Few-Shot Deployment of Pretrained MRI Transformers in Brain Imaging Tasks

Few-Shot Deployment of Pretrained MRI Transformers in Brain Imaging 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 Article Few-Shot Deployment of Pretrained MRI Transformers in Brain Imaging Tasks Mengyu Li, Guoyao Shen, Chad W. Farris, Xin Zhang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7483576/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 Machine learning using transformers has shown great potential in medical imaging, but its real-world applicability remains limited due to the scarcity of annotated data. In this study, we propose a practical framework for the few-shot deployment of pretrained MRI transformers in diverse brain imaging tasks. By utilizing the Masked Autoencoder (MAE) pretraining strategy on a large-scale, multi-cohort brain MRI dataset comprising over 31 million slices, we obtain highly transferable latent representations that generalize well across tasks and datasets. For high-level tasks such as classification, a frozen MAE encoder combined with a lightweight linear head achieves state-of-the-art accuracy in MRI sequence identification with minimal supervision. For low-level tasks such as segmentation, we propose MAE-FUnet, a hybrid architecture that fuses multiscale CNN features with pretrained MAE embeddings. This model consistently outperforms other strong baselines in both skull stripping and multi-class anatomical segmentation under data-limited conditions. With extensive quantitative and qualitative evaluations, our framework demonstrates efficiency, stability, and scalability, suggesting its suitability for low-resource clinical environments and broader neuroimaging applications. Biological sciences/Computational biology and bioinformatics Physical sciences/Engineering Health sciences/Health care Physical sciences/Mathematics and computing Health sciences/Medical research Biological sciences/Neuroscience Full Text Additional Declarations No competing interests reported. Supplementary Files ScientificreportnewsubmissionSupplementaryInformation.pdf 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-7483576","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":513251947,"identity":"38b34ac5-da6a-4ad0-8e19-9e3152f9af00","order_by":0,"name":"Mengyu Li","email":"","orcid":"","institution":"Boston University","correspondingAuthor":false,"prefix":"","firstName":"Mengyu","middleName":"","lastName":"Li","suffix":""},{"id":513251948,"identity":"ae5b4fbb-7e6b-4690-9d45-f3c38c76644f","order_by":1,"name":"Guoyao Shen","email":"","orcid":"","institution":"Boston University","correspondingAuthor":false,"prefix":"","firstName":"Guoyao","middleName":"","lastName":"Shen","suffix":""},{"id":513251952,"identity":"5b5412b3-14b6-425d-94d4-40409d189061","order_by":2,"name":"Chad W. Farris","email":"","orcid":"","institution":"Boston University Chobanian \u0026 Avedisian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Chad","middleName":"W.","lastName":"Farris","suffix":""},{"id":513251954,"identity":"06d34202-4bd0-4423-b73c-d6438b6b43a8","order_by":3,"name":"Xin Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAkUlEQVRIiWNgGAWjYDACZiD+2ABiJZCghXEmaVpAunhJ0mJwnPfwZ9sdhxn42XMMiNMi2cyXYJx75jCDZM8bIrXwM/MYJOe2HWYwuEGsLWxALYctgVrsidYCtMWwmRFkiwTxfuExZuw9k84jceZZAXFaDM6fMf7wc4e1HH978gbitMAAD2nKR8EoGAWjYBTgBwBkZiU3eJhGpQAAAABJRU5ErkJggg==","orcid":"","institution":"Boston University","correspondingAuthor":true,"prefix":"","firstName":"Xin","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2025-08-28 23:08:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7483576/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7483576/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97264709,"identity":"da6385da-3f90-469f-bc28-5605f6cb561c","added_by":"auto","created_at":"2025-12-02 14:24:31","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1530101,"visible":true,"origin":"","legend":"","description":"","filename":"CompleteManuscriptwithTitlePageScientificreport.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7483576/v1_covered_12d674b9-a741-43da-bc11-21fe01d7be2d.pdf"},{"id":91126803,"identity":"3b023533-cc31-403a-9005-36fc96649db8","added_by":"auto","created_at":"2025-09-11 21:57:01","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":242196,"visible":true,"origin":"","legend":"","description":"","filename":"ScientificreportnewsubmissionSupplementaryInformation.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7483576/v1/3c7bb074d6cfdb63dc7f0950.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Few-Shot Deployment of Pretrained MRI Transformers in Brain Imaging Tasks","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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-7483576/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7483576/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMachine learning using transformers has shown great potential in medical imaging, but its real-world applicability remains limited due to the scarcity of annotated data. In this study, we propose a practical framework for the few-shot deployment of pretrained MRI transformers in diverse brain imaging tasks. By utilizing the Masked Autoencoder (MAE) pretraining strategy on a large-scale, multi-cohort brain MRI dataset comprising over 31\u0026nbsp;million slices, we obtain highly transferable latent representations that generalize well across tasks and datasets. For high-level tasks such as classification, a frozen MAE encoder combined with a lightweight linear head achieves state-of-the-art accuracy in MRI sequence identification with minimal supervision. For low-level tasks such as segmentation, we propose MAE-FUnet, a hybrid architecture that fuses multiscale CNN features with pretrained MAE embeddings. This model consistently outperforms other strong baselines in both skull stripping and multi-class anatomical segmentation under data-limited conditions. With extensive quantitative and qualitative evaluations, our framework demonstrates efficiency, stability, and scalability, suggesting its suitability for low-resource clinical environments and broader neuroimaging applications.\u003c/p\u003e","manuscriptTitle":"Few-Shot Deployment of Pretrained MRI Transformers in Brain Imaging Tasks","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-11 21:56:56","doi":"10.21203/rs.3.rs-7483576/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":"c62f2961-4cb3-4ed5-8a31-8364fef94180","owner":[],"postedDate":"September 11th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":54526645,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":54526646,"name":"Physical sciences/Engineering"},{"id":54526647,"name":"Health sciences/Health care"},{"id":54526648,"name":"Physical sciences/Mathematics and computing"},{"id":54526649,"name":"Health sciences/Medical research"},{"id":54526650,"name":"Biological sciences/Neuroscience"}],"tags":[],"updatedAt":"2025-12-02T14:23:32+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-11 21:56:56","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7483576","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7483576","identity":"rs-7483576","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}