An Explainable Self-Attention Deep Neural Network for Detecting Mild Cognitive Impairment Using Multi-input Digital Drawing Tasks

An Explainable Self-Attention Deep Neural Network for Detecting Mild Cognitive Impairment Using Multi-input Digital Drawing Tasks | 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 Research Article An Explainable Self-Attention Deep Neural Network for Detecting Mild Cognitive Impairment Using Multi-input Digital Drawing Tasks Chaipat Chunharas, Natthanan Ruengchaijatuporn, Itthi Chatnuntawech, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-1363649/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Background: Mild cognitive impairment (MCI) is an early stage of cognitive decline beyond the normal aging process, which could develop into dementia. An early detection of MCI is a crucial step for timely prevention and intervention. Recent studies have developed deep learning models to detect MCI and dementia using a bedside task like the classic clock-drawing test (CDT). While these models succeed at distinguishing severe forms of dementia, it remains a challenge to predict the early stage of the disease using the CDT data alone. Moreover, the state-of-the-art deep learning techniques still face the black-box challenges, making it questionable to implement them in the clinical setting. Methods: We consecutively recruited 981 subjects from Geriatric Clinic at King Chulalongkorn Memorial Hospital (651 healthy subjects and 267 MCI patients according to Petersen’s criteria). We propose a novel deep learning modeling framework that incorporates data from multiple drawing tasks including the CDT, cube-copying, and trail-making tasks obtained from a digital platform. Soft-label method and self-attention were applied to improve the model performance and provide visual explanation that aids the interpretation of the deep learning model. The performance of the model was measured on an independent dataset. The visualization of our model and the traditional grad-CAM approach were scored by experienced medical personnels. Results: Using multiple drawings as inputs improves classification between healthy versus MCI in terms of accuracy (0.75 ± 0.01 vs. 0.80 ± 0.01), F1 score (0.36 ± 0.04 vs. 0.59 ± 0.02) and AUC (0.74 ± 0.01 vs. 0.81 ± 0.02). By adding soft-label and self-attention methods improve the accuracy (0.81 ± 0.01) F1 score (0.65 ± 0.01) and AUC (0.84 ± 0.01). The experienced medical personnels preferred the visualization of our model over the grad-CAM model (2.82 ± 0.49 vs. 1.52 ± 0.54). Conclusions: Our model achieves better classification performance at detecting MCI compared to a well-established convolutional neural network model. Moreover, our model can highlight features of the MCI data that considerably deviate from those of the healthy aging population, offering accurate predictions for detecting MCI along with visual explanation that aids the interpretation of the deep learning model. Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Major revision 24 Mar, 2022 Reviews received at journal 24 Mar, 2022 Reviewers agreed at journal 18 Mar, 2022 Reviews received at journal 27 Jan, 2022 Reviewers agreed at journal 25 Jan, 2022 Reviewers invited by journal 22 Jan, 2022 Editor assigned by journal 22 Jan, 2022 Submission checks completed at journal 20 Jan, 2022 First submitted to journal 06 Jan, 2022 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-1363649","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":84136635,"identity":"fc09c990-50f2-4678-b575-d131ef415aaa","order_by":0,"name":"Chaipat Chunharas","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA8ElEQVRIie3SsYrCQBCA4QkHsZlj2xEhvsKIoJ3PkuNAmxRbCieoBM7OOqJv4QsEBrTJAwh3jQSsLATB6jhuE8Fy1/KK/ZssYT6yYRfA5/uPEUBeLxqzEEDfX748QRgA87B+moLURaAepfhJotbpUfTkB9SqPJXIA1CLvFFq20e+dyzZjoG+hv0u8jtQEQdpZiFMMQuGDNxKwhay+fGD2Rhayegi+GtIs6jIFNpukrC8fhpCWBEBdhE6JFpWyy5SMew1N7zHTvE2X9uIykbbq75FkVrIic7jjyjaS361karqrB8zZhHMHMB1PXw+n8/3B+aYQV74h2JqAAAAAElFTkSuQmCC","orcid":"","institution":"Cognitive Clinical and Computational Neuroscience, Chula Neuroscience Center, King Chulalongkorn Memorial Hospital, Faculty of Medicine, Chulalongkorn University, Bangkok","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Chaipat","middleName":"","lastName":"Chunharas","suffix":""},{"id":84136625,"identity":"8146592c-f615-4840-babb-41d0e4b805d1","order_by":1,"name":"Natthanan Ruengchaijatuporn","email":"","orcid":"","institution":"Computational Molecular Biology Group, Faculty of Medicine, Chulalongkorn University, Bangkok","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Natthanan","middleName":"","lastName":"Ruengchaijatuporn","suffix":""},{"id":84136626,"identity":"3e098c32-5fdc-4da6-b756-475c9a4c5d88","order_by":2,"name":"Itthi Chatnuntawech","email":"","orcid":"","institution":"National Nanotechnology Center, National Science and Technology Development Agency, Pathum Thani","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Itthi","middleName":"","lastName":"Chatnuntawech","suffix":""},{"id":84136627,"identity":"bae11210-c89b-49c0-868e-b1b025944e0e","order_by":3,"name":"Surat Teerapittayanon","email":"","orcid":"","institution":"National Nanotechnology Center, National Science and Technology Development Agency, Pathum Thani","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Surat","middleName":"","lastName":"Teerapittayanon","suffix":""},{"id":84136628,"identity":"73b76db5-b4e6-4f51-bdff-3a0a1ebf0da5","order_by":4,"name":"Sira Sriswasdi","email":"","orcid":"","institution":"Center for Artificial Intelligence in Medicine, Research Affairs, Faculty of Medicine, Chulalongkorn University, Bangkok","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Sira","middleName":"","lastName":"Sriswasdi","suffix":""},{"id":84136629,"identity":"16b4041b-0392-4510-867d-155fe1736dd9","order_by":5,"name":"Sirawaj Itthipuripat","email":"","orcid":"","institution":"Neuroscience Center for Research and Innovation, Learning Institute, King Mongkut’s University of Technology Thonburi, Bangkok","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Sirawaj","middleName":"","lastName":"Itthipuripat","suffix":""},{"id":84136630,"identity":"d2f9d4fb-5d0e-4339-95c7-359ecf011092","order_by":6,"name":"Solaphat Hemrungrojn","email":"","orcid":"","institution":"Department of Psychiatry, Faculty of Medicine, Chulalongkorn University, Bangkok","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Solaphat","middleName":"","lastName":"Hemrungrojn","suffix":""},{"id":84136631,"identity":"84e617b8-c8d0-4afd-8bf0-27f9da3922e0","order_by":7,"name":"Prodpran Bunyabukkana","email":"","orcid":"","institution":"Department of Computer Engineering, Faculty of Engineering, Chulalongkorn University, Bangkok","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Prodpran","middleName":"","lastName":"Bunyabukkana","suffix":""},{"id":84136632,"identity":"0d863358-a6e6-48f3-b951-ec15ad6905f3","order_by":8,"name":"Aisawan Petchlorlian","email":"","orcid":"","institution":"Geriatric Excellence Center, King Chulalongkorn Memorial Hospital, Thai Red Cross Society, Faculty of Medicine, Chulalongkorn University, Bangkok","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Aisawan","middleName":"","lastName":"Petchlorlian","suffix":""},{"id":84136633,"identity":"b7b78780-af5c-4874-a558-785cab915138","order_by":9,"name":"Sedthapong Chunamchai","email":"","orcid":"","institution":"Cognitive Clinical and Computational Neuroscience, Chula Neuroscience Center, King Chulalongkorn Memorial Hospital, Faculty of Medicine, Chulalongkorn University, Bangkok","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Sedthapong","middleName":"","lastName":"Chunamchai","suffix":""},{"id":84136634,"identity":"51c0d779-89ce-4e94-a663-f9b8246b2e0d","order_by":10,"name":"Thiparat Chotibut","email":"","orcid":"","institution":"Chula Intelligent and Complex Systems Lab, Department of Physics, Faculty of Science, Chulalongkorn University, Bangkok","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Thiparat","middleName":"","lastName":"Chotibut","suffix":""}],"badges":[],"createdAt":"2022-02-16 01:14:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-1363649/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-1363649/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":18327032,"identity":"46bc9503-45ba-44e1-bccf-10a629552d56","added_by":"auto","created_at":"2022-02-17 15:18:46","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":606425,"visible":true,"origin":"","legend":"","description":"","filename":"CDMCIConvAttclassificationmanuscript012022CC.pdf","url":"https://assets-eu.researchsquare.com/files/rs-1363649/v1_covered.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"An Explainable Self-Attention Deep Neural Network for Detecting Mild Cognitive Impairment Using Multi-input Digital Drawing Tasks","fulltext":[{"header":"Full Text","content":"This preprint is available for \u003ca href='/article/rs-1363649/latest.pdf' target='_blank'\u003edownload as a PDF\u003c/a\u003e."}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"alzheimers-research-and-therapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"azrt","sideBox":"Learn more about [Alzheimer's Research and Therapy](http://alzres.biomedcentral.com/)","snPcode":"13195","submissionUrl":"https://submission.nature.com/new-submission/13195/3","title":"Alzheimer's Research \u0026 Therapy","twitterHandle":"@AlzheimersRes","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-1363649/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-1363649/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eMild cognitive impairment (MCI) is an early stage of cognitive decline beyond the normal aging process, which could develop into dementia. An early detection of MCI is a crucial step for timely prevention and intervention. Recent studies have developed deep learning models to detect MCI and dementia using a bedside task like the classic clock-drawing test (CDT). While these models succeed at distinguishing severe forms of dementia, it remains a challenge to predict the early stage of the disease using the CDT data alone. Moreover, the state-of-the-art deep learning techniques still face the black-box challenges, making it questionable to implement them in the clinical setting.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eWe consecutively recruited 981 subjects from Geriatric Clinic at King Chulalongkorn Memorial Hospital (651 healthy subjects and 267 MCI patients according to Petersen’s criteria).\u0026nbsp;We propose a novel deep learning modeling framework that incorporates data from multiple drawing tasks including the CDT, cube-copying, and trail-making tasks obtained from a digital platform. Soft-label method and self-attention were applied to improve the model performance and provide visual explanation that aids the interpretation of the deep learning model. The performance of the model was measured on an independent dataset. The visualization of our model and the traditional grad-CAM approach were scored by experienced medical personnels.\u0026nbsp;\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eUsing multiple drawings as inputs improves classification between healthy versus MCI in terms of accuracy (0.75 ± 0.01 vs. 0.80 ± 0.01), F1 score (0.36 ± 0.04 vs. 0.59 ± 0.02) and AUC (0.74 ± 0.01 vs. 0.81 ± 0.02). By adding soft-label and self-attention methods improve the accuracy (0.81 ± 0.01) F1 score (0.65 ± 0.01) and AUC (0.84 ± 0.01). The experienced medical personnels preferred the visualization of our model over the grad-CAM model (2.82\u0026nbsp;± 0.49 vs. 1.52\u0026nbsp;± 0.54).\u003c/p\u003e\u003cp\u003e \u003cstrong\u003eConclusions:\u0026nbsp;\u003c/strong\u003eOur model achieves better classification performance at detecting MCI compared to a well-established convolutional neural network model.\u0026nbsp;Moreover, our model can highlight features of the MCI data that considerably deviate from those of the healthy aging population, offering accurate predictions for detecting MCI along with visual explanation that aids the interpretation of the deep learning model.\u0026nbsp;\u003c/p\u003e","manuscriptTitle":"An Explainable Self-Attention Deep Neural Network for Detecting Mild Cognitive Impairment Using Multi-input Digital Drawing Tasks","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2022-02-17 15:18:33","doi":"10.21203/rs.3.rs-1363649/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major revision","date":"2022-03-24T20:39:08+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2022-03-24T19:54:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"2b00de5a-0655-46ee-8640-ea6d9e691f44","date":"2022-03-18T23:26:36+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2022-01-27T18:01:25+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"0733bc2f-ad64-41fa-9681-bd270b9beac1","date":"2022-01-25T16:11:15+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2022-01-22T15:57:44+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2022-01-22T15:51:27+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2022-01-21T02:57:40+00:00","index":"","fulltext":""},{"type":"submitted","content":"Alzheimer's Research \u0026 Therapy","date":"2022-01-06T05:36:56+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"alzheimers-research-and-therapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"azrt","sideBox":"Learn more about [Alzheimer's Research and Therapy](http://alzres.biomedcentral.com/)","snPcode":"13195","submissionUrl":"https://submission.nature.com/new-submission/13195/3","title":"Alzheimer's Research \u0026 Therapy","twitterHandle":"@AlzheimersRes","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c323d461-00fe-4644-a20f-2cb4b04acbf6","owner":[],"postedDate":"February 17th, 2022","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2022-07-06T13:14:23+00:00","versionOfRecord":[],"versionCreatedAt":"2022-02-17 15:18:33","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-1363649","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-1363649","identity":"rs-1363649","version":["v1"]},"buildId":"ehx78VzkSd0WSzXnipQa-","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}