EyePose: Pose-guided Saccadic Eye Movement Video Generation for Deep Learning-Based Neurologic Disease Phenotyping

EyePose: Pose-guided Saccadic Eye Movement Video Generation for Deep Learning-Based Neurologic Disease Phenotyping | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article EyePose: Pose-guided Saccadic Eye Movement Video Generation for Deep Learning-Based Neurologic Disease Phenotyping Tianyu Lin, Jooyoung Ryu, Puvada Sreevarsha, Rahul Srinivasaragavan, and 11 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6995265/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Eye movements, including saccades, are widely regarded as highly sensitive and objective biomarkers of neurophysiologic states. Detecting saccadic signatures in neurologic diseases offers a promising, rapid, portable, and non-invasive alternative to current diagnostic tools, such as brain imaging, while overcoming access limitations and cost barriers. Currently, no robust video-oculographic solutions exist for localizing brain abnormalities due to privacy concerns and the lack of large datasets required to train accurate, reliable models. In this work, we propose the first fully synthetic, patient data-free, multimodal eye movement generation pipeline for building a generalizable dataset for saccade analysis. Using this synthetic dataset, we trained a deep learning classifier to distinguish between normal and abnormal (hypometria and hypermetria) saccadic accuracies, and evaluated its performance on real-world clinical data. The model achieved an AUROC of 0.76 and sensitivity of 0.71, showing that the synthetic data has strong potential to generalize for clinical applications. This work highlights the potential of synthetic eye movement data to be used to develop screening tools for at-home and emergency room settings. Health sciences/Biomarkers Biological sciences/Computational biology and bioinformatics Health sciences/Diseases Physical sciences/Engineering Health sciences/Health care Physical sciences/Mathematics and computing Health sciences/Medical research Health sciences/Neurology Biological sciences/Neuroscience Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6995265","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":479177916,"identity":"84d81c93-95f4-46d3-9ee3-7374ac49b369","order_by":0,"name":"Tianyu Lin","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Tianyu","middleName":"","lastName":"Lin","suffix":""},{"id":479177917,"identity":"57bf5bfe-0e51-4d7d-9961-644cc21b3581","order_by":1,"name":"Jooyoung Ryu","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Jooyoung","middleName":"","lastName":"Ryu","suffix":""},{"id":479177919,"identity":"7425c520-065f-4930-b544-33e37ae30cee","order_by":2,"name":"Puvada Sreevarsha","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Puvada","middleName":"","lastName":"Sreevarsha","suffix":""},{"id":479177921,"identity":"45e04897-4f63-402b-ae77-72df0b2a6a1f","order_by":3,"name":"Rahul Srinivasaragavan","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Rahul","middleName":"","lastName":"Srinivasaragavan","suffix":""},{"id":479177922,"identity":"630d6261-891d-46f1-a28f-d49dae06038e","order_by":4,"name":"Riya Satavlekar","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Riya","middleName":"","lastName":"Satavlekar","suffix":""},{"id":479177923,"identity":"c4a422ac-f2da-4a52-981e-aac4c0e040fd","order_by":5,"name":"Susan Kim","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Susan","middleName":"","lastName":"Kim","suffix":""},{"id":479177924,"identity":"ad17ef1a-8568-4e5e-b490-f4f1a4985206","order_by":6,"name":"Nidhi Soley","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Nidhi","middleName":"","lastName":"Soley","suffix":""},{"id":479177925,"identity":"3d15d6e9-f1f1-40ab-9c61-fe5c01ca3a42","order_by":7,"name":"Yujie Yan","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Yujie","middleName":"","lastName":"Yan","suffix":""},{"id":479177927,"identity":"880eae60-354e-4703-b7c8-478fb67722da","order_by":8,"name":"Ishan Vatsaraj","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Ishan","middleName":"","lastName":"Vatsaraj","suffix":""},{"id":479177929,"identity":"493b96b6-ae57-4458-8910-d18f238f09da","order_by":9,"name":"Carl Harris","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Carl","middleName":"","lastName":"Harris","suffix":""},{"id":479177931,"identity":"c70962a2-25d7-43fd-8064-1cfbaf6a0429","order_by":10,"name":"Aimon Rahman","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Aimon","middleName":"","lastName":"Rahman","suffix":""},{"id":479177934,"identity":"0d72d400-1c6e-4626-ac71-8ed3b410f5c0","order_by":11,"name":"Vishal Patel","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Vishal","middleName":"","lastName":"Patel","suffix":""},{"id":479177937,"identity":"758b0787-ae5b-43fd-b176-fe7f76773cb2","order_by":12,"name":"Joseph Greenstein","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Joseph","middleName":"","lastName":"Greenstein","suffix":""},{"id":479177941,"identity":"b82c13a4-0b2b-4631-a64f-dd4be91eb9b8","order_by":13,"name":"Casey Taylor","email":"","orcid":"","institution":"Johns Hopkins University","correspondingAuthor":false,"prefix":"","firstName":"Casey","middleName":"","lastName":"Taylor","suffix":""},{"id":479177943,"identity":"eb57f0d7-36ab-4f9b-a42f-950d3e44c5e2","order_by":14,"name":"Kemar Green","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0klEQVRIiWNgGAWjYBACCQhlI8fPcIaBIYGBgbGBSC1pxpINJGo5lGhwgAfMIqxFsn/twwcfdxxIMD549uiGBww2shsOENAiLfHc2HDmmTt5ZgfOpd1IALqQoBY5iWNs0rxtz4rNDpwxA2o5nEiMFvbff9sOJ25uAGv5T1iLNH8bGzMjUMsGBrCWA4S1SM5gY5bsbUszlgA7zCDZeCYhLRLnjzF++NkGjMoZZ8xu/qiwk+0jpIVBIgHGACk1IKQcBPhhhvI3EKN8FIyCUTAKRiIAAHL4TWKDbWDOAAAAAElFTkSuQmCC","orcid":"","institution":"Johns Hopkins Medicine","correspondingAuthor":true,"prefix":"","firstName":"Kemar","middleName":"","lastName":"Green","suffix":""}],"badges":[],"createdAt":"2025-06-28 03:53:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6995265/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6995265/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":101365267,"identity":"edd32263-71e7-4cc7-a3ae-818428014645","added_by":"auto","created_at":"2026-01-29 00:54:08","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2333957,"visible":true,"origin":"","legend":"","description":"","filename":"NpjNatureMedicineSubmission.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6995265/v1_covered_88dc4e56-d55c-4eaf-b0c1-e430d098afa0.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"EyePose: Pose-guided Saccadic Eye Movement Video Generation for Deep Learning-Based Neurologic Disease Phenotyping","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6995265/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6995265/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Eye movements, including saccades, are widely regarded as highly sensitive and objective biomarkers of neurophysiologic states. Detecting saccadic signatures in neurologic diseases offers a promising, rapid, portable, and non-invasive alternative to current diagnostic tools, such as brain imaging, while overcoming access limitations and cost barriers. Currently, no robust video-oculographic solutions exist for localizing brain abnormalities due to privacy concerns and the lack of large datasets required to train accurate, reliable models. In this work, we propose the first fully synthetic, patient data-free, multimodal eye movement generation pipeline for building a generalizable dataset for saccade analysis. Using this synthetic dataset, we trained a deep learning classifier to distinguish between normal and abnormal (hypometria and hypermetria) saccadic accuracies, and evaluated its performance on real-world clinical data. The model achieved an AUROC of 0.76 and sensitivity of 0.71, showing that the synthetic data has strong potential to generalize for clinical applications. This work highlights the potential of synthetic eye movement data to be used to develop screening tools for at-home and emergency room settings.","manuscriptTitle":"EyePose: Pose-guided Saccadic Eye Movement Video Generation for Deep Learning-Based Neurologic Disease Phenotyping","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-02 04:55:11","doi":"10.21203/rs.3.rs-6995265/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"eb74e198-d392-4726-9fef-92e2cf0d6f7c","owner":[],"postedDate":"July 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":50881241,"name":"Health sciences/Biomarkers"},{"id":50881242,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":50881243,"name":"Health sciences/Diseases"},{"id":50881244,"name":"Physical sciences/Engineering"},{"id":50881245,"name":"Health sciences/Health care"},{"id":50881246,"name":"Physical sciences/Mathematics and computing"},{"id":50881247,"name":"Health sciences/Medical research"},{"id":50881248,"name":"Health sciences/Neurology"},{"id":50881249,"name":"Biological sciences/Neuroscience"}],"tags":[],"updatedAt":"2026-01-29T00:53:17+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-02 04:55:11","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6995265","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6995265","identity":"rs-6995265","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}