-nav .corporate__nav[data-v-575455fb]{order:3}@media screen and (max-width:59.625em){.top-nav .corporate__nav[data-v-575455fb]{display:none}}.top-nav .corporate__links[data-v-575455fb]{display:flex;flex-wrap:wrap;padding:0}.top-nav .corporate__link-item[data-v-575455fb]{list-style:none;outline:none;padding:10px 0;position:relative}.top-nav .corporate__link-item[data-v-575455fb]:not(:last-child){align-self:center;margin-right:15px}.top-nav .corporate__link-item[data-v-575455fb]:focus-within,.top-nav .corporate__link-item[data-v-575455fb]:hover{background-color:hsla(0,0%,100%,.2);transition:all .2s ease}.top-nav .corporate__link-item:focus-within .corporate__link[data-v-575455fb],.top-nav .corporate__link-item:hover .corporate__link[data-v-575455fb]{color:#1e70c2}.top-nav .corporate__link-item:focus-within .corporate__link-submenu[data-v-575455fb],.top-nav .corporate__link-item:hover .corporate__link-submenu[data-v-575455fb]{visibility:visible}.top-nav .corporate__link-item:first-child:hover span[data-v-575455fb]{color:#1e70c2;transition:all .2s ease}.top-nav .corporate__link-item:focus-within .icon[data-v-575455fb],.top-nav .corporate__link-item:hover .icon[data-v-575455fb]{color:#1e70c2;transform:rotate(180deg);transition:transform .2s ease}.top-nav .corporate__link-item .icon[data-v-575455fb]{color:#1a254c;transition:transform .2s ease}.top-nav .corporate__link-item .icon[data-v-575455fb]:hover{cursor:pointer}.top-nav .corporate__link[data-v-575455fb]{color:#1a254c;cursor:pointer;font-size:1.6rem;-webkit-text-decoration:none;text-decoration:none;transition:all .2s ease}.top-nav .corporate__link-submenu[data-v-575455fb]{background-color:#fff;box-shadow:0 0 10px 0 rgba(0,0,0,.2);display:block;padding:0;position:absolute;top:35px;visibility:hidden;white-space:nowrap;z-index:300}.top-nav .corporate__link-submenu[data-v-575455fb]:before{border:7px solid transparent;border-bottom-color:#fff;content:"";height:0;left:50%;position:absolute;top:-14px;transform:translateX(-50%);width:14px}.top-nav .corporate__link-submenu li[data-v-575455fb]{list-style:none}.top-nav .corporate__link-submenu li a[data-v-575455fb]{color:#1a254c;display:block;padding:10px 20px;-webkit-text-decoration:none!important;text-decoration:none!important}.top-nav .corporate__link-submenu li a[data-v-575455fb]:focus-within,.top-nav .corporate__link-submenu li a[data-v-575455fb]:hover{background-color:#eceff9;font-weight:400}.top-nav .corporate__link-submenu--my-roles[data-v-575455fb]:focus{outline:2px solid #f69038!important;outline-offset:6px!important}.top-nav .corporate__link-item--logged-in[data-v-575455fb]{list-style:none}.top-nav .corporate__logged-in-link[data-v-575455fb]{cursor:pointer;font-weight:400;-webkit-text-decoration:none;text-decoration:none}.top-nav .corporate__user-account[data-v-575455fb]{align-items:center;display:flex;flex-wrap:wrap;justify-content:space-between}.top-nav .corporate__user-info[data-v-575455fb]{margin-left:5px;margin-right:5px}.top-nav .corporate__user-info p[data-v-575455fb]{color:#1a254c;line-height:10px}.top-nav .corporate__user-info small[data-v-575455fb]{color:#1a254c;opacity:.7}.top-nav .corporate__user-img[data-v-575455fb]{border-radius:50px;overflow:hidden}.top-nav .corporate__user-img img[data-v-575455fb]{height:auto;width:35px}.top-nav .corporate__user-details .icon[data-v-575455fb]{color:#1a254c!important;font-size:1.4rem}.top-nav .corporate__link-submenu--user[data-v-575455fb]{min-width:180px;padding:0;top:50px}.top-nav .corporate__link-submenu--user p[data-v-575455fb]{font-weight:700;padding:0 20px}.bottom-nav-1[data-v-575455fb]{background-color:#1a254c}@media screen and (max-width:59.625em){.bottom-nav-1[data-v-575455fb]{display:none}}@media screen and (max-width:47.9375em){.bottom-nav-1 .container[data-v-575455fb]{margin-left:0;margin-right:0;max-width:none;padding:0}}@media screen and (max-width:61.9375em){.bottom-nav-1 .container[data-v-575455fb]{max-width:none}}@media screen and (max-width:59.6875em){.bottom-nav-2[data-v-575455fb]{background-color:#1a254c;padding:0}}@media screen and (min-width:59.6875em){.bottom-nav-2[data-v-575455fb]{display:none}}.bottom-nav-2 .journal ul[data-v-575455fb]{margin:0;padding:0;width:100%}.bottom-nav-2 .journal__link-item--journals[data-v-575455fb]{width:100%}@media screen and (max-width:37.5em){.bottom-nav-2 .journal__link-submenu-journals li a[data-v-575455fb]{padding:10px}}@media screen and (max-width:26.9375em){.bottom-nav-2 .journal__link-submenu-journals li a[data-v-575455fb]{padding:10px 10px 15px}}@media screen and (max-width:37.5em){.bottom-nav-2 .journal__link--home[data-v-575455fb]{padding:10px}}.journal[data-v-575455fb]{align-items:center;display:flex;flex-wrap:wrap;justify-content:flex-start}.journal__nav[data-v-575455fb]{background-color:#1a254c;display:flex}.journal__links[data-v-575455fb]{display:flex;flex-wrap:wrap;margin:0;padding:0}.journal__link-item[data-v-575455fb]{cursor:pointer;list-style:none;outline:none;padding:16px 0;position:relative;transition:all .2s ease}@media screen and (max-width:45.9375em){.journal__link-item[data-v-575455fb]{width:100%}}.journal__link-item:focus-within .journal__link[data-v-575455fb],.journal__link-item:hover .journal__link[data-v-575455fb]{color:#b3b3b3;transition:all .2s ease}.journal__link-item:focus-within .icon.fa-caret-down[data-v-575455fb],.journal__link-item:hover .icon.fa-caret-down[data-v-575455fb]{color:#b3b3b3;transform:rotate(180deg);transition:all .2s ease}.journal__link-item:focus-within .journal__link-submenu[data-v-575455fb],.journal__link-item:hover .journal__link-submenu[data-v-575455fb]{visibility:visible}.journal__link-item .icon[data-v-575455fb]{color:#fff;font-size:14px}.journal__link-item--journals[data-v-575455fb]{padding:0}.journal__journals-list[data-v-575455fb]{outline:none}.journal__journals-list:focus .icon[data-v-575455fb]{border:1px solid #fff;transform:rotate(1turn);transition:.5s}.journal__journals-list a[data-v-575455fb]{padding-right:10px}.journal__journals-list .icon[data-v-575455fb]{border-left:1px solid #fff;font-size:1.6rem}.journal__journals-list .icon.fas.fa-arrow-down[data-v-575455fb]{padding:16px}@media screen and (max-width:37.5em){.journal__journals-list .icon.fas.fa-arrow-down[data-v-575455fb]{padding:12px}}.journal__journals-list .icon.fas.fa-times[data-v-575455fb]{padding:16px 17.5px}@media screen and (max-width:37.5em){.journal__journals-list .icon.fas.fa-times[data-v-575455fb]{padding:12px 13.5px}}.journal__link-item-container[data-v-575455fb]{align-items:center;display:flex;justify-content:space-between}.journal__link[data-v-575455fb]{color:#fff;cursor:pointer;font-size:1.6rem;-webkit-text-decoration:none;text-decoration:none;transition:all .2s ease}@media screen and (max-width:45.9375em){.journal__link[data-v-575455fb]{padding-left:30px}}.journal__link--home[data-v-575455fb]{color:#fff;cursor:pointer;font-size:1.6rem;font-weight:400!important;padding:16px;-webkit-text-decoration:none;text-decoration:none;transition:.3s;width:100%}.journal__link--home[data-v-575455fb]:focus{text-indent:4px}.journal__link-submenu[data-v-575455fb]{background:#fff;box-shadow:0 0 10px 0 rgba(0,0,0,.2);display:block;padding:0;position:absolute;top:42px;visibility:hidden;white-space:nowrap;z-index:300}.journal__link-submenu[data-v-575455fb]:before{border:7px solid transparent;border-bottom-color:#fff;content:"";height:0;left:50%;position:absolute;top:-14px;transform:translateX(-50%);width:14px}@media screen and (max-width:45.9375em){.journal__link-submenu[data-v-575455fb]:before{display:none}.journal__link-submenu[data-v-575455fb]{position:static}}.journal__link-submenu li[data-v-575455fb]{list-style:none}.journal__link-submenu li a[data-v-575455fb]{color:#1a254c;display:block;padding:10px 20px;-webkit-text-decoration:none!important;text-decoration:none!important}.journal__link-submenu li a[data-v-575455fb]:focus,.journal__link-submenu li a[data-v-575455fb]:hover{background-color:#eceff9;font-weight:400}@media screen and (max-width:45.9375em){.journal__link-submenu li a[data-v-575455fb]{padding:10px 40px}}.journal__link-submenu-journals[data-v-575455fb]{background:#1e70c2;box-shadow:0 4px 10px 0 rgba(0,0,0,.2);display:none;height:400px;left:0;overflow-y:scroll;padding:0;position:absolute;top:51px;white-space:nowrap;width:437px;z-index:300}@media screen and (max-width:45.9375em){.journal__link-submenu-journals[data-v-575455fb]{position:static}}.journal__link-submenu-journals li[data-v-575455fb]{border-top:1px solid #fff;list-style:none;transition:all .3s ease}.journal__link-submenu-journals li[data-v-575455fb]:focus,.journal__link-submenu-journals li[data-v-575455fb]:hover{background-color:#2d77c6;font-weight:400;text-indent:8px}.journal__link-submenu-journals li a[data-v-575455fb]{color:#fff;display:flex;flex-wrap:wrap;font-size:1.6rem;justify-content:space-between;padding:10px 10px 10px 15px;position:relative;-webkit-text-decoration:none!important;text-decoration:none!important}@media screen and (max-width:22.1875em){.journal__link-submenu-journals li a[data-v-575455fb]{font-size:1.4rem}}.journal__link-submenu-journals li a span.articles-number[data-v-575455fb]{align-self:center;font-size:1rem}.journal__link-item--journal-info[data-v-575455fb]{margin-left:29px}.journal__link-item--browse[data-v-575455fb],.journal__link-item--journal-info[data-v-575455fb]{margin-right:29px}.journal__link-item--journals[data-v-575455fb]{background-color:#1e70c2;width:437px}.journal__link-submenu--journal-info[data-v-575455fb]{left:-50px}.journal__link-submenu--browse[data-v-575455fb]{left:-30px}.journal__link-submenu--select[data-v-575455fb]{align-items:center;display:flex;padding:10px 20px}@media screen and (max-width:45.9375em){.journal__link-submenu--select[data-v-575455fb]{padding:10px 40px!important}}.journal__link-submenu--select label[data-v-575455fb]{margin-right:10px}.journal__link-submenu--select select[data-v-575455fb]{width:-webkit-fill-available}.journal__link-submenu--select select[data-v-575455fb]:focus{outline:none}.journal__submit-article[data-v-575455fb]{font-size:1.6rem!important;-webkit-text-decoration:none!important;text-decoration:none!important}.journal__submit-article[data-v-575455fb]:focus{font-weight:400;outline:2px solid #f69038!important;outline-offset:6px!important}@media screen and (max-width:45.9375em){.journal__submit-article[data-v-575455fb]{margin-left:30px;padding:12px}}.mobile-nav[data-v-575455fb]{background-color:#1a254c}@media screen and (max-width:59.625em){.mobile-nav[data-v-575455fb]{display:block}}.mobile-nav a[data-v-575455fb]{-webkit-text-decoration:none;text-decoration:none}.mobile-nav__links[data-v-575455fb]{display:flex;flex-direction:column;flex-wrap:wrap}.mobile-nav__links a[data-v-575455fb]{border-bottom:1px solid #485170;color:#fff;padding:12px 15px}.mobile-nav__links[data-v-575455fb]:first-child{border-top:1px solid #485170}.mobile-nav__expandable[data-v-575455fb]{align-items:center;justify-content:space-between}.mobile-nav__expandable[data-v-575455fb],.mobile-nav__user-account[data-v-575455fb]{display:flex;flex-wrap:wrap}.mobile-nav__user-img[data-v-575455fb]{border:2px solid #fff;border-radius:50px;margin-right:10px;overflow:hidden}.mobile-nav__user-img img[data-v-575455fb]{height:auto;width:35px}.mobile-nav__user-info small[data-v-575455fb]{opacity:.7}.mobile-nav__user-submenu[data-v-575455fb]{background-color:#313b5e;display:flex;flex-basis:100%;flex-direction:column}.mobile-nav__user-submenu p[data-v-575455fb]{color:#fff;font-weight:700;opacity:.5;padding:0 15px}.mobile-nav__user-submenu a[data-v-575455fb]{border-bottom:none}.mobile-nav__social-media[data-v-575455fb]{padding:12px 15px}.mobile-nav__social-media p[data-v-575455fb]{color:#fff;font-weight:700}input[data-v-575455fb]{border-radius:0!important}.show-journals[data-v-575455fb]{display:block!important}.journal__link-submenu-journals[data-v-575455fb]::-webkit-scrollbar{width:7px}.journal__link-submenu-journals[data-v-575455fb]::-webkit-scrollbar-track{background-color:#fff;border:1px solid rgba(0,0,0,.15)}.journal__link-submenu-journals[data-v-575455fb]::-webkit-scrollbar-thumb{background:#2d77c6;border-radius:20px}.journal__link-submenu-journals[data-v-575455fb]::-webkit-scrollbar-thumb:hover{background:#1e60bc}.corporate__search-results[data-v-575455fb]::-webkit-scrollbar{width:5.5px}.corporate__search-results[data-v-575455fb]::-webkit-scrollbar-track{background-color:#fff;border:1px solid rgba(0,0,0,.15);border-radius:3px}.corporate__search-results[data-v-575455fb]::-webkit-scrollbar-thumb{background:#1e70c2;border-radius:20px}.overlay[data-v-575455fb]{background:rgba(17,26,55,.702);height:100vh;left:0;padding:200vh 200vw;position:fixed;top:0;width:100%;z-index:999}.remove-styling[data-v-575455fb]{border:0!important;padding:0!important} .mt-0{margin-top:0!important}.mt-5{margin-top:5px!important}.mt-10{margin-top:10px!important}.mt-15{margin-top:15px!important}.mt-20{margin-top:20px!important}.mt-25{margin-top:25px!important}.mt-30{margin-top:30px!important}.mt-35{margin-top:35px!important}.mt-40{margin-top:40px!important}.mt-45{margin-top:45px!important}.mt-50{margin-top:50px!important}.mt-55{margin-top:55px!important}.mt-60{margin-top:60px!important}.mt-65{margin-top:65px!important}.mt-70{margin-top:70px!important}.mt-75{margin-top:75px!important}.mt-80{margin-top:80px!important}.mt-85{margin-top:85px!important}.mt-90{margin-top:90px!important}.mt-95{margin-top:95px!important}.mt-100{margin-top:100px!important}.mb-0{margin-bottom:0!important}.mb-5{margin-bottom:5px!important}.mb-10{margin-bottom:10px!important}.mb-15{margin-bottom:15px!important}.mb-20{margin-bottom:20px!important}.mb-25{margin-bottom:25px!important}.mb-30{margin-bottom:30px!important}.mb-35{margin-bottom:35px!important}.mb-40{margin-bottom:40px!important}.mb-45{margin-bottom:45px!important}.mb-50{margin-bottom:50px!important}.mb-55{margin-bottom:55px!important}.mb-60{margin-bottom:60px!important}.mb-65{margin-bottom:65px!important}.mb-70{margin-bottom:70px!important}.mb-75{margin-bottom:75px!important}.mb-80{margin-bottom:80px!important}.mb-85{margin-bottom:85px!important}.mb-90{margin-bottom:90px!important}.mb-95{margin-bottom:95px!important}.mb-100{margin-bottom:100px!important}.ml-0{margin-left:0!important}.ml-5{margin-left:5px!important}.ml-10{margin-left:10px!important}.ml-15{margin-left:15px!important}.ml-20{margin-left:20px!important}.ml-25{margin-left:25px!important}.ml-30{margin-left:30px!important}.ml-35{margin-left:35px!important}.ml-40{margin-left:40px!important}.ml-45{margin-left:45px!important}.ml-50{margin-left:50px!important}.ml-55{margin-left:55px!important}.ml-60{margin-left:60px!important}.ml-65{margin-left:65px!important}.ml-70{margin-left:70px!important}.ml-75{margin-left:75px!important}.ml-80{margin-left:80px!important}.ml-85{margin-left:85px!important}.ml-90{margin-left:90px!important}.ml-95{margin-left:95px!important}.ml-100{margin-left:100px!important}.mr-0{margin-right:0!important}.mr-5{margin-right:5px!important}.mr-10{margin-right:10px!important}.mr-15{margin-right:15px!important}.mr-20{margin-right:20px!important}.mr-25{margin-right:25px!important}.mr-30{margin-right:30px!important}.mr-35{margin-right:35px!important}.mr-40{margin-right:40px!important}.mr-45{margin-right:45px!important}.mr-50{margin-right:50px!important}.mr-55{margin-right:55px!important}.mr-60{margin-right:60px!important}.mr-65{margin-right:65px!important}.mr-70{margin-right:70px!important}.mr-75{margin-right:75px!important}.mr-80{margin-right:80px!important}.mr-85{margin-right:85px!important}.mr-90{margin-right:90px!important}.mr-95{margin-right:95px!important}.mr-100{margin-right:100px!important}.pt-0{padding-top:0!important}.pt-5{padding-top:5px!important}.pt-10{padding-top:10px!important}.pt-15{padding-top:15px!important}.pt-20{padding-top:20px!important}.pt-25{padding-top:25px!important}.pt-30{padding-top:30px!important}.pt-35{padding-top:35px!important}.pt-40{padding-top:40px!important}.pt-45{padding-top:45px!important}.pt-50{padding-top:50px!important}.pt-55{padding-top:55px!important}.pt-60{padding-top:60px!important}.pt-65{padding-top:65px!important}.pt-70{padding-top:70px!important}.pt-75{padding-top:75px!important}.pt-80{padding-top:80px!important}.pt-85{padding-top:85px!important}.pt-90{padding-top:90px!important}.pt-95{padding-top:95px!important}.pt-100{padding-top:100px!important}.pb-0{padding-bottom:0!important}.pb-5{padding-bottom:5px!important}.pb-10{padding-bottom:10px!important}.pb-15{padding-bottom:15px!important}.pb-20{padding-bottom:20px!important}.pb-25{padding-bottom:25px!important}.pb-30{padding-bottom:30px!important}.pb-35{padding-bottom:35px!important}.pb-40{padding-bottom:40px!important}.pb-45{padding-bottom:45px!important}.pb-50{padding-bottom:50px!important}.pb-55{padding-bottom:55px!important}.pb-60{padding-bottom:60px!important}.pb-65{padding-bottom:65px!important}.pb-70{padding-bottom:70px!important}.pb-75{padding-bottom:75px!important}.pb-80{padding-bottom:80px!important}.pb-85{padding-bottom:85px!important}.pb-90{padding-bottom:90px!important}.pb-95{padding-bottom:95px!important}.pb-100{padding-bottom:100px!important}.pl-0{padding-left:0!important}.pl-5{padding-left:5px!important}.pl-10{padding-left:10px!important}.pl-15{padding-left:15px!important}.pl-20{padding-left:20px!important}.pl-25{padding-left:25px!important}.pl-30{padding-left:30px!important}.pl-35{padding-left:35px!important}.pl-40{padding-left:40px!important}.pl-45{padding-left:45px!important}.pl-50{padding-left:50px!important}.pl-55{padding-left:55px!important}.pl-60{padding-left:60px!important}.pl-65{padding-left:65px!important}.pl-70{padding-left:70px!important}.pl-75{padding-left:75px!important}.pl-80{padding-left:80px!important}.pl-85{padding-left:85px!important}.pl-90{padding-left:90px!important}.pl-95{padding-left:95px!important}.pl-100{padding-left:100px!important}.pr-0{padding-right:0!important}.pr-5{padding-right:5px!important}.pr-10{padding-right:10px!important}.pr-15{padding-right:15px!important}.pr-20{padding-right:20px!important}.pr-25{padding-right:25px!important}.pr-30{padding-right:30px!important}.pr-35{padding-right:35px!important}.pr-40{padding-right:40px!important}.pr-45{padding-right:45px!important}.pr-50{padding-right:50px!important}.pr-55{padding-right:55px!important}.pr-60{padding-right:60px!important}.pr-65{padding-right:65px!important}.pr-70{padding-right:70px!important}.pr-75{padding-right:75px!important}.pr-80{padding-right:80px!important}.pr-85{padding-right:85px!important}.pr-90{padding-right:90px!important}.pr-95{padding-right:95px!important}.pr-100{padding-right:100px!important}.logo-img{max-width:340px;width:100%} .mt-0{margin-top:0!important}.mt-5{margin-top:5px!important}.mt-10{margin-top:10px!important}.mt-15{margin-top:15px!important}.mt-20{margin-top:20px!important}.mt-25{margin-top:25px!important}.mt-30{margin-top:30px!important}.mt-35{margin-top:35px!important}.mt-40{margin-top:40px!important}.mt-45{margin-top:45px!important}.mt-50{margin-top:50px!important}.mt-55{margin-top:55px!important}.mt-60{margin-top:60px!important}.mt-65{margin-top:65px!important}.mt-70{margin-top:70px!important}.mt-75{margin-top:75px!important}.mt-80{margin-top:80px!important}.mt-85{margin-top:85px!important}.mt-90{margin-top:90px!important}.mt-95{margin-top:95px!important}.mt-100{margin-top:100px!important}.mb-0{margin-bottom:0!important}.mb-5{margin-bottom:5px!important}.mb-10{margin-bottom:10px!important}.mb-15{margin-bottom:15px!important}.mb-20{margin-bottom:20px!important}.mb-25{margin-bottom:25px!important}.mb-30{margin-bottom:30px!important}.mb-35{margin-bottom:35px!important}.mb-40{margin-bottom:40px!important}.mb-45{margin-bottom:45px!important}.mb-50{margin-bottom:50px!important}.mb-55{margin-bottom:55px!important}.mb-60{margin-bottom:60px!important}.mb-65{margin-bottom:65px!important}.mb-70{margin-bottom:70px!important}.mb-75{margin-bottom:75px!important}.mb-80{margin-bottom:80px!important}.mb-85{margin-bottom:85px!important}.mb-90{margin-bottom:90px!important}.mb-95{margin-bottom:95px!important}.mb-100{margin-bottom:100px!important}.ml-0{margin-left:0!important}.ml-5{margin-left:5px!important}.ml-10{margin-left:10px!important}.ml-15{margin-left:15px!important}.ml-20{margin-left:20px!important}.ml-25{margin-left:25px!important}.ml-30{margin-left:30px!important}.ml-35{margin-left:35px!important}.ml-40{margin-left:40px!important}.ml-45{margin-left:45px!important}.ml-50{margin-left:50px!important}.ml-55{margin-left:55px!important}.ml-60{margin-left:60px!important}.ml-65{margin-left:65px!important}.ml-70{margin-left:70px!important}.ml-75{margin-left:75px!important}.ml-80{margin-left:80px!important}.ml-85{margin-left:85px!important}.ml-90{margin-left:90px!important}.ml-95{margin-left:95px!important}.ml-100{margin-left:100px!important}.mr-0{margin-right:0!important}.mr-5{margin-right:5px!important}.mr-10{margin-right:10px!important}.mr-15{margin-right:15px!important}.mr-20{margin-right:20px!important}.mr-25{margin-right:25px!important}.mr-30{margin-right:30px!important}.mr-35{margin-right:35px!important}.mr-40{margin-right:40px!important}.mr-45{margin-right:45px!important}.mr-50{margin-right:50px!important}.mr-55{margin-right:55px!important}.mr-60{margin-right:60px!important}.mr-65{margin-right:65px!important}.mr-70{margin-right:70px!important}.mr-75{margin-right:75px!important}.mr-80{margin-right:80px!important}.mr-85{margin-right:85px!important}.mr-90{margin-right:90px!important}.mr-95{margin-right:95px!important}.mr-100{margin-right:100px!important}.pt-0{padding-top:0!important}.pt-5{padding-top:5px!important}.pt-10{padding-top:10px!important}.pt-15{padding-top:15px!important}.pt-20{padding-top:20px!important}.pt-25{padding-top:25px!important}.pt-30{padding-top:30px!important}.pt-35{padding-top:35px!important}.pt-40{padding-top:40px!important}.pt-45{padding-top:45px!important}.pt-50{padding-top:50px!important}.pt-55{padding-top:55px!important}.pt-60{padding-top:60px!important}.pt-65{padding-top:65px!important}.pt-70{padding-top:70px!important}.pt-75{padding-top:75px!important}.pt-80{padding-top:80px!important}.pt-85{padding-top:85px!important}.pt-90{padding-top:90px!important}.pt-95{padding-top:95px!important}.pt-100{padding-top:100px!important}.pb-0{padding-bottom:0!important}.pb-5{padding-bottom:5px!important}.pb-10{padding-bottom:10px!important}.pb-15{padding-bottom:15px!important}.pb-20{padding-bottom:20px!important}.pb-25{padding-bottom:25px!important}.pb-30{padding-bottom:30px!important}.pb-35{padding-bottom:35px!important}.pb-40{padding-bottom:40px!important}.pb-45{padding-bottom:45px!important}.pb-50{padding-bottom:50px!important}.pb-55{padding-bottom:55px!important}.pb-60{padding-bottom:60px!important}.pb-65{padding-bottom:65px!important}.pb-70{padding-bottom:70px!important}.pb-75{padding-bottom:75px!important}.pb-80{padding-bottom:80px!important}.pb-85{padding-bottom:85px!important}.pb-90{padding-bottom:90px!important}.pb-95{padding-bottom:95px!important}.pb-100{padding-bottom:100px!important}.pl-0{padding-left:0!important}.pl-5{padding-left:5px!important}.pl-10{padding-left:10px!important}.pl-15{padding-left:15px!important}.pl-20{padding-left:20px!important}.pl-25{padding-left:25px!important}.pl-30{padding-left:30px!important}.pl-35{padding-left:35px!important}.pl-40{padding-left:40px!important}.pl-45{padding-left:45px!important}.pl-50{padding-left:50px!important}.pl-55{padding-left:55px!important}.pl-60{padding-left:60px!important}.pl-65{padding-left:65px!important}.pl-70{padding-left:70px!important}.pl-75{padding-left:75px!important}.pl-80{padding-left:80px!important}.pl-85{padding-left:85px!important}.pl-90{padding-left:90px!important}.pl-95{padding-left:95px!important}.pl-100{padding-left:100px!important}.pr-0{padding-right:0!important}.pr-5{padding-right:5px!important}.pr-10{padding-right:10px!important}.pr-15{padding-right:15px!important}.pr-20{padding-right:20px!important}.pr-25{padding-right:25px!important}.pr-30{padding-right:30px!important}.pr-35{padding-right:35px!important}.pr-40{padding-right:40px!important}.pr-45{padding-right:45px!important}.pr-50{padding-right:50px!important}.pr-55{padding-right:55px!important}.pr-60{padding-right:60px!important}.pr-65{padding-right:65px!important}.pr-70{padding-right:70px!important}.pr-75{padding-right:75px!important}.pr-80{padding-right:80px!important}.pr-85{padding-right:85px!important}.pr-90{padding-right:90px!important}.pr-95{padding-right:95px!important}.pr-100{padding-right:100px!important}.tabs{border-bottom:2px solid #939ab1;display:flex;justify-content:space-evenly;margin:40px 0 30px}@media screen and (max-width:33.1875em){.tabs{text-align:center}}.tabs a{border-bottom:2px solid #939ab1;color:#1a254c;font-size:1.6rem;font-weight:700;margin-bottom:-2px;padding:10px 10px 5px}.tabs a:focus,.tabs a:hover{background-color:#b8d6f4;border-bottom:2px solid #1e70c2;outline:none;-webkit-text-decoration:none;text-decoration:none}.tabs .nuxt-link-exact-active{border-bottom:2px solid #1e70c2;color:#1e70c2}.tabs-metrics,.tabs-tweetations{align-items:baseline;display:flex;flex-wrap:wrap;justify-content:flex-start;margin-bottom:20px}.tabs-metrics a,.tabs-tweetations a{color:#1a254c;cursor:pointer;padding:10px 0 5px;text-align:center}.tabs-metrics a:focus,.tabs-metrics a:hover,.tabs-tweetations a:focus,.tabs-tweetations a:hover{color:#1e70c2}.tabs-metrics span,.tabs-tweetations span{margin:0 10px}.active{border-bottom:2px solid #1e70c2!important;color:#1e70c2!important}.mobile-show{display:none}@media screen and (max-width:47.9375em){.mobile-show{display:block}}.desktop-show{display:block}@media screen and (max-width:47.9375em){.desktop-show{display:none}}.advertisement{margin-top:25px}.advertisement:focus .advertisement__text,.advertisement:hover .advertisement__text{color:#1e70c2}.advertisement__link{background-color:#f1f3f5;border-radius:3px;display:flex;font-weight:700;justify-content:space-between;padding:20px;-webkit-text-decoration:none;text-decoration:none}@media screen and (max-width:33.1875em){.advertisement__link{flex-wrap:wrap}}.advertisement__text{color:#000;font-size:1.8rem;padding-right:10px}.advertisement__button{align-self:center;background-color:#1e70c2;border-radius:3px;color:#fff;font-size:2rem;padding:15px 20px;position:relative;transition:.3s;width:200px}@media screen and (max-width:74.9375em){.advertisement__button{width:150px}}@media screen and (max-width:47.9375em){.advertisement__button{width:200px}}@media screen and (max-width:33.1875em){.advertisement__button{margin-left:auto;margin-top:10px;padding:10px 20px;width:170px}}.advertisement__button .advertisement__icon{font-size:1.7rem;position:absolute;right:2.2rem;top:2rem;transition:.3s}@media screen and (max-width:33.1875em){.advertisement__button .advertisement__icon{top:1.5rem}}.advertisement__button:focus,.advertisement__button:hover{opacity:.9}.advertisement__button:focus .advertisement__icon,.advertisement__button:hover .advertisement__icon{right:1.5rem}.authors-for-screen-reader{height:1px;left:-10000px;overflow:hidden;position:absolute;top:auto;width:1px}.authors-for-screen-reader:focus{display:block;font-weight:700;height:auto;outline:2px solid #f69038!important;outline-offset:6px!important;overflow:hidden;position:static;width:auto}.main .details{background-color:#f1f3f5;border-radius:3px;margin-top:25px;padding:10px}.main .details p{font-size:1.6rem;font-weight:700;margin:0}.main .preprints-version{align-items:baseline;display:flex;margin-top:10px}.info{display:flex;flex-wrap:wrap}.info__article-img{cursor:pointer;flex:0 0 22%;height:-moz-fit-content;height:fit-content;margin-right:20px;position:relative}@media screen and (max-width:61.9375em){.info__article-img{display:none}}.info__article-img img{display:block;width:100%}.info__article-img-info{background-color:rgba(26,37,76,.749);bottom:0;color:#fff;height:0;left:0;margin:0;overflow:hidden;position:absolute;right:0;text-align:center;transition:.5s ease;width:100%}.info__article-img-info .icon{font-size:3rem;margin-top:30%}.info__article-img:hover .info__article-img-info{height:100%}.info__title-authors{flex:0 0 75%}@media screen and (max-width:61.9375em){.info__title-authors{flex:0 0 100%}}.info__title-authors h3:focus{outline:2px solid #f69038!important;outline-offset:6px!important}.info__hidden-title{height:1px;left:-10000px;overflow:hidden;position:absolute;top:auto;width:1px}.info__authors{display:inline-block}.info__orcid-img{height:15px}.tabs a{flex-grow:1;text-align:center}@media screen and (max-width:47.9375em){.tabs{flex-direction:column}}.sidebar-citation .export-metadata div{margin-bottom:10px}.sidebar-citation .collection h4:focus{outline:2px solid #f69038!important;outline-offset:6px!important}.sidebar-citation .collection__span{display:block}@media screen and (max-width:61.9375em){.sidebar-citation .collection__span{display:inline-block}}.sidebar-citation .collection__link{background:#1e70c2;border-radius:3px;color:#fff;display:block;font-size:1.2rem;margin-bottom:5px;padding:5px 10px;width:-moz-fit-content;width:fit-content}.sidebar-citation .download-btns{display:flex;flex-wrap:wrap;justify-content:space-between;margin-bottom:20px}@media screen and (max-width:61.9375em){.sidebar-citation .download-btns{justify-content:flex-start}.sidebar-citation .download-btns a{margin-right:10px}}.article-content h3{font-size:2.2rem;line-height:2.4rem}.article-content ol li,.article-content ul li{margin-bottom:10px}.main-article-content a{color:#1e70c2}.main-article-content a:hover{-webkit-text-decoration:underline;text-decoration:underline}.author-affiliation-details,.authors-container .authors.clearfix .clearfix,.corresponding-author-and-affiliations,.h4-original-paper{display:none}.authors .clearfix{display:flex;flex-wrap:wrap;list-style:none;padding:0}.authors .clearfix li{margin-right:10px}.authors .clearfix li a{color:#1e70c2}.article-content figure{background-color:#f1f3f5;border-radius:3px;margin:0;padding:20px}#Abstract{margin-top:10px}#Abstract,#Discussion,#Introduction,#Keywords,#Methods,#Results{border-bottom:1px solid #ccd1d5}.abstract-sub-heading{display:block;font-size:1.6rem;font-weight:700}.figure-table{background:#f1f3f5;border-radius:3px;height:auto;margin-bottom:20px;overflow-x:auto;overflow-y:hidden;padding:20px 20px 0}.figure-table::-webkit-scrollbar{-webkit-appearance:none}.figure-table::-webkit-scrollbar-track{border-radius:8px;box-shadow:inset 0 0 5px #5d6581}.figure-table::-webkit-scrollbar:vertical{width:8px}.figure-table::-webkit-scrollbar:horizontal{height:8px}.figure-table::-webkit-scrollbar-thumb{background-color:#1e70c2;border-radius:8px}.textbox-container{border:2px solid #333;padding:15px}.textbox-container h5{margin-bottom:0;margin-top:0}.footnotes ol{word-wrap:break-word}img.figure-image,img.graphic-image,img.inline-graphic-image{width:100%}table{margin:0}td,th{padding:5px 0 5px 15px}.careers{margin-top:25px}.career-widget{border-bottom:1px solid #ccc;cursor:pointer;padding:5px 20px 20px;transition:box-shadow .3s ease}.career-widget:hover{box-shadow:0 0 10px 0 rgba(0,0,0,.1)}.job,.job:hover{color:inherit;-webkit-text-decoration:none;text-decoration:none}.job:hover .job-title{-webkit-text-decoration:underline;text-decoration:underline}.job-title{font-weight:700;margin-top:20px} .mt-0{margin-top:0!important}.mt-5{margin-top:5px!important}.mt-10{margin-top:10px!important}.mt-15{margin-top:15px!important}.mt-20{margin-top:20px!important}.mt-25{margin-top:25px!important}.mt-30{margin-top:30px!important}.mt-35{margin-top:35px!important}.mt-40{margin-top:40px!important}.mt-45{margin-top:45px!important}.mt-50{margin-top:50px!important}.mt-55{margin-top:55px!important}.mt-60{margin-top:60px!important}.mt-65{margin-top:65px!important}.mt-70{margin-top:70px!important}.mt-75{margin-top:75px!important}.mt-80{margin-top:80px!important}.mt-85{margin-top:85px!important}.mt-90{margin-top:90px!important}.mt-95{margin-top:95px!important}.mt-100{margin-top:100px!important}.mb-0{margin-bottom:0!important}.mb-5{margin-bottom:5px!important}.mb-10{margin-bottom:10px!important}.mb-15{margin-bottom:15px!important}.mb-20{margin-bottom:20px!important}.mb-25{margin-bottom:25px!important}.mb-30{margin-bottom:30px!important}.mb-35{margin-bottom:35px!important}.mb-40{margin-bottom:40px!important}.mb-45{margin-bottom:45px!important}.mb-50{margin-bottom:50px!important}.mb-55{margin-bottom:55px!important}.mb-60{margin-bottom:60px!important}.mb-65{margin-bottom:65px!important}.mb-70{margin-bottom:70px!important}.mb-75{margin-bottom:75px!important}.mb-80{margin-bottom:80px!important}.mb-85{margin-bottom:85px!important}.mb-90{margin-bottom:90px!important}.mb-95{margin-bottom:95px!important}.mb-100{margin-bottom:100px!important}.ml-0{margin-left:0!important}.ml-5{margin-left:5px!important}.ml-10{margin-left:10px!important}.ml-15{margin-left:15px!important}.ml-20{margin-left:20px!important}.ml-25{margin-left:25px!important}.ml-30{margin-left:30px!important}.ml-35{margin-left:35px!important}.ml-40{margin-left:40px!important}.ml-45{margin-left:45px!important}.ml-50{margin-left:50px!important}.ml-55{margin-left:55px!important}.ml-60{margin-left:60px!important}.ml-65{margin-left:65px!important}.ml-70{margin-left:70px!important}.ml-75{margin-left:75px!important}.ml-80{margin-left:80px!important}.ml-85{margin-left:85px!important}.ml-90{margin-left:90px!important}.ml-95{margin-left:95px!important}.ml-100{margin-left:100px!important}.mr-0{margin-right:0!important}.mr-5{margin-right:5px!important}.mr-10{margin-right:10px!important}.mr-15{margin-right:15px!important}.mr-20{margin-right:20px!important}.mr-25{margin-right:25px!important}.mr-30{margin-right:30px!important}.mr-35{margin-right:35px!important}.mr-40{margin-right:40px!important}.mr-45{margin-right:45px!important}.mr-50{margin-right:50px!important}.mr-55{margin-right:55px!important}.mr-60{margin-right:60px!important}.mr-65{margin-right:65px!important}.mr-70{margin-right:70px!important}.mr-75{margin-right:75px!important}.mr-80{margin-right:80px!important}.mr-85{margin-right:85px!important}.mr-90{margin-right:90px!important}.mr-95{margin-right:95px!important}.mr-100{margin-right:100px!important}.pt-0{padding-top:0!important}.pt-5{padding-top:5px!important}.pt-10{padding-top:10px!important}.pt-15{padding-top:15px!important}.pt-20{padding-top:20px!important}.pt-25{padding-top:25px!important}.pt-30{padding-top:30px!important}.pt-35{padding-top:35px!important}.pt-40{padding-top:40px!important}.pt-45{padding-top:45px!important}.pt-50{padding-top:50px!important}.pt-55{padding-top:55px!important}.pt-60{padding-top:60px!important}.pt-65{padding-top:65px!important}.pt-70{padding-top:70px!important}.pt-75{padding-top:75px!important}.pt-80{padding-top:80px!important}.pt-85{padding-top:85px!important}.pt-90{padding-top:90px!important}.pt-95{padding-top:95px!important}.pt-100{padding-top:100px!important}.pb-0{padding-bottom:0!important}.pb-5{padding-bottom:5px!important}.pb-10{padding-bottom:10px!important}.pb-15{padding-bottom:15px!important}.pb-20{padding-bottom:20px!important}.pb-25{padding-bottom:25px!important}.pb-30{padding-bottom:30px!important}.pb-35{padding-bottom:35px!important}.pb-40{padding-bottom:40px!important}.pb-45{padding-bottom:45px!important}.pb-50{padding-bottom:50px!important}.pb-55{padding-bottom:55px!important}.pb-60{padding-bottom:60px!important}.pb-65{padding-bottom:65px!important}.pb-70{padding-bottom:70px!important}.pb-75{padding-bottom:75px!important}.pb-80{padding-bottom:80px!important}.pb-85{padding-bottom:85px!important}.pb-90{padding-bottom:90px!important}.pb-95{padding-bottom:95px!important}.pb-100{padding-bottom:100px!important}.pl-0{padding-left:0!important}.pl-5{padding-left:5px!important}.pl-10{padding-left:10px!important}.pl-15{padding-left:15px!important}.pl-20{padding-left:20px!important}.pl-25{padding-left:25px!important}.pl-30{padding-left:30px!important}.pl-35{padding-left:35px!important}.pl-40{padding-left:40px!important}.pl-45{padding-left:45px!important}.pl-50{padding-left:50px!important}.pl-55{padding-left:55px!important}.pl-60{padding-left:60px!important}.pl-65{padding-left:65px!important}.pl-70{padding-left:70px!important}.pl-75{padding-left:75px!important}.pl-80{padding-left:80px!important}.pl-85{padding-left:85px!important}.pl-90{padding-left:90px!important}.pl-95{padding-left:95px!important}.pl-100{padding-left:100px!important}.pr-0{padding-right:0!important}.pr-5{padding-right:5px!important}.pr-10{padding-right:10px!important}.pr-15{padding-right:15px!important}.pr-20{padding-right:20px!important}.pr-25{padding-right:25px!important}.pr-30{padding-right:30px!important}.pr-35{padding-right:35px!important}.pr-40{padding-right:40px!important}.pr-45{padding-right:45px!important}.pr-50{padding-right:50px!important}.pr-55{padding-right:55px!important}.pr-60{padding-right:60px!important}.pr-65{padding-right:65px!important}.pr-70{padding-right:70px!important}.pr-75{padding-right:75px!important}.pr-80{padding-right:80px!important}.pr-85{padding-right:85px!important}.pr-90{padding-right:90px!important}.pr-95{padding-right:95px!important}.pr-100{padding-right:100px!important}@media screen and (max-width:61.9375em){.sidebar-sections{display:none}}.sidebar-sections ul{color:#1a254c;padding:13px 10px 0 15px}.sidebar-sections ul li{margin-bottom:1rem}.sidebar-sections ul li a{color:#1a254c;font-weight:400!important;word-break:break-word}.sidebar-sections ul li a:hover{color:#1e70c2}.sidebar-sections ul li a:focus{-webkit-text-decoration:none!important;text-decoration:none!important}.sidebar-sections ul li:hover .active{-webkit-text-decoration:none;text-decoration:none}.sidebar-sections ul li .active{font-weight:400!important;-webkit-text-decoration:none;text-decoration:none}.sidebar-nav{height:100%;left:0;position:absolute;top:0;width:100%}.sidebar-nav-sticky{position:sticky;top:0}.article{padding:0}.tooltip{cursor:pointer;position:relative}.tooltip .tooltiptext{background-color:#000;border-radius:6px;color:#fff;font-family:"Helvetica","Sans-serif";font-size:.9rem;left:50%;margin-left:-125px;margin-top:2px;min-width:250px;padding:10px;position:absolute;text-align:center;text-transform:none;top:100%;visibility:hidden;z-index:1}@media(max-width:768px){.tooltip .tooltiptext{display:none}}.tooltip:hover .tooltiptext{visibility:visible}@media(min-width:768px){.visual-abstract{display:none}}@media(max-width:767px){.visual-abstract{display:block}}.figure-table ul{margin:auto}named-content{display:block;padding-left:1em;white-space:pre-wrap} .mt-0{margin-top:0!important}.mt-5{margin-top:5px!important}.mt-10{margin-top:10px!important}.mt-15{margin-top:15px!important}.mt-20{margin-top:20px!important}.mt-25{margin-top:25px!important}.mt-30{margin-top:30px!important}.mt-35{margin-top:35px!important}.mt-40{margin-top:40px!important}.mt-45{margin-top:45px!important}.mt-50{margin-top:50px!important}.mt-55{margin-top:55px!important}.mt-60{margin-top:60px!important}.mt-65{margin-top:65px!important}.mt-70{margin-top:70px!important}.mt-75{margin-top:75px!important}.mt-80{margin-top:80px!important}.mt-85{margin-top:85px!important}.mt-90{margin-top:90px!important}.mt-95{margin-top:95px!important}.mt-100{margin-top:100px!important}.mb-0{margin-bottom:0!important}.mb-5{margin-bottom:5px!important}.mb-10{margin-bottom:10px!important}.mb-15{margin-bottom:15px!important}.mb-20{margin-bottom:20px!important}.mb-25{margin-bottom:25px!important}.mb-30{margin-bottom:30px!important}.mb-35{margin-bottom:35px!important}.mb-40{margin-bottom:40px!important}.mb-45{margin-bottom:45px!important}.mb-50{margin-bottom:50px!important}.mb-55{margin-bottom:55px!important}.mb-60{margin-bottom:60px!important}.mb-65{margin-bottom:65px!important}.mb-70{margin-bottom:70px!important}.mb-75{margin-bottom:75px!important}.mb-80{margin-bottom:80px!important}.mb-85{margin-bottom:85px!important}.mb-90{margin-bottom:90px!important}.mb-95{margin-bottom:95px!important}.mb-100{margin-bottom:100px!important}.ml-0{margin-left:0!important}.ml-5{margin-left:5px!important}.ml-10{margin-left:10px!important}.ml-15{margin-left:15px!important}.ml-20{margin-left:20px!important}.ml-25{margin-left:25px!important}.ml-30{margin-left:30px!important}.ml-35{margin-left:35px!important}.ml-40{margin-left:40px!important}.ml-45{margin-left:45px!important}.ml-50{margin-left:50px!important}.ml-55{margin-left:55px!important}.ml-60{margin-left:60px!important}.ml-65{margin-left:65px!important}.ml-70{margin-left:70px!important}.ml-75{margin-left:75px!important}.ml-80{margin-left:80px!important}.ml-85{margin-left:85px!important}.ml-90{margin-left:90px!important}.ml-95{margin-left:95px!important}.ml-100{margin-left:100px!important}.mr-0{margin-right:0!important}.mr-5{margin-right:5px!important}.mr-10{margin-right:10px!important}.mr-15{margin-right:15px!important}.mr-20{margin-right:20px!important}.mr-25{margin-right:25px!important}.mr-30{margin-right:30px!important}.mr-35{margin-right:35px!important}.mr-40{margin-right:40px!important}.mr-45{margin-right:45px!important}.mr-50{margin-right:50px!important}.mr-55{margin-right:55px!important}.mr-60{margin-right:60px!important}.mr-65{margin-right:65px!important}.mr-70{margin-right:70px!important}.mr-75{margin-right:75px!important}.mr-80{margin-right:80px!important}.mr-85{margin-right:85px!important}.mr-90{margin-right:90px!important}.mr-95{margin-right:95px!important}.mr-100{margin-right:100px!important}.pt-0{padding-top:0!important}.pt-5{padding-top:5px!important}.pt-10{padding-top:10px!important}.pt-15{padding-top:15px!important}.pt-20{padding-top:20px!important}.pt-25{padding-top:25px!important}.pt-30{padding-top:30px!important}.pt-35{padding-top:35px!important}.pt-40{padding-top:40px!important}.pt-45{padding-top:45px!important}.pt-50{padding-top:50px!important}.pt-55{padding-top:55px!important}.pt-60{padding-top:60px!important}.pt-65{padding-top:65px!important}.pt-70{padding-top:70px!important}.pt-75{padding-top:75px!important}.pt-80{padding-top:80px!important}.pt-85{padding-top:85px!important}.pt-90{padding-top:90px!important}.pt-95{padding-top:95px!important}.pt-100{padding-top:100px!important}.pb-0{padding-bottom:0!important}.pb-5{padding-bottom:5px!important}.pb-10{padding-bottom:10px!important}.pb-15{padding-bottom:15px!important}.pb-20{padding-bottom:20px!important}.pb-25{padding-bottom:25px!important}.pb-30{padding-bottom:30px!important}.pb-35{padding-bottom:35px!important}.pb-40{padding-bottom:40px!important}.pb-45{padding-bottom:45px!important}.pb-50{padding-bottom:50px!important}.pb-55{padding-bottom:55px!important}.pb-60{padding-bottom:60px!important}.pb-65{padding-bottom:65px!important}.pb-70{padding-bottom:70px!important}.pb-75{padding-bottom:75px!important}.pb-80{padding-bottom:80px!important}.pb-85{padding-bottom:85px!important}.pb-90{padding-bottom:90px!important}.pb-95{padding-bottom:95px!important}.pb-100{padding-bottom:100px!important}.pl-0{padding-left:0!important}.pl-5{padding-left:5px!important}.pl-10{padding-left:10px!important}.pl-15{padding-left:15px!important}.pl-20{padding-left:20px!important}.pl-25{padding-left:25px!important}.pl-30{padding-left:30px!important}.pl-35{padding-left:35px!important}.pl-40{padding-left:40px!important}.pl-45{padding-left:45px!important}.pl-50{padding-left:50px!important}.pl-55{padding-left:55px!important}.pl-60{padding-left:60px!important}.pl-65{padding-left:65px!important}.pl-70{padding-left:70px!important}.pl-75{padding-left:75px!important}.pl-80{padding-left:80px!important}.pl-85{padding-left:85px!important}.pl-90{padding-left:90px!important}.pl-95{padding-left:95px!important}.pl-100{padding-left:100px!important}.pr-0{padding-right:0!important}.pr-5{padding-right:5px!important}.pr-10{padding-right:10px!important}.pr-15{padding-right:15px!important}.pr-20{padding-right:20px!important}.pr-25{padding-right:25px!important}.pr-30{padding-right:30px!important}.pr-35{padding-right:35px!important}.pr-40{padding-right:40px!important}.pr-45{padding-right:45px!important}.pr-50{padding-right:50px!important}.pr-55{padding-right:55px!important}.pr-60{padding-right:60px!important}.pr-65{padding-right:65px!important}.pr-70{padding-right:70px!important}.pr-75{padding-right:75px!important}.pr-80{padding-right:80px!important}.pr-85{padding-right:85px!important}.pr-90{padding-right:90px!important}.pr-95{padding-right:95px!important}.pr-100{padding-right:100px!important}.email-subscribtion-button{margin-left:auto}.email-subscribtion-button h3{margin-top:0}@media screen and (max-width:35.9375em){.email-subscribtion-button h3{margin-top:50px}}.email-subscribtion-button a{background-color:#fff;border:2px solid #fff;border-radius:100px;color:#1a254c;cursor:pointer;display:block;margin:auto;padding:10px;position:relative;text-align:center;transition:all .2s ease;width:205px}.email-subscribtion-button a span{font-size:1.6rem;font-weight:700;position:relative}.email-subscribtion-button a span.icon{position:absolute;right:38px;transition:all .2s ease}.email-subscribtion-button a:before{background-color:#1a254c;border-radius:28px;content:"";display:block;height:38px;left:0;position:absolute;top:0;transition:all .3s ease;width:50px}.email-subscribtion-button a:hover{color:#fff}.email-subscribtion-button a:hover span.icon{right:25px;transform:rotate(45deg)}.email-subscribtion-button a:hover:before{width:100%}.email-subscribtion-button a:active{color:#1a254c}.email-subscribtion-button a:active span.icon{right:20px}.email-subscribtion-button a:active:before{background-color:#fff}.footer-modal-window .modal-window-body{max-height:650px}.footer-modal-window .mc-field-group ul{padding:0}.footer-modal-window .mc-field-group ul li{list-style-type:none;margin-bottom:2px}.footer-modal-window .mc-field-group ul li input{cursor:pointer;height:15px;width:15px}.footer-modal-window .mc-field-group ul li label{font-size:1.6rem}.footer-modal-window .email-subscribtion label{display:block;font-weight:700;margin:5px 0;max-width:-moz-fit-content;max-width:fit-content}.footer-modal-window .email-subscribtion input{width:100%}.footer-modal-window .email-subscribtion input.error{background-color:#fcf1f1!important;border:1px solid red!important}.footer-modal-window .email-subscribtion small.error{color:red;display:block}.footer{background:#1a254c}.footer-journal-name{background:#111831;padding:7px 0}.footer-journal-name h2{color:#fff;font-size:1.8rem;line-height:2.4rem;margin:5px 0;text-align:center}.footer-title{color:#fff;font-size:1.6rem;font-weight:700;line-height:2.2rem;margin:50px 0 18px}.footer-title:focus{outline:2px solid #f69038!important;outline-offset:6px!important}.footer ul{padding:0}.footer ul li{list-style-type:none;margin-bottom:12px}.footer ul li a{color:#fff;font-size:1.6rem;font-weight:light;opacity:.6;-webkit-text-decoration:none;text-decoration:none}.footer ul li a:focus,.footer ul li a:hover{opacity:1}.footer-social .bluesky a{background-color:#0085ff;border-radius:50px;margin-right:1px;padding:4px 5px;text-align:center;transition:all .2s ease}.footer-social .bluesky a:hover{background-color:#0078e6;transition:all .2s ease}.footer-social .twitter a{border-radius:50px;margin-right:1px;padding:4px 5px;text-align:center}.footer-social .twitter a,.footer-social .twitter a:hover{background-color:#000;transition:all .2s ease}.footer-social .facebook a{background-color:#3b5a98;border-radius:50px;margin-right:1px;padding:4px 5px;text-align:center;transition:all .2s ease}.footer-social .facebook a:hover{background-color:#344f86;transition:all .2s ease}.footer-social .linkedin a{background-color:#0077b5;border-radius:50px;margin-right:1px;padding:4px 5px;text-align:center;transition:all .2s ease}.footer-social .linkedin a:hover{background-color:#00669c;transition:all .2s ease}.footer-social .youtube a{background-color:red;border-radius:50px;margin-right:1px;padding:4px 5px;text-align:center;transition:all .2s ease}.footer-social .youtube a:hover{background-color:#e60000;transition:all .2s ease}.footer-social a:hover{-webkit-text-decoration:none;text-decoration:none}.footer-social a i{color:#fff;text-align:center;width:14px}.footer-copyright{border-top:1px solid #fff;color:#fff;font-size:1.2rem;margin:20px 0 40px;opacity:.6;padding-top:20px}@media screen and (max-width:20.6875em){.footer .rss{margin-top:.25rem}} .mt-0{margin-top:0!important}.mt-5{margin-top:5px!important}.mt-10{margin-top:10px!important}.mt-15{margin-top:15px!important}.mt-20{margin-top:20px!important}.mt-25{margin-top:25px!important}.mt-30{margin-top:30px!important}.mt-35{margin-top:35px!important}.mt-40{margin-top:40px!important}.mt-45{margin-top:45px!important}.mt-50{margin-top:50px!important}.mt-55{margin-top:55px!important}.mt-60{margin-top:60px!important}.mt-65{margin-top:65px!important}.mt-70{margin-top:70px!important}.mt-75{margin-top:75px!important}.mt-80{margin-top:80px!important}.mt-85{margin-top:85px!important}.mt-90{margin-top:90px!important}.mt-95{margin-top:95px!important}.mt-100{margin-top:100px!important}.mb-0{margin-bottom:0!important}.mb-5{margin-bottom:5px!important}.mb-10{margin-bottom:10px!important}.mb-15{margin-bottom:15px!important}.mb-20{margin-bottom:20px!important}.mb-25{margin-bottom:25px!important}.mb-30{margin-bottom:30px!important}.mb-35{margin-bottom:35px!important}.mb-40{margin-bottom:40px!important}.mb-45{margin-bottom:45px!important}.mb-50{margin-bottom:50px!important}.mb-55{margin-bottom:55px!important}.mb-60{margin-bottom:60px!important}.mb-65{margin-bottom:65px!important}.mb-70{margin-bottom:70px!important}.mb-75{margin-bottom:75px!important}.mb-80{margin-bottom:80px!important}.mb-85{margin-bottom:85px!important}.mb-90{margin-bottom:90px!important}.mb-95{margin-bottom:95px!important}.mb-100{margin-bottom:100px!important}.ml-0{margin-left:0!important}.ml-5{margin-left:5px!important}.ml-10{margin-left:10px!important}.ml-15{margin-left:15px!important}.ml-20{margin-left:20px!important}.ml-25{margin-left:25px!important}.ml-30{margin-left:30px!important}.ml-35{margin-left:35px!important}.ml-40{margin-left:40px!important}.ml-45{margin-left:45px!important}.ml-50{margin-left:50px!important}.ml-55{margin-left:55px!important}.ml-60{margin-left:60px!important}.ml-65{margin-left:65px!important}.ml-70{margin-left:70px!important}.ml-75{margin-left:75px!important}.ml-80{margin-left:80px!important}.ml-85{margin-left:85px!important}.ml-90{margin-left:90px!important}.ml-95{margin-left:95px!important}.ml-100{margin-left:100px!important}.mr-0{margin-right:0!important}.mr-5{margin-right:5px!important}.mr-10{margin-right:10px!important}.mr-15{margin-right:15px!important}.mr-20{margin-right:20px!important}.mr-25{margin-right:25px!important}.mr-30{margin-right:30px!important}.mr-35{margin-right:35px!important}.mr-40{margin-right:40px!important}.mr-45{margin-right:45px!important}.mr-50{margin-right:50px!important}.mr-55{margin-right:55px!important}.mr-60{margin-right:60px!important}.mr-65{margin-right:65px!important}.mr-70{margin-right:70px!important}.mr-75{margin-right:75px!important}.mr-80{margin-right:80px!important}.mr-85{margin-right:85px!important}.mr-90{margin-right:90px!important}.mr-95{margin-right:95px!important}.mr-100{margin-right:100px!important}.pt-0{padding-top:0!important}.pt-5{padding-top:5px!important}.pt-10{padding-top:10px!important}.pt-15{padding-top:15px!important}.pt-20{padding-top:20px!important}.pt-25{padding-top:25px!important}.pt-30{padding-top:30px!important}.pt-35{padding-top:35px!important}.pt-40{padding-top:40px!important}.pt-45{padding-top:45px!important}.pt-50{padding-top:50px!important}.pt-55{padding-top:55px!important}.pt-60{padding-top:60px!important}.pt-65{padding-top:65px!important}.pt-70{padding-top:70px!important}.pt-75{padding-top:75px!important}.pt-80{padding-top:80px!important}.pt-85{padding-top:85px!important}.pt-90{padding-top:90px!important}.pt-95{padding-top:95px!important}.pt-100{padding-top:100px!important}.pb-0{padding-bottom:0!important}.pb-5{padding-bottom:5px!important}.pb-10{padding-bottom:10px!important}.pb-15{padding-bottom:15px!important}.pb-20{padding-bottom:20px!important}.pb-25{padding-bottom:25px!important}.pb-30{padding-bottom:30px!important}.pb-35{padding-bottom:35px!important}.pb-40{padding-bottom:40px!important}.pb-45{padding-bottom:45px!important}.pb-50{padding-bottom:50px!important}.pb-55{padding-bottom:55px!important}.pb-60{padding-bottom:60px!important}.pb-65{padding-bottom:65px!important}.pb-70{padding-bottom:70px!important}.pb-75{padding-bottom:75px!important}.pb-80{padding-bottom:80px!important}.pb-85{padding-bottom:85px!important}.pb-90{padding-bottom:90px!important}.pb-95{padding-bottom:95px!important}.pb-100{padding-bottom:100px!important}.pl-0{padding-left:0!important}.pl-5{padding-left:5px!important}.pl-10{padding-left:10px!important}.pl-15{padding-left:15px!important}.pl-20{padding-left:20px!important}.pl-25{padding-left:25px!important}.pl-30{padding-left:30px!important}.pl-35{padding-left:35px!important}.pl-40{padding-left:40px!important}.pl-45{padding-left:45px!important}.pl-50{padding-left:50px!important}.pl-55{padding-left:55px!important}.pl-60{padding-left:60px!important}.pl-65{padding-left:65px!important}.pl-70{padding-left:70px!important}.pl-75{padding-left:75px!important}.pl-80{padding-left:80px!important}.pl-85{padding-left:85px!important}.pl-90{padding-left:90px!important}.pl-95{padding-left:95px!important}.pl-100{padding-left:100px!important}.pr-0{padding-right:0!important}.pr-5{padding-right:5px!important}.pr-10{padding-right:10px!important}.pr-15{padding-right:15px!important}.pr-20{padding-right:20px!important}.pr-25{padding-right:25px!important}.pr-30{padding-right:30px!important}.pr-35{padding-right:35px!important}.pr-40{padding-right:40px!important}.pr-45{padding-right:45px!important}.pr-50{padding-right:50px!important}.pr-55{padding-right:55px!important}.pr-60{padding-right:60px!important}.pr-65{padding-right:65px!important}.pr-70{padding-right:70px!important}.pr-75{padding-right:75px!important}.pr-80{padding-right:80px!important}.pr-85{padding-right:85px!important}.pr-90{padding-right:90px!important}.pr-95{padding-right:95px!important}.pr-100{padding-right:100px!important}.scroll-to-very-top{background-color:#1e70c2;border-radius:3px 0 0 3px;bottom:.7vh;box-shadow:0 0 5px rgba(0,0,0,.2);position:fixed;right:0;z-index:2}.scroll-to-very-top:focus,.scroll-to-very-top:hover{opacity:.9}.scroll-to-very-top:focus{border:2px solid #f69038;margin:1px}.scroll-to-very-top .icon{color:#fff;font-size:2.5rem;padding:9px;transition:all .3s}</style></head><body ><div data-server-rendered="true" id="__nuxt"><div id="__layout"><div id="jmir-html"><span tabindex="-1"></span> <div id="skip-link"><a href="#main-content"> Skip to Main Content <span aria-hidden="true" class="icon fas fa-chevron-down"></span></a> <a href="#footer"> Skip to Footer <span aria-hidden="true" class="icon fas fa-chevron-down"></span></a></div>  <div data-v-49c694ee><span tabindex="0" aria-label="Accessibility settings" title="Accessibility settings" role="button" class="icon fas fa-universal-access universal-access-btn" data-v-49c694ee></span>  <style type="text/css" data-v-49c694ee> html { filter: none !important } html { font-weight: inherit !important } html { font-size: 0.625rem !important } html { text-align: initial !important } *:not svg { font-weight: inherit } </style></div> <div id="main-layout-container"><div style="background-color: white;"><div id="scroll-to-very-top"><header id="header"><section class="header"><nav><div data-v-575455fb><nav aria-label="Navigation" data-v-575455fb><section class="top-nav" data-v-575455fb><div class="container" data-v-575455fb> <div class="corporate" data-v-575455fb><div class="corporate__logo" data-v-575455fb><a href="https://jmirpublications.com" aria-label="JMIR Publications main website" data-v-575455fb><div data-v-575455fb><img src="https://asset.jmir.pub/resources/images/logos/JMIR-25-year-logo-less-white.png" alt="JMIR Publications" class="logo-img"></div></a></div> <div class="corporate__mobile-menu" data-v-575455fb><button data-v-575455fb><span aria-hidden="true" class="icon fas fa-bars" data-v-575455fb></span></button></div> <div class="corporate__search" data-v-575455fb><label for="corporate__search-select" class="screen-readers-only" data-v-575455fb> Select options </label> <select id="corporate__search-select" data-test="search-select" class="corporate__search-select" data-v-575455fb><option value="articles" selected="selected" data-v-575455fb> Articles </option> <option value="help" data-v-575455fb> Help </option></select> <input name="type" type="hidden" value="text" data-v-575455fb>  <button aria-label="Search articles" data-test="unisearch-button" class="btn btn-small btn-blue corporate__search-btn" data-v-575455fb><span aria-hidden="true" class="icon fas fa-search" data-v-575455fb></span></button></div> <div class="corporate__nav" data-v-575455fb><ul class="corporate__links" data-v-575455fb><li tabindex="0" aria-haspopup="true" data-test="resource-center" class="corporate__link-item" data-v-575455fb><a href="https://careers.jmir.org/" class="corporate__link" data-v-575455fb> Career Center </a></li> <li class="corporate__link-item" data-v-575455fb><a id="button-login" href="javascript:;" data-test="login-button" class="corporate__link" data-v-575455fb>Login</a></li> <li id="button-register" data-test="register-button" class="corporate__link-item" data-v-575455fb><a class="corporate__link" data-v-575455fb>Register</a></li> </ul></div></div></div></section> <section class="bottom-nav-1" data-v-575455fb><div class="container" data-v-575455fb><div class="journal" data-v-575455fb><div class="journal__nav" data-v-575455fb><ul class="journal__links" data-v-575455fb><li class="journal__link-item journal__link-item--journals" data-v-575455fb><div aria-haspopup="true" class="journal__link-item-container" data-v-575455fb><a href="/" aria-label="JMIR Neurotechnology home page" class="journal__link--home nuxt-link-active" data-v-575455fb><span aria-hidden="true" class="icon fas fa-home" style="font-size:16px;" data-v-575455fb></span> JMIR Neurotechnology </a> <span aria-expanded="false" aria-label="Other journals" tabindex="0" role="button" class="journal__journals-list" data-v-575455fb><span aria-hidden="true" data-test="journal-list-show-dropdown-button" class="icon fas fa-arrow-down" style="padding: 16px 16px;" data-v-575455fb></span></span></div> <ul aria-label="submenu" data-test="journal-list-dropdown" class="journal__link-submenu-journals" data-v-575455fb><li class="m-0" data-v-575455fb><a href="https://www.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Journal of Medical Internet Research</span> <span class="articles-number" data-v-575455fb>9060 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://www.researchprotocols.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Research Protocols</span> <span class="articles-number" data-v-575455fb>4343 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://formative.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Formative Research</span> <span class="articles-number" data-v-575455fb>3083 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://mhealth.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR mHealth and uHealth</span> <span class="articles-number" data-v-575455fb>2736 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://ojphi.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Online Journal of Public Health Informatics</span> <span class="articles-number" data-v-575455fb>1719 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://publichealth.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Public Health and Surveillance</span> <span class="articles-number" data-v-575455fb>1644 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://medinform.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Medical Informatics</span> <span class="articles-number" data-v-575455fb>1409 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://mental.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Mental Health</span> <span class="articles-number" data-v-575455fb>1076 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://humanfactors.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Human Factors</span> <span class="articles-number" data-v-575455fb>803 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://games.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Serious Games</span> <span class="articles-number" data-v-575455fb>635 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://mededu.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Medical Education</span> <span class="articles-number" data-v-575455fb>556 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://www.iproc.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Iproceedings</span> <span class="articles-number" data-v-575455fb>510 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://aging.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Aging</span> <span class="articles-number" data-v-575455fb>436 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://www.i-jmr.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Interactive Journal of Medical Research</span> <span class="articles-number" data-v-575455fb>433 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://xmed.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIRx Med</span> <span class="articles-number" data-v-575455fb>420 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://pediatrics.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Pediatrics and Parenting</span> <span class="articles-number" data-v-575455fb>406 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://cancer.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Cancer</span> <span class="articles-number" data-v-575455fb>386 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://derma.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Dermatology</span> <span class="articles-number" data-v-575455fb>308 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://diabetes.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Diabetes</span> <span class="articles-number" data-v-575455fb>268 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://rehab.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Rehabilitation and Assistive Technologies</span> <span class="articles-number" data-v-575455fb>263 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://cardio.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Cardio</span> <span class="articles-number" data-v-575455fb>198 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://infodemiology.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Infodemiology</span> <span class="articles-number" data-v-575455fb>150 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://ai.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR AI</span> <span class="articles-number" data-v-575455fb>119 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://periop.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Perioperative Medicine</span> <span class="articles-number" data-v-575455fb>115 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://nursing.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Nursing</span> <span class="articles-number" data-v-575455fb>114 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://jopm.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Journal of Participatory Medicine</span> <span class="articles-number" data-v-575455fb>98 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://biomedeng.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Biomedical Engineering</span> <span class="articles-number" data-v-575455fb>92 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://bioinform.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Bioinformatics and Biotechnology</span> <span class="articles-number" data-v-575455fb>51 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://www.medicine20.com" no-prefetch="" data-v-575455fb><span data-v-575455fb>Medicine 2.0</span> <span class="articles-number" data-v-575455fb>26 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://apinj.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Asian/Pacific Island Nursing Journal</span> <span class="articles-number" data-v-575455fb>25 articles </span></a></li><li class="m-0" data-v-575455fb><a href="/" class="nuxt-link-active" data-v-575455fb><span data-v-575455fb>JMIR Neurotechnology</span> <span class="articles-number" data-v-575455fb>24 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://xbio.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIRx Bio</span> <span class="articles-number" data-v-575455fb>22 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://xr.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR XR and Spatial Computing (JMXR)</span> <span class="articles-number" data-v-575455fb>18 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://data.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Data</span> </a></li><li class="m-0" data-v-575455fb><a href="https://challenges.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Challenges</span> </a></li><li class="m-0" data-v-575455fb><a href="https://preprints.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Preprints</span> </a></li></ul></li> <li tabindex="0" aria-haspopup="true" data-test="journal-information" class="journal__link-item journal__link-item--journal-info" data-v-575455fb><span class="journal__link" data-v-575455fb> Journal Information </span> <span aria-hidden="true" class="icon fas fa-caret-down" data-v-575455fb></span> <ul aria-label="submenu" data-test="journal-info-popover" class="journal__link-submenu journal__link-submenu--journal-info" data-v-575455fb><li data-v-575455fb><a href="/about-journal/focus-and-scope" data-v-575455fb> Focus and Scope </a></li><li data-v-575455fb><a href="/about-journal/editorial-board" data-v-575455fb> Editorial Board </a></li><li data-v-575455fb><a href="/author-information/instructions-for-authors" data-v-575455fb> Author Information </a></li><li data-v-575455fb><a href="/resource-centre/author-hub" data-v-575455fb> Resource Center </a></li><li data-v-575455fb><a href="/about-journal/article-processing-fees" data-v-575455fb> Article Processing Fees </a></li><li data-v-575455fb><a href="/publishing-policies/section-policies" data-v-575455fb> Publishing Policies </a></li><li data-v-575455fb><a href="/get-involved/new-journal-editor-in-chief-proposals" data-v-575455fb> Get Involved </a></li><li data-v-575455fb><a href="/top-articles/overview" data-v-575455fb> Top Articles </a></li><li data-v-575455fb><a href="/fees/institutional-partners" data-v-575455fb> Institutional Partners </a></li><li data-v-575455fb><a href="/about-journal/indexing-and-impact-factor" data-v-575455fb> Indexing and Impact Factor </a></li></ul></li> <li tabindex="0" aria-haspopup="true" class="journal__link-item journal__link-item--browse" data-v-575455fb><span class="journal__link" data-v-575455fb> Browse Journal </span> <span aria-hidden="true" class="icon fas fa-caret-down" data-v-575455fb></span> <ul aria-label="submenu" data-test="browse-journal-popover" class="journal__link-submenu journal__link-submenu--browse" data-v-575455fb><li class="m-0" data-v-575455fb><div class="journal__link-submenu--select" data-v-575455fb><label for="nav-year" data-v-575455fb> Year: </label> <select id="nav-year" data-v-575455fb><option disabled="disabled" value="" data-v-575455fb> Select... </option> <option value="2022" data-v-575455fb> 2022 </option><option value="2023" data-v-575455fb> 2023 </option><option value="2024" data-v-575455fb> 2024 </option></select></div></li> <li data-v-575455fb><a href="/announcements" data-v-575455fb> Latest Announcements </a></li><li data-v-575455fb><a href="/search/authors" data-v-575455fb> Authors </a></li> <li data-v-575455fb><a href="/themes" data-v-575455fb> Themes </a></li><li data-v-575455fb><a href="/issues" data-v-575455fb> Issues </a></li> <li data-v-575455fb><a href="https://blog.jmir.org/" data-v-575455fb> Blog </a></li></ul></li> <li class="journal__link-item journal__link-item--submit-article" data-v-575455fb><a href="/author" class="btn btn-small btn-blue journal__submit-article" data-v-575455fb>Submit Article</a></li></ul></div></div></div></section> <section class="bottom-nav-2" data-v-575455fb><div class="journal" data-v-575455fb><ul data-v-575455fb><li class="journal__link-item journal__link-item--journals" data-v-575455fb><div class="journal__link-item-container" data-v-575455fb><a href="/" class="journal__link--home nuxt-link-active" data-v-575455fb><span aria-hidden="true" class="icon fas fa-home" style="font-size:16px;" data-v-575455fb></span> JMIR Neurotechnology </a> <span class="journal__journals-list" data-v-575455fb><span aria-hidden="true" class="icon fas fa-arrow-down" data-v-575455fb></span></span></div> <ul data-test="journal-list" class="journal__link-submenu-journals" data-v-575455fb><li class="m-0" data-v-575455fb><a href="https://www.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Journal of Medical Internet Research</span> <span class="articles-number" data-v-575455fb>9060 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://www.researchprotocols.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Research Protocols</span> <span class="articles-number" data-v-575455fb>4343 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://formative.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Formative Research</span> <span class="articles-number" data-v-575455fb>3083 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://mhealth.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR mHealth and uHealth</span> <span class="articles-number" data-v-575455fb>2736 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://ojphi.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Online Journal of Public Health Informatics</span> <span class="articles-number" data-v-575455fb>1719 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://publichealth.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Public Health and Surveillance</span> <span class="articles-number" data-v-575455fb>1644 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://medinform.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Medical Informatics</span> <span class="articles-number" data-v-575455fb>1409 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://mental.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Mental Health</span> <span class="articles-number" data-v-575455fb>1076 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://humanfactors.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Human Factors</span> <span class="articles-number" data-v-575455fb>803 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://games.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Serious Games</span> <span class="articles-number" data-v-575455fb>635 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://mededu.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Medical Education</span> <span class="articles-number" data-v-575455fb>556 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://www.iproc.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Iproceedings</span> <span class="articles-number" data-v-575455fb>510 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://aging.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Aging</span> <span class="articles-number" data-v-575455fb>436 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://www.i-jmr.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Interactive Journal of Medical Research</span> <span class="articles-number" data-v-575455fb>433 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://xmed.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIRx Med</span> <span class="articles-number" data-v-575455fb>420 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://pediatrics.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Pediatrics and Parenting</span> <span class="articles-number" data-v-575455fb>406 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://cancer.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Cancer</span> <span class="articles-number" data-v-575455fb>386 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://derma.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Dermatology</span> <span class="articles-number" data-v-575455fb>308 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://diabetes.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Diabetes</span> <span class="articles-number" data-v-575455fb>268 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://rehab.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Rehabilitation and Assistive Technologies</span> <span class="articles-number" data-v-575455fb>263 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://cardio.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Cardio</span> <span class="articles-number" data-v-575455fb>198 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://infodemiology.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Infodemiology</span> <span class="articles-number" data-v-575455fb>150 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://ai.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR AI</span> <span class="articles-number" data-v-575455fb>119 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://periop.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Perioperative Medicine</span> <span class="articles-number" data-v-575455fb>115 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://nursing.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Nursing</span> <span class="articles-number" data-v-575455fb>114 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://jopm.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Journal of Participatory Medicine</span> <span class="articles-number" data-v-575455fb>98 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://biomedeng.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Biomedical Engineering</span> <span class="articles-number" data-v-575455fb>92 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://bioinform.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Bioinformatics and Biotechnology</span> <span class="articles-number" data-v-575455fb>51 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://www.medicine20.com" no-prefetch="" data-v-575455fb><span data-v-575455fb>Medicine 2.0</span> <span class="articles-number" data-v-575455fb>26 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://apinj.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>Asian/Pacific Island Nursing Journal</span> <span class="articles-number" data-v-575455fb>25 articles </span></a></li><li class="m-0" data-v-575455fb><a href="/" class="nuxt-link-active" data-v-575455fb><span data-v-575455fb>JMIR Neurotechnology</span> <span class="articles-number" data-v-575455fb>24 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://xbio.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIRx Bio</span> <span class="articles-number" data-v-575455fb>22 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://xr.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR XR and Spatial Computing (JMXR)</span> <span class="articles-number" data-v-575455fb>18 articles </span></a></li><li class="m-0" data-v-575455fb><a href="https://data.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Data</span> </a></li><li class="m-0" data-v-575455fb><a href="https://challenges.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Challenges</span> </a></li><li class="m-0" data-v-575455fb><a href="https://preprints.jmir.org" no-prefetch="" data-v-575455fb><span data-v-575455fb>JMIR Preprints</span> </a></li></ul></li></ul></div></section> </nav></div></nav></section></header></div> <div class="container-fluid" style="padding: 0px;"><div class="element-wrapper"><div data-test="main-content" class="container"><div class="sidebar-citation mobile-show"><div class="collection"><h2 tabindex="0" data-test="article-collection" class="h4 green-heading-underline width-fit-content"> This paper is in the following <span class="collection__span">e-collection/theme issue:</span></h2> <a href="/themes/1469" data-test="article-collection" aria-label="5 articles belongs to AI in Neurotechnology e-collection/theme issue" class="collection__link"> AI in Neurotechnology (5) </a><a href="/themes/805" data-test="article-collection" aria-label="577 articles belongs to Registered Report e-collection/theme issue" class="collection__link"> Registered Report (577) </a><a href="/themes/142" data-test="article-collection" aria-label="93 articles belongs to Reviews in Medical Informatics e-collection/theme issue" class="collection__link"> Reviews in Medical Informatics (93) </a><a href="/themes/1006" data-test="article-collection" aria-label="133 articles belongs to Computer-Aided Diagnosis e-collection/theme issue" class="collection__link"> Computer-Aided Diagnosis (133) </a><a href="/themes/797" data-test="article-collection" aria-label="1407 articles belongs to Artificial Intelligence e-collection/theme issue" class="collection__link"> Artificial Intelligence (1407) </a><a href="/themes/171" data-test="article-collection" aria-label="703 articles belongs to Natural Language Processing e-collection/theme issue" class="collection__link"> Natural Language Processing (703) </a><a href="/themes/181" data-test="article-collection" aria-label="621 articles belongs to mHealth in a Clinical Setting e-collection/theme issue" class="collection__link"> mHealth in a Clinical Setting (621) </a><a href="/themes/18" data-test="article-collection" aria-label="1315 articles belongs to Digital Health Reviews e-collection/theme issue" class="collection__link"> Digital Health Reviews (1315) </a><a href="/themes/1007" data-test="article-collection" aria-label="62 articles belongs to mHealth for Diagnosis e-collection/theme issue" class="collection__link"> mHealth for Diagnosis (62) </a><a href="/themes/1470" data-test="article-collection" aria-label="36 articles belongs to Neurotech Innovations, Diagnostic Tools and Techniques e-collection/theme issue" class="collection__link"> Neurotech Innovations, Diagnostic Tools and Techniques (36) </a><a href="/themes/102" data-test="article-collection" aria-label="189 articles belongs to Neurology and Neurosciences e-collection/theme issue" class="collection__link"> Neurology and Neurosciences (189) </a></div></div> <div class="row"><div class="main col-lg-9 mb-1"> <div data-test="details" class="details"><div><p id="main-content" tabindex="0"> Published on <time datetime="22.05.2024">22.05.2024 </time> in <span data-test="issue-info"><a href="/2024/1" class="nuxt-link-active"> Vol 3 (2024)</a></span></p> </div> <div class="preprints-version"><span aria-hidden="true" class="icon fas fa-thumbtack"></span> <div><span class="ml-2"> Preprints (earlier versions) of this paper are available at <a data-test="preprint-link" aria-label="'Preprints (earlier versions) of this paper are available at preprints.jmir.org/preprint/'51822" href="https://preprints.jmir.org/preprint/51822" target="_blank">https://preprints.jmir.org/preprint/51822</a>, first published <time datetime="October 02, 2023">October 02, 2023</time>. </span></div></div></div> <div class="info mt-3"><div class="info__article-img"><div data-v-10f10a3e><img data-srcset="https://asset.jmir.pub/assets/92d91f88ddb727d190767f7247b60cd4.png 480w,https://asset.jmir.pub/assets/92d91f88ddb727d190767f7247b60cd4.png 960w,https://asset.jmir.pub/assets/92d91f88ddb727d190767f7247b60cd4.png 1920w,https://asset.jmir.pub/assets/92d91f88ddb727d190767f7247b60cd4.png 2500w" alt="Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review" title="Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review" aria-label="Article Thumbnail Image" src="https://asset.jmir.pub/placeholder.svg" data-v-10f10a3e></div> <div data-test="article-img-info" class="info__article-img-info"><span aria-hidden="true" class="icon fas fa-search-plus"></span></div></div> <div class="info__title-authors"><h1 tabindex="0" aria-label="Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review" class="h3 mb-0 mt-0">Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review</h1> <h2 class="info__hidden-title"> Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review </h2> <div class="mt-3"><p tabindex="0" class="authors-for-screen-reader"> Authors of this article: </p> <span data-test="authors-info" class="info__authors"><a href="/search?term=Ilana%20Lefkovitz&type=author&precise=true" aria-label="Ilana Lefkovitz. Search more articles by this author."> Ilana Lefkovitz<sup>1</sup> </a> <span><a aria-label="Visit this author on ORCID website" data-test="orcid-link" target="_blank" href="https://orcid.org/0000-0002-8724-3798"><img src="https://asset.jmir.pub/assets/static/images/Orcid-ID-Logo-Colour.png" alt="Author Orcid Image" aria-label="Author Orcid Image" class="info__orcid-img"></a></span> <span style="margin-left: -2px;"> ; </span></span><span data-test="authors-info" class="info__authors"><a href="/search?term=Samantha%20Walsh&type=author&precise=true" aria-label="Samantha Walsh. Search more articles by this author."> Samantha Walsh<sup>2</sup> </a> <span><a aria-label="Visit this author on ORCID website" data-test="orcid-link" target="_blank" href="https://orcid.org/0000-0002-5040-6514"><img src="https://asset.jmir.pub/assets/static/images/Orcid-ID-Logo-Colour.png" alt="Author Orcid Image" aria-label="Author Orcid Image" class="info__orcid-img"></a></span> <span style="margin-left: -2px;"> ; </span></span><span data-test="authors-info" class="info__authors"><a href="/search?term=Leah%20J%20Blank&type=author&precise=true" aria-label="Leah J Blank. Search more articles by this author."> Leah J Blank<sup>1, 3</sup> </a> <span><a aria-label="Visit this author on ORCID website" data-test="orcid-link" target="_blank" href="https://orcid.org/0000-0001-8719-6752"><img src="https://asset.jmir.pub/assets/static/images/Orcid-ID-Logo-Colour.png" alt="Author Orcid Image" aria-label="Author Orcid Image" class="info__orcid-img"></a></span> <span style="margin-left: -2px;"> ; </span></span><span data-test="authors-info" class="info__authors"><a href="/search?term=Nathalie%20Jett%C3%A9&type=author&precise=true" aria-label="Nathalie Jetté. Search more articles by this author."> Nathalie Jetté<sup>1, 4</sup> </a> <span><a aria-label="Visit this author on ORCID website" data-test="orcid-link" target="_blank" href="https://orcid.org/0000-0003-1351-5866"><img src="https://asset.jmir.pub/assets/static/images/Orcid-ID-Logo-Colour.png" alt="Author Orcid Image" aria-label="Author Orcid Image" class="info__orcid-img"></a></span> <span style="margin-left: -2px;"> ; </span></span><span data-test="authors-info" class="info__authors"><a href="/search?term=Benjamin%20R%20Kummer&type=author&precise=true" aria-label="Benjamin R Kummer. Search more articles by this author."> Benjamin R Kummer<sup>1, 5, 6</sup> </a> <span><a aria-label="Visit this author on ORCID website" data-test="orcid-link" target="_blank" href="https://orcid.org/0000-0002-1413-8014"><img src="https://asset.jmir.pub/assets/static/images/Orcid-ID-Logo-Colour.png" alt="Author Orcid Image" aria-label="Author Orcid Image" class="info__orcid-img"></a></span> </span></div> </div></div> <div role="tablist" aria-label="Article" class="tabs"><a href="/2024/1/e51822/" aria-current="page" role="tab" aria-label="Article" data-test="tabs" class="nuxt-link-exact-active nuxt-link-active active"> Article </a><a href="/2024/1/e51822/authors" role="tab" aria-label="Authors" data-test="tabs"> Authors </a><a href="/2024/1/e51822/citations" role="tab" aria-label="Cited by " data-test="tabs"> Cited by </a><a href="/2024/1/e51822/tweetations" role="tab" aria-label="Tweetations (3)" data-test="tabs"> Tweetations (3) </a><a href="/2024/1/e51822/metrics" role="tab" aria-label="Metrics" data-test="tabs"> Metrics </a></div> <div class="container"><div class="row"><div class="col-lg-3 mb-5 sidebar-sections"><div class="sidebar-nav"><div class="sidebar-nav-sticky"><ul></ul></div></div></div> <div data-test="keyword-links" class="col-lg-9 article"><main id="wrapper" class="wrapper ArticleMain clearfix"><section class="inner-wrapper clearfix"><section class="main-article-content clearfix"><article class="ajax-article-content"><h4 class="h4-original-paper"><span class="typcn typcn-document-text"></span>Review</h4><div class="authors-container"><div class="authors clearfix"></div></div><div class="authors-container"><div class="authors clearfix"></div></div><div class="authors-container"><div class="authors clearfix"><ul class="clearfix"><li><a href="/search/searchResult?field%5B%5D=author&criteria%5B%5D=Ilana+Lefkovitz" class="btn-view-author-options">Ilana Lefkovitz<sup><small>1</small></sup>, MD</a><a class="author-orcid" href="https://orcid.org/0000-0002-8724-3798" target="_blank" title="ORCID"> </a>; </li><li><a href="/search/searchResult?field%5B%5D=author&criteria%5B%5D=Samantha+Walsh" class="btn-view-author-options">Samantha Walsh<sup><small>2</small></sup>, MLS, MA</a><a class="author-orcid" href="https://orcid.org/0000-0002-5040-6514" target="_blank" title="ORCID"> </a>; </li><li><a href="/search/searchResult?field%5B%5D=author&criteria%5B%5D=Leah%20J+Blank" class="btn-view-author-options">Leah J Blank<sup><small>1,</small></sup><sup><small>3</small></sup>, MD, MPH</a><a class="author-orcid" href="https://orcid.org/0000-0001-8719-6752" target="_blank" title="ORCID"> </a>; </li><li><a href="/search/searchResult?field%5B%5D=author&criteria%5B%5D=Nathalie+Jetté" class="btn-view-author-options">Nathalie Jetté<sup><small>1,</small></sup><sup><small>4</small></sup>, MD, MSc</a><a class="author-orcid" href="https://orcid.org/0000-0003-1351-5866" target="_blank" title="ORCID"> </a>; </li><li><a href="/search/searchResult?field%5B%5D=author&criteria%5B%5D=Benjamin%20R+Kummer" class="btn-view-author-options">Benjamin R Kummer<sup><small>1,</small></sup><sup><small>5,</small></sup><sup><small>6</small></sup>, MD</a><a class="author-orcid" href="https://orcid.org/0000-0002-1413-8014" target="_blank" title="ORCID"> </a></li></ul><div class="author-affiliation-details"><p><sup>1</sup>Department of Neurology, Icahn School of Medicine at Mount Sinai, New York, NY, United States</p><p><sup>2</sup>Hunter College Libraries, Hunter College, City University of New York, New York, NY, United States</p><p><sup>3</sup>Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, NY, United States</p><p><sup>4</sup>Department of Clinical Neurosciences, University of Calgary, Calgary, AB, Canada</p><p><sup>5</sup>Clinical Neuro-Informatics Program, Icahn School of Medicine at Mount Sinai, New York, NY, United States</p><p><sup>6</sup>Windreich Department of Artificial Intelligence and Human Health, Icahn School of Medicine at Mount Sinai, New York, NY, United States</p></div></div><div class="corresponding-author-and-affiliations clearfix"><div class="corresponding-author-details"><h3>Corresponding Author:</h3><p>Benjamin R Kummer, MD</p><p></p><p>Department of Neurology</p><p>Icahn School of Medicine at Mount Sinai</p><p>One Gustave Levy Place</p><p>Box 1137</p><p>New York, NY, 10029</p><p>United States</p><p>Phone: 1 212 241 5050</p><p>Email: <a href="mailto:benjamin.kummer@mountsinai.org">benjamin.kummer@mountsinai.org</a></p><br></div></div></div><section class="article-content clearfix"><article class="abstract"><h3 id="Abstract" class="navigation-heading" data-label="Abstract">Abstract</h3><p><span class="abstract-sub-heading">Background: </span>Natural language processing (NLP), a branch of artificial intelligence that analyzes unstructured language, is being increasingly used in health care. However, the extent to which NLP has been formally studied in neurological disorders remains unclear.<br></p><p><span class="abstract-sub-heading">Objective: </span>We sought to characterize studies that applied NLP to the diagnosis, prediction, or treatment of common neurological disorders.<br></p><p><span class="abstract-sub-heading">Methods: </span>This review followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) standards. The search was conducted using MEDLINE and Embase on May 11, 2022. Studies of NLP use in migraine, Parkinson disease, Alzheimer disease, stroke and transient ischemic attack, epilepsy, or multiple sclerosis were included. We excluded conference abstracts, review papers, as well as studies involving heterogeneous clinical populations or indirect clinical uses of NLP. Study characteristics were extracted and analyzed using descriptive statistics. We did not aggregate measurements of performance in our review due to the high variability in study outcomes, which is the main limitation of the study.<br></p><p><span class="abstract-sub-heading">Results: </span>In total, 916 studies were identified, of which 41 (4.5%) met all eligibility criteria and were included in the final review. Of the 41 included studies, the most frequently represented disorders were stroke and transient ischemic attack (n=20, 49%), followed by epilepsy (n=10, 24%), Alzheimer disease (n=6, 15%), and multiple sclerosis (n=5, 12%). We found no studies of NLP use in migraine or Parkinson disease that met our eligibility criteria. The main objective of NLP was diagnosis (n=20, 49%), followed by disease phenotyping (n=17, 41%), prognostication (n=9, 22%), and treatment (n=4, 10%). In total, 18 (44%) studies used only machine learning approaches, 6 (15%) used only rule-based methods, and 17 (41%) used both.<br></p><p><span class="abstract-sub-heading">Conclusions: </span>We found that NLP was most commonly applied for diagnosis, implying a potential role for NLP in augmenting diagnostic accuracy in settings with limited access to neurological expertise. We also found several gaps in neurological NLP research, with few to no studies addressing certain disorders, which may suggest additional areas of inquiry.<br></p><p><span class="abstract-sub-heading">Trial Registration: </span>Prospective Register of Systematic Reviews (PROSPERO) CRD42021228703; https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=228703<br></p><strong class="h4-article-volume-issue">JMIR Neurotech 2024;3:e51822</strong><br><br><span class="article-doi"><a href="https://doi.org/10.2196/51822">doi:10.2196/51822</a></span><br><br><h3 class="h3-main-heading" id="Keywords">Keywords</h3><div class="keywords"><span><a href="/search?type=keyword&term=natural%20language%20processing&precise=true">natural language processing</a>; </span><span><a href="/search?type=keyword&term=NLP&precise=true">NLP</a>; </span><span><a href="/search?type=keyword&term=unstructured&precise=true">unstructured</a>; </span><span><a href="/search?type=keyword&term=text&precise=true">text</a>; </span><span><a href="/search?type=keyword&term=machine%20learning&precise=true">machine learning</a>; </span><span><a href="/search?type=keyword&term=deep%20learning&precise=true">deep learning</a>; </span><span><a href="/search?type=keyword&term=neurology&precise=true">neurology</a>; </span><span><a href="/search?type=keyword&term=headache%20disorders&precise=true">headache disorders</a>; </span><span><a href="/search?type=keyword&term=migraine&precise=true">migraine</a>; </span><span><a href="/search?type=keyword&term=Parkinson%20disease&precise=true">Parkinson disease</a>; </span><span><a href="/search?type=keyword&term=cerebrovascular%20disease&precise=true">cerebrovascular disease</a>; </span><span><a href="/search?type=keyword&term=stroke&precise=true">stroke</a>; </span><span><a href="/search?type=keyword&term=transient%20ischemic%20attack&precise=true">transient ischemic attack</a>; </span><span><a href="/search?type=keyword&term=epilepsy&precise=true">epilepsy</a>; </span><span><a href="/search?type=keyword&term=multiple%20sclerosis&precise=true">multiple sclerosis</a>; </span><span><a href="/search?type=keyword&term=cardiovascular&precise=true">cardiovascular</a>; </span><span><a href="/search?type=keyword&term=artificial%20intelligence&precise=true">artificial intelligence</a>; </span><span><a href="/search?type=keyword&term=Parkinson&precise=true">Parkinson</a>; </span><span><a href="/search?type=keyword&term=neurological&precise=true">neurological</a>; </span><span><a href="/search?type=keyword&term=neurological%20disorder&precise=true">neurological disorder</a>; </span><span><a href="/search?type=keyword&term=scoping%20review&precise=true">scoping review</a>; </span><span><a href="/search?type=keyword&term=diagnosis&precise=true">diagnosis</a>; </span><span><a href="/search?type=keyword&term=treatment&precise=true">treatment</a>; </span><span><a href="/search?type=keyword&term=prediction&precise=true">prediction</a> </span></div><div id="trendmd-suggestions"></div></article><br><article class="main-article clearfix"><br><h3 class="navigation-heading h3-main-heading" id="Introduction" data-label="Introduction">Introduction</h3><p class="abstract-paragraph">The implementation of the electronic medical record (EMR) in health care systems has resulted in a remarkable increase in the amount of digital patient data [<span class="footers"><a class="citation-link" href="#ref1" rel="footnote">1</a></span>], much of which is text-based and stored in an unstructured, narrative format [<span class="footers"><a class="citation-link" href="#ref2" rel="footnote">2</a></span>-<span class="footers"><a class="citation-link" href="#ref4" rel="footnote">4</a></span>]. While unstructured text is a rich data source, analyses of these data often require time- and cost-intensive manual processing [<span class="footers"><a class="citation-link" href="#ref3" rel="footnote">3</a></span>]. Natural language processing (NLP), a type of artificial intelligence that automatically derives meaning from unstructured language, can significantly reduce costs and enhance the quality of health care systems by converting unstructured text into a structured form that can be processed by computers [<span class="footers"><a class="citation-link" href="#ref2" rel="footnote">2</a></span>,<span class="footers"><a class="citation-link" href="#ref4" rel="footnote">4</a></span>,<span class="footers"><a class="citation-link" href="#ref5" rel="footnote">5</a></span>].</p><p class="abstract-paragraph">Approaches to NLP can use rule-based techniques, machine learning (ML), or a combination of both [<span class="footers"><a class="citation-link" href="#ref6" rel="footnote">6</a></span>-<span class="footers"><a class="citation-link" href="#ref8" rel="footnote">8</a></span>]. Between the fifth and eighth decades of the 20th century, NLP approaches were predominantly rule-based, using a set of rules defined by human experts [<span class="footers"><a class="citation-link" href="#ref7" rel="footnote">7</a></span>,<span class="footers"><a class="citation-link" href="#ref9" rel="footnote">9</a></span>] to systematically extract meaning from unstructured text. Rule-based methods are comprehensible by humans but difficult to generalize [<span class="footers"><a class="citation-link" href="#ref7" rel="footnote">7</a></span>,<span class="footers"><a class="citation-link" href="#ref9" rel="footnote">9</a></span>]. Driven by recent advances in computing power and access to computing resources, contemporary approaches to NLP have increasingly incorporated ML, which possesses greater scalability [<span class="footers"><a class="citation-link" href="#ref7" rel="footnote">7</a></span>] than rule-based methods despite the need for greater computational power to construct ML-based NLP models. Most recently, complex ML methods such as deep learning (DL), which are based on neural networks and larger datasets than conventional ML approaches, have become popular approaches to address NLP tasks [<span class="footers"><a class="citation-link" href="#ref9" rel="footnote">9</a></span>,<span class="footers"><a class="citation-link" href="#ref10" rel="footnote">10</a></span>].</p><p class="abstract-paragraph">The high prevalence of unstructured text in EMR systems creates an ideal use case for NLP in health care. However, the majority of current NLP research remains focused on nonneurological conditions such as mental health, cancer, and pneumonia [<span class="footers"><a class="citation-link" href="#ref5" rel="footnote">5</a></span>]. The dearth of neurological NLP research is out of proportion to the worldwide importance of neurological conditions, both in terms of public health burden and cost. For instance, cerebrovascular disease occupies the second leading cause of death worldwide [<span class="footers"><a class="citation-link" href="#ref11" rel="footnote">11</a></span>], and in the United States, neurological and musculoskeletal disorders generate the greatest number of years lost to disability [<span class="footers"><a class="citation-link" href="#ref12" rel="footnote">12</a></span>]. Finally, the estimated annual cost of the most prevalent neurological diseases in the United States is nearly US $800 billion [<span class="footers"><a class="citation-link" href="#ref12" rel="footnote">12</a></span>].</p><p class="abstract-paragraph">Neurology is a specialty that is uniquely well suited to benefit from NLP approaches. The data used in the diagnosis and management of neurological conditions, such as examination findings or clinical impressions, are often recorded as narrative, unstructured text in clinical documentation. Aside from clinical notes containing the patient history and neurological examination, reports from radiology [<span class="footers"><a class="citation-link" href="#ref13" rel="footnote">13</a></span>,<span class="footers"><a class="citation-link" href="#ref14" rel="footnote">14</a></span>], sonography, or electrophysiology studies are integral to neurological practice and often are crucial for detection, prognosis, and treatment. Further, NLP analysis of spoken language may allow the detection of certain neurodegenerative conditions such as Alzheimer disease in their early stages [<span class="footers"><a class="citation-link" href="#ref15" rel="footnote">15</a></span>]. Given the unique position of neurology with respect to NLP and the relative lack of research on the applications of NLP in neurology, we sought to conduct a scoping review in order to quantify and characterize studies that directly applied NLP for clinical use in common neurological disorders.</p><br><h3 class="navigation-heading h3-main-heading" id="Methods" data-label="Methods">Methods</h3><h4>Literature Search Strategy and Eligibility Criteria</h4><p class="abstract-paragraph">This review was conducted using the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines (<span class="footers"><a class="citation-link" href="#app1" rel="footnote">Multimedia Appendix 1</a></span>) and was registered with the Prospective Register of Systematic Reviews (PROSPERO CRD42021228703). Our search was conducted using Ovid Embase and MEDLINE on May 11, 2022 (<span class="footers"><a class="citation-link" href="#app2" rel="footnote">Multimedia Appendix 2</a></span> [<span class="footers"><a class="citation-link" href="#ref16" rel="footnote">16</a></span>-<span class="footers"><a class="citation-link" href="#ref22" rel="footnote">22</a></span>]). Based on the most globally prevalent and costly neurological disorders [<span class="footers"><a class="citation-link" href="#ref11" rel="footnote">11</a></span>], studies investigating the use of NLP in Alzheimer disease (exclusive of Alzheimer disease–related disorders), Parkinson disease, stroke and transient ischemic attack, epilepsy, multiple sclerosis (MS), and migraine were included.</p><p class="abstract-paragraph">Studies that used NLP to analyze radiographic findings without any clinical correlation (eg, silent brain infarcts) or for purposes other than diagnosis, detection, phenotyping, subtyping, prognostication, risk stratification, or therapy were excluded. We excluded studies with populations comprised of patients with heterogeneous diseases or ambiguously defined populations (eg, we excluded studies that used a patient cohort consisting of patients with both Alzheimer dementia and mild cognitive impairment) as well as studies that did not use NLP for direct clinical applications. Examples of indirect clinical applications include the use of NLP to identify cohorts for subsequent model development or conduct epidemiological associations between cohorts without direct impact on clinical practice. We additionally excluded abstracts, conference proceedings, reviews, and editorials.</p><h4>Data Extraction</h4><p class="abstract-paragraph">A medical librarian (SW) with expertise in scoping reviews first conducted a literature search (<span class="footers"><a class="citation-link" href="#app2" rel="footnote">Multimedia Appendix 2</a></span>) based on our eligibility criteria to generate a list of abstracts, which were then imported into a web application (Covidence Ltd) for initial screening by 3 authors (BRK, LJB, and IL). After the abstract screening was completed, full-text papers for screened abstracts were reviewed by 2 authors (BRK and IL) to determine eligibility for inclusion. Disagreements at both stages were resolved by discussion and consensus.</p><p class="abstract-paragraph">Using the final list of full-text studies, study characteristics were manually extracted by 1 author (IL) and charted in a REDCap (Research Electronic Data Capture; REDCap Consortium) web database form, which was subsequently reviewed by a second author (BRK) for accuracy. The data charting form was initially tested by the data extractor (IL) and revised after feedback from all coauthors (BRK, NJ, LJB, and SW). We extracted study publication year, population size, country of origin, journal field (eg, medical informatics, clinical neurology, nonclinical neuroscience, clinical medicine, or other), neurological disorder, and target of NLP (eg, diagnosis or detection, phenotyping or subtyping and severity, prognostication or risk stratification, or disease management or therapy). Each study could have multiple targets whenever applicable.</p><p class="abstract-paragraph">For each study, the source language to which NLP techniques were applied was also extracted. For studies conducted in or authored by teams from non-English–speaking countries, the source language was extrapolated directly as described from the study methodology. If the source language was a publicly available research dataset or ontology (eg, MetaMap ontology or ADReSS dataset, both of which use English), the source language was reported as English. Source of language for NLP (eg, clinical notes, radiographic reports, speech audio, or other) and type of study (eg, model derivation, validation, or both) were also noted. Validation studies were defined as studies that specifically investigated the validation of a derived model in a population external to the original model derivation population. Our definition of validation studies did not include validation on held-out test sets as part of model derivation. If the NLP model was both derived and externally validated in the same study, the population size included the additional population used for validation. Simulated patients, who were used as a training set in one study, were included in the population size. If no population size was mentioned in the studies, the number of text instances (eg, clinical notes and radiographic reports) was recorded.</p><p class="abstract-paragraph">We additionally extracted the study’s NLP approaches (ie, rule-based methods, ML, or both). Rule-based NLP included any approaches that used keyword searches, pattern matching, regular expressions, or ontological systems for word-concept mapping, text preprocessing, or classification. ML-based NLP comprised both conventional ML and DL approaches and both were distinguished as dichotomous study characteristic variables but could co-occur in the studies. A study was characterized as including any of these methods if either ML or DL was used at any point in model development for the study.</p><p class="abstract-paragraph">Under the category of conventional ML methods, linear regression, logistic regression, support vector machines (SVMs), naïve Bayes classifiers, decision trees, random forest classifiers, k-nearest neighbor algorithms, gradient boosting techniques such as extreme gradient boosting, latent Dirichlet allocation, and shallow neural networks were included. Under the definition of shallow neural network, we included any approaches using Word2vec or other “-2vec” word-embedding techniques that use a neural network to construct word contexts and extract semantic and syntactic meaning from text [<span class="footers"><a class="citation-link" href="#ref23" rel="footnote">23</a></span>,<span class="footers"><a class="citation-link" href="#ref24" rel="footnote">24</a></span>]. We also included other types of regression, such as lasso regression, which is often used for dimensionality reduction, in the conventional ML category.</p><p class="abstract-paragraph">DL techniques included convolutional neural networks, recurrent neural networks (RNNs), long- and short-term memory networks, multilayer perceptrons, and transformers. Studies using long- and short-term memory networks were also categorized as using an RNN. We also note that neural networks of unspecified type and number of layers, which were not clearly referred to as DL in the study, were not included in this category.</p><br><h3 class="navigation-heading h3-main-heading" id="Results" data-label="Results">Results</h3><h4>Included Studies</h4><p class="abstract-paragraph">In total, 916 studies were identified from our search strategy, of which 271 were duplicates and were excluded. We then screened the resulting 645 abstracts, of which 565 were excluded due to not meeting initial eligibility criteria. Of the remaining 80 studies, 39 (49%) were excluded. The 2 most common reasons for exclusion were the use of NLP for nonclinical applications (n=15, 38%) and heterogeneous clinical populations (n=12, 31%). In total, 41 (4.5%) of the 916 studies from the original search results were ultimately included for review (<span class="footers"><a class="citation-link" href="#figure1" rel="footnote">Figure 1</a></span> and <span class="footers"><a class="citation-link" href="#table1" rel="footnote">Table 1</a></span>).</p><p class="abstract-paragraph">Of the 41 included studies, NLP was applied to stroke or transient ischemic attack in 20 (49%) studies, epilepsy in 10 (24%) studies, Alzheimer dementia in 6 (15%) studies, and MS in 5 (12%) studies. We found no studies applying NLP to Parkinson disease or migraine that met our eligibility criteria. Across all neurological conditions, NLP was most commonly applied for the purposes of detection or diagnosis (n=20, 49%), followed by clinical disease phenotyping or subtyping (n=17, 41%), prognostication or risk stratification (n=9, 22%), and management or therapy (n=4, 10%; <span class="footers"><a class="citation-link" href="#table2" rel="footnote">Table 2</a></span>).</p><figure><a name="figure1">‎</a><a class="fancybox" title="Figure 1. Study PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) diagram. NLP: natural language processing." href="https://asset.jmir.pub/assets/d2d99db4f7ccd60a8d1c2d68ea63db52.png" id="figure1"><img class="figure-image" src="https://asset.jmir.pub/assets/d2d99db4f7ccd60a8d1c2d68ea63db52.png"></a><figcaption><span class="typcn typcn-image"></span><b>Figure 1. </b> Study PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) diagram. NLP: natural language processing. </figcaption></figure><div class="figure-table"><figcaption><span class="typcn typcn-clipboard"></span><b>Table 1. </b>Included studies.</figcaption><table width="1000" cellpadding="5" cellspacing="0" border="1" rules="groups" frame="hsides"><col width="70" span="1"><col width="70" span="1"><col width="60" span="1"><col width="70" span="1"><col width="70" span="1"><col width="70" span="1"><col width="80" span="1"><col width="90" span="1"><col width="70" span="1"><col width="60" span="1"><col width="150" span="1"><col width="140" span="1"><thead><tr valign="top"><td rowspan="1" colspan="1">Paper authors</td><td rowspan="1" colspan="1">Publication date</td><td rowspan="1" colspan="1">Country</td><td rowspan="1" colspan="1">Source text</td><td rowspan="1" colspan="1">Journal field</td><td rowspan="1" colspan="1">External model validation</td><td rowspan="1" colspan="1">Condition being studied</td><td rowspan="1" colspan="1">Purpose of NLP<sup>a</sup></td><td rowspan="1" colspan="1">NLP method</td><td rowspan="1" colspan="1">Deep learning</td><td rowspan="1" colspan="1">Algorithms used</td><td rowspan="1" colspan="1">Study outcomes</td></tr></thead><tbody><tr valign="top"><td rowspan="1" colspan="1">Miller et al [<span class="footers"><a class="citation-link" href="#ref19" rel="footnote">19</a></span>]</td><td rowspan="1" colspan="1">May 9, 2022</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Radiology reports</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">Rule-based, ML<sup>b</sup></td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Random forest, linear regression, KNN<sup>c</sup>, lasso regression, MLP<sup>d</sup>, transformer</td><td rowspan="1" colspan="1">Radiographic complications of ischemic stroke (eg, hemorrhagic transformation)</td></tr><tr valign="top"><td rowspan="1" colspan="1">Lay et al [<span class="footers"><a class="citation-link" href="#ref25" rel="footnote">25</a></span>]</td><td rowspan="1" colspan="1">October 23, 2020</td><td rowspan="1" colspan="1">Australia</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Latent Dirichlet allocation</td><td rowspan="1" colspan="1">Identifying themes in medical records in patients with PNES<sup>e</sup>, congruency of themes</td></tr><tr valign="top"><td rowspan="1" colspan="1">Mayampurath et al [<span class="footers"><a class="citation-link" href="#ref26" rel="footnote">26</a></span>]</td><td rowspan="1" colspan="1">June 24, 2021</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Detection or diagnosis, clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM<sup>f</sup>, logistic regression</td><td rowspan="1" colspan="1">Acute stroke diagnosis, stroke severity and subtypes</td></tr><tr valign="top"><td rowspan="1" colspan="1">Li et al [<span class="footers"><a class="citation-link" href="#ref16" rel="footnote">16</a></span>]</td><td rowspan="1" colspan="1">March 1, 2021</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Radiology reports</td><td rowspan="1" colspan="1">Neuroradiology</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Random forest</td><td rowspan="1" colspan="1">Acute or subacute ischemic stroke cases before and during COVID-19</td></tr><tr valign="top"><td rowspan="1" colspan="1">Lineback et al [<span class="footers"><a class="citation-link" href="#ref27" rel="footnote">27</a></span>]</td><td rowspan="1" colspan="1">July 13, 2021</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Prognosis or risk stratification</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM, naïve Bayes, random forest, logistic regression, shallow neural network, lasso regression, ensemble, boosting</td><td rowspan="1" colspan="1">30-day stroke readmission, 30-day all-cause readmission</td></tr><tr valign="top"><td rowspan="1" colspan="1">Liu et al [<span class="footers"><a class="citation-link" href="#ref28" rel="footnote">28</a></span>]</td><td rowspan="1" colspan="1">April 13, 2022</td><td rowspan="1" colspan="1">China</td><td rowspan="1" colspan="1">Speech</td><td rowspan="1" colspan="1">Public health</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Alzheimer disease</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">SVM, random forest, logistic regression, boosting, CNN<sup>g</sup>, transformer</td><td rowspan="1" colspan="1">Detection of Alzheimer disease from speech</td></tr><tr valign="top"><td rowspan="1" colspan="1">Mahajan and Baths [<span class="footers"><a class="citation-link" href="#ref29" rel="footnote">29</a></span>]</td><td rowspan="1" colspan="1">February 5, 2021</td><td rowspan="1" colspan="1">India</td><td rowspan="1" colspan="1">Speech</td><td rowspan="1" colspan="1">Nonclinical neuroscience</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Alzheimer disease</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">CNN, RNN<sup>h</sup> (LSTM<sup>i</sup>)</td><td rowspan="1" colspan="1">Detection of Alzheimer disease from speech</td></tr><tr valign="top"><td rowspan="1" colspan="1">Bacchi et al [<span class="footers"><a class="citation-link" href="#ref30" rel="footnote">30</a></span>]</td><td rowspan="1" colspan="1">February 20, 2022</td><td rowspan="1" colspan="1">Australia</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Clinical medicine</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Random forest, decision tree, logistic regression, neural network with an unspecified number of layers</td><td rowspan="1" colspan="1">Extraction of stroke key performance indicators</td></tr><tr valign="top"><td rowspan="1" colspan="1">Hamid et al [<span class="footers"><a class="citation-link" href="#ref31" rel="footnote">31</a></span>]</td><td rowspan="1" colspan="1">October 14, 2013</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Naïve Bayes</td><td rowspan="1" colspan="1">Identification of patients with PNES</td></tr><tr valign="top"><td rowspan="1" colspan="1">Yu et al [<span class="footers"><a class="citation-link" href="#ref13" rel="footnote">13</a></span>]</td><td rowspan="1" colspan="1">September 16, 2020</td><td rowspan="1" colspan="1">Canada</td><td rowspan="1" colspan="1">Radiology reports</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Detection or diagnosis, clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">N/A<sup>j</sup></td><td rowspan="1" colspan="1">Identification of the presence and location of vascular occlusions and other stroke-related attributes</td></tr><tr valign="top"><td rowspan="1" colspan="1">Bacchi et al [<span class="footers"><a class="citation-link" href="#ref32" rel="footnote">32</a></span>]</td><td rowspan="1" colspan="1">January 17, 2019</td><td rowspan="1" colspan="1">Australia</td><td rowspan="1" colspan="1">Clinical notes and radiology reports</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Random forest, decision tree, CNN, RNN (LSTM)</td><td rowspan="1" colspan="1">Determining the cause of TIA<sup>k</sup>-like presentations (cerebrovascular vs noncerebrovascular)</td></tr><tr valign="top"><td rowspan="1" colspan="1">Garg et al [<span class="footers"><a class="citation-link" href="#ref33" rel="footnote">33</a></span>]</td><td rowspan="1" colspan="1">May 15, 2019</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes and radiology reports</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM, random forest, logistic regression, KNN, boosting, ensemble (stacking logistic regression, extra trees classifier)</td><td rowspan="1" colspan="1">Ischemic stroke subtypes</td></tr><tr valign="top"><td rowspan="1" colspan="1">Zhao et al [<span class="footers"><a class="citation-link" href="#ref21" rel="footnote">21</a></span>]</td><td rowspan="1" colspan="1">March 8, 2021</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Detection or diagnosis, clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Random forest, logistic regression</td><td rowspan="1" colspan="1">Incidence of stroke, stroke subtypes</td></tr><tr valign="top"><td rowspan="1" colspan="1">Pevy et al [<span class="footers"><a class="citation-link" href="#ref34" rel="footnote">34</a></span>]</td><td rowspan="1" colspan="1">October 1, 2021</td><td rowspan="1" colspan="1">United Kingdom</td><td rowspan="1" colspan="1">Speech</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Random forest</td><td rowspan="1" colspan="1">Distinguishing between PNES and epilepsy, hesitations and repetitions in descriptions of epileptic seizures versus PNES</td></tr><tr valign="top"><td rowspan="1" colspan="1">Guan et al [<span class="footers"><a class="citation-link" href="#ref35" rel="footnote">35</a></span>]</td><td rowspan="1" colspan="1">December 10, 2020</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Echocardiographic reports</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM, random forest, decision tree, logistic regression, KNN</td><td rowspan="1" colspan="1">Subtyping and phenotyping cardioembolic stroke</td></tr><tr valign="top"><td rowspan="1" colspan="1">Cui et al [<span class="footers"><a class="citation-link" href="#ref36" rel="footnote">36</a></span>]</td><td rowspan="1" colspan="1">June 26, 2014</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">N/A</td><td rowspan="1" colspan="1">Epilepsy phenotype extraction with correlated anatomic location</td></tr><tr valign="top"><td rowspan="1" colspan="1">Heo et al [<span class="footers"><a class="citation-link" href="#ref37" rel="footnote">37</a></span>]</td><td rowspan="1" colspan="1">December 16, 2020</td><td rowspan="1" colspan="1">South Korea</td><td rowspan="1" colspan="1">Radiology reports</td><td rowspan="1" colspan="1">Clinical medicine</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Prognosis or risk stratification</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">SVM, random forest, decision tree, shallow neural network, lasso regression, CNN, RNN (LSTM), MLP</td><td rowspan="1" colspan="1">Prediction of poor stroke outcome</td></tr><tr valign="top"><td rowspan="1" colspan="1">Zanotto et al [<span class="footers"><a class="citation-link" href="#ref38" rel="footnote">38</a></span>]</td><td rowspan="1" colspan="1">November 1, 2021</td><td rowspan="1" colspan="1">Brazil</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Prognosis or risk stratification, clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">SVM, naïve Bayes, random forest, KNN, CNN, transformer</td><td rowspan="1" colspan="1">Prediction of stroke outcome measurements and extraction of patient characteristics</td></tr><tr valign="top"><td rowspan="1" colspan="1">Barbour et al [<span class="footers"><a class="citation-link" href="#ref17" rel="footnote">17</a></span>]</td><td rowspan="1" colspan="1">May 21, 2019</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Prognosis or risk stratification</td><td rowspan="1" colspan="1">Rule-based</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">N/A</td><td rowspan="1" colspan="1">Risk factors for SUDEP<sup>l</sup></td></tr><tr valign="top"><td rowspan="1" colspan="1">Kim et al [<span class="footers"><a class="citation-link" href="#ref39" rel="footnote">39</a></span>]</td><td rowspan="1" colspan="1">February 28, 2019</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Radiology reports</td><td rowspan="1" colspan="1">Nonclinical neuroscience</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM, naïve Bayes, decision tree, logistic regression</td><td rowspan="1" colspan="1">Identification of acute ischemic stroke, features of acute ischemic stroke reports versus nonischemic stroke reports</td></tr><tr valign="top"><td rowspan="1" colspan="1">Davis et al [<span class="footers"><a class="citation-link" href="#ref40" rel="footnote">40</a></span>]</td><td rowspan="1" colspan="1">October 22, 2013</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes, letters, and problem lists</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">MS<sup>m</sup></td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">N/A</td><td rowspan="1" colspan="1">Extraction of clinical traits of patients with MS</td></tr><tr valign="top"><td rowspan="1" colspan="1">Glauser et al [<span class="footers"><a class="citation-link" href="#ref41" rel="footnote">41</a></span>]</td><td rowspan="1" colspan="1">January 22, 2020</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Speech</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM</td><td rowspan="1" colspan="1">Epilepsy psychiatric comorbidities</td></tr><tr valign="top"><td rowspan="1" colspan="1">Cohen et al [<span class="footers"><a class="citation-link" href="#ref42" rel="footnote">42</a></span>]</td><td rowspan="1" colspan="1">May 22, 2016</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Prognosis or risk stratification, management or therapy</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM, naïve Bayes</td><td rowspan="1" colspan="1">Identification of potential candidates for surgical intervention for pediatric drug–resistant epilepsy, performance of classification algorithm over time</td></tr><tr valign="top"><td rowspan="1" colspan="1">Alim-Marvasti et al [<span class="footers"><a class="citation-link" href="#ref43" rel="footnote">43</a></span>]</td><td rowspan="1" colspan="1">February 10, 2021</td><td rowspan="1" colspan="1">United Kingdom</td><td rowspan="1" colspan="1">Clinical notes and radiology reports</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity, prognosis or risk stratification</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM, naïve Bayes, random forest, logistic regression, boosting</td><td rowspan="1" colspan="1">Localizing the epileptogenic zone (temporal vs extra-temporal), postsurgical prognosis and outcome</td></tr><tr valign="top"><td rowspan="1" colspan="1">Balagopalan et al [<span class="footers"><a class="citation-link" href="#ref44" rel="footnote">44</a></span>]</td><td rowspan="1" colspan="1">April 27, 2021</td><td rowspan="1" colspan="1">Canada</td><td rowspan="1" colspan="1">Speech</td><td rowspan="1" colspan="1">Nonclinical neuroscience</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Alzheimer disease</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">SVM, naïve Bayes, random forest, linear regression, shallow neural network, ridge regression, transformer</td><td rowspan="1" colspan="1">Detection of Alzheimer disease from speech, prediction of MMSE<sup>n</sup></td></tr><tr valign="top"><td rowspan="1" colspan="1">Martinc et al [<span class="footers"><a class="citation-link" href="#ref45" rel="footnote">45</a></span>]</td><td rowspan="1" colspan="1">June 14, 2021</td><td rowspan="1" colspan="1">Slovenia</td><td rowspan="1" colspan="1">Speech</td><td rowspan="1" colspan="1">Nonclinical neuroscience</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Alzheimer disease</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">SVM, random forest, logistic regression, boosting, transformer</td><td rowspan="1" colspan="1">Detection of Alzheimer disease from speech</td></tr><tr valign="top"><td rowspan="1" colspan="1">Liu et al [<span class="footers"><a class="citation-link" href="#ref46" rel="footnote">46</a></span>]</td><td rowspan="1" colspan="1">April 5, 2022</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Speech</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Alzheimer disease</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Shallow neural network, transformer</td><td rowspan="1" colspan="1">Detection of Alzheimer disease from speech</td></tr><tr valign="top"><td rowspan="1" colspan="1">Nelson et al [<span class="footers"><a class="citation-link" href="#ref47" rel="footnote">47</a></span>]</td><td rowspan="1" colspan="1">December 22, 2016</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Pharmacy</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">MS</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">N/A</td><td rowspan="1" colspan="1">Identification of MS phenotype, percentages of each phenotype</td></tr><tr valign="top"><td rowspan="1" colspan="1">Deng et al [<span class="footers"><a class="citation-link" href="#ref18" rel="footnote">18</a></span>]</td><td rowspan="1" colspan="1">April 8, 2022</td><td rowspan="1" colspan="1">China</td><td rowspan="1" colspan="1">Clinical notes and radiology reports</td><td rowspan="1" colspan="1">Nonclinical neuroscience</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Management or therapy</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Transformer</td><td rowspan="1" colspan="1">Performance of system to generate ICH<sup>o</sup> treatment plan</td></tr><tr valign="top"><td rowspan="1" colspan="1">Chase et al [<span class="footers"><a class="citation-link" href="#ref48" rel="footnote">48</a></span>]</td><td rowspan="1" colspan="1">February 28, 2017</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">MS</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Naïve Bayes</td><td rowspan="1" colspan="1">Early detection of MS</td></tr><tr valign="top"><td rowspan="1" colspan="1">Wissel et al [<span class="footers"><a class="citation-link" href="#ref49" rel="footnote">49</a></span>]</td><td rowspan="1" colspan="1">November 29, 2019</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Prognosis or risk stratification, management or therapy</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM</td><td rowspan="1" colspan="1">Epilepsy surgery candidacy score</td></tr><tr valign="top"><td rowspan="1" colspan="1">Sung et al [<span class="footers"><a class="citation-link" href="#ref50" rel="footnote">50</a></span>]</td><td rowspan="1" colspan="1">February 28, 2020</td><td rowspan="1" colspan="1">Taiwan</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">SVM, random forest, decision tree, logistic regression, KNN, ensemble</td><td rowspan="1" colspan="1">Classification of ischemic stroke subtypes</td></tr><tr valign="top"><td rowspan="1" colspan="1">Sung et al [<span class="footers"><a class="citation-link" href="#ref20" rel="footnote">20</a></span>]</td><td rowspan="1" colspan="1">November 19, 2021</td><td rowspan="1" colspan="1">Taiwan</td><td rowspan="1" colspan="1">Clinical notes and radiology reports</td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Prognosis or risk stratification</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Random forest, logistic regression, transformer</td><td rowspan="1" colspan="1">Prediction of poor functional outcome after acute ischemic stroke</td></tr><tr valign="top"><td rowspan="1" colspan="1">Yang et al [<span class="footers"><a class="citation-link" href="#ref51" rel="footnote">51</a></span>]</td><td rowspan="1" colspan="1">October 20, 2020</td><td rowspan="1" colspan="1">Canada</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">MS</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Shallow neural network, CNN, RNN</td><td rowspan="1" colspan="1">Expanded disability status scale score, expanded disability status scale subscore</td></tr><tr valign="top"><td rowspan="1" colspan="1">Xie et al [<span class="footers"><a class="citation-link" href="#ref52" rel="footnote">52</a></span>]</td><td rowspan="1" colspan="1">February 22, 2022</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Transformer</td><td rowspan="1" colspan="1">Seizure freedom, seizure frequency, date of last seizure</td></tr><tr valign="top"><td rowspan="1" colspan="1">Sung et al [<span class="footers"><a class="citation-link" href="#ref53" rel="footnote">53</a></span>]</td><td rowspan="1" colspan="1">February 8, 2018</td><td rowspan="1" colspan="1">Taiwan</td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Management or therapy</td><td rowspan="1" colspan="1">Rule-based</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">N/A</td><td rowspan="1" colspan="1">Performance of EMR<sup>p</sup> interface that determines eligibility for intravenous thrombolytic therapy</td></tr><tr valign="top"><td rowspan="1" colspan="1">Sung et al [<span class="footers"><a class="citation-link" href="#ref54" rel="footnote">54</a></span>]</td><td rowspan="1" colspan="1">February 17, 2022</td><td rowspan="1" colspan="1">Taiwan</td><td rowspan="1" colspan="1">Clinical notes and radiology reports</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Prognosis or risk stratification</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Logistic regression, boosting, unspecified penalized logistic regression method, ensemble (extra trees classifier)</td><td rowspan="1" colspan="1">Prediction of poor functional outcome after acute ischemic stroke</td></tr><tr valign="top"><td rowspan="1" colspan="1">Xia et al [<span class="footers"><a class="citation-link" href="#ref55" rel="footnote">55</a></span>]</td><td rowspan="1" colspan="1">November 11, 2013</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Clinical notes and radiology reports</td><td rowspan="1" colspan="1">Nonclinical neuroscience</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">MS</td><td rowspan="1" colspan="1">Detection or diagnosis, clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Lasso regression, stepwise regression</td><td rowspan="1" colspan="1">Identification of patients with MS, severity of MS</td></tr><tr valign="top"><td rowspan="1" colspan="1">Ong et al [<span class="footers"><a class="citation-link" href="#ref22" rel="footnote">22</a></span>]</td><td rowspan="1" colspan="1">June 19, 2020</td><td rowspan="1" colspan="1">United States</td><td rowspan="1" colspan="1">Radiology reports</td><td rowspan="1" colspan="1">Nonclinical neuroscience</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Detection or diagnosis, clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Random forest, decision tree, logistic regression, KNN, RNN (LSTM)</td><td rowspan="1" colspan="1">Ischemic stroke presence, location, and acuity</td></tr><tr valign="top"><td rowspan="1" colspan="1">Roshanzamir et al [<span class="footers"><a class="citation-link" href="#ref56" rel="footnote">56</a></span>]</td><td rowspan="1" colspan="1">March 9, 2021</td><td rowspan="1" colspan="1">Iran</td><td rowspan="1" colspan="1">Speech</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Alzheimer disease</td><td rowspan="1" colspan="1">Detection or diagnosis</td><td rowspan="1" colspan="1">ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">Logistic regression, shallow neural network, CNN, RNN (LSTM) transformer</td><td rowspan="1" colspan="1">Detection of Alzheimer disease from speech</td></tr><tr valign="top"><td rowspan="1" colspan="1">Rannikmäe et al [<span class="footers"><a class="citation-link" href="#ref57" rel="footnote">57</a></span>]</td><td rowspan="1" colspan="1">June 15, 2021</td><td rowspan="1" colspan="1">United Kingdom</td><td rowspan="1" colspan="1">Radiology reports</td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">Clinical disease phenotyping or severity</td><td rowspan="1" colspan="1">Rule-based, ML</td><td rowspan="1" colspan="1">Yes</td><td rowspan="1" colspan="1">RNN</td><td rowspan="1" colspan="1">Stroke subtypes</td></tr></tbody></table><fn id="table1fn1"><p><sup>a</sup>NLP: natural language processing.</p></fn><fn id="table1fn2"><p><sup>b</sup>ML: machine learning.</p></fn><fn id="table1fn3"><p><sup>c</sup>KNN: k-nearest neighbor.</p></fn><fn id="table1fn4"><p><sup>d</sup>MLP: multilayer perceptron.</p></fn><fn id="table1fn5"><p><sup>e</sup>PNES: psychogenic nonepileptic seizures.</p></fn><fn id="table1fn6"><p><sup>f</sup>SVM: support vector machine.</p></fn><fn id="table1fn7"><p><sup>g</sup>CNN: convolutional neural network.</p></fn><fn id="table1fn8"><p><sup>h</sup>RNN: recurrent neural network.</p></fn><fn id="table1fn9"><p><sup>i</sup>LSTM: long- and short-term memory network.</p></fn><fn id="table1fn10"><p><sup>j</sup>N/A: Not applicable.</p></fn><fn id="table1fn11"><p><sup>k</sup>TIA: transient ischemic attack.</p></fn><fn id="table1fn12"><p><sup>l</sup>SUDEP: sudden unexpected death in epilepsy.</p></fn><fn id="table1fn13"><p><sup>m</sup>MS: multiple sclerosis.</p></fn><fn id="table1fn14"><p><sup>n</sup>MMSE: Mini-Mental Status Examination.</p></fn><fn id="table1fn15"><p><sup>o</sup>ICH: intracerebral hemorrhage.</p></fn><fn id="table1fn16"><p><sup>p</sup>EMR: electronic medical record.</p></fn></div><div class="figure-table"><figcaption><span class="typcn typcn-clipboard"></span><b>Table 2. </b>Overall study characteristics: journal field, target of NLP<sup>a</sup>, and neurological condition.</figcaption><table width="1000" cellpadding="5" cellspacing="0" border="1" rules="groups" frame="hsides"><col width="30" span="1"><col width="670" span="1"><col width="300" span="1"><thead><tr valign="top"><td colspan="2" rowspan="1">Study characteristics</td><td rowspan="1" colspan="1">Studies (n=41), n (%)</td></tr></thead><tbody><tr valign="top"><td colspan="3" rowspan="1"><b>Condition</b></td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Stroke</td><td rowspan="1" colspan="1">20 (49)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Epilepsy</td><td rowspan="1" colspan="1">10 (24)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Alzheimer disease</td><td rowspan="1" colspan="1">6 (15)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Multiple sclerosis</td><td rowspan="1" colspan="1">5 (12)</td></tr><tr valign="top"><td colspan="3" rowspan="1"><b>Target of NLP</b></td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Diagnosis</td><td rowspan="1" colspan="1">20 (49)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Phenotyping</td><td rowspan="1" colspan="1">17 (42)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Prognosis</td><td rowspan="1" colspan="1">9 (22)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Therapy</td><td rowspan="1" colspan="1">4 (10)</td></tr><tr valign="top"><td colspan="3" rowspan="1"><b>Journal field</b></td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Medical informatics</td><td rowspan="1" colspan="1">15 (37)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Clinical neurology</td><td rowspan="1" colspan="1">14 (34)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Nonclinical neuroscience</td><td rowspan="1" colspan="1">7 (17)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Clinical medicine</td><td rowspan="1" colspan="1">2 (5)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Other<sup>b</sup></td><td rowspan="1" colspan="1">3 (7)</td></tr></tbody></table><fn id="table2fn1"><p><sup>a</sup>NLP: natural language processing.</p></fn><fn id="table2fn2"><p><sup>b</sup>Other includes studies published in pharmacy, public health, and neuroradiology journals.</p></fn></div><p class="abstract-paragraph">Of the 41 studies, the language sources for NLP comprised clinical notes (n=25, 61%); radiology reports (n=14, 34%); speech (n=8, 20%); and other sources (n=2, 5%) that included echocardiography reports, letters to referring providers, and problem lists (<span class="footers"><a class="citation-link" href="#table3" rel="footnote">Table 3</a></span>). Of studies with speech as the language source, half (4/8, 50%) analyzed transcripts only, whereas half additionally incorporated acoustic features from the audio files themselves. These transcripts and audio files were largely from research datasets (eg, ADReSS and Pitt corpus). Two studies analyzed transcripts from interviews with patients. In the study including problem lists, it is unknown who reported the problems.</p><div class="figure-table"><figcaption><span class="typcn typcn-clipboard"></span><b>Table 3. </b>Overall study characteristics: NLP<sup>a</sup> methods and language sources.</figcaption><table width="1000" cellpadding="5" cellspacing="0" border="1" rules="groups" frame="hsides"><col width="30" span="1"><col width="670" span="1"><col width="300" span="1"><thead><tr valign="top"><td colspan="2" rowspan="1">Study characteristics</td><td rowspan="1" colspan="1">Studies (n=41), n (%)</td></tr></thead><tbody><tr valign="top"><td colspan="3" rowspan="1"><b>NLP method</b></td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Rule-based</td><td rowspan="1" colspan="1">23 (56)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Machine learning</td><td rowspan="1" colspan="1">35 (85)</td></tr><tr valign="top"><td colspan="3" rowspan="1"><b>Type of</b><b>machine learning</b></td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Conventional machine learning</td><td rowspan="1" colspan="1">31 (76)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Deep learning</td><td rowspan="1" colspan="1">16 (39)</td></tr><tr valign="top"><td colspan="3" rowspan="1"><b>Source text</b></td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Clinical notes</td><td rowspan="1" colspan="1">25 (61)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Radiology reports</td><td rowspan="1" colspan="1">14 (34)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Speech</td><td rowspan="1" colspan="1">8 (20)</td></tr><tr valign="top"><td rowspan="1" colspan="1"><br></td><td rowspan="1" colspan="1">Other<sup>b</sup></td><td rowspan="1" colspan="1">2 (5)</td></tr></tbody></table><fn id="table3fn1"><p><sup>a</sup>NLP: natural language processing.</p></fn><fn id="table3fn2"><p><sup>b</sup>Other includes echocardiography reports, problem lists, and letters to referring providers.</p></fn></div><p class="abstract-paragraph">Of the 41 studies, the most common source language for NLP was English (n=39, 95%), Portuguese in 1 (2%) study, and unspecified in the remaining 1 study (which was of Chinese nationality, not multicentric). When patient population size was recorded, the median was 1091 (IQR 188-4211). In studies that did not specify a population size (n=4, 10%), the median number of clinical or radiographic notes was 2172 (IQR 1155.5-22,018.0).</p><p class="abstract-paragraph">Papers were most commonly published in medical informatics (n=15, 37%) journals, followed closely by clinical neurology (n=14, 34%) journals. Seven (17%) studies were published in nonclinical neuroscience journals; 2 (5%) in clinical medicine journals; and 1 (2%) each in neuroradiology, public health, and pharmacy journals. Studies were mostly conducted in the United States (n=21, 51%), followed by Taiwan (n=4, 10%) and the United Kingdom, Canada, and Australia (n=3, 7% each). Two (5%) studies were conducted in China, and 1 (2%) study was conducted in each of South Korea, Brazil, Iran, India, and Slovenia (<span class="footers"><a class="citation-link" href="#figure2" rel="footnote">Figure 2</a></span>).</p><figure><a name="figure2">‎</a><a class="fancybox" title="Figure 2. Proportion of included studies (n=41), organized according to country of origin: the United States (n=21, 51%); Taiwan (n=4, 10%); the United Kingdom, Canada, and Australia (n=3, 7% each); China (n=2, 5%); and South Korea, Brazil, Iran, India, and Slovenia (n=1, 2% each)." href="https://asset.jmir.pub/assets/13d0fcca15153fd67444eab827f368fb.png" id="figure2"><img class="figure-image" src="https://asset.jmir.pub/assets/13d0fcca15153fd67444eab827f368fb.png"></a><figcaption><span class="typcn typcn-image"></span><b>Figure 2. </b> Proportion of included studies (n=41), organized according to country of origin: the United States (n=21, 51%); Taiwan (n=4, 10%); the United Kingdom, Canada, and Australia (n=3, 7% each); China (n=2, 5%); and South Korea, Brazil, Iran, India, and Slovenia (n=1, 2% each). </figcaption></figure><p class="abstract-paragraph">Only 6 (15%) studies used strictly rule-based methods. The majority of studies incorporated ML (n=35, 85%), either exclusively (n=18, 44%) or in combination with rule-based methods (n=17, 41%). Of the studies that used ML, most (n=31, 89%) used conventional ML methods, whereas 16 (46%) used DL approaches (<span class="footers"><a class="citation-link" href="#table3" rel="footnote">Table 3</a></span>), and 12 (34%) used a combination of both conventional ML and DL approaches.</p><p class="abstract-paragraph">As shown in <span class="footers"><a class="citation-link" href="#figure3" rel="footnote">Figure 3</a></span>, the most frequently used conventional ML algorithms were random forest (n=18, 58%), SVM (n=15, 48%), and logistic regression (n=15, 48%) models. Among studies using DL approaches, transformers (n=10, 63%) were the most commonly used algorithm, followed by convolutional neural networks and RNNs (each n=7, 44%). The co-occurrence of random forest and transformer algorithms was a prevalent trend in research combining traditional ML with DL methodologies (n=6, 15%). Studies that used DL only began to appear in 2019 and later (<span class="footers"><a class="citation-link" href="#figure4" rel="footnote">Figure 4</a></span>). The most often reported performance metrics for ML models were precision or recall (n=31, 76%), accuracy (n=22, 54%), area under the receiver operating curve (n=20, 49%), and <i>F</i><sub>1</sub>-score (n=19, 46%).</p><figure><a name="figure3">‎</a><a class="fancybox" title="Figure 3. Relative proportions of machine learning algorithms used by the included NLP models. CNN: convolutional neural network; KNN: k-nearest neighbor; LSTM: long- and short-term memory networks; MLP: multilayer perceptron; RNN: recurrent neural network; SVM: support vector machine. *Other includes stepwise regression, ridge regression, an unspecified penalized regression method, latent Dirichlet allocation, and an unspecified neural network with an unspecified number of layers." href="https://asset.jmir.pub/assets/7f2f9de6933564bb7d73b0d8e8710712.png" id="figure3"><img class="figure-image" src="https://asset.jmir.pub/assets/7f2f9de6933564bb7d73b0d8e8710712.png"></a><figcaption><span class="typcn typcn-image"></span><b>Figure 3. </b> Relative proportions of machine learning algorithms used by the included NLP models. CNN: convolutional neural network; KNN: k-nearest neighbor; LSTM: long- and short-term memory networks; MLP: multilayer perceptron; RNN: recurrent neural network; SVM: support vector machine. *Other includes stepwise regression, ridge regression, an unspecified penalized regression method, latent Dirichlet allocation, and an unspecified neural network with an unspecified number of layers. </figcaption></figure><figure><a name="figure4">‎</a><a class="fancybox" title="Figure 4. Number of studies applying natural language processing (NLP) to neurological conditions, stratified by NLP methodology and publication year." href="https://asset.jmir.pub/assets/7fc5c1273a1b60a24f3f10d91c1fca43.png" id="figure4"><img class="figure-image" src="https://asset.jmir.pub/assets/7fc5c1273a1b60a24f3f10d91c1fca43.png"></a><figcaption><span class="typcn typcn-image"></span><b>Figure 4. </b> Number of studies applying natural language processing (NLP) to neurological conditions, stratified by NLP methodology and publication year. </figcaption></figure><p class="abstract-paragraph">All 41 studies were model derivation studies, with only 7 (17%) studies conducting additional external validation (<span class="footers"><a class="citation-link" href="#app2" rel="footnote">Multimedia Appendix 2</a></span>). Furthermore, nearly all the study models were developed retrospectively and were not applied in practice or deployed in real-world environments, except for 3 studies. A study by Li et al [<span class="footers"><a class="citation-link" href="#ref16" rel="footnote">16</a></span>] developed a model for stroke detection from imaging reports and then applied it to quantify the change in stroke cases before and during the COVID-19 pandemic. A second by Sung et al [<span class="footers"><a class="citation-link" href="#ref53" rel="footnote">53</a></span>], also in the stroke category, evaluated the deployment of a user-interface system to determine intravenous thrombolysis eligibility built on the NLP model devised. A third study by Wissel et al [<span class="footers"><a class="citation-link" href="#ref49" rel="footnote">49</a></span>] created a model to identify surgical resection candidates in adult patients with epilepsy. The model was retrained prospectively to incorporate new information.</p><h4>Study Characteristics, Stratified by Condition</h4><p class="abstract-paragraph">In studies focused on Alzheimer dementia, diagnosis and detection was the only target of NLP (6/6, 100%). Disease phenotyping and subtyping was the most common purpose of NLP in stroke (10/20, 50%) and MS (4/5, 80%), whereas prognostication was seen as often as diagnosis in epilepsy studies (4/10, 40%; Figure S9 in <span class="footers"><a class="citation-link" href="#app2" rel="footnote">Multimedia Appendix 2</a></span>). Studies that applied NLP for the purpose of disease treatment or management were limited to stroke and epilepsy (Figure S9 in <span class="footers"><a class="citation-link" href="#app2" rel="footnote">Multimedia Appendix 2</a></span>).</p><p class="abstract-paragraph">Rule-based methods were used across all studies, except for Alzheimer dementia, in which only ML approaches were used (Figure S10 in <span class="footers"><a class="citation-link" href="#app2" rel="footnote">Multimedia Appendix 2</a></span>). Conventional ML methods were used most often by Alzheimer dementia studies (5/6, 83%), followed by stroke (16/20, 80%). Similarly, DL methods were used predominantly by Alzheimer dementia (6/6, 100%) and stroke (8/20, 40%) studies (Figure S10 in <span class="footers"><a class="citation-link" href="#app2" rel="footnote">Multimedia Appendix 2</a></span>). The transformer was the DL method used most frequently in Alzheimer disease-related studies (5/6, 83%).</p><br><h3 class="navigation-heading h3-main-heading" id="Discussion" data-label="Discussion">Discussion</h3><h4>Principal Findings</h4><p class="abstract-paragraph">In this scoping review, 41 studies [<span class="footers"><a class="citation-link" href="#ref13" rel="footnote">13</a></span>,<span class="footers"><a class="citation-link" href="#ref16" rel="footnote">16</a></span>-<span class="footers"><a class="citation-link" href="#ref22" rel="footnote">22</a></span>,<span class="footers"><a class="citation-link" href="#ref25" rel="footnote">25</a></span>-<span class="footers"><a class="citation-link" href="#ref57" rel="footnote">57</a></span>] that investigated direct clinical applications of NLP to common neurological disorders were identified. We found that the majority of these studies focused on detection and diagnosis and applied NLP to stroke, whereas we found no studies of NLP that met our eligibility criteria in the clinical areas of migraine or Parkinson disease. Methodologically, ML techniques were used more often than rule-based methods, but a considerable number of studies still relied on rule-based approaches in combination with ML. While we observed that DL began to emerge as a methodology for NLP in 2019, we found that the transformer was the most commonly used DL algorithm overall.</p><p class="abstract-paragraph">At the time of writing, we believe our scoping review to be the first to examine direct clinical NLP applications in common neurological conditions. One prior review [<span class="footers"><a class="citation-link" href="#ref58" rel="footnote">58</a></span>] investigated NLP applications across the combined clinical specialties of neurosurgery, spine surgery, and neurology, whereas another evaluated the use of NLP in both psychiatry and clinical neuroscience [<span class="footers"><a class="citation-link" href="#ref59" rel="footnote">59</a></span>]. However, neither reviews analyzed studies and NLP applications according to neurological condition. More importantly, these reviews included many studies where NLP was not applied for direct clinical use, instead aiming to perform tasks such as characterizing patient cohorts [<span class="footers"><a class="citation-link" href="#ref58" rel="footnote">58</a></span>], analyzing information extraction, or determining causal inference between concepts [<span class="footers"><a class="citation-link" href="#ref59" rel="footnote">59</a></span>]. In contrast to this prior work, our review focused on direct clinical applications of NLP.</p><p class="abstract-paragraph">Of note, we found no studies applying NLP to migraine or Parkinson disease that met our eligibility criteria, thereby highlighting a potential gap in NLP research focusing on these disorders. This is perhaps unexpected, as the combined prevalence of migraine and Parkinson disease in the United States exceeds that of both stroke and MS [<span class="footers"><a class="citation-link" href="#ref12" rel="footnote">12</a></span>]. Two explanations may account for this finding. One is that migraine and Parkinson disease may rely less on radiographic imaging studies and their reports to establish a diagnosis than stroke, Alzheimer dementia, or MS. Given that many ML applications in stroke have focused on neuroimaging [<span class="footers"><a class="citation-link" href="#ref60" rel="footnote">60</a></span>], it is plausible that stroke imaging reports could represent an important source of data for NLP analyses. Indeed, the results of our review demonstrate that stroke-related NLP studies made use of radiographic reports as often as clinical notes for source text, which could have resulted in a relatively higher number of NLP studies within stroke than in other neurological conditions.</p><p class="abstract-paragraph">A second explanation may be that Alzheimer disease is a more common cause of dementia worldwide than dementing syndromes associated with Parkinson disease [<span class="footers"><a class="citation-link" href="#ref61" rel="footnote">61</a></span>] and has in turn garnered a larger proportion of research funding. National Institutes of Health [<span class="footers"><a class="citation-link" href="#ref62" rel="footnote">62</a></span>] research funding for Alzheimer dementia was approximately US $3 billion in 2022, as compared to US $259 million for Parkinson disease.</p><p class="abstract-paragraph">Our finding that NLP was most frequently applied to diagnostic problems is expected, given that clinical decision support is a common focus of artificial intelligence in medicine [<span class="footers"><a class="citation-link" href="#ref63" rel="footnote">63</a></span>]. Historically, clinical decision support has also played an important role in medical informatics by constituting the main focus of archetypal systems such as MYCIN, INTERNIST-1, and DXplain, which were first developed in the 1970s and 1980s [<span class="footers"><a class="citation-link" href="#ref64" rel="footnote">64</a></span>]. An alternative explanation is that the shortage of neurologists that already exists worldwide [<span class="footers"><a class="citation-link" href="#ref65" rel="footnote">65</a></span>] may have potentially created a more urgent need for detection-oriented NLP applications rather than NLP applications targeting therapeutic management or prognostication.</p><p class="abstract-paragraph">Though diagnosis was the most common target of NLP overall, we found that epilepsy-related studies focused as much on prognostication as they did on diagnostic tasks. Given that roughly one-third of all patients with epilepsy are drug resistant [<span class="footers"><a class="citation-link" href="#ref66" rel="footnote">66</a></span>], determining good surgical resection candidates as well as predicting surgical outcomes are important objectives that have been the focus of considerable research [<span class="footers"><a class="citation-link" href="#ref67" rel="footnote">67</a></span>]. Consistent with this, the epilepsy-related studies in the prognostication category were directed toward identifying adult [<span class="footers"><a class="citation-link" href="#ref49" rel="footnote">49</a></span>] and pediatric [<span class="footers"><a class="citation-link" href="#ref42" rel="footnote">42</a></span>] surgical candidates, predicting postsurgical outcomes [<span class="footers"><a class="citation-link" href="#ref43" rel="footnote">43</a></span>], and detecting risk factors for sudden unexpected death in epilepsy [<span class="footers"><a class="citation-link" href="#ref17" rel="footnote">17</a></span>].</p><p class="abstract-paragraph">With respect to the types of ML models we found in our review, the relatively high proportion of conventional ML-based studies using random forest and SVM (18/31, 58% and 15/31, 48%, respectively) may have been related to the fact that SVM together with random forest models generally represented the dominant ML techniques prior to the advent of neural networks [<span class="footers"><a class="citation-link" href="#ref68" rel="footnote">68</a></span>] in diagnostic and clinical decision support applications [<span class="footers"><a class="citation-link" href="#ref63" rel="footnote">63</a></span>,<span class="footers"><a class="citation-link" href="#ref69" rel="footnote">69</a></span>,<span class="footers"><a class="citation-link" href="#ref70" rel="footnote">70</a></span>]. Despite its position as a potentially more basic classification method than either SVM or random forest, logistic regression was used as commonly as SVM in our analysis.</p><p class="abstract-paragraph">Furthermore, while we found that SVM and random forest models were common in ML-based NLP approaches, the optimal problems these models address are fundamentally different. SVM generally works best as a binary classifier, whereas random forest models are best used for classification tasks involving multiple categories [<span class="footers"><a class="citation-link" href="#ref71" rel="footnote">71</a></span>]. We found that the most frequently used ML algorithms in stroke-related NLP studies were random forest models. This matches the most frequent target of NLP in stroke-related studies, which was disease subtyping (a multiple classification problem).</p><p class="abstract-paragraph">Among DL algorithms, which are becoming increasingly widespread in NLP [<span class="footers"><a class="citation-link" href="#ref72" rel="footnote">72</a></span>], the transformer was the most commonly used technique we identified. Unlike other word embedding methods, a transformer processes a whole sequence of text while preserving the context and meaning of words [<span class="footers"><a class="citation-link" href="#ref59" rel="footnote">59</a></span>,<span class="footers"><a class="citation-link" href="#ref73" rel="footnote">73</a></span>]. Another significant advantage of transformers is that they can use transfer learning, which first trains a model on a learning task and then applies the model to a separate but closely related task [<span class="footers"><a class="citation-link" href="#ref58" rel="footnote">58</a></span>,<span class="footers"><a class="citation-link" href="#ref74" rel="footnote">74</a></span>]. A prevalent example of transfer learning in our results is Bidirectional Encoder Representations From Transformers (BERT), a transformer model that was originally trained using publicly available text from Wikipedia and BookCorpus, a collection of free, unpublished novels consisting of over 50 million sentences [<span class="footers"><a class="citation-link" href="#ref75" rel="footnote">75</a></span>,<span class="footers"><a class="citation-link" href="#ref76" rel="footnote">76</a></span>]. BERT can then be further refined on a target training task and dataset before being passed to a separate classification algorithm [<span class="footers"><a class="citation-link" href="#ref28" rel="footnote">28</a></span>]. This is helpful in situations where the target training set is small [<span class="footers"><a class="citation-link" href="#ref28" rel="footnote">28</a></span>]. The high frequency of Alzheimer disease–related NLP studies we found using BERT is expected within this context, as these studies often used the ADReSS speech dataset that consists of only 78 healthy controls and 78 patients with Alzheimer disease [<span class="footers"><a class="citation-link" href="#ref28" rel="footnote">28</a></span>,<span class="footers"><a class="citation-link" href="#ref45" rel="footnote">45</a></span>].</p><p class="abstract-paragraph">A particularly important finding of our review is that although many of the NLP studies leveraged powerful and sophisticated computational tools, most studies constitute research work rather than reports of operationalization or evaluation in practical settings. This is consistent with the current state of clinical NLP outside of neurology, wherein real-world deployment of NLP models continues to be limited [<span class="footers"><a class="citation-link" href="#ref7" rel="footnote">7</a></span>,<span class="footers"><a class="citation-link" href="#ref77" rel="footnote">77</a></span>,<span class="footers"><a class="citation-link" href="#ref78" rel="footnote">78</a></span>].</p><p class="abstract-paragraph">One major obstacle to the implementation of NLP in clinical practice is model generalizability [<span class="footers"><a class="citation-link" href="#ref7" rel="footnote">7</a></span>]. Published NLP models are usually internally validated rather than externally validated [<span class="footers"><a class="citation-link" href="#ref7" rel="footnote">7</a></span>,<span class="footers"><a class="citation-link" href="#ref17" rel="footnote">17</a></span>], limiting the understanding of model accuracy beyond the model’s original training environment [<span class="footers"><a class="citation-link" href="#ref60" rel="footnote">60</a></span>]. We found this to be true for the majority of studies identified in our review. The lack of EMR standardization, including note formatting [<span class="footers"><a class="citation-link" href="#ref17" rel="footnote">17</a></span>,<span class="footers"><a class="citation-link" href="#ref78" rel="footnote">78</a></span>], documentation styles, and radiographic report structures across different medical institutions [<span class="footers"><a class="citation-link" href="#ref7" rel="footnote">7</a></span>] and between clinicians, may partly account for our observations. Furthermore, the preponderance of English language as source text in NLP [<span class="footers"><a class="citation-link" href="#ref79" rel="footnote">79</a></span>], as demonstrated by the single study in our review using non-English (Portuguese) text for analysis, suggests that the generalizability of NLP within neurology is most likely limited outside the English language.</p><p class="abstract-paragraph">Another major obstacle impeding the adoption of NLP tools is the inherent lack of transparency of ML-based algorithms [<span class="footers"><a class="citation-link" href="#ref60" rel="footnote">60</a></span>], particularly artificial neural networks and other forms of DL approaches [<span class="footers"><a class="citation-link" href="#ref80" rel="footnote">80</a></span>]. These approaches have low transparency because the computational methods they use to characterize relationships between inputs and outputs are not readily intelligible to humans [<span class="footers"><a class="citation-link" href="#ref7" rel="footnote">7</a></span>,<span class="footers"><a class="citation-link" href="#ref78" rel="footnote">78</a></span>,<span class="footers"><a class="citation-link" href="#ref80" rel="footnote">80</a></span>] acting as a black box that could undermine clinicians’ trust in their performance.</p><p class="abstract-paragraph">The lack of well-defined regulatory guidelines and standards overseeing the artificial intelligence space [<span class="footers"><a class="citation-link" href="#ref81" rel="footnote">81</a></span>] has furthered this mistrust. Compromise of personal health data, algorithmic bias, and the question of how to attribute culpability when diagnostic errors arise [<span class="footers"><a class="citation-link" href="#ref82" rel="footnote">82</a></span>,<span class="footers"><a class="citation-link" href="#ref83" rel="footnote">83</a></span>] are all ethical concerns that may serve to explain the relative paucity of studies across all neurological conditions that externally validated DL models.</p><p class="abstract-paragraph">Finally, the lack of portability of NLP applications into external EMRs is another factor that has restricted the development of NLP models to the research arena. External software modules containing ML and DL models are challenging to integrate into EMRs [<span class="footers"><a class="citation-link" href="#ref1" rel="footnote">1</a></span>,<span class="footers"><a class="citation-link" href="#ref84" rel="footnote">84</a></span>], as most implementations require a high level of computing infrastructure and technical expertise that many hospital information technology systems and personnel may lack [<span class="footers"><a class="citation-link" href="#ref84" rel="footnote">84</a></span>]. Recent work suggests few EMR-integrated aggregative tools exist to display NLP findings to clinicians in a digestible format [<span class="footers"><a class="citation-link" href="#ref85" rel="footnote">85</a></span>]. To address these barriers, some authors have advocated for collaborations between NLP researchers and EMR companies [<span class="footers"><a class="citation-link" href="#ref77" rel="footnote">77</a></span>].</p><h4>Limitations and Future Work</h4><p class="abstract-paragraph">Our scoping review has several limitations. First, we note that the target of NLP was categorized according to author experience and interpretation of the literature, which may have underreported the application of the published NLP algorithms. Second, due to the variable performance metrics and outcomes across studies, we did not aggregate measurements of performance in our review, and we therefore could not reliably provide summary performance metrics for NLP models within individual diseases, applications, or outcomes. Future work should focus on individual outcomes within a clinical disorder for a more exact appraisal of NLP model performance than this review.</p><p class="abstract-paragraph">Third, this review only included studies based on common neurological disorders, direct clinical applications of NLP, and homogeneous clinical populations, which limited the number of studies we identified. It is therefore important to note that this review cannot be used to make definitive conclusions on the state of NLP research across all neurological disorders. Future efforts can be directed at characterizing the use of NLP across less common neurological disorders as well as in heterogeneous or ambiguously defined clinical populations. As NLP technologies continue to advance, it will also be critically important to evaluate studies that use newer transformers, such as GPT3, which have better performance than BERT models [<span class="footers"><a class="citation-link" href="#ref59" rel="footnote">59</a></span>].</p><h4>Conclusions</h4><p class="abstract-paragraph">The abundance of unstructured text data in modern-day EMRs as well as the emphasis in neurology on narrative history and physical examination and heavy reliance on ancillary information such as radiographic reports and speech, all create an optimal use case for applying NLP for the diagnosis, management, or prognostication of neurological disorders. To our knowledge, this is the first attempt to systematically characterize research efforts to investigate direct NLP applications to common neurological conditions. Our review reveals gaps in neurological NLP research, showing a relative deficiency of NLP studies in subspecialties outside of stroke or epilepsy, and underlines the need to actualize NLP models outside of the research phase. Moreover, the current emphasis of NLP on diagnostic tasks suggests that NLP may be particularly useful in settings that lack access to neurological expertise.</p></article><h4>Funding</h4><p class="abstract-paragraph">None.</p><h4 class="h4-border-top">Conflicts of Interest</h4><p><p class="abstract-paragraph">NJ receives an honorarium for her work as an associate editor of Epilepsia. There are no other conflicts of interest to report.</p></p><div id="app1" name="app1">Multimedia Appendix 1<p class="abstract-paragraph">PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews): checklist and explanation.</p><a href="https://jmir.org/api/download?alt_name=neuro_v3i1e51822_app1.pdf&filename=b857986a2f57ecb484f5e9759b0c26b0.pdf" target="_blank">PDF File (Adobe PDF File), 546 KB</a></div><hr><div id="app2" name="app2">Multimedia Appendix 2<p class="abstract-paragraph">Search strategy and additional data.</p><a href="https://jmir.org/api/download?alt_name=neuro_v3i1e51822_app2.docx&filename=6b7610b4e5c37871e8cca3eedfe7720c.docx" target="_blank">DOCX File , 756 KB</a></div><hr><div class="footnotes"><h4 id="References" class="h4-border-top navigation-heading" data-label="References">References</h4><ol><li><span id="ref1">Pivovarov R, Elhadad N. Automated methods for the summarization of electronic health records. J Am Med Inform Assoc. 2015;22(5):938-947. [<a href="https://europepmc.org/abstract/MED/25882031" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1093/jamia/ocv032">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=25882031&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref2">Locke S, Bashall A, Al-Adely S, Moore J, Wilson A, Kitchen GB. Natural language processing in medicine: a review. Trends Anaesth Crit Care. 2021;38:4-9. [<a target="_blank" href="https://dx.doi.org/10.1016/j.tacc.2021.02.007">CrossRef</a>]</span></li><li><span id="ref3">Kimia AA, Savova G, Landschaft A, Harper MB. An introduction to natural language processing: how you can get more from those electronic notes you are generating. Pediatr Emerg Care. 2015;31(7):536-541. [<a target="_blank" href="https://dx.doi.org/10.1097/PEC.0000000000000484">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=26148107&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref4">Iroju OG, Olaleke JO. A systematic review of natural language processing in healthcare. Int J Inf Technol Comput Sci. 2015;7(8):44-50. [<a target="_blank" href="https://dx.doi.org/10.5815/ijitcs.2015.08.07">CrossRef</a>]</span></li><li><span id="ref5">Wang J, Deng H, Liu B, Hu A, Liang J, Fan L, et al. Systematic evaluation of research progress on natural language processing in medicine over the past 20 years: bibliometric study on PubMed. J Med Internet Res. 2020;22(1):e16816. [<a href="https://www.jmir.org/2020/1/e16816/" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.2196/16816">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=32012074&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref6">Gunter D, Puac-Polanco P, Miguel O, Thornhill RE, Yu AYX, Liu ZA, et al. Rule-based natural language processing for automation of stroke data extraction: a validation study. Neuroradiology. 2022;64(12):2357-2362. [<a target="_blank" href="https://dx.doi.org/10.1007/s00234-022-03029-1">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35913525&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref7">Pons E, Braun LMM, Hunink MGM, Kors JA. Natural language processing in radiology: a systematic review. Radiology. 2016;279(2):329-343. [<a target="_blank" href="https://dx.doi.org/10.1148/radiol.16142770">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=27089187&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref8">Ohno-Machado L, Nadkarni P, Johnson K. Natural language processing: algorithms and tools to extract computable information from EHRs and from the biomedical literature. J Am Med Inform Assoc. 2013;20(5):805. [<a href="https://europepmc.org/abstract/MED/23935077" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1136/amiajnl-2013-002214">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23935077&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref9">Zhou B, Yang G, Shi Z, Ma S. Natural language processing for smart healthcare. IEEE Rev Biomed Eng. 2024;17:4-18. [<a target="_blank" href="https://dx.doi.org/10.1109/RBME.2022.3210270">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=36170385&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref10">Le Glaz A, Haralambous Y, Kim-Dufor DH, Lenca P, Billot R, Ryan TC, et al. Machine learning and natural language processing in mental health: systematic review. J Med Internet Res. 2021;23(5):e15708. [<a href="https://www.jmir.org/2021/5/e15708/" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.2196/15708">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33944788&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref11">Chin JH, Vora N. The global burden of neurologic diseases. Neurology. 2014;83(4):349-351. [<a href="https://europepmc.org/abstract/MED/25049303" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1212/WNL.0000000000000610">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=25049303&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref12">Gooch CL, Pracht E, Borenstein AR. The burden of neurological disease in the United States: a summary report and call to action. Ann Neurol. 2017;81(4):479-484. [<a target="_blank" href="https://dx.doi.org/10.1002/ana.24897">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=28198092&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref13">Yu AYX, Liu ZA, Pou-Prom C, Lopes K, Kapral MK, Aviv RI, et al. Automating stroke data extraction from free-text radiology reports using natural language processing: instrument validation study. JMIR Med Inform. 2021;9(5):e24381. [<a href="https://medinform.jmir.org/2021/5/e24381/" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.2196/24381">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33944791&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref14">Pinter NK, Fritz JV. Neuroimaging for the neurologist: clinical MRI and future trends. Neurol Clin. 2020;38(1):1-35. [<a href="https://www.sciencedirect.com/science/article/abs/pii/S0733861919300738?via%3Dihub" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1016/j.ncl.2019.08.002">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=31761054&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref15">de la Fuente Garcia S, Ritchie CW, Luz S. Artificial intelligence, speech, and language processing approaches to monitoring Alzheimer's disease: a systematic review. J Alzheimers Dis. 2020;78(4):1547-1574. [<a href="https://europepmc.org/abstract/MED/33185605" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3233/JAD-200888">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33185605&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref16">Li MD, Lang M, Deng F, Chang K, Buch K, Rincon S, et al. Analysis of stroke detection during the COVID-19 pandemic using natural language processing of radiology reports. AJNR Am J Neuroradiol. 2021;42(3):429-434. [<a href="http://www.ajnr.org/cgi/pmidlookup?view=long&pmid=33334851" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3174/ajnr.A6961">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33334851&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref17">Barbour K, Hesdorffer DC, Tian N, Yozawitz EG, McGoldrick PE, Wolf S, et al. Automated detection of sudden unexpected death in epilepsy risk factors in electronic medical records using natural language processing. Epilepsia. 2019;60(6):1209-1220. [<a href="https://onlinelibrary.wiley.com/doi/10.1111/epi.15966" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1111/epi.15966">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=31111463&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref18">Deng B, Zhu W, Sun X, Xie Y, Dan W, Zhan Y, et al. Development and validation of an automatic system for intracerebral hemorrhage medical text recognition and treatment plan output. Front Aging Neurosci. 2022;14:798132. [<a href="https://europepmc.org/abstract/MED/35462698" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3389/fnagi.2022.798132">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35462698&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref19">Miller MI, Orfanoudaki A, Cronin M, Saglam H, So Yeon Kim I, Balogun O, et al. Natural language processing of radiology reports to detect complications of ischemic stroke. Neurocrit Care. 2022;37(Suppl 2):291-302. [<a href="https://europepmc.org/abstract/MED/35534660" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1007/s12028-022-01513-3">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35534660&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref20">Sung SF, Chen CH, Pan RC, Hu YH, Jeng JS. Natural language processing enhances prediction of functional outcome after acute ischemic stroke. JAHA. 2021;10(24):e023486. [<a href="https://www.ahajournals.org/doi/10.1161/JAHA.121.023486" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1161/jaha.121.023486">CrossRef</a>]</span></li><li><span id="ref21">Zhao Y, Fu S, Bielinski SJ, Decker PA, Chamberlain AM, Roger VL, et al. Natural language processing and machine learning for identifying incident stroke from electronic health records: algorithm development and validation. J Med Internet Res. 2021;23(3):e22951. [<a href="https://www.jmir.org/2021/3/e22951/" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.2196/22951">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33683212&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref22">Ong CJ, Orfanoudaki A, Zhang R, Caprasse FPM, Hutch M, Ma L, et al. Machine learning and natural language processing methods to identify ischemic stroke, acuity and location from radiology reports. PLoS One. 2020;15(6):e0234908. [<a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0234908" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1371/journal.pone.0234908">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=32559211&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref23">Skansi S. Introduction to Deep Learning: From Logical Calculus to Artificial Intelligence. 1st Edition. Cham. Springer International Publishing; 2018. </span></li><li><span id="ref24">Mikolov T, Chen K, Corrado G, Dean J. Efficient estimation of word representations in vector space. ArXiv. Preprint posted online on January 16, 2013. . [<a href="https://arxiv.org/abs/1301.3781" target="_blank">FREE Full text</a>]</span></li><li><span id="ref25">Lay J, Seneviratne U, Fok A, Roberts H, Phan T. Discovering themes in medical records of patients with psychogenic non-epileptic seizures. BMJ Neurol Open. 2020;2(2):e000087. [<a href="https://europepmc.org/abstract/MED/33681804" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1136/bmjno-2020-000087">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33681804&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref26">Mayampurath A, Parnianpour Z, Richards CT, Meurer WJ, Lee J, Ankenman B, et al. Improving prehospital stroke diagnosis using natural language processing of paramedic reports. Stroke. 2021;52(8):2676-2679. [<a href="https://europepmc.org/abstract/MED/34162217" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1161/STROKEAHA.120.033580">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=34162217&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref27">Lineback CM, Garg R, Oh E, Naidech AM, Holl JL, Prabhakaran S. Prediction of 30-day readmission after stroke using machine learning and natural language processing. Front Neurol. 2021;12:649521. [<a href="https://europepmc.org/abstract/MED/34326805" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3389/fneur.2021.649521">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=34326805&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref28">Liu N, Luo K, Yuan Z, Chen Y. A transfer learning method for detecting Alzheimer's disease based on speech and natural language processing. Front Public Health. 2022;10:772592. [<a href="https://europepmc.org/abstract/MED/35493375" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3389/fpubh.2022.772592">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35493375&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref29">Mahajan P, Baths V. Acoustic and language based deep learning approaches for Alzheimer's Dementia detection from spontaneous speech. Front Aging Neurosci. 2021;13:623607. [<a href="https://europepmc.org/abstract/MED/33613269" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3389/fnagi.2021.623607">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33613269&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref30">Bacchi S, Gluck S, Koblar S, Jannes J, Kleinig T. Automated information extraction from free-text medical documents for stroke key performance indicators: a pilot study. Intern Med J. 2022;52(2):315-317. [<a target="_blank" href="https://dx.doi.org/10.1111/imj.15678">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35187820&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref31">Hamid H, Fodeh SJ, Lizama AG, Czlapinski R, Pugh MJ, LaFrance WC, et al. Validating a natural language processing tool to exclude psychogenic nonepileptic seizures in electronic medical record-based epilepsy research. Epilepsy Behav. 2013;29(3):578-580. [<a target="_blank" href="https://dx.doi.org/10.1016/j.yebeh.2013.09.025">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24135384&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref32">Bacchi S, Oakden-Rayner L, Zerner T, Kleinig T, Patel S, Jannes J. Deep learning natural language processing successfully predicts the cerebrovascular cause of transient ischemic attack-like presentations. Stroke. 2019;50(3):758-760. [<a href="https://www.ahajournals.org/doi/10.1161/STROKEAHA.118.024124" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1161/STROKEAHA.118.024124">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=30653397&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref33">Garg R, Oh E, Naidech A, Kording K, Prabhakaran S. Automating ischemic stroke subtype classification using machine learning and natural language processing. J Stroke Cerebrovasc Dis. 2019;28(7):2045-2051. [<a target="_blank" href="https://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2019.02.004">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=31103549&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref34">Pevy N, Christensen H, Walker T, Reuber M. Feasibility of using an automated analysis of formulation effort in patients' spoken seizure descriptions in the differential diagnosis of epileptic and nonepileptic seizures. Seizure. 2021;91:141-145. [<a href="https://linkinghub.elsevier.com/retrieve/pii/S1059-1311(21)00193-X" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1016/j.seizure.2021.06.009">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=34157636&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref35">Guan W, Ko D, Khurshid S, Trisini Lipsanopoulos AT, Ashburner JM, Harrington LX, et al. Automated electronic phenotyping of cardioembolic stroke. Stroke. 2021;52(1):181-189. [<a href="https://europepmc.org/abstract/MED/33297865" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1161/STROKEAHA.120.030663">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33297865&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref36">Cui L, Sahoo SS, Lhatoo SD, Garg G, Rai P, Bozorgi A, et al. Complex epilepsy phenotype extraction from narrative clinical discharge summaries. J Biomed Inform. 2014;51:272-279. [<a href="https://linkinghub.elsevier.com/retrieve/pii/S1532-0464(14)00144-0" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1016/j.jbi.2014.06.006">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24973735&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref37">Heo TS, Kim YS, Choi JM, Jeong YS, Seo SY, Lee JH, et al. Prediction of stroke outcome using natural language processing-based machine learning of radiology report of brain MRI. J Pers Med. 2020;10(4):286. [<a href="https://www.mdpi.com/resolver?pii=jpm10040286" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3390/jpm10040286">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33339385&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref38">Zanotto BS, da Silva Etges APB, Dal Bosco A, Cortes EG, Ruschel R, De Souza AC, et al. Stroke outcome measurements from electronic medical records: cross-sectional study on the effectiveness of neural and nonneural classifiers. JMIR Med Inform. 2021;9(11):e29120. [<a href="https://medinform.jmir.org/2021/11/e29120/" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.2196/29120">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=34723829&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref39">Kim C, Zhu V, Obeid J, Lenert L. Natural language processing and machine learning algorithm to identify brain MRI reports with acute ischemic stroke. PLoS One. 2019;14(2):e0212778. [<a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0212778" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1371/journal.pone.0212778">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=30818342&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref40">Davis MF, Sriram S, Bush WS, Denny JC, Haines JL. Automated extraction of clinical traits of multiple sclerosis in electronic medical records. J Am Med Inform Assoc. 2013;20(e2):e334-e340. [<a href="https://europepmc.org/abstract/MED/24148554" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1136/amiajnl-2013-001999">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24148554&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref41">Glauser T, Santel D, DelBello M, Faist R, Toon T, Clark P, et al. Identifying epilepsy psychiatric comorbidities with machine learning. Acta Neurol Scand. 2020;141(5):388-396. [<a href="https://europepmc.org/abstract/MED/31889296" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1111/ane.13216">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=31889296&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref42">Cohen KB, Glass B, Greiner HM, Holland-Bouley K, Standridge S, Arya R, et al. Methodological issues in predicting pediatric epilepsy surgery candidates through natural language processing and machine learning. Biomed Inform Insights. 2016;8:11-18. [<a href="https://journals.sagepub.com/doi/10.4137/BII.S38308?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%20%200pubmed" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.4137/BII.S38308">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=27257386&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref43">Alim-Marvasti A, Pérez-García F, Dahele K, Romagnoli G, Diehl B, Sparks R, et al. Machine learning for localizing epileptogenic-zone in the temporal lobe: quantifying the value of multimodal clinical-semiology and imaging concordance. Front Digit Health. 2021;3:559103. [<a href="https://europepmc.org/abstract/MED/34713078" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3389/fdgth.2021.559103">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=34713078&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref44">Balagopalan A, Eyre B, Robin J, Rudzicz F, Novikova J. Comparing pre-trained and feature-based models for prediction of Alzheimer's disease based on speech. Front Aging Neurosci. 2021;13:635945. [<a href="https://europepmc.org/abstract/MED/33986655" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3389/fnagi.2021.635945">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33986655&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref45">Martinc M, Haider F, Pollak S, Luz S. Temporal integration of text transcripts and acoustic features for Alzheimer's diagnosis based on spontaneous speech. Front Aging Neurosci. 2021;13:642647. [<a href="https://europepmc.org/abstract/MED/34194313" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3389/fnagi.2021.642647">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=34194313&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref46">Liu Z, Paek EJ, Yoon SO, Casenhiser D, Zhou W, Zhao X. Detecting Alzheimer's disease using natural language processing of referential communication task transcripts. J Alzheimers Dis. 2022;86(3):1385-1398. [<a target="_blank" href="https://dx.doi.org/10.3233/JAD-215137">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35213368&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref47">Nelson RE, Butler J, LaFleur J, Knippenberg K, Kamauu AWC, DuVall SL. Determining multiple sclerosis phenotype from electronic medical records. J Manag Care Spec Pharm. 2016;22(12):1377-1382. [<a href="https://europepmc.org/abstract/MED/27882837" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.18553/jmcp.2016.22.12.1377">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=27882837&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref48">Chase HS, Mitrani LR, Lu GG, Fulgieri DJ. Early recognition of multiple sclerosis using natural language processing of the electronic health record. BMC Med Inform Decis Mak. 2017;17(1):24. [<a href="https://bmcmedinformdecismak.biomedcentral.com/articles/10.1186/s12911-017-0418-4" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1186/s12911-017-0418-4">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=28241760&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref49">Wissel BD, Greiner HM, Glauser TA, Holland-Bouley KD, Mangano FT, Santel D, et al. Prospective validation of a machine learning model that uses provider notes to identify candidates for resective epilepsy surgery. Epilepsia. 2020;61(1):39-48. [<a href="https://europepmc.org/abstract/MED/31784992" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1111/epi.16398">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=31784992&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref50">Sung SF, Lin CY, Hu YH. EMR-based phenotyping of ischemic stroke using supervised machine learning and text mining techniques. IEEE J Biomed Health Inform. 2020;24(10):2922-2931. [<a target="_blank" href="https://dx.doi.org/10.1109/JBHI.2020.2976931">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=32142458&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref51">Yang Z, Pou-Prom C, Jones A, Banning M, Dai D, Mamdani M, et al. Assessment of natural language processing methods for ascertaining the expanded disability status scale score from the electronic health records of patients with multiple sclerosis: algorithm development and validation study. JMIR Med Inform. 2022;10(1):e25157. [<a href="https://medinform.jmir.org/2022/1/e25157/" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.2196/25157">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35019849&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref52">Xie K, Gallagher RS, Conrad EC, Garrick CO, Baldassano SN, Bernabei JM, et al. Extracting seizure frequency from epilepsy clinic notes: a machine reading approach to natural language processing. J Am Med Inform Assoc. 2022;29(5):873-881. [<a href="https://europepmc.org/abstract/MED/35190834" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1093/jamia/ocac018">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35190834&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref53">Sung SF, Chen K, Wu DP, Hung LC, Su YH, Hu YH. Applying natural language processing techniques to develop a task-specific EMR interface for timely stroke thrombolysis: a feasibility study. Int J Med Inform. 2018;112:149-157. [<a target="_blank" href="https://dx.doi.org/10.1016/j.ijmedinf.2018.02.005">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=29500013&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref54">Sung SF, Hsieh CY, Hu YH. Early prediction of functional outcomes after acute ischemic stroke using unstructured clinical text: retrospective cohort study. JMIR Med Inform. 2022;10(2):e29806. [<a href="https://medinform.jmir.org/2022/2/e29806/" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.2196/29806">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35175201&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref55">Xia Z, Secor E, Chibnik LB, Bove RM, Cheng S, Chitnis T, et al. Modeling disease severity in multiple sclerosis using electronic health records. PLoS One. 2013;8(11):e78927. [<a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0078927" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1371/journal.pone.0078927">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24244385&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref56">Roshanzamir A, Aghajan H, Baghshah MS. Transformer-based deep neural network language models for Alzheimer's disease risk assessment from targeted speech. BMC Med Inform Decis Mak. 2021;21(1):92. [<a href="https://bmcmedinformdecismak.biomedcentral.com/articles/10.1186/s12911-021-01456-3" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1186/s12911-021-01456-3">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33750385&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref57">Rannikmäe K, Wu H, Tominey S, Whiteley W, Allen N, Sudlow C, et al. Developing automated methods for disease subtyping in UK Biobank: an exemplar study on stroke. BMC Med Inform Decis Mak. 2021;21(1):191. [<a href="https://bmcmedinformdecismak.biomedcentral.com/articles/10.1186/s12911-021-01556-0" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1186/s12911-021-01556-0">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=34130677&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref58">Buchlak Q, Esmaili N, Bennett C, Farrokhi F. Natural language processing applications in the clinical neurosciences: a machine learning augmented systematic review. Acta Neurochir Suppl. 2022;134:277-289. [<a target="_blank" href="https://dx.doi.org/10.1007/978-3-030-85292-4_32">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=34862552&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref59">Crema C, Attardi G, Sartiano D, Redolfi A. Natural language processing in clinical neuroscience and psychiatry: a review. Front Psychiatry. 2022;13:946387. [<a href="https://europepmc.org/abstract/MED/36186874" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3389/fpsyt.2022.946387">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=36186874&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref60">Wang W, Kiik M, Peek N, Curcin V, Marshall IJ, Rudd AG, et al. A systematic review of machine learning models for predicting outcomes of stroke with structured data. PLoS One. 2020;15(6):e0234722. [<a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0234722" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1371/journal.pone.0234722">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=32530947&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref61">Weller J, Budson A. Current understanding of Alzheimer's disease diagnosis and treatment. F1000Res. 2018;7:1-9. [<a href="https://europepmc.org/abstract/MED/30135715" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.12688/f1000research.14506.1">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=30135715&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref62">Estimates of funding for various Research, Condition, and Disease Categories (RCDC). National Institutes of Health. URL: <a target="_blank" href="https://report.nih.gov/funding/categorical-spending#/">https://report.nih.gov/funding/categorical-spending#/</a> [accessed 2024-02-15] </span></li><li><span id="ref63">Ahsan MM, Luna SA, Siddique Z. Machine-learning-based disease diagnosis: a comprehensive review. Healthcare (Basel). 2022;10(3):541. [<a href="https://www.mdpi.com/resolver?pii=healthcare10030541" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.3390/healthcare10030541">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=35327018&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref64">Kaul V, Enslin S, Gross SA. History of artificial intelligence in medicine. Gastrointest Endosc. 2020;92(4):807-812. [<a href="https://www.giejournal.org/article/S0016-5107(20)34466-7/fulltext" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1016/j.gie.2020.06.040">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=32565184&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref65">Burton A. How do we fix the shortage of neurologists? Lancet Neurol. 2018;17(6):502-503. [<a href="https://www.thelancet.com/journals/laneur/article/PIIS1474-4422(18)30143-1/fulltext" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1016/S1474-4422(18)30143-1">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=29680206&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref66">Löscher W, Potschka H, Sisodiya SM, Vezzani A. Drug resistance in epilepsy: clinical impact, potential mechanisms, and new innovative treatment options. Pharmacol Rev. 2020;72(3):606-638. [<a href="https://europepmc.org/abstract/MED/32540959" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1124/pr.120.019539">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=32540959&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref67">Dlugos DJ. The early identification of candidates for epilepsy surgery. Arch Neurol. 2001;58(10):1543-1546. [<a target="_blank" href="https://dx.doi.org/10.1001/archneur.58.10.1543">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11594910&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref68">Myszczynska MA, Ojamies PN, Lacoste AMB, Neil D, Saffari A, Mead R, et al. Applications of machine learning to diagnosis and treatment of neurodegenerative diseases. Nat Rev Neurol. 2020;16(8):440-456. [<a href="https://eprints.whiterose.ac.uk/163905/" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1038/s41582-020-0377-8">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=32669685&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref69">Jain V, Chatterjee JM. Machine Learning with Health Care Perspective: Machine Learning and Healthcare. 1st Edition. Cham. Springer International Publishing; 2020. </span></li><li><span id="ref70">Ortiz-Posadas MR. Pattern Recognition Techniques Applied to Biomedical Problems. 1st Edition. Cham. Springer International Publishing; 2020. </span></li><li><span id="ref71">Gray KR, Aljabar P, Heckemann RA, Hammers A, Rueckert D, Alzheimer's Disease Neuroimaging Initiative. Random forest-based similarity measures for multi-modal classification of Alzheimer's disease. Neuroimage. 2013;65:167-175. [<a href="https://europepmc.org/abstract/MED/23041336" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1016/j.neuroimage.2012.09.065">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23041336&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref72">Sorin V, Barash Y, Konen E, Klang E. Deep learning for natural language processing in radiology—fundamentals and a systematic review. J Am Coll Radiol. 2020;17(5):639-648. [<a target="_blank" href="https://dx.doi.org/10.1016/j.jacr.2019.12.026">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=32004480&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref73">Amrutha K, Prabu P. Effortless and beneficial processing of natural languages using transformers. J Discret Math Sci Cryptogr. 2022;25(7):1987-2005. [<a target="_blank" href="https://dx.doi.org/10.1080/09720529.2022.2133239">CrossRef</a>]</span></li><li><span id="ref74">Weiss K, Khoshgoftaar TM, Wang D. A survey of transfer learning. J Big Data. 2016;3(1):9. [<a href="https://journalofbigdata.springeropen.com/articles/10.1186/s40537-016-0043-6#citeas" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1186/s40537-016-0043-6">CrossRef</a>]</span></li><li><span id="ref75">Devlin J, Chang MW, Lee K, Toutanova K. BERT: pre-training of deep bidirectional transformers for language understanding. ArXiv. Preprint posted online on October 11, 2018. . [<a href="https://ui.adsabs.harvard.edu/abs/2018arXiv181004805D" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.48550/arXiv.1810.04805">CrossRef</a>]</span></li><li><span id="ref76">Russell D, Li L, Tian F. Generating text using generative adversarial networks and quick-thought vectors. IEEE; 2019. Presented at: 2019 IEEE 2nd International Conference on Computer and Communication Engineering Technology (CCET); August 16-18, 2019:16-18; Beijing, China. [<a target="_blank" href="https://dx.doi.org/10.1109/ccet48361.2019.8989352">CrossRef</a>]</span></li><li><span id="ref77">Demner-Fushman D, Elhadad N. Aspiring to unintended consequences of natural language processing: a review of recent developments in clinical and consumer-generated text processing. Yearb Med Inform. 2016;(1):224-233. [<a href="https://www.thieme-connect.de/products/ejournals/abstract/10.15265/IY-2016-017" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.15265/IY-2016-017">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=27830255&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref78">Bitterman DS, Miller TA, Mak RH, Savova GK. Clinical natural language processing for radiation oncology: a review and practical primer. Int J Radiat Oncol Biol Phys. 2021;110(3):641-655. [<a href="https://www.redjournal.org/article/S0360-3016(21)00118-8/fulltext" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1016/j.ijrobp.2021.01.044">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=33545300&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref79">Magueresse A, Carles V, Heetderks E. Low-resource languages: a review of past work and future challenges. ArXiv. Preprint posted online on June 12, 2020. . [<a href="https://arxiv.org/abs/2006.07264" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.48550/arXiv.2006.07264">CrossRef</a>]</span></li><li><span id="ref80">Sajda P. Machine learning for detection and diagnosis of disease. Annu Rev Biomed Eng. 2006;8:537-565. [<a target="_blank" href="https://dx.doi.org/10.1146/annurev.bioeng.8.061505.095802">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=16834566&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref81">Khan B, Fatima H, Qureshi A, Kumar S, Hanan A, Hussain J, et al. Drawbacks of artificial intelligence and their potential solutions in the healthcare sector. Biomed Mater Devices. 2023:1-8. [<a href="https://europepmc.org/abstract/MED/36785697" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.1007/s44174-023-00063-2">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=36785697&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref82">Habli I, Lawton T, Porter Z. Artificial intelligence in health care: accountability and safety. Bull World Health Organ. 2020;98(4):251-256. [<a href="https://europepmc.org/abstract/MED/32284648" target="_blank">FREE Full text</a>] [<a target="_blank" href="https://dx.doi.org/10.2471/BLT.19.237487">CrossRef</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=32284648&dopt=Abstract" target="_blank">Medline</a>]</span></li><li><span id="ref83">Amann J. Machine learning (ML) in stroke medicine: opportunities and challenges for risk prediction and prevention. In: Jotterand F, Ienca M, editors. Artificial Intelligence in Brain and Mental Health: Philosophical, Ethical & Policy Issues. Cham. Springer International Publishing; 2021:57-71.</span></li><li><span id="ref84">Elbattah M, Arnaud É, Gignon M, Dequen G. The role of text analytics in healthcare: a review of recent developments and applications. 2021. Presented at: Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021); February 11-13, 2021:825-832; Vienna, Austria. URL: <a target="_blank" href="https://www.scitepress.org/PublishedPapers/2021/104145/104145.pdf">https://www.scitepress.org/PublishedPapers/2021/104145/104145.pdf</a> [<a target="_blank" href="https://dx.doi.org/10.5220/0010414508250832">CrossRef</a>]</span></li><li><span id="ref85">Chard K, Russell M, Lussier YA, Mendonça EA, Silverstein JC. A cloud-based approach to medical NLP. AMIA Annu Symp Proc. 2011;2011:207-216. [<a href="https://europepmc.org/abstract/MED/22195072" target="_blank">FREE Full text</a>] [<a href="https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22195072&dopt=Abstract" target="_blank">Medline</a>]</span></li></ol></div><br><hr><a name="Abbreviations">‎</a><h4 class="navigation-heading" id="Abbreviations" data-label="Abbreviations">Abbreviations</h4><table width="80%" border="0" align="center"><tr><td><b>BERT:</b> Bidirectional Encoder Representations From Transformers</td></tr><tr><td><b>DL:</b> deep learning</td></tr><tr><td><b>EMR:</b> electronic medical record</td></tr><tr><td><b>ML:</b> machine learning</td></tr><tr><td><b>MS:</b> multiple sclerosis</td></tr><tr><td><b>NLP:</b> natural language processing</td></tr><tr><td><b>PRISMA-ScR:</b> Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews</td></tr><tr><td><b>PROSPERO:</b> Prospective Register of Systematic Reviews</td></tr><tr><td><b>REDCap:</b> Research Electronic Data Capture</td></tr><tr><td><b>RNN:</b> recurrent neural network</td></tr><tr><td><b>SVM:</b> support vector machine</td></tr></table><br><hr><p style="font-style: italic">Edited by P Kubben; submitted 02.10.23; peer-reviewed by DH Kim-Dufor, P Gazerani; comments to author 13.12.23; revised version received 08.01.24; accepted 10.01.24; published 22.05.24.</p><a href="https://support.jmir.org/hc/en-us/articles/115002955531" id="Copyright" target="_blank" class="navigation-heading h4 d-block" aria-label="Copyright - what is a Creative Commons License?" data-label="Copyright">Copyright <span class="fas fa-question-circle"></span></a><p class="article-copyright">©Ilana Lefkovitz, Samantha Walsh, Leah J Blank, Nathalie Jetté, Benjamin R Kummer. Originally published in JMIR Neurotechnology (https://neuro.jmir.org), 22.05.2024.</p><small class="article-license"><p class="abstract-paragraph">This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Neurotechnology, is properly cited. The complete bibliographic information, a link to the original publication on https://neuro.jmir.org, as well as this copyright and license information must be included.</p></small><br></section></article></section></section></main> </div></div></div></div> <aside data-test="sidebar-exists" class="sidebar-citation col-lg-3 mb-5"> <div><h2 class="h4 green-heading-underline width-fit-content"> Citation </h2> <p class="fw-bold"> Please cite as: </p> <p><span> Lefkovitz I<span>,</span></span><span> Walsh S<span>,</span></span><span> Blank LJ<span>,</span></span><span> Jetté N<span>,</span></span><span> Kummer BR</span> <br> <span>Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review</span> <br> <span>JMIR Neurotech 2024;3:e51822</span> <br> <span>doi: <span><a aria-label="DOI number 10.2196/51822" data-test="article-doi" target="_blank" href="https://doi.org/10.2196/51822"> 10.2196/51822 </a></span></span>  </p> <button title="Copy Citation" data-test="copy-to-clipboard-button" class="btn btn-small btn-grey"><span aria-hidden="true" class="icon fas fa-paste"></span> Copy Citation to Clipboard </button> </div> <div class="export-metadata"><h2 class="h4 green-heading-underline width-fit-content"> Export Metadata </h2> <div><a aria-label="Export metadata in END" target="_blank" data-test="test-end-link" href="https://neuro.jmir.org/article/export/end/neuro_v3i1e51822" rel="noreferrer"> END </a><span> for: Endnote</span></div> <div><a aria-label="Export metadata in BibTeX" target="_blank" data-test="test-bib-link" href="https://neuro.jmir.org/article/export/bib/neuro_v3i1e51822" rel="noreferrer"> BibTeX </a><span> for: BibDesk, LaTeX</span></div> <div><a aria-label="Export metadata in RIS" target="_blank" data-test="test-ris-link" href="https://neuro.jmir.org/article/export/ris/neuro_v3i1e51822" rel="noreferrer"> RIS </a><span> for: RefMan, Procite, Endnote, RefWorks</span></div> <div><a target="_blank" data-test="doi-link" href="http://www.mendeley.com/import/?doi=10.2196/51822"> Add this article to your Mendeley library </a></div></div> <div class="collection desktop-show"><h2 tabindex="0" data-test="article-collection" class="h4 green-heading-underline width-fit-content"> This paper is in the following <span class="collection__span">e-collection/theme issue:</span></h2> <a href="/themes/1469" data-test="article-collection" aria-label="5 articles belongs to AI in Neurotechnology e-collection/theme issue" class="collection__link"> AI in Neurotechnology (5) </a><a href="/themes/805" data-test="article-collection" aria-label="577 articles belongs to Registered Report e-collection/theme issue" class="collection__link"> Registered Report (577) </a><a href="/themes/142" data-test="article-collection" aria-label="93 articles belongs to Reviews in Medical Informatics e-collection/theme issue" class="collection__link"> Reviews in Medical Informatics (93) </a><a href="/themes/1006" data-test="article-collection" aria-label="133 articles belongs to Computer-Aided Diagnosis e-collection/theme issue" class="collection__link"> Computer-Aided Diagnosis (133) </a><a href="/themes/797" data-test="article-collection" aria-label="1407 articles belongs to Artificial Intelligence e-collection/theme issue" class="collection__link"> Artificial Intelligence (1407) </a><a href="/themes/171" data-test="article-collection" aria-label="703 articles belongs to Natural Language Processing e-collection/theme issue" class="collection__link"> Natural Language Processing (703) </a><a href="/themes/181" data-test="article-collection" aria-label="621 articles belongs to mHealth in a Clinical Setting e-collection/theme issue" class="collection__link"> mHealth in a Clinical Setting (621) </a><a href="/themes/18" data-test="article-collection" aria-label="1315 articles belongs to Digital Health Reviews e-collection/theme issue" class="collection__link"> Digital Health Reviews (1315) </a><a href="/themes/1007" data-test="article-collection" aria-label="62 articles belongs to mHealth for Diagnosis e-collection/theme issue" class="collection__link"> mHealth for Diagnosis (62) </a><a href="/themes/1470" data-test="article-collection" aria-label="36 articles belongs to Neurotech Innovations, Diagnostic Tools and Techniques e-collection/theme issue" class="collection__link"> Neurotech Innovations, Diagnostic Tools and Techniques (36) </a><a href="/themes/102" data-test="article-collection" aria-label="189 articles belongs to Neurology and Neurosciences e-collection/theme issue" class="collection__link"> Neurology and Neurosciences (189) </a></div> <div><h2 class="h4 green-heading-underline width-fit-content"> Download </h2> <div class="download-btns"><a target="_blank" href="https://neuro.jmir.org/2024/1/e51822/PDF" aria-label="Download PDF" data-test="pdf-button" class="btn btn-small btn-grey mt-1"><span aria-hidden="true" class="icon fas fa-download"></span> Download PDF</a> <a target="_blank" href="https://neuro.jmir.org/2024/1/e51822/XML" aria-label="Download XML" data-test="xml-button" class="btn btn-small btn-grey mt-1"><span aria-hidden="true" class="icon fas fa-download"></span> Download XML</a></div></div> <div><h2 class="h4 green-heading-underline width-fit-content"> Share Article </h2> <span class="sm-icons"><a title="share-on-Bluesky" href="" aria-label="Share this item on Bluesky" data-test="bluesky-button" rel="noreferrer" target="_blank" class="bluesky small"></a> <a title="share-on-Twitter" href="https://neuro.jmir.org/2024/1/e51822?text=Check out this fascinating article" aria-label="Share this item on Twitter" data-test="twitter-button" rel="noreferrer" target="_blank" class="twitter small"></a> <a title="share-on-Facebook" href="https://neuro.jmir.org/2024/1/e51822?quote=Check out this fascinating article" aria-label="Share this item on Facebook" data-test="facebook-button" rel="noreferrer" target="_blank" class="facebook small"></a> <a title="Linkedin" href="https://www.linkedin.com/sharing/share-offsite/?url=https://neuro.jmir.org/2024/1/e51822" aria-label="Share this item on Linkedin" data-test="linkedin-button" rel="noreferrer" target="_blank" class="linkedin small"></a></span></div> <h2 class="h4 green-heading-underline width-fit-content careers"> Career Opportunities </h2> <div class="career-widget"><a href="https://careers.jmir.org/jobs/json/20885813/recreation-therapist" data-test="job-link" target="_blank" class="job"><p class="job-title"> Recreation Therapist </p> <p> California Correction Health Care Services </p> <p> Stockton, California </p></a></div><div class="career-widget"><a href="https://careers.jmir.org/jobs/json/20885641/recreation-therapist" data-test="job-link" target="_blank" class="job"><p class="job-title"> Recreation Therapist </p> <p> California Correction Health Care Services </p> <p> Delano, California </p></a></div><div class="career-widget"><a href="https://careers.jmir.org/jobs/json/20885816/recreation-therapist" data-test="job-link" target="_blank" class="job"><p class="job-title"> Recreation Therapist </p> <p> California Correction Health Care Services </p> <p> Tehachapi, California </p></a></div><div class="career-widget"><a href="https://careers.jmir.org/jobs/json/20885639/recreation-therapist" data-test="job-link" target="_blank" class="job"><p class="job-title"> Recreation Therapist </p> <p> California Correction Health Care Services </p> <p> Crescent City, California </p></a></div><div class="career-widget"><a href="https://careers.jmir.org/jobs/json/20885638/recreation-therapist" data-test="job-link" target="_blank" class="job"><p class="job-title"> Recreation Therapist </p> <p> California Correction Health Care Services </p> <p> Corcoran, California </p></a></div></aside></div></div></div></div></div> <div><section data-test="footer" class="footer"><footer id="footer"><div class="container-fluid footer-journal-name"><div class="col-12"><h2 data-test="journal-name"> JMIR Neurotechnology <span> ISSN 2817-092X </span></h2></div></div> <div class="container"><div class="row"><div class="col-lg-3 col-6"><h3 tabindex="0" class="footer-title"> Resource Centre </h3> <ul><li data-test="resource-links"><a href="/resource-centre/author-hub"> Author Hub </a></li><li data-test="resource-links"><a href="/resource-centre/editor-hub"> Editor Hub </a></li><li data-test="resource-links"><a href="/resource-centre/reviewer-hub"> Reviewer Hub </a></li><li data-test="resource-links"><a href="/resource-centre/librarian-hub"> Librarian Hub </a></li></ul></div> <div class="col-lg-3 col-6"><h3 tabindex="0" class="footer-title"> Browse Journal </h3> <ul><li data-test="journal-links"><a href="/announcements"> Latest Announcements </a></li><li data-test="journal-links"><a href="/search/authors"> Authors </a></li> <li data-test="journal-links"><a href="/themes"> Themes </a></li><li data-test="journal-links"><a href="/issues"> Issues </a></li> <li data-test="journal-links"><a href="https://blog.jmir.org "> Blog </a></li></ul></div> <div class="col-lg-3 col-6"><h3 tabindex="0" class="footer-title"> About </h3> <ul><li data-test="about-links"><a href="/privacy-statement"> Privacy Statement </a></li><li data-test="about-links"><a href="/contact-us"> Contact Us </a></li> <li><a href="/sitemap.xml" target="_blank"> Sitemap </a></li></ul></div> <div class="col-lg-3 col-6"><h3 tabindex="0" class="footer-title"> Connect With Us </h3> <span class="sm-icons"><a aria-label="JMIR Publications Bluesky account" title="Bluesky" href="https://bsky.app/profile/jmirpub.bsky.social " target="_blank" rel="noreferrer" data-test="social-links" class="bluesky mr-1"></a><a aria-label="JMIR Publications Tweeter account" title="Twitter" href="https://twitter.com/jmirpub" target="_blank" rel="noreferrer" data-test="social-links" class="twitter mr-1"></a><a aria-label="JMIR Publications Facebook account" title="Facebook" href="https://www.facebook.com/JMedInternetRes" target="_blank" rel="noreferrer" data-test="social-links" class="facebook mr-1"></a><a aria-label="JMIR Publications Linkedin account" title="Linkedin" href="https://www.linkedin.com/company/jmir-publications" target="_blank" rel="noreferrer" data-test="social-links" class="linkedin mr-1"></a><a aria-label="JMIR Publications YouTube account" title="YouTube" href="https://www.youtube.com/c/JMIRPublications" target="_blank" rel="noreferrer" data-test="social-links" class="youtube mr-1"></a><a aria-label="JMIR Publications Instagram account" title="Instagram" href="https://www.instagram.com/jmirpub" target="_blank" rel="noreferrer" data-test="social-links" class="instagram"></a> <a target="_blank" rel="noreferrer" aria-label="RSS Subscription" title="RSS Subscription" href="https://neuro.jmir.org/feed/atom" class="rss"></a></span></div> <div class="email-subscribtion-button col-lg-3 col-md-6 col-sm-6 col-12"><h3 tabindex="0" class="footer-title"> Get Table of Contents Alerts </h3> <a target="_blank" rel="noopener noreferrer" aria-label="Newsletter Subscription" title="Newsletter Subscription" href="https://landingpage.jmirpublications.com/journal-preference-selection"><span>Get Alerts</span> <span class="icon fas fa-paper-plane"></span></a></div> <div class="col-12 text-center mt-5"><p class="footer-copyright"> Copyright © <time datetime="2024"> 2024 </time> JMIR Publications </p></div></div></div></footer></section></div></div>  <div><a tabindex="0" href="javascript:;" title="Scroll to the top of the page" role="button" class="scroll-to-very-top"><span aria-hidden="true" class="icon fas fa-chevron-up"></span></a></div> </div></div></div><script>window.__NUXT__=(function(a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,A,B,C,D,E,F,G,H,I,J,K,L,M,N,O,P,Q,R,S,T,U,V,W,X,Y,Z,_,$,aa,ab,ac,ad,ae,af,ag,ah,ai,aj,ak,al,am,an,ao,ap,aq,ar,as,at,au,av,aw,ax,ay,az,aA,aB,aC,aD,aE,aF,aG,aH,aI,aJ,aK,aL,aM,aN,aO,aP,aQ,aR,aS,aT,aU,aV,aW,aX,aY,aZ,a_,a$,ba,bb,bc,bd,be,bf,bg,bh,bi,bj,bk,bl){return {layout:"front",data:[{tabs:{html:{name:"Article",route:c,path:"articleHtml"},authors:{name:"Authors",route:"authors",path:"articleAuthors"},citations:{name:"Cited by ",route:"citations",path:"articleCitations"},tweetations:{name:"Tweetations (3)",route:"tweetations",path:"articleTweetations"},metrics:{name:"Metrics",route:"metrics",path:"articleMetrics"}},registeredReport:L,jobs:[{title:o,employer:p,city:"Stockton",state_province:q,url:"https:\u002F\u002Fcareers.jmir.org\u002Fjobs\u002Fjson\u002F20885813\u002Frecreation-therapist"},{title:o,employer:p,city:"Delano",state_province:q,url:"https:\u002F\u002Fcareers.jmir.org\u002Fjobs\u002Fjson\u002F20885641\u002Frecreation-therapist"},{title:o,employer:p,city:"Tehachapi",state_province:q,url:"https:\u002F\u002Fcareers.jmir.org\u002Fjobs\u002Fjson\u002F20885816\u002Frecreation-therapist"},{title:o,employer:p,city:"Crescent City",state_province:q,url:"https:\u002F\u002Fcareers.jmir.org\u002Fjobs\u002Fjson\u002F20885639\u002Frecreation-therapist"},{title:o,employer:p,city:"Corcoran",state_province:q,url:"https:\u002F\u002Fcareers.jmir.org\u002Fjobs\u002Fjson\u002F20885638\u002Frecreation-therapist"}]},{html:"\u003Cmain id=\"wrapper\" class=\"wrapper ArticleMain clearfix\"\u003E\u003Csection class=\"inner-wrapper clearfix\"\u003E\u003Csection class=\"main-article-content clearfix\"\u003E\u003Carticle class=\"ajax-article-content\"\u003E\u003Ch4 class=\"h4-original-paper\"\u003E\u003Cspan class=\"typcn typcn-document-text\"\u003E\u003C\u002Fspan\u003EReview\u003C\u002Fh4\u003E\u003Cdiv class=\"authors-container\"\u003E\u003Cdiv class=\"authors clearfix\"\u003E\u003C\u002Fdiv\u003E\u003C\u002Fdiv\u003E\u003Cdiv class=\"authors-container\"\u003E\u003Cdiv class=\"authors clearfix\"\u003E\u003C\u002Fdiv\u003E\u003C\u002Fdiv\u003E\u003Cdiv class=\"authors-container\"\u003E\u003Cdiv class=\"authors clearfix\"\u003E\u003Cul class=\"clearfix\"\u003E\u003Cli\u003E\u003Ca href=\"\u002Fsearch\u002FsearchResult?field%5B%5D=author&criteria%5B%5D=Ilana+Lefkovitz\" class=\"btn-view-author-options\"\u003EIlana Lefkovitz\u003Csup\u003E\u003Csmall\u003E1\u003C\u002Fsmall\u003E\u003C\u002Fsup\u003E, MD\u003C\u002Fa\u003E\u003Ca class=\"author-orcid\" href=\"https:\u002F\u002Forcid.org\u002F0000-0002-8724-3798\" target=\"_blank\" title=\"ORCID\"\u003E \u003C\u002Fa\u003E; \u003C\u002Fli\u003E\u003Cli\u003E\u003Ca href=\"\u002Fsearch\u002FsearchResult?field%5B%5D=author&criteria%5B%5D=Samantha+Walsh\" class=\"btn-view-author-options\"\u003ESamantha Walsh\u003Csup\u003E\u003Csmall\u003E2\u003C\u002Fsmall\u003E\u003C\u002Fsup\u003E, MLS, MA\u003C\u002Fa\u003E\u003Ca class=\"author-orcid\" href=\"https:\u002F\u002Forcid.org\u002F0000-0002-5040-6514\" target=\"_blank\" title=\"ORCID\"\u003E \u003C\u002Fa\u003E; \u003C\u002Fli\u003E\u003Cli\u003E\u003Ca href=\"\u002Fsearch\u002FsearchResult?field%5B%5D=author&criteria%5B%5D=Leah%20J+Blank\" class=\"btn-view-author-options\"\u003ELeah J Blank\u003Csup\u003E\u003Csmall\u003E1,\u003C\u002Fsmall\u003E\u003C\u002Fsup\u003E\u003Csup\u003E\u003Csmall\u003E3\u003C\u002Fsmall\u003E\u003C\u002Fsup\u003E, MD, MPH\u003C\u002Fa\u003E\u003Ca class=\"author-orcid\" href=\"https:\u002F\u002Forcid.org\u002F0000-0001-8719-6752\" target=\"_blank\" title=\"ORCID\"\u003E \u003C\u002Fa\u003E; \u003C\u002Fli\u003E\u003Cli\u003E\u003Ca href=\"\u002Fsearch\u002FsearchResult?field%5B%5D=author&criteria%5B%5D=Nathalie+Jetté\" class=\"btn-view-author-options\"\u003ENathalie Jetté\u003Csup\u003E\u003Csmall\u003E1,\u003C\u002Fsmall\u003E\u003C\u002Fsup\u003E\u003Csup\u003E\u003Csmall\u003E4\u003C\u002Fsmall\u003E\u003C\u002Fsup\u003E, MD, MSc\u003C\u002Fa\u003E\u003Ca class=\"author-orcid\" href=\"https:\u002F\u002Forcid.org\u002F0000-0003-1351-5866\" target=\"_blank\" title=\"ORCID\"\u003E \u003C\u002Fa\u003E; \u003C\u002Fli\u003E\u003Cli\u003E\u003Ca href=\"\u002Fsearch\u002FsearchResult?field%5B%5D=author&criteria%5B%5D=Benjamin%20R+Kummer\" class=\"btn-view-author-options\"\u003EBenjamin R Kummer\u003Csup\u003E\u003Csmall\u003E1,\u003C\u002Fsmall\u003E\u003C\u002Fsup\u003E\u003Csup\u003E\u003Csmall\u003E5,\u003C\u002Fsmall\u003E\u003C\u002Fsup\u003E\u003Csup\u003E\u003Csmall\u003E6\u003C\u002Fsmall\u003E\u003C\u002Fsup\u003E, MD\u003C\u002Fa\u003E\u003Ca class=\"author-orcid\" href=\"https:\u002F\u002Forcid.org\u002F0000-0002-1413-8014\" target=\"_blank\" title=\"ORCID\"\u003E \u003C\u002Fa\u003E\u003C\u002Fli\u003E\u003C\u002Ful\u003E\u003Cdiv class=\"author-affiliation-details\"\u003E\u003Cp\u003E\u003Csup\u003E1\u003C\u002Fsup\u003EDepartment of Neurology, Icahn School of Medicine at Mount Sinai, New York, NY, United States\u003C\u002Fp\u003E\u003Cp\u003E\u003Csup\u003E2\u003C\u002Fsup\u003EHunter College Libraries, Hunter College, City University of New York, New York, NY, United States\u003C\u002Fp\u003E\u003Cp\u003E\u003Csup\u003E3\u003C\u002Fsup\u003EDepartment of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, NY, United States\u003C\u002Fp\u003E\u003Cp\u003E\u003Csup\u003E4\u003C\u002Fsup\u003EDepartment of Clinical Neurosciences, University of Calgary, Calgary, AB, Canada\u003C\u002Fp\u003E\u003Cp\u003E\u003Csup\u003E5\u003C\u002Fsup\u003EClinical Neuro-Informatics Program, Icahn School of Medicine at Mount Sinai, New York, NY, United States\u003C\u002Fp\u003E\u003Cp\u003E\u003Csup\u003E6\u003C\u002Fsup\u003EWindreich Department of Artificial Intelligence and Human Health, Icahn School of Medicine at Mount Sinai, New York, NY, United States\u003C\u002Fp\u003E\u003C\u002Fdiv\u003E\u003C\u002Fdiv\u003E\u003Cdiv class=\"corresponding-author-and-affiliations clearfix\"\u003E\u003Cdiv class=\"corresponding-author-details\"\u003E\u003Ch3\u003ECorresponding Author:\u003C\u002Fh3\u003E\u003Cp\u003EBenjamin R Kummer, MD\u003C\u002Fp\u003E\u003Cp\u003E\u003C\u002Fp\u003E\u003Cp\u003EDepartment of Neurology\u003C\u002Fp\u003E\u003Cp\u003EIcahn School of Medicine at Mount Sinai\u003C\u002Fp\u003E\u003Cp\u003EOne Gustave Levy Place\u003C\u002Fp\u003E\u003Cp\u003EBox 1137\u003C\u002Fp\u003E\u003Cp\u003ENew York, NY, 10029\u003C\u002Fp\u003E\u003Cp\u003EUnited States\u003C\u002Fp\u003E\u003Cp\u003EPhone: 1 212 241 5050\u003C\u002Fp\u003E\u003Cp\u003EEmail: \u003Ca href=\"mailto:benjamin.kummer@mountsinai.org\"\u003Ebenjamin.kummer@mountsinai.org\u003C\u002Fa\u003E\u003C\u002Fp\u003E\u003Cbr\u003E\u003C\u002Fdiv\u003E\u003C\u002Fdiv\u003E\u003C\u002Fdiv\u003E\u003Csection class=\"article-content clearfix\"\u003E\u003Carticle class=\"abstract\"\u003E\u003Ch3 id=\"Abstract\" class=\"navigation-heading\" data-label=\"Abstract\"\u003EAbstract\u003C\u002Fh3\u003E\u003Cp\u003E\u003Cspan class=\"abstract-sub-heading\"\u003EBackground: \u003C\u002Fspan\u003ENatural language processing (NLP), a branch of artificial intelligence that analyzes unstructured language, is being increasingly used in health care. However, the extent to which NLP has been formally studied in neurological disorders remains unclear.\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003E\u003Cspan class=\"abstract-sub-heading\"\u003EObjective: \u003C\u002Fspan\u003EWe sought to characterize studies that applied NLP to the diagnosis, prediction, or treatment of common neurological disorders.\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003E\u003Cspan class=\"abstract-sub-heading\"\u003EMethods: \u003C\u002Fspan\u003EThis review followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) standards. The search was conducted using MEDLINE and Embase on May 11, 2022. Studies of NLP use in migraine, Parkinson disease, Alzheimer disease, stroke and transient ischemic attack, epilepsy, or multiple sclerosis were included. We excluded conference abstracts, review papers, as well as studies involving heterogeneous clinical populations or indirect clinical uses of NLP. Study characteristics were extracted and analyzed using descriptive statistics. We did not aggregate measurements of performance in our review due to the high variability in study outcomes, which is the main limitation of the study.\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003E\u003Cspan class=\"abstract-sub-heading\"\u003EResults: \u003C\u002Fspan\u003EIn total, 916 studies were identified, of which 41 (4.5%) met all eligibility criteria and were included in the final review. Of the 41 included studies, the most frequently represented disorders were stroke and transient ischemic attack (n=20, 49%), followed by epilepsy (n=10, 24%), Alzheimer disease (n=6, 15%), and multiple sclerosis (n=5, 12%). We found no studies of NLP use in migraine or Parkinson disease that met our eligibility criteria. The main objective of NLP was diagnosis (n=20, 49%), followed by disease phenotyping (n=17, 41%), prognostication (n=9, 22%), and treatment (n=4, 10%). In total, 18 (44%) studies used only machine learning approaches, 6 (15%) used only rule-based methods, and 17 (41%) used both.\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003E\u003Cspan class=\"abstract-sub-heading\"\u003EConclusions: \u003C\u002Fspan\u003EWe found that NLP was most commonly applied for diagnosis, implying a potential role for NLP in augmenting diagnostic accuracy in settings with limited access to neurological expertise. We also found several gaps in neurological NLP research, with few to no studies addressing certain disorders, which may suggest additional areas of inquiry.\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003E\u003Cspan class=\"abstract-sub-heading\"\u003ETrial Registration: \u003C\u002Fspan\u003EProspective Register of Systematic Reviews (PROSPERO) CRD42021228703; https:\u002F\u002Fwww.crd.york.ac.uk\u002FPROSPERO\u002Fdisplay_record.php?RecordID=228703\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cstrong class=\"h4-article-volume-issue\"\u003EJMIR Neurotech 2024;3:e51822\u003C\u002Fstrong\u003E\u003Cbr\u003E\u003Cbr\u003E\u003Cspan class=\"article-doi\"\u003E\u003Ca href=\"https:\u002F\u002Fdoi.org\u002F10.2196\u002F51822\"\u003Edoi:10.2196\u002F51822\u003C\u002Fa\u003E\u003C\u002Fspan\u003E\u003Cbr\u003E\u003Cbr\u003E\u003Ch3 class=\"h3-main-heading\" id=\"Keywords\"\u003EKeywords\u003C\u002Fh3\u003E\u003Cdiv class=\"keywords\"\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=natural%20language%20processing&precise=true\"\u003Enatural language processing\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=NLP&precise=true\"\u003ENLP\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=unstructured&precise=true\"\u003Eunstructured\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=text&precise=true\"\u003Etext\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=machine%20learning&precise=true\"\u003Emachine learning\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=deep%20learning&precise=true\"\u003Edeep learning\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=neurology&precise=true\"\u003Eneurology\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=headache%20disorders&precise=true\"\u003Eheadache disorders\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=migraine&precise=true\"\u003Emigraine\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=Parkinson%20disease&precise=true\"\u003EParkinson disease\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=cerebrovascular%20disease&precise=true\"\u003Ecerebrovascular disease\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=stroke&precise=true\"\u003Estroke\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=transient%20ischemic%20attack&precise=true\"\u003Etransient ischemic attack\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=epilepsy&precise=true\"\u003Eepilepsy\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=multiple%20sclerosis&precise=true\"\u003Emultiple sclerosis\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=cardiovascular&precise=true\"\u003Ecardiovascular\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=artificial%20intelligence&precise=true\"\u003Eartificial intelligence\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=Parkinson&precise=true\"\u003EParkinson\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=neurological&precise=true\"\u003Eneurological\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=neurological%20disorder&precise=true\"\u003Eneurological disorder\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=scoping%20review&precise=true\"\u003Escoping review\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=diagnosis&precise=true\"\u003Ediagnosis\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=treatment&precise=true\"\u003Etreatment\u003C\u002Fa\u003E; \u003C\u002Fspan\u003E\u003Cspan\u003E\u003Ca href=\"\u002Fsearch?type=keyword&term=prediction&precise=true\"\u003Eprediction\u003C\u002Fa\u003E \u003C\u002Fspan\u003E\u003C\u002Fdiv\u003E\u003Cdiv id=\"trendmd-suggestions\"\u003E\u003C\u002Fdiv\u003E\u003C\u002Farticle\u003E\u003Cbr\u003E\u003Carticle class=\"main-article clearfix\"\u003E\u003Cbr\u003E\u003Ch3 class=\"navigation-heading h3-main-heading\" id=\"Introduction\" data-label=\"Introduction\"\u003EIntroduction\u003C\u002Fh3\u003E\u003Cp class=\"abstract-paragraph\"\u003EThe implementation of the electronic medical record (EMR) in health care systems has resulted in a remarkable increase in the amount of digital patient data [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref1\" rel=\"footnote\"\u003E1\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], much of which is text-based and stored in an unstructured, narrative format [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref2\" rel=\"footnote\"\u003E2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E-\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref4\" rel=\"footnote\"\u003E4\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. While unstructured text is a rich data source, analyses of these data often require time- and cost-intensive manual processing [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref3\" rel=\"footnote\"\u003E3\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Natural language processing (NLP), a type of artificial intelligence that automatically derives meaning from unstructured language, can significantly reduce costs and enhance the quality of health care systems by converting unstructured text into a structured form that can be processed by computers [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref2\" rel=\"footnote\"\u003E2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref4\" rel=\"footnote\"\u003E4\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref5\" rel=\"footnote\"\u003E5\u003C\u002Fa\u003E\u003C\u002Fspan\u003E].\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EApproaches to NLP can use rule-based techniques, machine learning (ML), or a combination of both [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref6\" rel=\"footnote\"\u003E6\u003C\u002Fa\u003E\u003C\u002Fspan\u003E-\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref8\" rel=\"footnote\"\u003E8\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Between the fifth and eighth decades of the 20th century, NLP approaches were predominantly rule-based, using a set of rules defined by human experts [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref7\" rel=\"footnote\"\u003E7\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref9\" rel=\"footnote\"\u003E9\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] to systematically extract meaning from unstructured text. Rule-based methods are comprehensible by humans but difficult to generalize [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref7\" rel=\"footnote\"\u003E7\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref9\" rel=\"footnote\"\u003E9\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Driven by recent advances in computing power and access to computing resources, contemporary approaches to NLP have increasingly incorporated ML, which possesses greater scalability [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref7\" rel=\"footnote\"\u003E7\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] than rule-based methods despite the need for greater computational power to construct ML-based NLP models. Most recently, complex ML methods such as deep learning (DL), which are based on neural networks and larger datasets than conventional ML approaches, have become popular approaches to address NLP tasks [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref9\" rel=\"footnote\"\u003E9\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref10\" rel=\"footnote\"\u003E10\u003C\u002Fa\u003E\u003C\u002Fspan\u003E].\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EThe high prevalence of unstructured text in EMR systems creates an ideal use case for NLP in health care. However, the majority of current NLP research remains focused on nonneurological conditions such as mental health, cancer, and pneumonia [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref5\" rel=\"footnote\"\u003E5\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. The dearth of neurological NLP research is out of proportion to the worldwide importance of neurological conditions, both in terms of public health burden and cost. For instance, cerebrovascular disease occupies the second leading cause of death worldwide [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref11\" rel=\"footnote\"\u003E11\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], and in the United States, neurological and musculoskeletal disorders generate the greatest number of years lost to disability [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref12\" rel=\"footnote\"\u003E12\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Finally, the estimated annual cost of the most prevalent neurological diseases in the United States is nearly US $800 billion [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref12\" rel=\"footnote\"\u003E12\u003C\u002Fa\u003E\u003C\u002Fspan\u003E].\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003ENeurology is a specialty that is uniquely well suited to benefit from NLP approaches. The data used in the diagnosis and management of neurological conditions, such as examination findings or clinical impressions, are often recorded as narrative, unstructured text in clinical documentation. Aside from clinical notes containing the patient history and neurological examination, reports from radiology [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref13\" rel=\"footnote\"\u003E13\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref14\" rel=\"footnote\"\u003E14\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], sonography, or electrophysiology studies are integral to neurological practice and often are crucial for detection, prognosis, and treatment. Further, NLP analysis of spoken language may allow the detection of certain neurodegenerative conditions such as Alzheimer disease in their early stages [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref15\" rel=\"footnote\"\u003E15\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Given the unique position of neurology with respect to NLP and the relative lack of research on the applications of NLP in neurology, we sought to conduct a scoping review in order to quantify and characterize studies that directly applied NLP for clinical use in common neurological disorders.\u003C\u002Fp\u003E\u003Cbr\u003E\u003Ch3 class=\"navigation-heading h3-main-heading\" id=\"Methods\" data-label=\"Methods\"\u003EMethods\u003C\u002Fh3\u003E\u003Ch4\u003ELiterature Search Strategy and Eligibility Criteria\u003C\u002Fh4\u003E\u003Cp class=\"abstract-paragraph\"\u003EThis review was conducted using the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines (\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#app1\" rel=\"footnote\"\u003EMultimedia Appendix 1\u003C\u002Fa\u003E\u003C\u002Fspan\u003E) and was registered with the Prospective Register of Systematic Reviews (PROSPERO CRD42021228703). Our search was conducted using Ovid Embase and MEDLINE on May 11, 2022 (\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#app2\" rel=\"footnote\"\u003EMultimedia Appendix 2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref16\" rel=\"footnote\"\u003E16\u003C\u002Fa\u003E\u003C\u002Fspan\u003E-\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref22\" rel=\"footnote\"\u003E22\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]). Based on the most globally prevalent and costly neurological disorders [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref11\" rel=\"footnote\"\u003E11\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], studies investigating the use of NLP in Alzheimer disease (exclusive of Alzheimer disease–related disorders), Parkinson disease, stroke and transient ischemic attack, epilepsy, multiple sclerosis (MS), and migraine were included.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EStudies that used NLP to analyze radiographic findings without any clinical correlation (eg, silent brain infarcts) or for purposes other than diagnosis, detection, phenotyping, subtyping, prognostication, risk stratification, or therapy were excluded. We excluded studies with populations comprised of patients with heterogeneous diseases or ambiguously defined populations (eg, we excluded studies that used a patient cohort consisting of patients with both Alzheimer dementia and mild cognitive impairment) as well as studies that did not use NLP for direct clinical applications. Examples of indirect clinical applications include the use of NLP to identify cohorts for subsequent model development or conduct epidemiological associations between cohorts without direct impact on clinical practice. We additionally excluded abstracts, conference proceedings, reviews, and editorials.\u003C\u002Fp\u003E\u003Ch4\u003EData Extraction\u003C\u002Fh4\u003E\u003Cp class=\"abstract-paragraph\"\u003EA medical librarian (SW) with expertise in scoping reviews first conducted a literature search (\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#app2\" rel=\"footnote\"\u003EMultimedia Appendix 2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E) based on our eligibility criteria to generate a list of abstracts, which were then imported into a web application (Covidence Ltd) for initial screening by 3 authors (BRK, LJB, and IL). After the abstract screening was completed, full-text papers for screened abstracts were reviewed by 2 authors (BRK and IL) to determine eligibility for inclusion. Disagreements at both stages were resolved by discussion and consensus.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EUsing the final list of full-text studies, study characteristics were manually extracted by 1 author (IL) and charted in a REDCap (Research Electronic Data Capture; REDCap Consortium) web database form, which was subsequently reviewed by a second author (BRK) for accuracy. The data charting form was initially tested by the data extractor (IL) and revised after feedback from all coauthors (BRK, NJ, LJB, and SW). We extracted study publication year, population size, country of origin, journal field (eg, medical informatics, clinical neurology, nonclinical neuroscience, clinical medicine, or other), neurological disorder, and target of NLP (eg, diagnosis or detection, phenotyping or subtyping and severity, prognostication or risk stratification, or disease management or therapy). Each study could have multiple targets whenever applicable.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EFor each study, the source language to which NLP techniques were applied was also extracted. For studies conducted in or authored by teams from non-English–speaking countries, the source language was extrapolated directly as described from the study methodology. If the source language was a publicly available research dataset or ontology (eg, MetaMap ontology or ADReSS dataset, both of which use English), the source language was reported as English. Source of language for NLP (eg, clinical notes, radiographic reports, speech audio, or other) and type of study (eg, model derivation, validation, or both) were also noted. Validation studies were defined as studies that specifically investigated the validation of a derived model in a population external to the original model derivation population. Our definition of validation studies did not include validation on held-out test sets as part of model derivation. If the NLP model was both derived and externally validated in the same study, the population size included the additional population used for validation. Simulated patients, who were used as a training set in one study, were included in the population size. If no population size was mentioned in the studies, the number of text instances (eg, clinical notes and radiographic reports) was recorded.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EWe additionally extracted the study’s NLP approaches (ie, rule-based methods, ML, or both). Rule-based NLP included any approaches that used keyword searches, pattern matching, regular expressions, or ontological systems for word-concept mapping, text preprocessing, or classification. ML-based NLP comprised both conventional ML and DL approaches and both were distinguished as dichotomous study characteristic variables but could co-occur in the studies. A study was characterized as including any of these methods if either ML or DL was used at any point in model development for the study.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EUnder the category of conventional ML methods, linear regression, logistic regression, support vector machines (SVMs), naïve Bayes classifiers, decision trees, random forest classifiers, k-nearest neighbor algorithms, gradient boosting techniques such as extreme gradient boosting, latent Dirichlet allocation, and shallow neural networks were included. Under the definition of shallow neural network, we included any approaches using Word2vec or other “-2vec” word-embedding techniques that use a neural network to construct word contexts and extract semantic and syntactic meaning from text [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref23\" rel=\"footnote\"\u003E23\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref24\" rel=\"footnote\"\u003E24\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. We also included other types of regression, such as lasso regression, which is often used for dimensionality reduction, in the conventional ML category.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EDL techniques included convolutional neural networks, recurrent neural networks (RNNs), long- and short-term memory networks, multilayer perceptrons, and transformers. Studies using long- and short-term memory networks were also categorized as using an RNN. We also note that neural networks of unspecified type and number of layers, which were not clearly referred to as DL in the study, were not included in this category.\u003C\u002Fp\u003E\u003Cbr\u003E\u003Ch3 class=\"navigation-heading h3-main-heading\" id=\"Results\" data-label=\"Results\"\u003EResults\u003C\u002Fh3\u003E\u003Ch4\u003EIncluded Studies\u003C\u002Fh4\u003E\u003Cp class=\"abstract-paragraph\"\u003EIn total, 916 studies were identified from our search strategy, of which 271 were duplicates and were excluded. We then screened the resulting 645 abstracts, of which 565 were excluded due to not meeting initial eligibility criteria. Of the remaining 80 studies, 39 (49%) were excluded. The 2 most common reasons for exclusion were the use of NLP for nonclinical applications (n=15, 38%) and heterogeneous clinical populations (n=12, 31%). In total, 41 (4.5%) of the 916 studies from the original search results were ultimately included for review (\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#figure1\" rel=\"footnote\"\u003EFigure 1\u003C\u002Fa\u003E\u003C\u002Fspan\u003E and \u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#table1\" rel=\"footnote\"\u003ETable 1\u003C\u002Fa\u003E\u003C\u002Fspan\u003E).\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EOf the 41 included studies, NLP was applied to stroke or transient ischemic attack in 20 (49%) studies, epilepsy in 10 (24%) studies, Alzheimer dementia in 6 (15%) studies, and MS in 5 (12%) studies. We found no studies applying NLP to Parkinson disease or migraine that met our eligibility criteria. Across all neurological conditions, NLP was most commonly applied for the purposes of detection or diagnosis (n=20, 49%), followed by clinical disease phenotyping or subtyping (n=17, 41%), prognostication or risk stratification (n=9, 22%), and management or therapy (n=4, 10%; \u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#table2\" rel=\"footnote\"\u003ETable 2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E).\u003C\u002Fp\u003E\u003Cfigure\u003E\u003Ca name=\"figure1\"\u003E‎\u003C\u002Fa\u003E\u003Ca class=\"fancybox\" title=\"Figure 1. Study PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) diagram. NLP: natural language processing.\" href=\"https:\u002F\u002Fasset.jmir.pub\u002Fassets\u002Fd2d99db4f7ccd60a8d1c2d68ea63db52.png\" id=\"figure1\"\u003E\u003Cimg class=\"figure-image\" src=\"https:\u002F\u002Fasset.jmir.pub\u002Fassets\u002Fd2d99db4f7ccd60a8d1c2d68ea63db52.png\"\u003E\u003C\u002Fa\u003E\u003Cfigcaption\u003E\u003Cspan class=\"typcn typcn-image\"\u003E\u003C\u002Fspan\u003E\u003Cb\u003EFigure 1. \u003C\u002Fb\u003E Study PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) diagram. NLP: natural language processing. \u003C\u002Ffigcaption\u003E\u003C\u002Ffigure\u003E\u003Cdiv class=\"figure-table\"\u003E\u003Cfigcaption\u003E\u003Cspan class=\"typcn typcn-clipboard\"\u003E\u003C\u002Fspan\u003E\u003Cb\u003ETable 1. \u003C\u002Fb\u003EIncluded studies.\u003C\u002Ffigcaption\u003E\u003Ctable width=\"1000\" cellpadding=\"5\" cellspacing=\"0\" border=\"1\" rules=\"groups\" frame=\"hsides\"\u003E\u003Ccol width=\"70\" span=\"1\"\u003E\u003Ccol width=\"70\" span=\"1\"\u003E\u003Ccol width=\"60\" span=\"1\"\u003E\u003Ccol width=\"70\" span=\"1\"\u003E\u003Ccol width=\"70\" span=\"1\"\u003E\u003Ccol width=\"70\" span=\"1\"\u003E\u003Ccol width=\"80\" span=\"1\"\u003E\u003Ccol width=\"90\" span=\"1\"\u003E\u003Ccol width=\"70\" span=\"1\"\u003E\u003Ccol width=\"60\" span=\"1\"\u003E\u003Ccol width=\"150\" span=\"1\"\u003E\u003Ccol width=\"140\" span=\"1\"\u003E\u003Cthead\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPaper authors\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPublication date\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ECountry\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESource text\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EJournal field\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EExternal model validation\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ECondition being studied\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPurpose of NLP\u003Csup\u003Ea\u003C\u002Fsup\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENLP method\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDeep learning\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAlgorithms used\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStudy outcomes\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003C\u002Fthead\u003E\u003Ctbody\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMiller et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref19\" rel=\"footnote\"\u003E19\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMay 9, 2022\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERadiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003Csup\u003Eb\u003C\u002Fsup\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERandom forest, linear regression, KNN\u003Csup\u003Ec\u003C\u002Fsup\u003E, lasso regression, MLP\u003Csup\u003Ed\u003C\u002Fsup\u003E, transformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERadiographic complications of ischemic stroke (eg, hemorrhagic transformation)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELay et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref25\" rel=\"footnote\"\u003E25\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EOctober 23, 2020\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAustralia\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELatent Dirichlet allocation\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIdentifying themes in medical records in patients with PNES\u003Csup\u003Ee\u003C\u002Fsup\u003E, congruency of themes\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMayampurath et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref26\" rel=\"footnote\"\u003E26\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EJune 24, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis, clinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM\u003Csup\u003Ef\u003C\u002Fsup\u003E, logistic regression\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAcute stroke diagnosis, stroke severity and subtypes\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELi et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref16\" rel=\"footnote\"\u003E16\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMarch 1, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERadiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENeuroradiology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERandom forest\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAcute or subacute ischemic stroke cases before and during COVID-19\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELineback et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref27\" rel=\"footnote\"\u003E27\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EJuly 13, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrognosis or risk stratification\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, naïve Bayes, random forest, logistic regression, shallow neural network, lasso regression, ensemble, boosting\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E30-day stroke readmission, 30-day all-cause readmission\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELiu et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref28\" rel=\"footnote\"\u003E28\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EApril 13, 2022\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EChina\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESpeech\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPublic health\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAlzheimer disease\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, random forest, logistic regression, boosting, CNN\u003Csup\u003Eg\u003C\u002Fsup\u003E, transformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection of Alzheimer disease from speech\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMahajan and Baths [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref29\" rel=\"footnote\"\u003E29\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EFebruary 5, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIndia\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESpeech\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENonclinical neuroscience\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAlzheimer disease\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ECNN, RNN\u003Csup\u003Eh\u003C\u002Fsup\u003E (LSTM\u003Csup\u003Ei\u003C\u002Fsup\u003E)\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection of Alzheimer disease from speech\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EBacchi et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref30\" rel=\"footnote\"\u003E30\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EFebruary 20, 2022\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAustralia\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical medicine\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERandom forest, decision tree, logistic regression, neural network with an unspecified number of layers\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EExtraction of stroke key performance indicators\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EHamid et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref31\" rel=\"footnote\"\u003E31\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EOctober 14, 2013\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENaïve Bayes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIdentification of patients with PNES\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYu et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref13\" rel=\"footnote\"\u003E13\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESeptember 16, 2020\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ECanada\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERadiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis, clinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EN\u002FA\u003Csup\u003Ej\u003C\u002Fsup\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIdentification of the presence and location of vascular occlusions and other stroke-related attributes\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EBacchi et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref32\" rel=\"footnote\"\u003E32\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EJanuary 17, 2019\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAustralia\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes and radiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERandom forest, decision tree, CNN, RNN (LSTM)\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetermining the cause of TIA\u003Csup\u003Ek\u003C\u002Fsup\u003E-like presentations (cerebrovascular vs noncerebrovascular)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EGarg et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref33\" rel=\"footnote\"\u003E33\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMay 15, 2019\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes and radiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, random forest, logistic regression, KNN, boosting, ensemble (stacking logistic regression, extra trees classifier)\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIschemic stroke subtypes\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EZhao et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref21\" rel=\"footnote\"\u003E21\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMarch 8, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis, clinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERandom forest, logistic regression\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIncidence of stroke, stroke subtypes\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPevy et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref34\" rel=\"footnote\"\u003E34\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EOctober 1, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited Kingdom\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESpeech\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERandom forest\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDistinguishing between PNES and epilepsy, hesitations and repetitions in descriptions of epileptic seizures versus PNES\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EGuan et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref35\" rel=\"footnote\"\u003E35\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDecember 10, 2020\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEchocardiographic reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, random forest, decision tree, logistic regression, KNN\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESubtyping and phenotyping cardioembolic stroke\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ECui et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref36\" rel=\"footnote\"\u003E36\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EJune 26, 2014\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EN\u002FA\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy phenotype extraction with correlated anatomic location\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EHeo et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref37\" rel=\"footnote\"\u003E37\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDecember 16, 2020\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESouth Korea\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERadiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical medicine\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrognosis or risk stratification\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, random forest, decision tree, shallow neural network, lasso regression, CNN, RNN (LSTM), MLP\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrediction of poor stroke outcome\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EZanotto et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref38\" rel=\"footnote\"\u003E38\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENovember 1, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EBrazil\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrognosis or risk stratification, clinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, naïve Bayes, random forest, KNN, CNN, transformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrediction of stroke outcome measurements and extraction of patient characteristics\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EBarbour et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref17\" rel=\"footnote\"\u003E17\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMay 21, 2019\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrognosis or risk stratification\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EN\u002FA\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERisk factors for SUDEP\u003Csup\u003El\u003C\u002Fsup\u003E\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EKim et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref39\" rel=\"footnote\"\u003E39\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EFebruary 28, 2019\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERadiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENonclinical neuroscience\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, naïve Bayes, decision tree, logistic regression\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIdentification of acute ischemic stroke, features of acute ischemic stroke reports versus nonischemic stroke reports\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDavis et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref40\" rel=\"footnote\"\u003E40\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EOctober 22, 2013\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes, letters, and problem lists\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMS\u003Csup\u003Em\u003C\u002Fsup\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EN\u002FA\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EExtraction of clinical traits of patients with MS\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EGlauser et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref41\" rel=\"footnote\"\u003E41\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EJanuary 22, 2020\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESpeech\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy psychiatric comorbidities\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ECohen et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref42\" rel=\"footnote\"\u003E42\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMay 22, 2016\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrognosis or risk stratification, management or therapy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, naïve Bayes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIdentification of potential candidates for surgical intervention for pediatric drug–resistant epilepsy, performance of classification algorithm over time\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAlim-Marvasti et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref43\" rel=\"footnote\"\u003E43\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EFebruary 10, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited Kingdom\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes and radiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity, prognosis or risk stratification\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, naïve Bayes, random forest, logistic regression, boosting\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELocalizing the epileptogenic zone (temporal vs extra-temporal), postsurgical prognosis and outcome\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EBalagopalan et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref44\" rel=\"footnote\"\u003E44\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EApril 27, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ECanada\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESpeech\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENonclinical neuroscience\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAlzheimer disease\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, naïve Bayes, random forest, linear regression, shallow neural network, ridge regression, transformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection of Alzheimer disease from speech, prediction of MMSE\u003Csup\u003En\u003C\u002Fsup\u003E\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMartinc et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref45\" rel=\"footnote\"\u003E45\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EJune 14, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESlovenia\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESpeech\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENonclinical neuroscience\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAlzheimer disease\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, random forest, logistic regression, boosting, transformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection of Alzheimer disease from speech\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELiu et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref46\" rel=\"footnote\"\u003E46\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EApril 5, 2022\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESpeech\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAlzheimer disease\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EShallow neural network, transformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection of Alzheimer disease from speech\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENelson et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref47\" rel=\"footnote\"\u003E47\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDecember 22, 2016\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPharmacy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMS\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EN\u002FA\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIdentification of MS phenotype, percentages of each phenotype\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDeng et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref18\" rel=\"footnote\"\u003E18\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EApril 8, 2022\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EChina\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes and radiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENonclinical neuroscience\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EManagement or therapy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ETransformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPerformance of system to generate ICH\u003Csup\u003Eo\u003C\u002Fsup\u003E treatment plan\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EChase et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref48\" rel=\"footnote\"\u003E48\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EFebruary 28, 2017\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMS\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENaïve Bayes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEarly detection of MS\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EWissel et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref49\" rel=\"footnote\"\u003E49\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENovember 29, 2019\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrognosis or risk stratification, management or therapy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy surgery candidacy score\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESung et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref50\" rel=\"footnote\"\u003E50\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EFebruary 28, 2020\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ETaiwan\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESVM, random forest, decision tree, logistic regression, KNN, ensemble\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClassification of ischemic stroke subtypes\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESung et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref20\" rel=\"footnote\"\u003E20\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENovember 19, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ETaiwan\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes and radiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrognosis or risk stratification\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERandom forest, logistic regression, transformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrediction of poor functional outcome after acute ischemic stroke\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYang et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref51\" rel=\"footnote\"\u003E51\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EOctober 20, 2020\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ECanada\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMS\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EShallow neural network, CNN, RNN\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EExpanded disability status scale score, expanded disability status scale subscore\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EXie et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref52\" rel=\"footnote\"\u003E52\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EFebruary 22, 2022\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ETransformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESeizure freedom, seizure frequency, date of last seizure\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESung et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref53\" rel=\"footnote\"\u003E53\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EFebruary 8, 2018\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ETaiwan\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EManagement or therapy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EN\u002FA\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPerformance of EMR\u003Csup\u003Ep\u003C\u002Fsup\u003E interface that determines eligibility for intravenous thrombolytic therapy\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESung et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref54\" rel=\"footnote\"\u003E54\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EFebruary 17, 2022\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ETaiwan\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes and radiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrognosis or risk stratification\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELogistic regression, boosting, unspecified penalized logistic regression method, ensemble (extra trees classifier)\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrediction of poor functional outcome after acute ischemic stroke\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EXia et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref55\" rel=\"footnote\"\u003E55\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENovember 11, 2013\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes and radiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENonclinical neuroscience\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMS\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis, clinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELasso regression, stepwise regression\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIdentification of patients with MS, severity of MS\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EOng et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref22\" rel=\"footnote\"\u003E22\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EJune 19, 2020\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited States\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERadiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENonclinical neuroscience\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis, clinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERandom forest, decision tree, logistic regression, KNN, RNN (LSTM)\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIschemic stroke presence, location, and acuity\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERoshanzamir et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref56\" rel=\"footnote\"\u003E56\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMarch 9, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EIran\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESpeech\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAlzheimer disease\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection or diagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ELogistic regression, shallow neural network, CNN, RNN (LSTM) transformer\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDetection of Alzheimer disease from speech\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERannikmäe et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref57\" rel=\"footnote\"\u003E57\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EJune 15, 2021\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EUnited Kingdom\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERadiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENo\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical disease phenotyping or severity\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based, ML\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EYes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERNN\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke subtypes\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003C\u002Ftbody\u003E\u003C\u002Ftable\u003E\u003Cfn id=\"table1fn1\"\u003E\u003Cp\u003E\u003Csup\u003Ea\u003C\u002Fsup\u003ENLP: natural language processing.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn2\"\u003E\u003Cp\u003E\u003Csup\u003Eb\u003C\u002Fsup\u003EML: machine learning.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn3\"\u003E\u003Cp\u003E\u003Csup\u003Ec\u003C\u002Fsup\u003EKNN: k-nearest neighbor.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn4\"\u003E\u003Cp\u003E\u003Csup\u003Ed\u003C\u002Fsup\u003EMLP: multilayer perceptron.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn5\"\u003E\u003Cp\u003E\u003Csup\u003Ee\u003C\u002Fsup\u003EPNES: psychogenic nonepileptic seizures.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn6\"\u003E\u003Cp\u003E\u003Csup\u003Ef\u003C\u002Fsup\u003ESVM: support vector machine.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn7\"\u003E\u003Cp\u003E\u003Csup\u003Eg\u003C\u002Fsup\u003ECNN: convolutional neural network.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn8\"\u003E\u003Cp\u003E\u003Csup\u003Eh\u003C\u002Fsup\u003ERNN: recurrent neural network.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn9\"\u003E\u003Cp\u003E\u003Csup\u003Ei\u003C\u002Fsup\u003ELSTM: long- and short-term memory network.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn10\"\u003E\u003Cp\u003E\u003Csup\u003Ej\u003C\u002Fsup\u003EN\u002FA: Not applicable.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn11\"\u003E\u003Cp\u003E\u003Csup\u003Ek\u003C\u002Fsup\u003ETIA: transient ischemic attack.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn12\"\u003E\u003Cp\u003E\u003Csup\u003El\u003C\u002Fsup\u003ESUDEP: sudden unexpected death in epilepsy.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn13\"\u003E\u003Cp\u003E\u003Csup\u003Em\u003C\u002Fsup\u003EMS: multiple sclerosis.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn14\"\u003E\u003Cp\u003E\u003Csup\u003En\u003C\u002Fsup\u003EMMSE: Mini-Mental Status Examination.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn15\"\u003E\u003Cp\u003E\u003Csup\u003Eo\u003C\u002Fsup\u003EICH: intracerebral hemorrhage.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table1fn16\"\u003E\u003Cp\u003E\u003Csup\u003Ep\u003C\u002Fsup\u003EEMR: electronic medical record.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003C\u002Fdiv\u003E\u003Cdiv class=\"figure-table\"\u003E\u003Cfigcaption\u003E\u003Cspan class=\"typcn typcn-clipboard\"\u003E\u003C\u002Fspan\u003E\u003Cb\u003ETable 2. \u003C\u002Fb\u003EOverall study characteristics: journal field, target of NLP\u003Csup\u003Ea\u003C\u002Fsup\u003E, and neurological condition.\u003C\u002Ffigcaption\u003E\u003Ctable width=\"1000\" cellpadding=\"5\" cellspacing=\"0\" border=\"1\" rules=\"groups\" frame=\"hsides\"\u003E\u003Ccol width=\"30\" span=\"1\"\u003E\u003Ccol width=\"670\" span=\"1\"\u003E\u003Ccol width=\"300\" span=\"1\"\u003E\u003Cthead\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd colspan=\"2\" rowspan=\"1\"\u003EStudy characteristics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStudies (n=41), n (%)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003C\u002Fthead\u003E\u003Ctbody\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd colspan=\"3\" rowspan=\"1\"\u003E\u003Cb\u003ECondition\u003C\u002Fb\u003E\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStroke\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E20 (49)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EEpilepsy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E10 (24)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EAlzheimer disease\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E6 (15)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMultiple sclerosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E5 (12)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd colspan=\"3\" rowspan=\"1\"\u003E\u003Cb\u003ETarget of NLP\u003C\u002Fb\u003E\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDiagnosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E20 (49)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPhenotyping\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E17 (42)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EPrognosis\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E9 (22)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ETherapy\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E4 (10)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd colspan=\"3\" rowspan=\"1\"\u003E\u003Cb\u003EJournal field\u003C\u002Fb\u003E\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMedical informatics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E15 (37)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical neurology\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E14 (34)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ENonclinical neuroscience\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E7 (17)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical medicine\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E2 (5)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EOther\u003Csup\u003Eb\u003C\u002Fsup\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E3 (7)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003C\u002Ftbody\u003E\u003C\u002Ftable\u003E\u003Cfn id=\"table2fn1\"\u003E\u003Cp\u003E\u003Csup\u003Ea\u003C\u002Fsup\u003ENLP: natural language processing.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table2fn2\"\u003E\u003Cp\u003E\u003Csup\u003Eb\u003C\u002Fsup\u003EOther includes studies published in pharmacy, public health, and neuroradiology journals.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003C\u002Fdiv\u003E\u003Cp class=\"abstract-paragraph\"\u003EOf the 41 studies, the language sources for NLP comprised clinical notes (n=25, 61%); radiology reports (n=14, 34%); speech (n=8, 20%); and other sources (n=2, 5%) that included echocardiography reports, letters to referring providers, and problem lists (\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#table3\" rel=\"footnote\"\u003ETable 3\u003C\u002Fa\u003E\u003C\u002Fspan\u003E). Of studies with speech as the language source, half (4\u002F8, 50%) analyzed transcripts only, whereas half additionally incorporated acoustic features from the audio files themselves. These transcripts and audio files were largely from research datasets (eg, ADReSS and Pitt corpus). Two studies analyzed transcripts from interviews with patients. In the study including problem lists, it is unknown who reported the problems.\u003C\u002Fp\u003E\u003Cdiv class=\"figure-table\"\u003E\u003Cfigcaption\u003E\u003Cspan class=\"typcn typcn-clipboard\"\u003E\u003C\u002Fspan\u003E\u003Cb\u003ETable 3. \u003C\u002Fb\u003EOverall study characteristics: NLP\u003Csup\u003Ea\u003C\u002Fsup\u003E methods and language sources.\u003C\u002Ffigcaption\u003E\u003Ctable width=\"1000\" cellpadding=\"5\" cellspacing=\"0\" border=\"1\" rules=\"groups\" frame=\"hsides\"\u003E\u003Ccol width=\"30\" span=\"1\"\u003E\u003Ccol width=\"670\" span=\"1\"\u003E\u003Ccol width=\"300\" span=\"1\"\u003E\u003Cthead\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd colspan=\"2\" rowspan=\"1\"\u003EStudy characteristics\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EStudies (n=41), n (%)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003C\u002Fthead\u003E\u003Ctbody\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd colspan=\"3\" rowspan=\"1\"\u003E\u003Cb\u003ENLP method\u003C\u002Fb\u003E\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERule-based\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E23 (56)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EMachine learning\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E35 (85)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd colspan=\"3\" rowspan=\"1\"\u003E\u003Cb\u003EType of\u003C\u002Fb\u003E\u003Cb\u003Emachine learning\u003C\u002Fb\u003E\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EConventional machine learning\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E31 (76)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EDeep learning\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E16 (39)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd colspan=\"3\" rowspan=\"1\"\u003E\u003Cb\u003ESource text\u003C\u002Fb\u003E\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EClinical notes\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E25 (61)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ERadiology reports\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E14 (34)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003ESpeech\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E8 (20)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr valign=\"top\"\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E\u003Cbr\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003EOther\u003Csup\u003Eb\u003C\u002Fsup\u003E\u003C\u002Ftd\u003E\u003Ctd rowspan=\"1\" colspan=\"1\"\u003E2 (5)\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003C\u002Ftbody\u003E\u003C\u002Ftable\u003E\u003Cfn id=\"table3fn1\"\u003E\u003Cp\u003E\u003Csup\u003Ea\u003C\u002Fsup\u003ENLP: natural language processing.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003Cfn id=\"table3fn2\"\u003E\u003Cp\u003E\u003Csup\u003Eb\u003C\u002Fsup\u003EOther includes echocardiography reports, problem lists, and letters to referring providers.\u003C\u002Fp\u003E\u003C\u002Ffn\u003E\u003C\u002Fdiv\u003E\u003Cp class=\"abstract-paragraph\"\u003EOf the 41 studies, the most common source language for NLP was English (n=39, 95%), Portuguese in 1 (2%) study, and unspecified in the remaining 1 study (which was of Chinese nationality, not multicentric). When patient population size was recorded, the median was 1091 (IQR 188-4211). In studies that did not specify a population size (n=4, 10%), the median number of clinical or radiographic notes was 2172 (IQR 1155.5-22,018.0).\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EPapers were most commonly published in medical informatics (n=15, 37%) journals, followed closely by clinical neurology (n=14, 34%) journals. Seven (17%) studies were published in nonclinical neuroscience journals; 2 (5%) in clinical medicine journals; and 1 (2%) each in neuroradiology, public health, and pharmacy journals. Studies were mostly conducted in the United States (n=21, 51%), followed by Taiwan (n=4, 10%) and the United Kingdom, Canada, and Australia (n=3, 7% each). Two (5%) studies were conducted in China, and 1 (2%) study was conducted in each of South Korea, Brazil, Iran, India, and Slovenia (\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#figure2\" rel=\"footnote\"\u003EFigure 2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E).\u003C\u002Fp\u003E\u003Cfigure\u003E\u003Ca name=\"figure2\"\u003E‎\u003C\u002Fa\u003E\u003Ca class=\"fancybox\" title=\"Figure 2. Proportion of included studies (n=41), organized according to country of origin: the United States (n=21, 51%); Taiwan (n=4, 10%); the United Kingdom, Canada, and Australia (n=3, 7% each); China (n=2, 5%); and South Korea, Brazil, Iran, India, and Slovenia (n=1, 2% each).\" href=\"https:\u002F\u002Fasset.jmir.pub\u002Fassets\u002F13d0fcca15153fd67444eab827f368fb.png\" id=\"figure2\"\u003E\u003Cimg class=\"figure-image\" src=\"https:\u002F\u002Fasset.jmir.pub\u002Fassets\u002F13d0fcca15153fd67444eab827f368fb.png\"\u003E\u003C\u002Fa\u003E\u003Cfigcaption\u003E\u003Cspan class=\"typcn typcn-image\"\u003E\u003C\u002Fspan\u003E\u003Cb\u003EFigure 2. \u003C\u002Fb\u003E Proportion of included studies (n=41), organized according to country of origin: the United States (n=21, 51%); Taiwan (n=4, 10%); the United Kingdom, Canada, and Australia (n=3, 7% each); China (n=2, 5%); and South Korea, Brazil, Iran, India, and Slovenia (n=1, 2% each). \u003C\u002Ffigcaption\u003E\u003C\u002Ffigure\u003E\u003Cp class=\"abstract-paragraph\"\u003EOnly 6 (15%) studies used strictly rule-based methods. The majority of studies incorporated ML (n=35, 85%), either exclusively (n=18, 44%) or in combination with rule-based methods (n=17, 41%). Of the studies that used ML, most (n=31, 89%) used conventional ML methods, whereas 16 (46%) used DL approaches (\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#table3\" rel=\"footnote\"\u003ETable 3\u003C\u002Fa\u003E\u003C\u002Fspan\u003E), and 12 (34%) used a combination of both conventional ML and DL approaches.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EAs shown in \u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#figure3\" rel=\"footnote\"\u003EFigure 3\u003C\u002Fa\u003E\u003C\u002Fspan\u003E, the most frequently used conventional ML algorithms were random forest (n=18, 58%), SVM (n=15, 48%), and logistic regression (n=15, 48%) models. Among studies using DL approaches, transformers (n=10, 63%) were the most commonly used algorithm, followed by convolutional neural networks and RNNs (each n=7, 44%). The co-occurrence of random forest and transformer algorithms was a prevalent trend in research combining traditional ML with DL methodologies (n=6, 15%). Studies that used DL only began to appear in 2019 and later (\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#figure4\" rel=\"footnote\"\u003EFigure 4\u003C\u002Fa\u003E\u003C\u002Fspan\u003E). The most often reported performance metrics for ML models were precision or recall (n=31, 76%), accuracy (n=22, 54%), area under the receiver operating curve (n=20, 49%), and \u003Ci\u003EF\u003C\u002Fi\u003E\u003Csub\u003E1\u003C\u002Fsub\u003E-score (n=19, 46%).\u003C\u002Fp\u003E\u003Cfigure\u003E\u003Ca name=\"figure3\"\u003E‎\u003C\u002Fa\u003E\u003Ca class=\"fancybox\" title=\"Figure 3. Relative proportions of machine learning algorithms used by the included NLP models. CNN: convolutional neural network; KNN: k-nearest neighbor; LSTM: long- and short-term memory networks; MLP: multilayer perceptron; RNN: recurrent neural network; SVM: support vector machine. *Other includes stepwise regression, ridge regression, an unspecified penalized regression method, latent Dirichlet allocation, and an unspecified neural network with an unspecified number of layers.\" href=\"https:\u002F\u002Fasset.jmir.pub\u002Fassets\u002F7f2f9de6933564bb7d73b0d8e8710712.png\" id=\"figure3\"\u003E\u003Cimg class=\"figure-image\" src=\"https:\u002F\u002Fasset.jmir.pub\u002Fassets\u002F7f2f9de6933564bb7d73b0d8e8710712.png\"\u003E\u003C\u002Fa\u003E\u003Cfigcaption\u003E\u003Cspan class=\"typcn typcn-image\"\u003E\u003C\u002Fspan\u003E\u003Cb\u003EFigure 3. \u003C\u002Fb\u003E Relative proportions of machine learning algorithms used by the included NLP models. CNN: convolutional neural network; KNN: k-nearest neighbor; LSTM: long- and short-term memory networks; MLP: multilayer perceptron; RNN: recurrent neural network; SVM: support vector machine. *Other includes stepwise regression, ridge regression, an unspecified penalized regression method, latent Dirichlet allocation, and an unspecified neural network with an unspecified number of layers. \u003C\u002Ffigcaption\u003E\u003C\u002Ffigure\u003E\u003Cfigure\u003E\u003Ca name=\"figure4\"\u003E‎\u003C\u002Fa\u003E\u003Ca class=\"fancybox\" title=\"Figure 4. Number of studies applying natural language processing (NLP) to neurological conditions, stratified by NLP methodology and publication year.\" href=\"https:\u002F\u002Fasset.jmir.pub\u002Fassets\u002F7fc5c1273a1b60a24f3f10d91c1fca43.png\" id=\"figure4\"\u003E\u003Cimg class=\"figure-image\" src=\"https:\u002F\u002Fasset.jmir.pub\u002Fassets\u002F7fc5c1273a1b60a24f3f10d91c1fca43.png\"\u003E\u003C\u002Fa\u003E\u003Cfigcaption\u003E\u003Cspan class=\"typcn typcn-image\"\u003E\u003C\u002Fspan\u003E\u003Cb\u003EFigure 4. \u003C\u002Fb\u003E Number of studies applying natural language processing (NLP) to neurological conditions, stratified by NLP methodology and publication year. \u003C\u002Ffigcaption\u003E\u003C\u002Ffigure\u003E\u003Cp class=\"abstract-paragraph\"\u003EAll 41 studies were model derivation studies, with only 7 (17%) studies conducting additional external validation (\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#app2\" rel=\"footnote\"\u003EMultimedia Appendix 2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E). Furthermore, nearly all the study models were developed retrospectively and were not applied in practice or deployed in real-world environments, except for 3 studies. A study by Li et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref16\" rel=\"footnote\"\u003E16\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] developed a model for stroke detection from imaging reports and then applied it to quantify the change in stroke cases before and during the COVID-19 pandemic. A second by Sung et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref53\" rel=\"footnote\"\u003E53\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], also in the stroke category, evaluated the deployment of a user-interface system to determine intravenous thrombolysis eligibility built on the NLP model devised. A third study by Wissel et al [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref49\" rel=\"footnote\"\u003E49\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] created a model to identify surgical resection candidates in adult patients with epilepsy. The model was retrained prospectively to incorporate new information.\u003C\u002Fp\u003E\u003Ch4\u003EStudy Characteristics, Stratified by Condition\u003C\u002Fh4\u003E\u003Cp class=\"abstract-paragraph\"\u003EIn studies focused on Alzheimer dementia, diagnosis and detection was the only target of NLP (6\u002F6, 100%). Disease phenotyping and subtyping was the most common purpose of NLP in stroke (10\u002F20, 50%) and MS (4\u002F5, 80%), whereas prognostication was seen as often as diagnosis in epilepsy studies (4\u002F10, 40%; Figure S9 in \u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#app2\" rel=\"footnote\"\u003EMultimedia Appendix 2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E). Studies that applied NLP for the purpose of disease treatment or management were limited to stroke and epilepsy (Figure S9 in \u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#app2\" rel=\"footnote\"\u003EMultimedia Appendix 2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E).\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003ERule-based methods were used across all studies, except for Alzheimer dementia, in which only ML approaches were used (Figure S10 in \u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#app2\" rel=\"footnote\"\u003EMultimedia Appendix 2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E). Conventional ML methods were used most often by Alzheimer dementia studies (5\u002F6, 83%), followed by stroke (16\u002F20, 80%). Similarly, DL methods were used predominantly by Alzheimer dementia (6\u002F6, 100%) and stroke (8\u002F20, 40%) studies (Figure S10 in \u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#app2\" rel=\"footnote\"\u003EMultimedia Appendix 2\u003C\u002Fa\u003E\u003C\u002Fspan\u003E). The transformer was the DL method used most frequently in Alzheimer disease-related studies (5\u002F6, 83%).\u003C\u002Fp\u003E\u003Cbr\u003E\u003Ch3 class=\"navigation-heading h3-main-heading\" id=\"Discussion\" data-label=\"Discussion\"\u003EDiscussion\u003C\u002Fh3\u003E\u003Ch4\u003EPrincipal Findings\u003C\u002Fh4\u003E\u003Cp class=\"abstract-paragraph\"\u003EIn this scoping review, 41 studies [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref13\" rel=\"footnote\"\u003E13\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref16\" rel=\"footnote\"\u003E16\u003C\u002Fa\u003E\u003C\u002Fspan\u003E-\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref22\" rel=\"footnote\"\u003E22\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref25\" rel=\"footnote\"\u003E25\u003C\u002Fa\u003E\u003C\u002Fspan\u003E-\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref57\" rel=\"footnote\"\u003E57\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] that investigated direct clinical applications of NLP to common neurological disorders were identified. We found that the majority of these studies focused on detection and diagnosis and applied NLP to stroke, whereas we found no studies of NLP that met our eligibility criteria in the clinical areas of migraine or Parkinson disease. Methodologically, ML techniques were used more often than rule-based methods, but a considerable number of studies still relied on rule-based approaches in combination with ML. While we observed that DL began to emerge as a methodology for NLP in 2019, we found that the transformer was the most commonly used DL algorithm overall.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EAt the time of writing, we believe our scoping review to be the first to examine direct clinical NLP applications in common neurological conditions. One prior review [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref58\" rel=\"footnote\"\u003E58\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] investigated NLP applications across the combined clinical specialties of neurosurgery, spine surgery, and neurology, whereas another evaluated the use of NLP in both psychiatry and clinical neuroscience [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref59\" rel=\"footnote\"\u003E59\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. However, neither reviews analyzed studies and NLP applications according to neurological condition. More importantly, these reviews included many studies where NLP was not applied for direct clinical use, instead aiming to perform tasks such as characterizing patient cohorts [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref58\" rel=\"footnote\"\u003E58\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], analyzing information extraction, or determining causal inference between concepts [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref59\" rel=\"footnote\"\u003E59\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. In contrast to this prior work, our review focused on direct clinical applications of NLP.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EOf note, we found no studies applying NLP to migraine or Parkinson disease that met our eligibility criteria, thereby highlighting a potential gap in NLP research focusing on these disorders. This is perhaps unexpected, as the combined prevalence of migraine and Parkinson disease in the United States exceeds that of both stroke and MS [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref12\" rel=\"footnote\"\u003E12\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Two explanations may account for this finding. One is that migraine and Parkinson disease may rely less on radiographic imaging studies and their reports to establish a diagnosis than stroke, Alzheimer dementia, or MS. Given that many ML applications in stroke have focused on neuroimaging [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref60\" rel=\"footnote\"\u003E60\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], it is plausible that stroke imaging reports could represent an important source of data for NLP analyses. Indeed, the results of our review demonstrate that stroke-related NLP studies made use of radiographic reports as often as clinical notes for source text, which could have resulted in a relatively higher number of NLP studies within stroke than in other neurological conditions.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EA second explanation may be that Alzheimer disease is a more common cause of dementia worldwide than dementing syndromes associated with Parkinson disease [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref61\" rel=\"footnote\"\u003E61\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] and has in turn garnered a larger proportion of research funding. National Institutes of Health [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref62\" rel=\"footnote\"\u003E62\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] research funding for Alzheimer dementia was approximately US $3 billion in 2022, as compared to US $259 million for Parkinson disease.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EOur finding that NLP was most frequently applied to diagnostic problems is expected, given that clinical decision support is a common focus of artificial intelligence in medicine [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref63\" rel=\"footnote\"\u003E63\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Historically, clinical decision support has also played an important role in medical informatics by constituting the main focus of archetypal systems such as MYCIN, INTERNIST-1, and DXplain, which were first developed in the 1970s and 1980s [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref64\" rel=\"footnote\"\u003E64\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. An alternative explanation is that the shortage of neurologists that already exists worldwide [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref65\" rel=\"footnote\"\u003E65\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] may have potentially created a more urgent need for detection-oriented NLP applications rather than NLP applications targeting therapeutic management or prognostication.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EThough diagnosis was the most common target of NLP overall, we found that epilepsy-related studies focused as much on prognostication as they did on diagnostic tasks. Given that roughly one-third of all patients with epilepsy are drug resistant [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref66\" rel=\"footnote\"\u003E66\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], determining good surgical resection candidates as well as predicting surgical outcomes are important objectives that have been the focus of considerable research [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref67\" rel=\"footnote\"\u003E67\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Consistent with this, the epilepsy-related studies in the prognostication category were directed toward identifying adult [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref49\" rel=\"footnote\"\u003E49\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] and pediatric [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref42\" rel=\"footnote\"\u003E42\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] surgical candidates, predicting postsurgical outcomes [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref43\" rel=\"footnote\"\u003E43\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], and detecting risk factors for sudden unexpected death in epilepsy [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref17\" rel=\"footnote\"\u003E17\u003C\u002Fa\u003E\u003C\u002Fspan\u003E].\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EWith respect to the types of ML models we found in our review, the relatively high proportion of conventional ML-based studies using random forest and SVM (18\u002F31, 58% and 15\u002F31, 48%, respectively) may have been related to the fact that SVM together with random forest models generally represented the dominant ML techniques prior to the advent of neural networks [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref68\" rel=\"footnote\"\u003E68\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] in diagnostic and clinical decision support applications [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref63\" rel=\"footnote\"\u003E63\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref69\" rel=\"footnote\"\u003E69\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref70\" rel=\"footnote\"\u003E70\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Despite its position as a potentially more basic classification method than either SVM or random forest, logistic regression was used as commonly as SVM in our analysis.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EFurthermore, while we found that SVM and random forest models were common in ML-based NLP approaches, the optimal problems these models address are fundamentally different. SVM generally works best as a binary classifier, whereas random forest models are best used for classification tasks involving multiple categories [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref71\" rel=\"footnote\"\u003E71\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. We found that the most frequently used ML algorithms in stroke-related NLP studies were random forest models. This matches the most frequent target of NLP in stroke-related studies, which was disease subtyping (a multiple classification problem).\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EAmong DL algorithms, which are becoming increasingly widespread in NLP [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref72\" rel=\"footnote\"\u003E72\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], the transformer was the most commonly used technique we identified. Unlike other word embedding methods, a transformer processes a whole sequence of text while preserving the context and meaning of words [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref59\" rel=\"footnote\"\u003E59\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref73\" rel=\"footnote\"\u003E73\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Another significant advantage of transformers is that they can use transfer learning, which first trains a model on a learning task and then applies the model to a separate but closely related task [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref58\" rel=\"footnote\"\u003E58\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref74\" rel=\"footnote\"\u003E74\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. A prevalent example of transfer learning in our results is Bidirectional Encoder Representations From Transformers (BERT), a transformer model that was originally trained using publicly available text from Wikipedia and BookCorpus, a collection of free, unpublished novels consisting of over 50 million sentences [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref75\" rel=\"footnote\"\u003E75\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref76\" rel=\"footnote\"\u003E76\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. BERT can then be further refined on a target training task and dataset before being passed to a separate classification algorithm [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref28\" rel=\"footnote\"\u003E28\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. This is helpful in situations where the target training set is small [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref28\" rel=\"footnote\"\u003E28\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. The high frequency of Alzheimer disease–related NLP studies we found using BERT is expected within this context, as these studies often used the ADReSS speech dataset that consists of only 78 healthy controls and 78 patients with Alzheimer disease [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref28\" rel=\"footnote\"\u003E28\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref45\" rel=\"footnote\"\u003E45\u003C\u002Fa\u003E\u003C\u002Fspan\u003E].\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EA particularly important finding of our review is that although many of the NLP studies leveraged powerful and sophisticated computational tools, most studies constitute research work rather than reports of operationalization or evaluation in practical settings. This is consistent with the current state of clinical NLP outside of neurology, wherein real-world deployment of NLP models continues to be limited [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref7\" rel=\"footnote\"\u003E7\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref77\" rel=\"footnote\"\u003E77\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref78\" rel=\"footnote\"\u003E78\u003C\u002Fa\u003E\u003C\u002Fspan\u003E].\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EOne major obstacle to the implementation of NLP in clinical practice is model generalizability [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref7\" rel=\"footnote\"\u003E7\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Published NLP models are usually internally validated rather than externally validated [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref7\" rel=\"footnote\"\u003E7\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref17\" rel=\"footnote\"\u003E17\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], limiting the understanding of model accuracy beyond the model’s original training environment [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref60\" rel=\"footnote\"\u003E60\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. We found this to be true for the majority of studies identified in our review. The lack of EMR standardization, including note formatting [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref17\" rel=\"footnote\"\u003E17\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref78\" rel=\"footnote\"\u003E78\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], documentation styles, and radiographic report structures across different medical institutions [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref7\" rel=\"footnote\"\u003E7\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] and between clinicians, may partly account for our observations. Furthermore, the preponderance of English language as source text in NLP [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref79\" rel=\"footnote\"\u003E79\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], as demonstrated by the single study in our review using non-English (Portuguese) text for analysis, suggests that the generalizability of NLP within neurology is most likely limited outside the English language.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EAnother major obstacle impeding the adoption of NLP tools is the inherent lack of transparency of ML-based algorithms [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref60\" rel=\"footnote\"\u003E60\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], particularly artificial neural networks and other forms of DL approaches [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref80\" rel=\"footnote\"\u003E80\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. These approaches have low transparency because the computational methods they use to characterize relationships between inputs and outputs are not readily intelligible to humans [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref7\" rel=\"footnote\"\u003E7\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref78\" rel=\"footnote\"\u003E78\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref80\" rel=\"footnote\"\u003E80\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] acting as a black box that could undermine clinicians’ trust in their performance.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EThe lack of well-defined regulatory guidelines and standards overseeing the artificial intelligence space [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref81\" rel=\"footnote\"\u003E81\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] has furthered this mistrust. Compromise of personal health data, algorithmic bias, and the question of how to attribute culpability when diagnostic errors arise [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref82\" rel=\"footnote\"\u003E82\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref83\" rel=\"footnote\"\u003E83\u003C\u002Fa\u003E\u003C\u002Fspan\u003E] are all ethical concerns that may serve to explain the relative paucity of studies across all neurological conditions that externally validated DL models.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EFinally, the lack of portability of NLP applications into external EMRs is another factor that has restricted the development of NLP models to the research arena. External software modules containing ML and DL models are challenging to integrate into EMRs [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref1\" rel=\"footnote\"\u003E1\u003C\u002Fa\u003E\u003C\u002Fspan\u003E,\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref84\" rel=\"footnote\"\u003E84\u003C\u002Fa\u003E\u003C\u002Fspan\u003E], as most implementations require a high level of computing infrastructure and technical expertise that many hospital information technology systems and personnel may lack [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref84\" rel=\"footnote\"\u003E84\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. Recent work suggests few EMR-integrated aggregative tools exist to display NLP findings to clinicians in a digestible format [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref85\" rel=\"footnote\"\u003E85\u003C\u002Fa\u003E\u003C\u002Fspan\u003E]. To address these barriers, some authors have advocated for collaborations between NLP researchers and EMR companies [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref77\" rel=\"footnote\"\u003E77\u003C\u002Fa\u003E\u003C\u002Fspan\u003E].\u003C\u002Fp\u003E\u003Ch4\u003ELimitations and Future Work\u003C\u002Fh4\u003E\u003Cp class=\"abstract-paragraph\"\u003EOur scoping review has several limitations. First, we note that the target of NLP was categorized according to author experience and interpretation of the literature, which may have underreported the application of the published NLP algorithms. Second, due to the variable performance metrics and outcomes across studies, we did not aggregate measurements of performance in our review, and we therefore could not reliably provide summary performance metrics for NLP models within individual diseases, applications, or outcomes. Future work should focus on individual outcomes within a clinical disorder for a more exact appraisal of NLP model performance than this review.\u003C\u002Fp\u003E\u003Cp class=\"abstract-paragraph\"\u003EThird, this review only included studies based on common neurological disorders, direct clinical applications of NLP, and homogeneous clinical populations, which limited the number of studies we identified. It is therefore important to note that this review cannot be used to make definitive conclusions on the state of NLP research across all neurological disorders. Future efforts can be directed at characterizing the use of NLP across less common neurological disorders as well as in heterogeneous or ambiguously defined clinical populations. As NLP technologies continue to advance, it will also be critically important to evaluate studies that use newer transformers, such as GPT3, which have better performance than BERT models [\u003Cspan class=\"footers\"\u003E\u003Ca class=\"citation-link\" href=\"#ref59\" rel=\"footnote\"\u003E59\u003C\u002Fa\u003E\u003C\u002Fspan\u003E].\u003C\u002Fp\u003E\u003Ch4\u003EConclusions\u003C\u002Fh4\u003E\u003Cp class=\"abstract-paragraph\"\u003EThe abundance of unstructured text data in modern-day EMRs as well as the emphasis in neurology on narrative history and physical examination and heavy reliance on ancillary information such as radiographic reports and speech, all create an optimal use case for applying NLP for the diagnosis, management, or prognostication of neurological disorders. To our knowledge, this is the first attempt to systematically characterize research efforts to investigate direct NLP applications to common neurological conditions. Our review reveals gaps in neurological NLP research, showing a relative deficiency of NLP studies in subspecialties outside of stroke or epilepsy, and underlines the need to actualize NLP models outside of the research phase. Moreover, the current emphasis of NLP on diagnostic tasks suggests that NLP may be particularly useful in settings that lack access to neurological expertise.\u003C\u002Fp\u003E\u003C\u002Farticle\u003E\u003Ch4\u003EFunding\u003C\u002Fh4\u003E\u003Cp class=\"abstract-paragraph\"\u003ENone.\u003C\u002Fp\u003E\u003Ch4 class=\"h4-border-top\"\u003EConflicts of Interest\u003C\u002Fh4\u003E\u003Cp\u003E\u003Cp class=\"abstract-paragraph\"\u003ENJ receives an honorarium for her work as an associate editor of Epilepsia. There are no other conflicts of interest to report.\u003C\u002Fp\u003E\u003C\u002Fp\u003E\u003Cdiv id=\"app1\" name=\"app1\"\u003EMultimedia Appendix 1\u003Cp class=\"abstract-paragraph\"\u003EPRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews): checklist and explanation.\u003C\u002Fp\u003E\u003Ca href=\"https:\u002F\u002Fjmir.org\u002Fapi\u002Fdownload?alt_name=neuro_v3i1e51822_app1.pdf&filename=b857986a2f57ecb484f5e9759b0c26b0.pdf\" target=\"_blank\"\u003EPDF File (Adobe PDF File), 546 KB\u003C\u002Fa\u003E\u003C\u002Fdiv\u003E\u003Chr\u003E\u003Cdiv id=\"app2\" name=\"app2\"\u003EMultimedia Appendix 2\u003Cp class=\"abstract-paragraph\"\u003ESearch strategy and additional data.\u003C\u002Fp\u003E\u003Ca href=\"https:\u002F\u002Fjmir.org\u002Fapi\u002Fdownload?alt_name=neuro_v3i1e51822_app2.docx&filename=6b7610b4e5c37871e8cca3eedfe7720c.docx\" target=\"_blank\"\u003EDOCX File , 756 KB\u003C\u002Fa\u003E\u003C\u002Fdiv\u003E\u003Chr\u003E\u003Cdiv class=\"footnotes\"\u003E\u003Ch4 id=\"References\" class=\"h4-border-top navigation-heading\" data-label=\"References\"\u003EReferences\u003C\u002Fh4\u003E\u003Col\u003E\u003Cli\u003E\u003Cspan id=\"ref1\"\u003EPivovarov R, Elhadad N. Automated methods for the summarization of electronic health records. J Am Med Inform Assoc. 2015;22(5):938-947. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F25882031\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1093\u002Fjamia\u002Focv032\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=25882031&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref2\"\u003ELocke S, Bashall A, Al-Adely S, Moore J, Wilson A, Kitchen GB. Natural language processing in medicine: a review. Trends Anaesth Crit Care. 2021;38:4-9. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.tacc.2021.02.007\"\u003ECrossRef\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref3\"\u003EKimia AA, Savova G, Landschaft A, Harper MB. An introduction to natural language processing: how you can get more from those electronic notes you are generating. Pediatr Emerg Care. 2015;31(7):536-541. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1097\u002FPEC.0000000000000484\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=26148107&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref4\"\u003EIroju OG, Olaleke JO. A systematic review of natural language processing in healthcare. Int J Inf Technol Comput Sci. 2015;7(8):44-50. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.5815\u002Fijitcs.2015.08.07\"\u003ECrossRef\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref5\"\u003EWang J, Deng H, Liu B, Hu A, Liang J, Fan L, et al. Systematic evaluation of research progress on natural language processing in medicine over the past 20 years: bibliometric study on PubMed. J Med Internet Res. 2020;22(1):e16816. [\u003Ca href=\"https:\u002F\u002Fwww.jmir.org\u002F2020\u002F1\u002Fe16816\u002F\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.2196\u002F16816\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=32012074&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref6\"\u003EGunter D, Puac-Polanco P, Miguel O, Thornhill RE, Yu AYX, Liu ZA, et al. Rule-based natural language processing for automation of stroke data extraction: a validation study. Neuroradiology. 2022;64(12):2357-2362. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1007\u002Fs00234-022-03029-1\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35913525&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref7\"\u003EPons E, Braun LMM, Hunink MGM, Kors JA. Natural language processing in radiology: a systematic review. Radiology. 2016;279(2):329-343. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1148\u002Fradiol.16142770\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=27089187&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref8\"\u003EOhno-Machado L, Nadkarni P, Johnson K. Natural language processing: algorithms and tools to extract computable information from EHRs and from the biomedical literature. J Am Med Inform Assoc. 2013;20(5):805. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F23935077\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1136\u002Famiajnl-2013-002214\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=23935077&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref9\"\u003EZhou B, Yang G, Shi Z, Ma S. Natural language processing for smart healthcare. IEEE Rev Biomed Eng. 2024;17:4-18. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1109\u002FRBME.2022.3210270\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=36170385&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref10\"\u003ELe Glaz A, Haralambous Y, Kim-Dufor DH, Lenca P, Billot R, Ryan TC, et al. Machine learning and natural language processing in mental health: systematic review. J Med Internet Res. 2021;23(5):e15708. [\u003Ca href=\"https:\u002F\u002Fwww.jmir.org\u002F2021\u002F5\u002Fe15708\u002F\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.2196\u002F15708\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33944788&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref11\"\u003EChin JH, Vora N. The global burden of neurologic diseases. Neurology. 2014;83(4):349-351. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F25049303\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1212\u002FWNL.0000000000000610\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=25049303&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref12\"\u003EGooch CL, Pracht E, Borenstein AR. The burden of neurological disease in the United States: a summary report and call to action. Ann Neurol. 2017;81(4):479-484. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1002\u002Fana.24897\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=28198092&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref13\"\u003EYu AYX, Liu ZA, Pou-Prom C, Lopes K, Kapral MK, Aviv RI, et al. Automating stroke data extraction from free-text radiology reports using natural language processing: instrument validation study. JMIR Med Inform. 2021;9(5):e24381. [\u003Ca href=\"https:\u002F\u002Fmedinform.jmir.org\u002F2021\u002F5\u002Fe24381\u002F\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.2196\u002F24381\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33944791&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref14\"\u003EPinter NK, Fritz JV. Neuroimaging for the neurologist: clinical MRI and future trends. Neurol Clin. 2020;38(1):1-35. [\u003Ca href=\"https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fabs\u002Fpii\u002FS0733861919300738?via%3Dihub\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.ncl.2019.08.002\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=31761054&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref15\"\u003Ede la Fuente Garcia S, Ritchie CW, Luz S. Artificial intelligence, speech, and language processing approaches to monitoring Alzheimer's disease: a systematic review. J Alzheimers Dis. 2020;78(4):1547-1574. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F33185605\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3233\u002FJAD-200888\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33185605&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref16\"\u003ELi MD, Lang M, Deng F, Chang K, Buch K, Rincon S, et al. Analysis of stroke detection during the COVID-19 pandemic using natural language processing of radiology reports. AJNR Am J Neuroradiol. 2021;42(3):429-434. [\u003Ca href=\"http:\u002F\u002Fwww.ajnr.org\u002Fcgi\u002Fpmidlookup?view=long&pmid=33334851\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3174\u002Fajnr.A6961\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33334851&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref17\"\u003EBarbour K, Hesdorffer DC, Tian N, Yozawitz EG, McGoldrick PE, Wolf S, et al. Automated detection of sudden unexpected death in epilepsy risk factors in electronic medical records using natural language processing. Epilepsia. 2019;60(6):1209-1220. [\u003Ca href=\"https:\u002F\u002Fonlinelibrary.wiley.com\u002Fdoi\u002F10.1111\u002Fepi.15966\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1111\u002Fepi.15966\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=31111463&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref18\"\u003EDeng B, Zhu W, Sun X, Xie Y, Dan W, Zhan Y, et al. Development and validation of an automatic system for intracerebral hemorrhage medical text recognition and treatment plan output. Front Aging Neurosci. 2022;14:798132. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F35462698\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3389\u002Ffnagi.2022.798132\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35462698&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref19\"\u003EMiller MI, Orfanoudaki A, Cronin M, Saglam H, So Yeon Kim I, Balogun O, et al. Natural language processing of radiology reports to detect complications of ischemic stroke. Neurocrit Care. 2022;37(Suppl 2):291-302. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F35534660\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1007\u002Fs12028-022-01513-3\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35534660&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref20\"\u003ESung SF, Chen CH, Pan RC, Hu YH, Jeng JS. Natural language processing enhances prediction of functional outcome after acute ischemic stroke. JAHA. 2021;10(24):e023486. [\u003Ca href=\"https:\u002F\u002Fwww.ahajournals.org\u002Fdoi\u002F10.1161\u002FJAHA.121.023486\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1161\u002Fjaha.121.023486\"\u003ECrossRef\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref21\"\u003EZhao Y, Fu S, Bielinski SJ, Decker PA, Chamberlain AM, Roger VL, et al. Natural language processing and machine learning for identifying incident stroke from electronic health records: algorithm development and validation. J Med Internet Res. 2021;23(3):e22951. [\u003Ca href=\"https:\u002F\u002Fwww.jmir.org\u002F2021\u002F3\u002Fe22951\u002F\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.2196\u002F22951\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33683212&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref22\"\u003EOng CJ, Orfanoudaki A, Zhang R, Caprasse FPM, Hutch M, Ma L, et al. Machine learning and natural language processing methods to identify ischemic stroke, acuity and location from radiology reports. PLoS One. 2020;15(6):e0234908. [\u003Ca href=\"https:\u002F\u002Fjournals.plos.org\u002Fplosone\u002Farticle?id=10.1371\u002Fjournal.pone.0234908\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1371\u002Fjournal.pone.0234908\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=32559211&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref23\"\u003ESkansi S. Introduction to Deep Learning: From Logical Calculus to Artificial Intelligence. 1st Edition. Cham. Springer International Publishing; 2018. \u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref24\"\u003EMikolov T, Chen K, Corrado G, Dean J. Efficient estimation of word representations in vector space. ArXiv. Preprint posted online on January 16, 2013. . [\u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F1301.3781\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref25\"\u003ELay J, Seneviratne U, Fok A, Roberts H, Phan T. Discovering themes in medical records of patients with psychogenic non-epileptic seizures. BMJ Neurol Open. 2020;2(2):e000087. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F33681804\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1136\u002Fbmjno-2020-000087\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33681804&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref26\"\u003EMayampurath A, Parnianpour Z, Richards CT, Meurer WJ, Lee J, Ankenman B, et al. Improving prehospital stroke diagnosis using natural language processing of paramedic reports. Stroke. 2021;52(8):2676-2679. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F34162217\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1161\u002FSTROKEAHA.120.033580\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=34162217&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref27\"\u003ELineback CM, Garg R, Oh E, Naidech AM, Holl JL, Prabhakaran S. Prediction of 30-day readmission after stroke using machine learning and natural language processing. Front Neurol. 2021;12:649521. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F34326805\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3389\u002Ffneur.2021.649521\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=34326805&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref28\"\u003ELiu N, Luo K, Yuan Z, Chen Y. A transfer learning method for detecting Alzheimer's disease based on speech and natural language processing. Front Public Health. 2022;10:772592. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F35493375\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3389\u002Ffpubh.2022.772592\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35493375&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref29\"\u003EMahajan P, Baths V. Acoustic and language based deep learning approaches for Alzheimer's Dementia detection from spontaneous speech. Front Aging Neurosci. 2021;13:623607. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F33613269\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3389\u002Ffnagi.2021.623607\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33613269&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref30\"\u003EBacchi S, Gluck S, Koblar S, Jannes J, Kleinig T. Automated information extraction from free-text medical documents for stroke key performance indicators: a pilot study. Intern Med J. 2022;52(2):315-317. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1111\u002Fimj.15678\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35187820&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref31\"\u003EHamid H, Fodeh SJ, Lizama AG, Czlapinski R, Pugh MJ, LaFrance WC, et al. Validating a natural language processing tool to exclude psychogenic nonepileptic seizures in electronic medical record-based epilepsy research. Epilepsy Behav. 2013;29(3):578-580. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.yebeh.2013.09.025\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=24135384&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref32\"\u003EBacchi S, Oakden-Rayner L, Zerner T, Kleinig T, Patel S, Jannes J. Deep learning natural language processing successfully predicts the cerebrovascular cause of transient ischemic attack-like presentations. Stroke. 2019;50(3):758-760. [\u003Ca href=\"https:\u002F\u002Fwww.ahajournals.org\u002Fdoi\u002F10.1161\u002FSTROKEAHA.118.024124\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1161\u002FSTROKEAHA.118.024124\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=30653397&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref33\"\u003EGarg R, Oh E, Naidech A, Kording K, Prabhakaran S. Automating ischemic stroke subtype classification using machine learning and natural language processing. J Stroke Cerebrovasc Dis. 2019;28(7):2045-2051. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.jstrokecerebrovasdis.2019.02.004\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=31103549&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref34\"\u003EPevy N, Christensen H, Walker T, Reuber M. Feasibility of using an automated analysis of formulation effort in patients' spoken seizure descriptions in the differential diagnosis of epileptic and nonepileptic seizures. Seizure. 2021;91:141-145. [\u003Ca href=\"https:\u002F\u002Flinkinghub.elsevier.com\u002Fretrieve\u002Fpii\u002FS1059-1311(21)00193-X\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.seizure.2021.06.009\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=34157636&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref35\"\u003EGuan W, Ko D, Khurshid S, Trisini Lipsanopoulos AT, Ashburner JM, Harrington LX, et al. Automated electronic phenotyping of cardioembolic stroke. Stroke. 2021;52(1):181-189. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F33297865\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1161\u002FSTROKEAHA.120.030663\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33297865&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref36\"\u003ECui L, Sahoo SS, Lhatoo SD, Garg G, Rai P, Bozorgi A, et al. Complex epilepsy phenotype extraction from narrative clinical discharge summaries. J Biomed Inform. 2014;51:272-279. [\u003Ca href=\"https:\u002F\u002Flinkinghub.elsevier.com\u002Fretrieve\u002Fpii\u002FS1532-0464(14)00144-0\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.jbi.2014.06.006\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=24973735&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref37\"\u003EHeo TS, Kim YS, Choi JM, Jeong YS, Seo SY, Lee JH, et al. Prediction of stroke outcome using natural language processing-based machine learning of radiology report of brain MRI. J Pers Med. 2020;10(4):286. [\u003Ca href=\"https:\u002F\u002Fwww.mdpi.com\u002Fresolver?pii=jpm10040286\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3390\u002Fjpm10040286\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33339385&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref38\"\u003EZanotto BS, da Silva Etges APB, Dal Bosco A, Cortes EG, Ruschel R, De Souza AC, et al. Stroke outcome measurements from electronic medical records: cross-sectional study on the effectiveness of neural and nonneural classifiers. JMIR Med Inform. 2021;9(11):e29120. [\u003Ca href=\"https:\u002F\u002Fmedinform.jmir.org\u002F2021\u002F11\u002Fe29120\u002F\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.2196\u002F29120\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=34723829&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref39\"\u003EKim C, Zhu V, Obeid J, Lenert L. Natural language processing and machine learning algorithm to identify brain MRI reports with acute ischemic stroke. PLoS One. 2019;14(2):e0212778. [\u003Ca href=\"https:\u002F\u002Fjournals.plos.org\u002Fplosone\u002Farticle?id=10.1371\u002Fjournal.pone.0212778\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1371\u002Fjournal.pone.0212778\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=30818342&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref40\"\u003EDavis MF, Sriram S, Bush WS, Denny JC, Haines JL. Automated extraction of clinical traits of multiple sclerosis in electronic medical records. J Am Med Inform Assoc. 2013;20(e2):e334-e340. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F24148554\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1136\u002Famiajnl-2013-001999\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=24148554&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref41\"\u003EGlauser T, Santel D, DelBello M, Faist R, Toon T, Clark P, et al. Identifying epilepsy psychiatric comorbidities with machine learning. Acta Neurol Scand. 2020;141(5):388-396. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F31889296\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1111\u002Fane.13216\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=31889296&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref42\"\u003ECohen KB, Glass B, Greiner HM, Holland-Bouley K, Standridge S, Arya R, et al. Methodological issues in predicting pediatric epilepsy surgery candidates through natural language processing and machine learning. Biomed Inform Insights. 2016;8:11-18. [\u003Ca href=\"https:\u002F\u002Fjournals.sagepub.com\u002Fdoi\u002F10.4137\u002FBII.S38308?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%20%200pubmed\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.4137\u002FBII.S38308\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=27257386&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref43\"\u003EAlim-Marvasti A, Pérez-García F, Dahele K, Romagnoli G, Diehl B, Sparks R, et al. Machine learning for localizing epileptogenic-zone in the temporal lobe: quantifying the value of multimodal clinical-semiology and imaging concordance. Front Digit Health. 2021;3:559103. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F34713078\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3389\u002Ffdgth.2021.559103\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=34713078&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref44\"\u003EBalagopalan A, Eyre B, Robin J, Rudzicz F, Novikova J. Comparing pre-trained and feature-based models for prediction of Alzheimer's disease based on speech. Front Aging Neurosci. 2021;13:635945. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F33986655\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3389\u002Ffnagi.2021.635945\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33986655&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref45\"\u003EMartinc M, Haider F, Pollak S, Luz S. Temporal integration of text transcripts and acoustic features for Alzheimer's diagnosis based on spontaneous speech. Front Aging Neurosci. 2021;13:642647. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F34194313\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3389\u002Ffnagi.2021.642647\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=34194313&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref46\"\u003ELiu Z, Paek EJ, Yoon SO, Casenhiser D, Zhou W, Zhao X. Detecting Alzheimer's disease using natural language processing of referential communication task transcripts. J Alzheimers Dis. 2022;86(3):1385-1398. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3233\u002FJAD-215137\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35213368&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref47\"\u003ENelson RE, Butler J, LaFleur J, Knippenberg K, Kamauu AWC, DuVall SL. Determining multiple sclerosis phenotype from electronic medical records. J Manag Care Spec Pharm. 2016;22(12):1377-1382. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F27882837\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.18553\u002Fjmcp.2016.22.12.1377\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=27882837&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref48\"\u003EChase HS, Mitrani LR, Lu GG, Fulgieri DJ. Early recognition of multiple sclerosis using natural language processing of the electronic health record. BMC Med Inform Decis Mak. 2017;17(1):24. [\u003Ca href=\"https:\u002F\u002Fbmcmedinformdecismak.biomedcentral.com\u002Farticles\u002F10.1186\u002Fs12911-017-0418-4\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1186\u002Fs12911-017-0418-4\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=28241760&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref49\"\u003EWissel BD, Greiner HM, Glauser TA, Holland-Bouley KD, Mangano FT, Santel D, et al. Prospective validation of a machine learning model that uses provider notes to identify candidates for resective epilepsy surgery. Epilepsia. 2020;61(1):39-48. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F31784992\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1111\u002Fepi.16398\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=31784992&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref50\"\u003ESung SF, Lin CY, Hu YH. EMR-based phenotyping of ischemic stroke using supervised machine learning and text mining techniques. IEEE J Biomed Health Inform. 2020;24(10):2922-2931. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1109\u002FJBHI.2020.2976931\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=32142458&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref51\"\u003EYang Z, Pou-Prom C, Jones A, Banning M, Dai D, Mamdani M, et al. Assessment of natural language processing methods for ascertaining the expanded disability status scale score from the electronic health records of patients with multiple sclerosis: algorithm development and validation study. JMIR Med Inform. 2022;10(1):e25157. [\u003Ca href=\"https:\u002F\u002Fmedinform.jmir.org\u002F2022\u002F1\u002Fe25157\u002F\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.2196\u002F25157\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35019849&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref52\"\u003EXie K, Gallagher RS, Conrad EC, Garrick CO, Baldassano SN, Bernabei JM, et al. Extracting seizure frequency from epilepsy clinic notes: a machine reading approach to natural language processing. J Am Med Inform Assoc. 2022;29(5):873-881. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F35190834\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1093\u002Fjamia\u002Focac018\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35190834&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref53\"\u003ESung SF, Chen K, Wu DP, Hung LC, Su YH, Hu YH. Applying natural language processing techniques to develop a task-specific EMR interface for timely stroke thrombolysis: a feasibility study. Int J Med Inform. 2018;112:149-157. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.ijmedinf.2018.02.005\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=29500013&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref54\"\u003ESung SF, Hsieh CY, Hu YH. Early prediction of functional outcomes after acute ischemic stroke using unstructured clinical text: retrospective cohort study. JMIR Med Inform. 2022;10(2):e29806. [\u003Ca href=\"https:\u002F\u002Fmedinform.jmir.org\u002F2022\u002F2\u002Fe29806\u002F\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.2196\u002F29806\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35175201&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref55\"\u003EXia Z, Secor E, Chibnik LB, Bove RM, Cheng S, Chitnis T, et al. Modeling disease severity in multiple sclerosis using electronic health records. PLoS One. 2013;8(11):e78927. [\u003Ca href=\"https:\u002F\u002Fjournals.plos.org\u002Fplosone\u002Farticle?id=10.1371\u002Fjournal.pone.0078927\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1371\u002Fjournal.pone.0078927\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=24244385&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref56\"\u003ERoshanzamir A, Aghajan H, Baghshah MS. Transformer-based deep neural network language models for Alzheimer's disease risk assessment from targeted speech. BMC Med Inform Decis Mak. 2021;21(1):92. [\u003Ca href=\"https:\u002F\u002Fbmcmedinformdecismak.biomedcentral.com\u002Farticles\u002F10.1186\u002Fs12911-021-01456-3\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1186\u002Fs12911-021-01456-3\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33750385&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref57\"\u003ERannikmäe K, Wu H, Tominey S, Whiteley W, Allen N, Sudlow C, et al. Developing automated methods for disease subtyping in UK Biobank: an exemplar study on stroke. BMC Med Inform Decis Mak. 2021;21(1):191. [\u003Ca href=\"https:\u002F\u002Fbmcmedinformdecismak.biomedcentral.com\u002Farticles\u002F10.1186\u002Fs12911-021-01556-0\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1186\u002Fs12911-021-01556-0\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=34130677&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref58\"\u003EBuchlak Q, Esmaili N, Bennett C, Farrokhi F. Natural language processing applications in the clinical neurosciences: a machine learning augmented systematic review. Acta Neurochir Suppl. 2022;134:277-289. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1007\u002F978-3-030-85292-4_32\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=34862552&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref59\"\u003ECrema C, Attardi G, Sartiano D, Redolfi A. Natural language processing in clinical neuroscience and psychiatry: a review. Front Psychiatry. 2022;13:946387. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F36186874\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3389\u002Ffpsyt.2022.946387\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=36186874&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref60\"\u003EWang W, Kiik M, Peek N, Curcin V, Marshall IJ, Rudd AG, et al. A systematic review of machine learning models for predicting outcomes of stroke with structured data. PLoS One. 2020;15(6):e0234722. [\u003Ca href=\"https:\u002F\u002Fjournals.plos.org\u002Fplosone\u002Farticle?id=10.1371\u002Fjournal.pone.0234722\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1371\u002Fjournal.pone.0234722\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=32530947&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref61\"\u003EWeller J, Budson A. Current understanding of Alzheimer's disease diagnosis and treatment. F1000Res. 2018;7:1-9. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F30135715\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.12688\u002Ff1000research.14506.1\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=30135715&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref62\"\u003EEstimates of funding for various Research, Condition, and Disease Categories (RCDC). National Institutes of Health. URL: \u003Ca target=\"_blank\" href=\"https:\u002F\u002Freport.nih.gov\u002Ffunding\u002Fcategorical-spending#\u002F\"\u003Ehttps:\u002F\u002Freport.nih.gov\u002Ffunding\u002Fcategorical-spending#\u002F\u003C\u002Fa\u003E [accessed 2024-02-15]\n \u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref63\"\u003EAhsan MM, Luna SA, Siddique Z. Machine-learning-based disease diagnosis: a comprehensive review. Healthcare (Basel). 2022;10(3):541. [\u003Ca href=\"https:\u002F\u002Fwww.mdpi.com\u002Fresolver?pii=healthcare10030541\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.3390\u002Fhealthcare10030541\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=35327018&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref64\"\u003EKaul V, Enslin S, Gross SA. History of artificial intelligence in medicine. Gastrointest Endosc. 2020;92(4):807-812. [\u003Ca href=\"https:\u002F\u002Fwww.giejournal.org\u002Farticle\u002FS0016-5107(20)34466-7\u002Ffulltext\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.gie.2020.06.040\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=32565184&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref65\"\u003EBurton A. How do we fix the shortage of neurologists? Lancet Neurol. 2018;17(6):502-503. [\u003Ca href=\"https:\u002F\u002Fwww.thelancet.com\u002Fjournals\u002Flaneur\u002Farticle\u002FPIIS1474-4422(18)30143-1\u002Ffulltext\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002FS1474-4422(18)30143-1\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=29680206&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref66\"\u003ELöscher W, Potschka H, Sisodiya SM, Vezzani A. Drug resistance in epilepsy: clinical impact, potential mechanisms, and new innovative treatment options. Pharmacol Rev. 2020;72(3):606-638. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F32540959\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1124\u002Fpr.120.019539\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=32540959&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref67\"\u003EDlugos DJ. The early identification of candidates for epilepsy surgery. Arch Neurol. 2001;58(10):1543-1546. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1001\u002Farchneur.58.10.1543\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=11594910&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref68\"\u003EMyszczynska MA, Ojamies PN, Lacoste AMB, Neil D, Saffari A, Mead R, et al. Applications of machine learning to diagnosis and treatment of neurodegenerative diseases. Nat Rev Neurol. 2020;16(8):440-456. [\u003Ca href=\"https:\u002F\u002Feprints.whiterose.ac.uk\u002F163905\u002F\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1038\u002Fs41582-020-0377-8\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=32669685&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref69\"\u003EJain V, Chatterjee JM. Machine Learning with Health Care Perspective: Machine Learning and Healthcare. 1st Edition. Cham. Springer International Publishing; 2020. \u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref70\"\u003EOrtiz-Posadas MR. Pattern Recognition Techniques Applied to Biomedical Problems. 1st Edition. Cham. Springer International Publishing; 2020. \u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref71\"\u003EGray KR, Aljabar P, Heckemann RA, Hammers A, Rueckert D, Alzheimer's Disease Neuroimaging Initiative. Random forest-based similarity measures for multi-modal classification of Alzheimer's disease. Neuroimage. 2013;65:167-175. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F23041336\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.neuroimage.2012.09.065\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=23041336&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref72\"\u003ESorin V, Barash Y, Konen E, Klang E. Deep learning for natural language processing in radiology—fundamentals and a systematic review. J Am Coll Radiol. 2020;17(5):639-648. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.jacr.2019.12.026\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=32004480&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref73\"\u003EAmrutha K, Prabu P. Effortless and beneficial processing of natural languages using transformers. J Discret Math Sci Cryptogr. 2022;25(7):1987-2005. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1080\u002F09720529.2022.2133239\"\u003ECrossRef\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref74\"\u003EWeiss K, Khoshgoftaar TM, Wang D. A survey of transfer learning. J Big Data. 2016;3(1):9. [\u003Ca href=\"https:\u002F\u002Fjournalofbigdata.springeropen.com\u002Farticles\u002F10.1186\u002Fs40537-016-0043-6#citeas\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1186\u002Fs40537-016-0043-6\"\u003ECrossRef\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref75\"\u003EDevlin J, Chang MW, Lee K, Toutanova K. BERT: pre-training of deep bidirectional transformers for language understanding. ArXiv. Preprint posted online on October 11, 2018. . [\u003Ca href=\"https:\u002F\u002Fui.adsabs.harvard.edu\u002Fabs\u002F2018arXiv181004805D\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.48550\u002FarXiv.1810.04805\"\u003ECrossRef\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref76\"\u003ERussell D, Li L, Tian F. Generating text using generative adversarial networks and quick-thought vectors. IEEE; 2019. Presented at: 2019 IEEE 2nd International Conference on Computer and Communication Engineering Technology (CCET); August 16-18, 2019:16-18; Beijing, China. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1109\u002Fccet48361.2019.8989352\"\u003ECrossRef\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref77\"\u003EDemner-Fushman D, Elhadad N. Aspiring to unintended consequences of natural language processing: a review of recent developments in clinical and consumer-generated text processing. Yearb Med Inform. 2016;(1):224-233. [\u003Ca href=\"https:\u002F\u002Fwww.thieme-connect.de\u002Fproducts\u002Fejournals\u002Fabstract\u002F10.15265\u002FIY-2016-017\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.15265\u002FIY-2016-017\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=27830255&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref78\"\u003EBitterman DS, Miller TA, Mak RH, Savova GK. Clinical natural language processing for radiation oncology: a review and practical primer. Int J Radiat Oncol Biol Phys. 2021;110(3):641-655. [\u003Ca href=\"https:\u002F\u002Fwww.redjournal.org\u002Farticle\u002FS0360-3016(21)00118-8\u002Ffulltext\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1016\u002Fj.ijrobp.2021.01.044\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=33545300&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref79\"\u003EMagueresse A, Carles V, Heetderks E. Low-resource languages: a review of past work and future challenges. ArXiv. Preprint posted online on June 12, 2020. . [\u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2006.07264\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.48550\u002FarXiv.2006.07264\"\u003ECrossRef\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref80\"\u003ESajda P. Machine learning for detection and diagnosis of disease. Annu Rev Biomed Eng. 2006;8:537-565. [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1146\u002Fannurev.bioeng.8.061505.095802\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=16834566&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref81\"\u003EKhan B, Fatima H, Qureshi A, Kumar S, Hanan A, Hussain J, et al. Drawbacks of artificial intelligence and their potential solutions in the healthcare sector. Biomed Mater Devices. 2023:1-8. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F36785697\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.1007\u002Fs44174-023-00063-2\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=36785697&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref82\"\u003EHabli I, Lawton T, Porter Z. Artificial intelligence in health care: accountability and safety. Bull World Health Organ. 2020;98(4):251-256. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F32284648\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.2471\u002FBLT.19.237487\"\u003ECrossRef\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=32284648&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref83\"\u003EAmann J. Machine learning (ML) in stroke medicine: opportunities and challenges for risk prediction and prevention. In: Jotterand F, Ienca M, editors. Artificial Intelligence in Brain and Mental Health: Philosophical, Ethical & Policy Issues. Cham. Springer International Publishing; 2021:57-71.\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref84\"\u003EElbattah M, Arnaud É, Gignon M, Dequen G. The role of text analytics in healthcare: a review of recent developments and applications. 2021. Presented at: Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021); February 11-13, 2021:825-832; Vienna, Austria. URL: \u003Ca target=\"_blank\" href=\"https:\u002F\u002Fwww.scitepress.org\u002FPublishedPapers\u002F2021\u002F104145\u002F104145.pdf\"\u003Ehttps:\u002F\u002Fwww.scitepress.org\u002FPublishedPapers\u002F2021\u002F104145\u002F104145.pdf\u003C\u002Fa\u003E [\u003Ca target=\"_blank\" href=\"https:\u002F\u002Fdx.doi.org\u002F10.5220\u002F0010414508250832\"\u003ECrossRef\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003Cli\u003E\u003Cspan id=\"ref85\"\u003EChard K, Russell M, Lussier YA, Mendonça EA, Silverstein JC. A cloud-based approach to medical NLP. AMIA Annu Symp Proc. 2011;2011:207-216. [\u003Ca href=\"https:\u002F\u002Feuropepmc.org\u002Fabstract\u002FMED\u002F22195072\" target=\"_blank\"\u003EFREE Full text\u003C\u002Fa\u003E] [\u003Ca href=\"https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fentrez\u002Fquery.fcgi?cmd=Retrieve&db=PubMed&list_uids=22195072&dopt=Abstract\" target=\"_blank\"\u003EMedline\u003C\u002Fa\u003E]\u003C\u002Fspan\u003E\u003C\u002Fli\u003E\u003C\u002Fol\u003E\u003C\u002Fdiv\u003E\u003Cbr\u003E\u003Chr\u003E\u003Ca name=\"Abbreviations\"\u003E‎\u003C\u002Fa\u003E\u003Ch4 class=\"navigation-heading\" id=\"Abbreviations\" data-label=\"Abbreviations\"\u003EAbbreviations\u003C\u002Fh4\u003E\u003Ctable width=\"80%\" border=\"0\" align=\"center\"\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003EBERT:\u003C\u002Fb\u003E Bidirectional Encoder Representations From Transformers\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003EDL:\u003C\u002Fb\u003E deep learning\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003EEMR:\u003C\u002Fb\u003E electronic medical record\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003EML:\u003C\u002Fb\u003E machine learning\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003EMS:\u003C\u002Fb\u003E multiple sclerosis\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003ENLP:\u003C\u002Fb\u003E natural language processing\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003EPRISMA-ScR:\u003C\u002Fb\u003E Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003EPROSPERO:\u003C\u002Fb\u003E Prospective Register of Systematic Reviews\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003EREDCap:\u003C\u002Fb\u003E Research Electronic Data Capture\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003ERNN:\u003C\u002Fb\u003E recurrent neural network\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cb\u003ESVM:\u003C\u002Fb\u003E support vector machine\u003C\u002Ftd\u003E\u003C\u002Ftr\u003E\u003C\u002Ftable\u003E\u003Cbr\u003E\u003Chr\u003E\u003Cp style=\"font-style: italic\"\u003EEdited by P Kubben; submitted 02.10.23; peer-reviewed by DH Kim-Dufor, P Gazerani; comments to author 13.12.23; revised version received 08.01.24; accepted 10.01.24; published 22.05.24.\u003C\u002Fp\u003E\u003Ca href=\"https:\u002F\u002Fsupport.jmir.org\u002Fhc\u002Fen-us\u002Farticles\u002F115002955531\" id=\"Copyright\" target=\"_blank\" class=\"navigation-heading h4 d-block\" aria-label=\"Copyright - what is a Creative Commons License?\" data-label=\"Copyright\"\u003ECopyright \u003Cspan class=\"fas fa-question-circle\"\u003E\u003C\u002Fspan\u003E\u003C\u002Fa\u003E\u003Cp class=\"article-copyright\"\u003E©Ilana Lefkovitz, Samantha Walsh, Leah J Blank, Nathalie Jetté, Benjamin R Kummer. Originally published in JMIR Neurotechnology (https:\u002F\u002Fneuro.jmir.org), 22.05.2024.\u003C\u002Fp\u003E\u003Csmall class=\"article-license\"\u003E\u003Cp class=\"abstract-paragraph\"\u003EThis is an open-access article distributed under the terms of the Creative Commons Attribution License (https:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby\u002F4.0\u002F), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Neurotechnology, is properly cited. The complete bibliographic information, a link to the original publication on https:\u002F\u002Fneuro.jmir.org, as well as this copyright and license information must be included.\u003C\u002Fp\u003E\u003C\u002Fsmall\u003E\u003Cbr\u003E\u003C\u002Fsection\u003E\u003C\u002Farticle\u003E\u003C\u002Fsection\u003E\u003C\u002Fsection\u003E\u003C\u002Fmain\u003E\n"}],fetch:{},error:a,state:{host:a,environment:d,journalPath:r,keys:{},domains:{},screensize:"desktop",accessibility:{filter:"none","font-weight":"inherit","font-size":.625,"text-align":"initial"},announcements:{data:[{announcement_id:515,title:"JMIR Neurotechnology Now Included In The Directory Of Open Access Journals",description_short:"\u003Cp\u003EJMIR Publications is happy to announce that JMIR Neurotechnology has been accepted and indexed with the \u003Ca href=\"https:\u002F\u002Fdoaj.org\u002Ftoc\u002F2817-092X\"\u003EDirectory of Open Access Journals\u003C\u002Fa\u003E (DOAJ). The DOAJ applies strict criteria to review and index Open Access journals, which include licensing and copyright criteria, quality control processes, journal website technical and usability setups, and editorial evaluation.\u003C\u002Fp\u003E",date_posted:"2024-10-16T18:46:36.000Z",journal_id:f},{announcement_id:445,title:"Call for Papers: Theme Issue: Brain-Computer Interfaces (BCIs)",description_short:"\u003Cp\u003E\u003Cem\u003EJMIR Neurotechnology\u003C\u002Fem\u003E, a peer-reviewed journal, invites submissions on advancements in the field of brain-computer interfaces (BCIs) that represent the transformative convergence of neuroscience, engineering, and technology. \u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E",date_posted:"2024-02-23T12:46:44.000Z",journal_id:f},{announcement_id:413,title:"Call for Papers: JMIR Neurotechnology",description_short:"\u003Cp\u003EThe journal \u003Cem\u003EJMIR Neurotechnology\u003C\u002Fem\u003E invites researchers, clinicians, caregivers, and technologists to submit manuscripts that explore novel diagnostic and treatment tools for neurological disorders, particularly those leveraging the potential of neurotechnology.\u003C\u002Fp\u003E",date_posted:"2023-10-20T14:00:02.000Z",journal_id:f}],pagination:{from:b,to:z,total:g,perPage:z,firstPage:b,lastPage:b}},article:{data:{article_id:51822,published_at:"2024-05-22T14:00:05.000Z",submitted_at:aj,section_id:ak,journal_id:f,year:al,issue:"1",volume:g,identifier:"51822",url:am,pdf_url:"https:\u002F\u002Fneuro.jmir.org\u002F2024\u002F1\u002Fe51822\u002FPDF",html_url:"https:\u002F\u002Fneuro.jmir.org\u002F2024\u002F1\u002Fe51822",xml_url:"https:\u002F\u002Fneuro.jmir.org\u002F2024\u002F1\u002Fe51822\u002FXML",title:"Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review",public_id:"JMIR Neurotech 2024;3:e51822",thumbnail:"https:\u002F\u002Fasset.jmir.pub\u002Fassets\u002F92d91f88ddb727d190767f7247b60cd4.png",doi:"10.2196\u002F51822",pmid:a,pmcid:a,issue_title:"Jan-Dec",pages:[],transfer:a,authors:[{first_name:"Ilana",last_name:"Lefkovitz",degrees:M,deceased:a,orcid:"0000-0002-8724-3798",equal_contrib:s,matchedAffiliations:[b]},{first_name:"Samantha",last_name:"Walsh",degrees:"MLS, MA",deceased:a,orcid:"0000-0002-5040-6514",equal_contrib:s,matchedAffiliations:[e]},{first_name:"Leah J",last_name:"Blank",degrees:"MD, MPH",deceased:a,orcid:"0000-0001-8719-6752",equal_contrib:s,matchedAffiliations:[b,g]},{first_name:"Nathalie",last_name:"Jetté",degrees:"MD, MSc",deceased:a,orcid:"0000-0003-1351-5866",equal_contrib:s,matchedAffiliations:[b,A]},{first_name:an,last_name:ao,degrees:M,deceased:a,orcid:"0000-0002-1413-8014",equal_contrib:s,matchedAffiliations:[b,i,B]}],affiliations:[{aff_id:11147318,author_id:369015,phone:a,fax:c,corresp_aff:j,aff_type:a,seq:b,article_id:a,institution_line_1:ap,institution_line_2:t,institution_line_3:c,address_line_1:a,address_line_2:a,city:k,prov_state:l,postal_code:a,country:m},{aff_id:11147319,author_id:369019,phone:a,fax:c,corresp_aff:j,aff_type:a,seq:b,article_id:a,institution_line_1:"Hunter College Libraries",institution_line_2:"Hunter College, City University of New York",institution_line_3:c,address_line_1:a,address_line_2:a,city:k,prov_state:l,postal_code:a,country:m},{aff_id:11147324,author_id:369023,phone:a,fax:c,corresp_aff:j,aff_type:a,seq:e,article_id:a,institution_line_1:"Department of Population Health Science and Policy",institution_line_2:t,institution_line_3:c,address_line_1:a,address_line_2:a,city:k,prov_state:l,postal_code:a,country:m},{aff_id:11147326,author_id:369024,phone:a,fax:c,corresp_aff:j,aff_type:a,seq:e,article_id:a,institution_line_1:"Department of Clinical Neurosciences",institution_line_2:"University of Calgary",institution_line_3:c,address_line_1:a,address_line_2:a,city:"Calgary",prov_state:"AB",postal_code:a,country:"Canada"},{aff_id:11147328,author_id:aq,phone:a,fax:c,corresp_aff:j,aff_type:a,seq:e,article_id:a,institution_line_1:"Clinical Neuro-Informatics Program",institution_line_2:t,institution_line_3:c,address_line_1:a,address_line_2:a,city:k,prov_state:l,postal_code:a,country:m},{aff_id:11147329,author_id:aq,phone:a,fax:c,corresp_aff:j,aff_type:a,seq:g,article_id:a,institution_line_1:"Windreich Department of Artificial Intelligence and Human Health",institution_line_2:t,institution_line_3:c,address_line_1:a,address_line_2:a,city:k,prov_state:l,postal_code:a,country:m}],primaryAuthor:{first_name:an,last_name:ao,email:"benjamin.kummer@mountsinai.org",degrees:M,primaryAffiliation:{fax:c,phone:"1 212 241 5050",country:m,postal_code:"10029",prov_state:l,city:k,address_line_1:"One Gustave Levy Place",address_line_2:"Box 1137",institution_line_1:ap,institution_line_2:t,institution_line_3:c}},abstract:"Background: Natural language processing (NLP), a branch of artificial intelligence that analyzes unstructured language, is being increasingly used in health care. However, the extent to which NLP has been formally studied in neurological disorders remains unclear.\nObjective: We sought to characterize studies that applied NLP to the diagnosis, prediction, or treatment of common neurological disorders.\nMethods: This review followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) standards. The search was conducted using MEDLINE and Embase on May 11, 2022. Studies of NLP use in migraine, Parkinson disease, Alzheimer disease, stroke and transient ischemic attack, epilepsy, or multiple sclerosis were included. We excluded conference abstracts, review papers, as well as studies involving heterogeneous clinical populations or indirect clinical uses of NLP. Study characteristics were extracted and analyzed using descriptive statistics. We did not aggregate measurements of performance in our review due to the high variability in study outcomes, which is the main limitation of the study.\nResults: In total, 916 studies were identified, of which 41 (4.5%) met all eligibility criteria and were included in the final review. Of the 41 included studies, the most frequently represented disorders were stroke and transient ischemic attack (n=20, 49%), followed by epilepsy (n=10, 24%), Alzheimer disease (n=6, 15%), and multiple sclerosis (n=5, 12%). We found no studies of NLP use in migraine or Parkinson disease that met our eligibility criteria. The main objective of NLP was diagnosis (n=20, 49%), followed by disease phenotyping (n=17, 41%), prognostication (n=9, 22%), and treatment (n=4, 10%). In total, 18 (44%) studies used only machine learning approaches, 6 (15%) used only rule-based methods, and 17 (41%) used both.\nConclusions: We found that NLP was most commonly applied for diagnosis, implying a potential role for NLP in augmenting diagnostic accuracy in settings with limited access to neurological expertise. We also found several gaps in neurological NLP research, with few to no studies addressing certain disorders, which may suggest additional areas of inquiry.\nTrial Registration: Prospective Register of Systematic Reviews (PROSPERO) CRD42021228703; https:\u002F\u002Fwww.crd.york.ac.uk\u002FPROSPERO\u002Fdisplay_record.php?RecordID=228703\n",keywords:"artificial intelligence; machine learning; neurology; stroke; parkinson; deep learning; cardiovascular; multiple sclerosis; natural language processing; treatment; epilepsy; scoping review; parkinson disease; prediction; diagnosis; migraine; cerebrovascular disease; neurological; neurological disorder; unstructured; transient ischemic attack; nlp; text; headache disorders",date_submitted:aj,title_html:a,sections:[{title:"AI in Neurotechnology",section_id:ak,journal_id:f,colour:u,count:i},{title:"Registered Report",section_id:805,journal_id:b,colour:C,count:577},{title:"Reviews in Medical Informatics",section_id:142,journal_id:n,colour:D,count:93},{title:"Computer-Aided Diagnosis",section_id:1006,journal_id:n,colour:D,count:133},{title:"Artificial Intelligence",section_id:797,journal_id:b,colour:C,count:1407},{title:"Natural Language Processing",section_id:171,journal_id:n,colour:D,count:703},{title:"mHealth in a Clinical Setting",section_id:181,journal_id:v,colour:E,count:621},{title:"Digital Health Reviews",section_id:w,journal_id:b,colour:C,count:1315},{title:"mHealth for Diagnosis",section_id:1007,journal_id:v,colour:E,count:62},{title:"Neurotech Innovations, Diagnostic Tools and Techniques",section_id:1470,journal_id:f,colour:u,count:F},{title:"Neurology and Neurosciences",section_id:102,journal_id:g,colour:ar,count:189}],preprint:h,articleKD:L,isOldOjphiMigrated:L}},articles:{recent:[],openReview:[]},articleTypes:{},authentication:{data:a,jwt:a},countries:{data:[]},departments:{data:[]},help:{data:{}},journal:{data:{journal_id:f,title:N,tag:as,description:at,path:r,slug:r,seq:F,enabled:b,environment:d,url:au,batch:e,year:O,colour:u,impact:a,order:G,published:H,transfers:a,cite_score:a,settings:{aboutJournal:"\u003Cp\u003EJMIR Neurotechnology is a premier, open-access journal indexed in Sherpa\u002FRomeo, DOAJ and EBSCO\u002FEBSCO Essentials. The journal opens a space for the publication of research exploring how technologies (e.g. information technology, neural engineering, neural interfacing, clinical data science, robotics, eHealth\u002FmHealth) can be applied in clinical neuroscience (e.g., neurology, neurosurgery, neuroradiology) to prevent, diagnose, and treat neurological disorders. The journal also aims to serve patients, caregivers, and others challenged by neurological disorders by supporting deeply translational medicine, stimulating connections from\u003Cem\u003E byte to bedside\u003C\u002Fem\u003E.\u003C\u002Fp\u003E\r\n\u003Cblockquote\u003E\r\n\u003Cp\u003E\u003Cem\u003E\"Neurotechnology can ameliorate or even eliminate some of the impairments that come with neurological disorders, by helping the patients to regain lost functions and participate in society, while reducing the cost of healthcare.\"\u003C\u002Fem\u003E - Prof. Dr. Pieter Roelfsema, Director of the Netherlands Institute for Neuroscience\u003C\u002Fp\u003E\r\n\u003C\u002Fblockquote\u003E",announcementLink:c,carouselVideoId:"VBMfn9DMJ8o",carouselVideoTabTitle:"Video: Focus and scope of JNT",copyrightNotice:c,focusScopeDesc:"\u003Cul\u003E\u003C\u002Ful\u003E\r\n\u003Cul\u003E\u003C\u002Ful\u003E\r\n\u003Cp\u003E\u003Cspan style=\"font-weight: 400;\"\u003EJMIR Neurotechnology (JNT, Editor-in-Chief: Pieter Kubben, MD, PhD) is a premier, open-access journal indexed in Sherpa\u002FRomeo, DOAJ and EBSCO\u002FEBSCO Essentials. JNT publishes research that links the work of neuroscientists and clinicians with innovative technologists, fostering new and strengthened connections between clinical neuroscience practice and cutting-edge technology. \u003C\u002Fspan\u003E\u003Cspan style=\"font-weight: 400;\"\u003EThe journal aims to serve patients, caregivers, and others challenged by neurological disorders by supporting deeply translational medicine and stimulating connections from\u003C\u002Fspan\u003E\u003Ci style=\"font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, 'Open Sans', 'Helvetica Neue', sans-serif;\"\u003E\u003Cspan\u003E byte to bedside\u003C\u002Fspan\u003E\u003C\u002Fi\u003E\u003Cspan style=\"font-weight: 400;\"\u003E.\u003C\u002Fspan\u003E\u003C\u002Fp\u003E\r\n\u003Cp\u003EWe are looking for papers covering, for example, the following themes: \u003C\u002Fp\u003E\r\n\u003Cul\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003ENeuroradiology\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003EAdvancements in neurosurgery\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003EInnovative diagnostic tools and techniques in neurotechnology\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003ECutting-edge neurotechnology for therapeutics\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003EData sharing and open science in neurotechnology\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003ECode transparency and reproducibility in neurotechnology\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003ENeurorehabilitation\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003ECognitive enhancement\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003EChallenges and ethical considerations in neurotechnology\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003ENeuroimaging and brain-machine interfaces\u003C\u002Fli\u003E\r\n\u003Cli dir=\"ltr\" aria-level=\"1\"\u003ENeurotechnology and artificial Iintelligence (AI)\u003C\u002Fli\u003E\r\n\u003C\u002Ful\u003E\r\n\u003Cp\u003E\u003Cspan style=\"font-weight: 400;\"\u003EJMIR Neurotechnology offers authors rapid and thorough peer-review, professional copyediting, professional production of PDF, XHTML, and XML proofs. This journal adheres to the same quality standards as our flagship journal, Journal of Medical Internet Research.\u003C\u002Fspan\u003E\u003C\u002Fp\u003E",googleAnalyticsId:"UA-186918-43",impactFactor:c,journalDescription:"\u003Cp\u003E\u003Cstrong\u003EThe intersection between clinical neuroscience and technology to prevent, diagnose, and treat neurological disorders.\u003C\u002Fstrong\u003E\u003C\u002Fp\u003E",journalInitials:"JNT",footer:"\u003Cdiv\u003E\r\n\u003Cul style=\"display: flex; flex-wrap: wrap; list-style: none; justify-content: center;\"\u003E\r\n\u003Cli style=\"margin: 10px;\"\u003E\u003Cimg src=\"https:\u002F\u002Fasset.jmir.pub\u002Fresources\u002Fimages\u002Fpartners\u002Fopen-access.jpg\" alt=\"Open Access\" \u002F\u003E\u003C\u002Fli\u003E\r\n\u003Cli style=\"margin: 10px;\"\u003E\u003Ca href=\"https:\u002F\u002Foaspa.org\u002Fmember-record-jmir-publications-inc\" target=\"_blank\" rel=\"noopener\"\u003E\u003Cimg src=\"https:\u002F\u002Fasset.jmir.pub\u002Fresources\u002Fimages\u002Fpartners\u002Foaspa.jpg\" alt=\"Open Access Scholarly Publishers Association\" \u002F\u003E\u003C\u002Fa\u003E\u003C\u002Fli\u003E\r\n\u003Cli style=\"margin: 10px;\"\u003E\u003Ca href=\"https:\u002F\u002Fwww.trendmd.com\" target=\"_blank\" rel=\"noopener\"\u003E\u003Cimg src=\"https:\u002F\u002Fasset.jmir.pub\u002Fresources\u002Fimages\u002Fpartners\u002Ftrend-MD.jpg\" alt=\"TrendMD Member\" \u002F\u003E\u003C\u002Fa\u003E\u003C\u002Fli\u003E\r\n\u003Cli style=\"margin: 10px;\"\u003E\u003Ca href=\"https:\u002F\u002Forcid.org\" target=\"_blank\" rel=\"noopener\"\u003E\u003Cimg src=\"https:\u002F\u002Fasset.jmir.pub\u002Fresources\u002Fimages\u002Fpartners\u002FORCID.jpg\" alt=\"ORCID Member\" \u002F\u003E\u003C\u002Fa\u003E\u003C\u002Fli\u003E\r\n\u003Cli style=\"margin: 10px;\"\u003E\u003C\u002Fli\u003E\r\n\u003C\u002Ful\u003E\r\n\u003C\u002Fdiv\u003E",onlineIssn:"2817-092X",searchDescription:N,searchKeywords:"neurotechnology; applied neuroscience; clinical neuroscience; neuro; neurotech; neural engineering; neural interfacing; brain computer interfacing; neuromodulation; bci",articlesWidget:{enabled:h,count:I,label:"Recent Articles"},openReviewWidget:{enabled:h,count:I,label:"\u003Ca href=\"https:\u002F\u002Fpreprints.jmir.org\"\u003EPreprints\u003C\u002Fa\u003E Open for Peer-Review"},searchWidget:{enabled:h},partnershipsWidget:{enabled:h},submitButton:{enabled:h,label:"Submit Article"},editorInChief:"\u003Cp\u003EPieter Kubben, MD, PhD, Neurosurgeon, Maastricht University Medical Center, Netherlands\u003C\u002Fp\u003E"}}},journals:{data:[{journal_id:b,title:"Journal of Medical Internet Research",tag:"The leading peer-reviewed journal for digital medicine and health and health care in the internet age. June 2024 - Journal Impact Factor: 5.8. Q1 journal in \"Medical Informatics\" and \"Health Care Sciences & Services\" categories.(Source: Journal Citation Reports™ 2024 from Clarivate™)",description:a,path:av,slug:av,seq:b,enabled:b,environment:d,url:"https:\u002F\u002Fwww.jmir.org",batch:b,year:1999,colour:C,impact:"5.8",order:b,published:9060,transfers:a,cite_score:"14.4"},{journal_id:i,title:"JMIR Research Protocols",tag:"Ongoing trials, grant proposals, formative research, methods, early results. June 2024 - Journal Impact Factor: 1.4 (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:"JMIR Res Protoc publishes research protocols, current and ongoing trials, and grant proposals in all areas of medicine (with an initial focus on ehealth\u002Fmhealth). Publish your work in this journal to let others know what you are working on, to facilitate collaboration and\u002For recruitment, to avoid duplication of efforts, to create a citable record of a research design idea, and to aid systematic reviewers in compiling evidence. Research protocols or grant proposals that are funded and have undergone peer-review will receive an expedited review if you upload peer-review reports as supplementary files.",path:"resprot",slug:"researchprotocols",seq:e,enabled:b,environment:d,url:"https:\u002F\u002Fwww.researchprotocols.org",batch:b,year:P,colour:"#837a7a",impact:"1.4",order:Q,published:4343,transfers:a,cite_score:"2.4"},{journal_id:aw,title:"JMIR Formative Research",tag:"Process evaluations, early results and feasibility\u002Fpilot studies of digital and non-digital interventions. June 2024 - Journal Impact Factor: 2.0 (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:c,path:ax,slug:ax,seq:J,enabled:b,environment:d,url:"https:\u002F\u002Fformative.jmir.org",batch:e,year:R,colour:"#605959",impact:"2.0",order:v,published:3083,transfers:a,cite_score:"2.7"},{journal_id:v,title:"JMIR mHealth and uHealth",tag:"Focused on health and biomedical applications in mobile and tablet computing, pervasive and ubiquitous computing, wearable computing and domotics. June 2024 - Journal Impact Factor: 5.4. Q1 journal in \"Health Care Sciences & Services\" and \"Medical Informatics\" categories. (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:"JMIR mhealth and uhealth is a new journal focussing on mobile and ubiquitous health technologies, including smartphones, augmented reality (Google Glasses), intelligent domestic devices, implantable devices, and other technologies designed to maintain health and improve life.",path:ay,slug:ay,seq:g,enabled:b,environment:d,url:"https:\u002F\u002Fmhealth.jmir.org",batch:e,year:S,colour:E,impact:"5.4",order:e,published:2736,transfers:a,cite_score:"12.6"},{journal_id:45,title:"Online Journal of Public Health Informatics",tag:"A leading peer-reviewed, open access journal dedicated to the dissemination of high-quality research and innovation in the field of public health informatics.",description:a,path:az,slug:az,seq:T,enabled:b,environment:d,url:"https:\u002F\u002Fojphi.jmir.org",batch:a,year:aA,colour:"#3399FF",impact:c,order:T,published:1719,transfers:a,cite_score:a},{journal_id:U,title:"JMIR Public Health and Surveillance",tag:"A multidisciplinary journal that focuses on the intersection of public health and technology, public health informatics, mass media campaigns, surveillance, participatory epidemiology, and innovation in public health practice and research. June 2024 - Journal Impact Factor: 3.5. Q1 journal in \"Public, Environmental & Occupational Health\" category (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:"Innovations in Public Health practice and research",path:aB,slug:aB,seq:B,enabled:b,environment:d,url:"https:\u002F\u002Fpublichealth.jmir.org",batch:b,year:x,colour:"#01538A",impact:V,order:B,published:1644,transfers:a,cite_score:"13.7"},{journal_id:n,title:"JMIR Medical Informatics",tag:"Clinical informatics, decision support for health professionals, electronic health records, and eHealth infrastructures. June 2024 - Journal Impact Factor: 3.1 (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:"Clinical informatics",path:aC,slug:aC,seq:i,enabled:b,environment:d,url:"https:\u002F\u002Fmedinform.jmir.org",batch:e,year:S,colour:D,impact:"3.1",order:z,published:1409,transfers:a,cite_score:"7.9"},{journal_id:W,title:"JMIR Mental Health",tag:"A journal focused on Internet interventions, technologies, and digital innovations for mental health and behavior change. Official journal of the Society for Digital Psychiatry. June 2024 - Journal Impact Factor: 4.8. Q1 journal in \"Psychiatry\" category. (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:c,path:aD,slug:aD,seq:n,enabled:b,environment:d,url:"https:\u002F\u002Fmental.jmir.org",batch:b,year:X,colour:"#45936C",impact:aE,order:A,published:1076,transfers:a,cite_score:"10.8"},{journal_id:B,title:"JMIR Human Factors",tag:"Making health care interventions and technologies usable, safe, and effective. June 2024 - Journal Impact Factor: 2.6 (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:"Usability Studies and Ergonomics",path:aF,slug:aF,seq:aG,enabled:b,environment:d,url:"https:\u002F\u002Fhumanfactors.jmir.org",batch:e,year:X,colour:"#008C9E",impact:"2.6",order:aH,published:803,transfers:a,cite_score:"3.4"},{journal_id:Q,title:"JMIR Serious Games",tag:"A multidisciplinary journal on gaming and gamification including simulation and immersive virtual reality for health education\u002Fpromotion, teaching, medicine, rehabilitation, and social change. June 2024 - Journal Impact Factor: 3.8. Q1 journal in \"Health Care Sciences & Services\" category. (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:"Serious games for health and social change",path:aI,slug:aI,seq:U,enabled:b,environment:d,url:"https:\u002F\u002Fgames.jmir.org",batch:b,year:S,colour:"#4A5A67",impact:"3.8",order:i,published:635,transfers:a,cite_score:"7.3"},{journal_id:Y,title:"JMIR Medical Education",tag:"Technology, innovation and openess in medical education in the information age. June 2024 - Journal Impact Factor: 3.2. Q1 journal in \"Education, Scientific Disciplines\" category. (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:c,path:aJ,slug:aJ,seq:aH,enabled:b,environment:d,url:"https:\u002F\u002Fmededu.jmir.org",batch:e,year:x,colour:"#6678A6",impact:aK,order:U,published:556,transfers:a,cite_score:"6.9"},{journal_id:A,title:"Iproceedings",tag:"Electronic Proceedings, Presentations and Posters of Leading Conferences",description:c,path:aL,slug:aL,seq:v,enabled:b,environment:d,url:"https:\u002F\u002Fwww.iproc.org",batch:e,year:x,colour:"#6F7D80",impact:a,order:Z,published:510,transfers:a,cite_score:a},{journal_id:_,title:"JMIR Aging",tag:"Digital health technologies, apps, and informatics for patient education, medicine and nursing, preventative interventions, and clinical care \u002F home care for elderly populations. June 2024 - Journal Impact Factor: 5.0. Q1 journal in \"Geriatrics & Gerontology\", \"Gerontology\" and \"Medical Informatics\" categories. (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:c,path:aM,slug:aM,seq:Z,enabled:b,environment:d,url:"https:\u002F\u002Faging.jmir.org",batch:e,year:y,colour:"#979bc4",impact:"5.0",order:g,published:436,transfers:a,cite_score:"6.5"},{journal_id:g,title:"Interactive Journal of Medical Research",tag:"A new general medical journal for the 21st centrury, focusing on innovation in health and medical research. June 2024 - Journal Impact Factor: 1.9 (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:c,path:"ijmr",slug:"i-jmr",seq:A,enabled:b,environment:d,url:"https:\u002F\u002Fwww.i-jmr.org",batch:e,year:P,colour:ar,impact:"1.9",order:aN,published:433,transfers:a,cite_score:a},{journal_id:aO,title:"JMIRx Med",tag:aP,description:a,path:aQ,slug:aQ,seq:$,enabled:b,environment:d,url:"https:\u002F\u002Fxmed.jmir.org",batch:a,year:aa,colour:"#3187df",impact:c,order:W,published:420,transfers:a,cite_score:a},{journal_id:Z,title:"JMIR Pediatrics and Parenting",tag:"Improving pediatric and adolescent health outcomes and empowering and educating parents. June 2024 - Journal Impact Factor: 2.1. Q1 journal in \"Pediatrics\" category. (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:c,path:aR,slug:aR,seq:ab,enabled:b,environment:d,url:"https:\u002F\u002Fpediatrics.jmir.org",batch:e,year:y,colour:"#d2a9ad",impact:"2.1",order:I,published:406,transfers:a,cite_score:"5"},{journal_id:ac,title:"JMIR Cancer",tag:"Patient-centered innovations, education and technology for cancer care, cancer survivorship, and cancer research. June 2024 - Journal Impact Factor: 3.3 (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:c,path:aS,slug:aS,seq:I,enabled:b,environment:d,url:"https:\u002F\u002Fcancer.jmir.org",batch:e,year:x,colour:"#584677",impact:"3.3",order:aG,published:386,transfers:a,cite_score:"4.1"},{journal_id:ab,title:"JMIR Dermatology",tag:"Technologies, devices, apps, and informatics applications for patient education in dermatology, including preventative interventions, and clinical care for dermatological populations",description:c,path:aT,slug:aT,seq:K,enabled:b,environment:d,url:"https:\u002F\u002Fderma.jmir.org",batch:e,year:y,colour:"#ecac7d",impact:a,order:ad,published:308,transfers:a,cite_score:"1.2"},{journal_id:ad,title:"JMIR Diabetes",tag:"Emerging Technologies, Medical Devices, Apps, Sensors, and Informatics to Help People with Diabetes.",description:c,path:aU,slug:aU,seq:Q,enabled:b,environment:d,url:"https:\u002F\u002Fdiabetes.jmir.org",batch:e,year:ae,colour:"#5c89c7",impact:c,order:Y,published:268,transfers:a,cite_score:"4"},{journal_id:af,title:"JMIR Rehabilitation and Assistive Technologies",tag:"Development and Evaluation of Rehabilitation, Physiotherapy and Assistive Technologies, Robotics, Prosthetics and Implants, Mobility and Communication Tools, Home Automation and Telerehabilitation",description:c,path:aV,slug:aV,seq:aN,enabled:b,environment:d,url:"https:\u002F\u002Frehab.jmir.org",batch:e,year:X,colour:"#15638E",impact:a,order:ac,published:263,transfers:a,cite_score:"4.2"},{journal_id:K,title:"JMIR Cardio",tag:"Electronic, mobile, digital health approaches in cardiology and for cardiovascular health.\r \r",description:c,path:aW,slug:aW,seq:ac,enabled:b,environment:d,url:"https:\u002F\u002Fcardio.jmir.org",batch:e,year:R,colour:"#791f20",impact:a,order:J,published:198,transfers:a,cite_score:V},{journal_id:38,title:"JMIR Infodemiology",tag:"Focusing on determinants and distribution of health information and misinformation on the internet, and its effect on public and individual health. June 2024 - Journal Impact Factor: 3.5. Q1 journal in \"Health Care Sciences & Services\" and \"Public, Environmental, and Occupational Health\" categories. (Source: Journal Citation Reports™ 2024 from Clarivate™)",description:"Focusing on determinants and distribution of health information and misinformation on the internet, and its effect on public and individual health.",path:aX,slug:aX,seq:37,enabled:b,environment:d,url:"https:\u002F\u002Finfodemiology.jmir.org",batch:e,year:2021,colour:"#32A852",impact:V,order:n,published:150,transfers:a,cite_score:aE},{journal_id:aY,title:"JMIR AI",tag:"The leading peer-reviewed journal for digital medicine, and health & health care in the internet age",description:"JMIR AI is a new journal that focuses on the applications of AI in health settings. This includes contemporary developments as well as historical examples, with an emphasis on sound methodological evaluations of AI techniques and authoritative analyses. It is intended to be the main source of reliable information for health informatics professionals to learn about how AI techniques can be applied and evaluated.",path:aZ,slug:aZ,seq:T,enabled:b,environment:d,url:"https:\u002F\u002Fai.jmir.org",batch:e,year:O,colour:ag,impact:c,order:w,published:119,transfers:a,cite_score:a},{journal_id:ah,title:"JMIR Perioperative Medicine",tag:"Technologies for pre- and post-operative education, preventative interventions and clinical care for surgery and anaesthesiology patients, as well as informatics applications in anesthesia, surgery, critical care and pain medicine",description:c,path:a_,slug:a_,seq:_,enabled:b,environment:d,url:"https:\u002F\u002Fperiop.jmir.org",batch:e,year:y,colour:"#187662",impact:a,order:ai,published:115,transfers:a,cite_score:"0.5"},{journal_id:$,title:"JMIR Nursing",tag:"Modern nursing in the age of information technology, patient empowerment and preventative, predictive, personal medicine",description:a,path:a$,slug:a$,seq:ah,enabled:b,environment:d,url:"https:\u002F\u002Fnursing.jmir.org",batch:e,year:y,colour:"#429a99",impact:a,order:aw,published:114,transfers:a,cite_score:"5.2"},{journal_id:ba,title:"Journal of Participatory Medicine",tag:"The Journal of Participatory Medicine is a peer-reviewed, open access journal with the mission to advance the understanding and practice of participatory medicine among health care professionals and patients.\n\nIt is the Official Journal of the Society for Participatory Medicine.",description:c,path:bb,slug:bb,seq:ad,enabled:b,environment:d,url:"https:\u002F\u002Fjopm.jmir.org",batch:e,year:aA,colour:"#2ea3f2",impact:a,order:H,published:98,transfers:a,cite_score:aK},{journal_id:H,title:"JMIR Biomedical Engineering",tag:"Engineering for health technologies, medical devices, and innovative medical treatments and procedures.",description:c,path:bc,slug:bc,seq:G,enabled:b,environment:d,url:"https:\u002F\u002Fbiomedeng.jmir.org",batch:e,year:ae,colour:bd,impact:a,order:K,published:92,transfers:a,cite_score:a},{journal_id:G,title:"JMIR Bioinformatics and Biotechnology",tag:"Methods, devices, web-based platforms, open data and open software tools for big data analytics, understanding biological\u002Fmedical data, and information retrieval in biology and medicine.",description:c,path:be,slug:be,seq:w,enabled:b,environment:d,url:"https:\u002F\u002Fbioinform.jmir.org",batch:e,year:aa,colour:bd,impact:a,order:ba,published:51,transfers:a,cite_score:"2.9"},{journal_id:e,title:"Medicine 2.0",tag:"Official proceedings publication of the Medicine 2.0 Congress",description:a,path:"med20",slug:"medicine20",seq:z,enabled:b,environment:d,url:"https:\u002F\u002Fwww.medicine20.com",batch:b,year:P,colour:E,impact:a,order:_,published:K,transfers:a,cite_score:a},{journal_id:43,title:"Asian\u002FPacific Island Nursing Journal",tag:"The official journal of the Asian American \u002F Pacific Islander Nurses Association (AAPINA), devoted to the exchange of knowledge in relation to Asian and Pacific Islander health and nursing care. Created to fill the gap between nursing science and behavioral\u002Fsocial sciences, the journal offers a forum for empirical, theoretical and methodological issues related to Asian American \u002F Pacific Islander ethnic, cultural values and beliefs and biological and physiological phenomena that can affect nursing care.",description:"The official journal of the Asian American \u002F Pacific Islander Nurses Association, this is a peer-reviewed, open access journal for the exchange of knowledge in relation to Asian and Pacific Islander health and nursing care. It will serve as a voice for nursing and other health care providers for research, education, and practice.",path:bf,slug:bf,seq:aY,enabled:b,environment:d,url:"https:\u002F\u002Fapinj.jmir.org",batch:e,year:R,colour:u,impact:c,order:F,published:ai,transfers:a,cite_score:"1.8"},{journal_id:f,title:N,tag:as,description:at,path:r,slug:r,seq:F,enabled:b,environment:d,url:au,batch:e,year:O,colour:u,impact:a,order:G,published:H,transfers:a,cite_score:a},{journal_id:35,title:"JMIRx Bio",tag:aP,description:c,path:bg,slug:bg,seq:aO,enabled:b,environment:d,url:"https:\u002F\u002Fxbio.jmir.org",batch:e,year:2023,colour:"#bf2433",impact:a,order:af,published:J,transfers:a,cite_score:a},{journal_id:46,title:"JMIR XR and Spatial Computing (JMXR)",tag:"A new peer-reviewed journal for extended reality and spatial computing in health and health care. ",description:a,path:bh,slug:bh,seq:bi,enabled:b,environment:d,url:"https:\u002F\u002Fxr.jmir.org",batch:e,year:al,colour:"#887ECB",impact:a,order:bi,published:w,transfers:a,cite_score:a},{journal_id:ai,title:"JMIR Data",tag:"A muldisciplinary journal to publish open datasets for analysis and re-analysis",description:c,path:bj,slug:bj,seq:Y,enabled:b,environment:d,url:"https:\u002F\u002Fdata.jmir.org",batch:e,year:aa,colour:"#5A6672",impact:a,order:$,published:i,transfers:a,cite_score:a},{journal_id:J,title:"JMIR Challenges",tag:"JMIR Challenges is a new platform connecting \"solution-seekers\" (sponsors such as companies or other researchers) with \"solution-providers\" (entrants, such as innovators, researchers, or developers in the ehealth space)",description:c,path:bk,slug:bk,seq:af,enabled:b,environment:d,url:"https:\u002F\u002Fchallenges.jmir.org",batch:e,year:ae,colour:ag,impact:a,order:ah,published:e,transfers:a,cite_score:a},{journal_id:w,title:"JMIR Preprints",tag:"A preprint server for pre-publication\u002Fpre-peer-review preprints intended for community review as well as ahead-of-print (accepted) manuscripts",description:"Publish your paper for open peer-review.",path:bl,slug:bl,seq:W,enabled:b,environment:d,url:"https:\u002F\u002Fpreprints.jmir.org",batch:b,year:x,colour:ag,impact:a,order:ab,published:b,transfers:a,cite_score:a}]},license:{},page:{data:a},pages:{data:{}},preprint:{data:a},searchArticles:{data:{}},searchAuthors:{data:[],pagination:{}},searchHelp:{data:{}},sections:{data:[],allSections:[],journalSections:[]},submission:{data:{id:a,files:{toc:[],figures:[],appendicies:[],other:[]}},file:a,affiliation:a,suggestedAffiliations:a,events:{extraction:[]},errors:{}},subscription:{},theme:{data:[]},themes:{random:a,data:[],sortType:"pubDate",sortOrder:"desc"}},serverRendered:h,routePath:am,config:{environment:d,_app:{basePath:"\u002F",assetsPath:"\u002F_nuxt\u002F",cdnURL:a}}}}(null,1,"","production",2,42,3,true,5,0,"New York","NY","United States",7,"Recreation Therapist","California Correction Health Care Services","California","neuro",-128,"Icahn School of Medicine at Mount Sinai","#3399ff",13,18,2015,2018,10,4,6,"#247CB3","#82ABB9","#65AD8C",36,19,24,12,22,26,false,"MD","JMIR Neurotechnology",2022,2012,15,2017,2013,39,9,"3.5",16,2014,20,30,31,33,2020,29,21,23,2016,17,"#666666",32,25,"2023-10-02T18:06:32.000Z",1469,2024,"\u002F2024\u002F1\u002Fe51822","Benjamin R","Kummer","Department of Neurology",369025,"#22B2C1","Cross-disciplinary journal that connects the broad domains of clinical neuroscience and all related technologies. The journal provides a space for the publication of research exploring how technologies can be applied in clinical neuroscience (e.g., neurology, neurosurgery, neuroradiology) to prevent, diagnose, and treat neurological disorders.","JMIR Neuro is an innovative new journal which aims to bridge clinical neurology & neurosurgery with advances in the web space and digital technologies","https:\u002F\u002Fneuro.jmir.org","jmir",27,"formative","mhealth","ojphi",2009,"publichealth","medinform","mental","4.8","humanfactors",8,11,"games","mededu","3.2","iproc","aging",14,34,"Overlay journal for preprints with post-review manuscript marketplace","xmed","pediatrics","cancer","derma","diabetes","rehab","cardio","infodemiology",41,"ai","periop","nursing",28,"jopm","biomedeng","#474760","bioinform","apinj","xbio","xr",40,"data","challenges","preprints"));</script><script src="/_nuxt/d762be0.js" defer></script><script src="/_nuxt/f3339c8.js" defer></script><script src="/_nuxt/76715db.js" defer></script><script src="/_nuxt/114bf28.js" defer></script><script src="/_nuxt/557a776.js" defer></script><script src="/_nuxt/f65c9d0.js" defer></script><script src="/_nuxt/b7cb769.js" defer></script><script src="/_nuxt/65075b6.js" defer></script><script src="/_nuxt/8499a81.js" defer></script><script src="/_nuxt/60620ca.js" defer></script><script src="/_nuxt/7bab8a9.js" defer></script><script src="/_nuxt/6c336c8.js" defer></script><script src="/_nuxt/e63f738.js" defer></script><script src="/_nuxt/2f3783d.js" defer></script><script src="/_nuxt/8679c54.js" defer></script><script src="/_nuxt/84ba7a4.js" defer></script><script src="/_nuxt/2042386.js" defer></script><script src="/_nuxt/7aa873e.js" defer></script><script src="/_nuxt/4f5e72f.js" defer></script><script src="/_nuxt/512387a.js" defer></script><script src="/_nuxt/69df2f4.js" defer></script><script src="/_nuxt/c8abcfd.js" defer></script><script src="/_nuxt/ef88041.js" defer></script><script src="/_nuxt/4e67ec6.js" defer></script><script src="/_nuxt/edec9c1.js" defer></script><script src="/_nuxt/a0160a8.js" defer></script><script src="/_nuxt/88d69a9.js" defer></script><script src="/_nuxt/456ce06.js" defer></script><script src="/_nuxt/3049f10.js" defer></script><script src="/_nuxt/fec3470.js" defer></script><script src="/_nuxt/48052c9.js" defer></script><script src="/_nuxt/f76b897.js" defer></script><script src="/_nuxt/f2fb177.js" defer></script><script src="/_nuxt/fdd6c0c.js" defer></script><script src="/_nuxt/51c3064.js" defer></script><script src="/_nuxt/67892e3.js" defer></script><script src="/_nuxt/6c2c095.js" defer></script><script src="/_nuxt/9a11cd2.js" defer></script><script src="/_nuxt/e3a1965.js" defer></script><script src="/_nuxt/c14e3a6.js" defer></script><script src="/_nuxt/ceb5b90.js" defer></script><script src="/_nuxt/2715511.js" defer></script><script src="/_nuxt/7819e44.js" defer></script><script src="/_nuxt/8d00881.js" defer></script><script src="/_nuxt/6171ac4.js" defer></script><script src="/_nuxt/9a8e2c4.js" defer></script><script src="/_nuxt/0f313a0.js" defer></script><script src="/_nuxt/90db445.js" defer></script><script src="/_nuxt/ee7c71d.js" defer></script><script src="/_nuxt/0c77ea6.js" defer></script><script src="/_nuxt/c4453ca.js" defer></script><script data-n-head="ssr" src="https://d1bxh8uas1mnw7.cloudfront.net/assets/embed.js" data-body="true" defer></script><script data-n-head="ssr" src="https://badge.dimensions.ai/badge.js" data-body="true" defer></script><script data-n-head="ssr" type="text/javascript" charset="utf-8" data-body="true">_linkedin_partner_id = "3149908"; window._linkedin_data_partner_ids = window._linkedin_data_partner_ids || []; window._linkedin_data_partner_ids.push(_linkedin_partner_id);</script><script data-n-head="ssr" type="text/javascript" charset="utf-8" data-body="true">(function(l) { if (!l){window.lintrk = function(a,b){window.lintrk.q.push([a,b])}; window.lintrk.q=[]} var s = document.getElementsByTagName("script")[0]; var b = document.createElement("script"); b.type = "text/javascript";b.async = true; b.src = "https://snap.licdn.com/li.lms-analytics/insight.min.js"; s.parentNode.insertBefore(b, s);})(window.lintrk);</script></body></html>