Interpretable Feature-Transformer Framework for Cross-Subject MCI Detection Using Nonlinear Dynamical and Graph-Theoretic EEG Features

Interpretable Feature-Transformer Framework for Cross-Subject MCI Detection Using Nonlinear Dynamical and Graph-Theoretic EEG Features | 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 Interpretable Feature-Transformer Framework for Cross-Subject MCI Detection Using Nonlinear Dynamical and Graph-Theoretic EEG Features Hadi Azizpour Lindi, Reza Shalbaf, Ahmad Shalbaf, Mohsen Sadat Shahabi, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8744978/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 Early and accurate detection of Mild Cognitive Impairment (MCI) is essential for preventing progression toward Alzheimer's disease (AD). In this cross-subject study, we investigate the effectiveness of entropy- and graph-based EEG features for distinguishing MCI from healthy controls (HC), using two modeling approaches: (1) a Transformer network applied to the engineered feature set, and (2) an EEGNet model trained on the same feature representation for comparison. The dataset consists of resting-state, eyes-closed EEG recordings from 183 participants (127 HC, 56 MCI), collected using a 20-channel STAT™ X24 wireless system and segmented into 3-second epochs. EEG data underwent standard preprocessing, including band-pass filtering, downsampling, normalization, and class-balancing augmentation applied to the minority class. From each channel, nonlinear dynamical measures (e.g., sample and fuzzy entropy, Higuchi fractal dimension, Lyapunov exponent) and graph-theoretic connectivity descriptors derived from coherence matrices across five frequency bands were extracted, yielding a structured 19×77 feature representation. The feature-based Transformer achieved the best performance (97.04% ± 0.72), outperforming the feature-based EEGNet baseline and highlighting the benefits of combining rich handcrafted features with attention-based modeling. SHAP (SHapley Additive exPlanations) analysis provided global and local interpretability, revealing the most influential nonlinear and connectivity features as well as the EEG channels contributing most to classification. Overall, these results demonstrate the effectiveness of feature-Transformer integration and support the potential of interpretable feature-driven deep learning models for early MCI detection. MCI Graph Theory Nonlinear Dynamics Entropy Transformer EEG Deep Learning Alzheimer’s disease EEGNet SHAP Resting State Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 07 Mar, 2026 Reviews received at journal 04 Mar, 2026 Reviews received at journal 01 Mar, 2026 Reviewers agreed at journal 09 Feb, 2026 Reviewers agreed at journal 09 Feb, 2026 Reviewers invited by journal 09 Feb, 2026 Editor assigned by journal 02 Feb, 2026 Submission checks completed at journal 02 Feb, 2026 First submitted to journal 30 Jan, 2026 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-8744978","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":588643966,"identity":"0f0955f9-d597-41cc-bdf8-9f079a170287","order_by":0,"name":"Hadi Azizpour Lindi","email":"","orcid":"","institution":"Institute for Cognitive Science Studies","correspondingAuthor":false,"prefix":"","firstName":"Hadi","middleName":"Azizpour","lastName":"Lindi","suffix":""},{"id":588643967,"identity":"bea6a426-0559-49fc-8d11-c121f3c7ca08","order_by":1,"name":"Reza Shalbaf","email":"","orcid":"","institution":"Institute for Cognitive Science Studies","correspondingAuthor":false,"prefix":"","firstName":"Reza","middleName":"","lastName":"Shalbaf","suffix":""},{"id":588643968,"identity":"8055ab08-7c29-49ef-af4e-af3d4617d9e3","order_by":2,"name":"Ahmad Shalbaf","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1ElEQVRIie3PMQuCQBTA8edykzkfCN5XeOHS5GdRhFwU/AAOQeBSORsGfQUlaBYO3NpvaAj6Ak5OIRlF0NLZ1nD/8bgf7z0AleoPYzloHMCygL5eiIygABiIbf9A6JN4Cyr7+ibm8sLjBIN9sZpTSBwwzPo7YbsGed5gVJ5PRwqND8RwJWOEi1wnt6ik0UBIDUSXbSaClus9Bix/kH4UCZFPUnRBDERLRxAUYcy3GU6HWw4zL/N1KWF5UF3jDhkrNpVoO8dia9liH7kA0ktUKpVKNaI7UTVAeKdu/LMAAAAASUVORK5CYII=","orcid":"","institution":"Shahid Beheshti University of Medical Sciences","correspondingAuthor":true,"prefix":"","firstName":"Ahmad","middleName":"","lastName":"Shalbaf","suffix":""},{"id":588643969,"identity":"19c74504-2642-4525-b35b-e66347098852","order_by":3,"name":"Mohsen Sadat Shahabi","email":"","orcid":"","institution":"Shahid Beheshti University of Medical Sciences","correspondingAuthor":false,"prefix":"","firstName":"Mohsen","middleName":"Sadat","lastName":"Shahabi","suffix":""},{"id":588643970,"identity":"dc9ad72a-bfde-4e9b-89d1-42d67a74ca4d","order_by":4,"name":"Peyman Abharian","email":"","orcid":"","institution":"Institute for Cognitive Science Studies","correspondingAuthor":false,"prefix":"","firstName":"Peyman","middleName":"","lastName":"Abharian","suffix":""}],"badges":[],"createdAt":"2026-01-30 21:38:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8744978/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8744978/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102746048,"identity":"4421cc94-7a73-49ab-976d-4370d38e7294","added_by":"auto","created_at":"2026-02-16 08:55:26","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3262494,"visible":true,"origin":"","legend":"","description":"","filename":"HadiAzizpourLindisnarticle.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8744978/v1_covered_6db72de6-c8d6-4887-ae8d-82125ce27154.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Interpretable Feature-Transformer Framework for Cross-Subject MCI Detection Using Nonlinear Dynamical and Graph-Theoretic EEG Features","fulltext":[],"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":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"cognitive-neurodynamics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"cody","sideBox":"Learn more about [Cognitive Neurodynamics](http://link.springer.com/journal/11571)","snPcode":"11571","submissionUrl":"https://submission.nature.com/new-submission/11571/3","title":"Cognitive Neurodynamics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"MCI, Graph Theory, Nonlinear Dynamics, Entropy, Transformer, EEG, Deep Learning, Alzheimer’s disease, EEGNet, SHAP, Resting State","lastPublishedDoi":"10.21203/rs.3.rs-8744978/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8744978/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Early and accurate detection of Mild Cognitive Impairment (MCI) is essential for preventing progression toward Alzheimer's disease (AD). In this cross-subject study, we investigate the effectiveness of entropy- and graph-based EEG features for distinguishing MCI from healthy controls (HC), using two modeling approaches: (1) a Transformer network applied to the engineered feature set, and (2) an EEGNet model trained on the same feature representation for comparison. The dataset consists of resting-state, eyes-closed EEG recordings from 183 participants (127 HC, 56 MCI), collected using a 20-channel STAT™ X24 wireless system and segmented into 3-second epochs. EEG data underwent standard preprocessing, including band-pass filtering, downsampling, normalization, and class-balancing augmentation applied to the minority class. From each channel, nonlinear dynamical measures (e.g., sample and fuzzy entropy, Higuchi fractal dimension, Lyapunov exponent) and graph-theoretic connectivity descriptors derived from coherence matrices across five frequency bands were extracted, yielding a structured 19×77 feature representation. The feature-based Transformer achieved the best performance (97.04% ± 0.72), outperforming the feature-based EEGNet baseline and highlighting the benefits of combining rich handcrafted features with attention-based modeling. SHAP (SHapley Additive exPlanations) analysis provided global and local interpretability, revealing the most influential nonlinear and connectivity features as well as the EEG channels contributing most to classification. Overall, these results demonstrate the effectiveness of feature-Transformer integration and support the potential of interpretable feature-driven deep learning models for early MCI detection.","manuscriptTitle":"Interpretable Feature-Transformer Framework for Cross-Subject MCI Detection Using Nonlinear Dynamical and Graph-Theoretic EEG Features","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-11 18:21:39","doi":"10.21203/rs.3.rs-8744978/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-08T02:55:59+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-04T17:53:21+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-01T13:34:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"33785026223604456894645111812086933391","date":"2026-02-10T02:29:20+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"177846588051985529549007197237114457080","date":"2026-02-09T10:00:49+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-09T09:58:25+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-02T10:07:24+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-02T10:04:10+00:00","index":"","fulltext":""},{"type":"submitted","content":"Cognitive Neurodynamics","date":"2026-01-30T21:32:40+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"cognitive-neurodynamics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"cody","sideBox":"Learn more about [Cognitive Neurodynamics](http://link.springer.com/journal/11571)","snPcode":"11571","submissionUrl":"https://submission.nature.com/new-submission/11571/3","title":"Cognitive Neurodynamics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"8b08d2b8-bc00-4ce6-9ebc-5bf93deb6485","owner":[],"postedDate":"February 11th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-14T02:54:05+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-11 18:21:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8744978","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8744978","identity":"rs-8744978","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}