WiseMind: Recontextualizing AI with a Knowledge-Guided, Theory-Informed Multi-Agent Framework for Instrumental and Humanistic Benefits

WiseMind: Recontextualizing AI with a Knowledge-Guided, Theory-Informed Multi-Agent Framework for Instrumental and Humanistic Benefits | 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 WiseMind: Recontextualizing AI with a Knowledge-Guided, Theory-Informed Multi-Agent Framework for Instrumental and Humanistic Benefits Yuqi Wu, Guangya Wan, Jingjing Li, Shengming Zhao, Lingfeng Ma, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7689405/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 25 Mar, 2026 Read the published version in npj Digital Medicine → Version 1 posted 9 You are reading this latest preprint version Abstract Purpose: Translating state-of-the-art NLP into real-world psychiatric diagnosis is challenging due to a lack of domain-specific contextualization across knowledge, processes, and evaluation. We introduce WiseMind, an interdisciplinary framework designed to improve both diagnostic precision and humanistic care. WiseMind combines: (i) structured proactive reasoning using a DSM-5 knowledge graph; (ii) a theory-driven dual-agent system inspired by Dialectical Behavior Therapy, coordinating reasoning and empathy agents; and (iii) a comprehensive evaluation pipeline including simulated patients, user studies, expert reviews, and ethical audits. Applied to depression, anxiety, and bipolar disorder, WiseMind achieves up to 84.2% diagnostic accuracy—on par with clinicians—while significantly improving perceived empathy and trustworthiness compared to single-agent baselines. Methods: WiseMind integrates a DSM-guided knowledge graph, a Dialectical Behavior Therapy–inspired dual-agent design, and a multi-perspective evaluation involving users, clinicians, and ethical reviewers. Results: The system achieves 84.2\% diagnostic accuracy and significantly improves trust and empathy ratings over baselines. Conclusion: Contextualizing NLP systems across clinical knowledge, reasoning processes, and evaluation methods is critical for bridging the gap between benchmarks and meaningful clinical deployment. While promising, further work is needed to extend WiseMind across the full spectrum of psychiatric conditions and diverse patient populations. Scientific community and society/Business and industry Health sciences/Health care Physical sciences/Mathematics and computing Biological sciences/Psychology Social science/Psychology Large Language Models Psychiatry Differential Diagnosis Conversational Diagnosis Full Text Additional Declarations No competing interests reported. Supplementary Files supplementary.pdf Cite Share Download PDF Status: Published Journal Publication published 25 Mar, 2026 Read the published version in npj Digital Medicine → Version 1 posted Editorial decision: Revision requested 10 Nov, 2025 Reviews received at journal 06 Nov, 2025 Reviews received at journal 27 Oct, 2025 Reviewers agreed at journal 25 Oct, 2025 Reviewers agreed at journal 23 Oct, 2025 Reviewers invited by journal 23 Oct, 2025 Editor assigned by journal 12 Oct, 2025 Submission checks completed at journal 11 Oct, 2025 First submitted to journal 23 Sep, 2025 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-7689405","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":538598476,"identity":"5cd7dd85-2794-4905-803c-7e0ebeb62d02","order_by":0,"name":"Yuqi Wu","email":"","orcid":"","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Yuqi","middleName":"","lastName":"Wu","suffix":""},{"id":538598478,"identity":"63c5f787-f89e-4d30-b214-ecd7612bf72c","order_by":1,"name":"Guangya Wan","email":"","orcid":"","institution":"University of Virginia","correspondingAuthor":false,"prefix":"","firstName":"Guangya","middleName":"","lastName":"Wan","suffix":""},{"id":538598479,"identity":"bde94da4-c631-4fbe-834c-d57c2ad262f2","order_by":2,"name":"Jingjing Li","email":"","orcid":"","institution":"University of Virginia","correspondingAuthor":false,"prefix":"","firstName":"Jingjing","middleName":"","lastName":"Li","suffix":""},{"id":538598483,"identity":"61168232-0b7e-4543-8a3f-3b4792539de8","order_by":3,"name":"Shengming Zhao","email":"","orcid":"","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Shengming","middleName":"","lastName":"Zhao","suffix":""},{"id":538598484,"identity":"586ea9ee-2a08-4713-a5fa-68455844c8a6","order_by":4,"name":"Lingfeng Ma","email":"","orcid":"","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Lingfeng","middleName":"","lastName":"Ma","suffix":""},{"id":538598486,"identity":"b7a071de-c06c-4677-8b11-116472c47ec0","order_by":5,"name":"Tianyi Ye","email":"","orcid":"","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Tianyi","middleName":"","lastName":"Ye","suffix":""},{"id":538598488,"identity":"9350f7e3-9eed-4ec9-a4b1-05bc0d6d3046","order_by":6,"name":"Mike Zhang","email":"","orcid":"","institution":"University of Alberta","correspondingAuthor":false,"prefix":"","firstName":"Mike","middleName":"","lastName":"Zhang","suffix":""},{"id":538598489,"identity":"9396f17f-f62f-4af2-9242-10c2f7ee5e6c","order_by":7,"name":"Ion Pop","email":"","orcid":"","institution":"University of Alberta","correspondingAuthor":false,"prefix":"","firstName":"Ion","middleName":"","lastName":"Pop","suffix":""},{"id":538598492,"identity":"8ecfab8a-2f7c-4172-9fcd-80d90f19042e","order_by":8,"name":"Yanbo Zhang","email":"","orcid":"","institution":"University of Alberta","correspondingAuthor":false,"prefix":"","firstName":"Yanbo","middleName":"","lastName":"Zhang","suffix":""},{"id":538598494,"identity":"b0b8bfe8-748d-4f20-b28b-cf2456edd5da","order_by":9,"name":"Jie Chen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyElEQVRIiWNgGAWjYBACxgbm4x8+MDDIGJCghS2NcQYDA48BiE2kJh41Zh6StDC3n2F7bNtmx2PO3vv8wQ8GO3nCDuvJPW6c25bMY9lz3LCxhyHZkKBVjDP4EqRz25h5DG6kMTbwMBwg7DrGGTwG0pZt9TwG958xNv5hOGBPjBYzaca2w0Bb2BibgbYkEtbSk5Zs2HPuONAvaYyzZQySkwlqMWw/fPDBj7JqOXP2Ywwf31TY2RLWgqqCmDRAOBpGwSgYBaNgFAAAIVE4u0M6fdsAAAAASUVORK5CYII=","orcid":"","institution":"Fudan University","correspondingAuthor":true,"prefix":"","firstName":"Jie","middleName":"","lastName":"Chen","suffix":""}],"badges":[],"createdAt":"2025-09-23 04:23:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7689405/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7689405/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41746-026-02559-9","type":"published","date":"2026-03-25T16:12:10+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":94992568,"identity":"48bfcf61-c16e-48e4-922f-a1622838f9bb","added_by":"auto","created_at":"2025-11-03 07:21:46","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10561,"visible":true,"origin":"","legend":"","description":"","filename":"0df7cc8acfcd4f1b9ad3ae12117c8627.json","url":"https://assets-eu.researchsquare.com/files/rs-7689405/v1/6431d2a836bfd0f3477f7363.json"},{"id":105756060,"identity":"382264bc-47e0-4f41-bc62-e6c8eb6f1a58","added_by":"auto","created_at":"2026-03-30 16:34:58","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1080927,"visible":true,"origin":"","legend":"","description":"","filename":"fzksbwghdjxqvmjmnxrvbyxmqpskshjz.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7689405/v1_covered_8de72a44-e080-49e7-ba51-f83baacfad98.pdf"},{"id":94992581,"identity":"c7348738-5cfc-4cd9-b766-daae13bb386e","added_by":"auto","created_at":"2025-11-03 07:21:48","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":3210900,"visible":true,"origin":"","legend":"","description":"","filename":"supplementary.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7689405/v1/77d37517a64bc7e154a07775.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eWiseMind: Recontextualizing AI with a Knowledge-Guided, Theory-Informed Multi-Agent Framework for Instrumental and Humanistic Benefits\u003c/p\u003e","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":"npj-digital-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjdigitalmed","sideBox":"Learn more about [npj Digital Medicine](http://www.nature.com/npjdigitalmed/)","snPcode":"41746","submissionUrl":"https://submission.springernature.com/new-submission/41746/3","title":"npj Digital Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Large Language Models, Psychiatry, Differential Diagnosis, Conversational Diagnosis","lastPublishedDoi":"10.21203/rs.3.rs-7689405/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7689405/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003ePurpose: \u003c/strong\u003eTranslating state-of-the-art NLP into real-world psychiatric diagnosis is challenging due to a lack of domain-specific contextualization across knowledge, processes, and evaluation. We introduce WiseMind, an interdisciplinary framework designed to improve both diagnostic precision and humanistic care. WiseMind combines: (i) structured proactive reasoning using a DSM-5 knowledge graph; (ii) a theory-driven dual-agent system inspired by Dialectical Behavior Therapy, coordinating reasoning and empathy agents; and (iii) a comprehensive evaluation pipeline including simulated patients, user studies, expert reviews, and ethical audits. Applied to depression, anxiety, and bipolar disorder, WiseMind achieves up to 84.2% diagnostic accuracy—on par with clinicians—while significantly improving perceived empathy and trustworthiness compared to single-agent baselines.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e WiseMind integrates a DSM-guided knowledge graph, a Dialectical Behavior Therapy–inspired dual-agent design, and a multi-perspective evaluation involving users, clinicians, and ethical reviewers.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e The system achieves 84.2\\% diagnostic accuracy and significantly improves trust and empathy ratings over baselines.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003e Contextualizing NLP systems across clinical knowledge, reasoning processes, and evaluation methods is critical for bridging the gap between benchmarks and meaningful clinical deployment. While promising, further work is needed to extend WiseMind across the full spectrum of psychiatric conditions and diverse patient populations.\u003c/p\u003e","manuscriptTitle":"WiseMind: Recontextualizing AI with a Knowledge-Guided, Theory-Informed Multi-Agent Framework for Instrumental and Humanistic Benefits","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-03 07:12:57","doi":"10.21203/rs.3.rs-7689405/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-11-10T14:27:35+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-06T17:47:02+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-27T17:09:56+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"254632142403887089553801223694402333141","date":"2025-10-25T07:59:23+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"263285346360971303770884095536627805533","date":"2025-10-23T05:51:41+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-23T04:41:25+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-10-13T00:44:16+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-11T14:37:33+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Digital Medicine","date":"2025-09-23T04:08:07+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-digital-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjdigitalmed","sideBox":"Learn more about [npj Digital Medicine](http://www.nature.com/npjdigitalmed/)","snPcode":"41746","submissionUrl":"https://submission.springernature.com/new-submission/41746/3","title":"npj Digital Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"b4672574-4a15-4a88-b1be-2540f8e8dfef","owner":[],"postedDate":"November 3rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":57314058,"name":"Scientific community and society/Business and industry"},{"id":57314059,"name":"Health sciences/Health care"},{"id":57314060,"name":"Physical sciences/Mathematics and computing"},{"id":57314061,"name":"Biological sciences/Psychology"},{"id":57314062,"name":"Social science/Psychology"}],"tags":[],"updatedAt":"2026-03-30T16:33:22+00:00","versionOfRecord":{"articleIdentity":"rs-7689405","link":"https://doi.org/10.1038/s41746-026-02559-9","journal":{"identity":"npj-digital-medicine","isVorOnly":false,"title":"npj Digital Medicine"},"publishedOn":"2026-03-25 16:12:10","publishedOnDateReadable":"March 25th, 2026"},"versionCreatedAt":"2025-11-03 07:12:57","video":"","vorDoi":"10.1038/s41746-026-02559-9","vorDoiUrl":"https://doi.org/10.1038/s41746-026-02559-9","workflowStages":[]},"version":"v1","identity":"rs-7689405","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7689405","identity":"rs-7689405","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}