Automatic Recognition and Classification of Patient–Ventilator Asynchrony Using Deep Learning with Generative Adversarial Networks

Automatic Recognition and Classification of Patient–Ventilator Asynchrony Using Deep Learning with Generative Adversarial Networks | 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 Automatic Recognition and Classification of Patient–Ventilator Asynchrony Using Deep Learning with Generative Adversarial Networks Shuai Ren, Rongheng Zhao, Xiaohan Wang, Xiaoqian Shi, Tao Wang, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8145560/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 Purpose: Patient–ventilator asynchrony (PVA) is frequent in mechanically ventilated patients and is associated with poor outcomes. Automated recognition has the potential to improve clinical care, but existing methods are limited by data scarcity and class imbalance. We aimed to develop and validate a deep learning model enhanced by synthetic data generation to improve the accuracy of PVA detection and classification. Methods: We developed a two-step framework: (1) a self-attention CNN–LSTM generative adversarial network (SACL-GAN) to generate realistic respiratory cycles, and (2) a hybrid multi-head attention CNN–BiLSTM classifier (MHACBL) model trained on both real and synthetic data. 4,907 respiratory cycles from 32 ICU patients were annotated into normal breathing (NB) and four asynchrony categories: ineffective effort (IE), early cycling (EC), double triggering (DT), and airway oscillation (AO). Model performance was assessed using accuracy, precision, recall, and F1-score and other metrics. Results: SACL-GAN successfully generated high-quality synthetic data with low mean squared error (MSE: 0.58-2.84 for pressure, 1.68-8.07 for flow) and dynamic time warping (DTW: 3.03-6.47 for pressure, 9.61-12.24 for flow). Without augmentation, the MHA-CBL classifier achieved high accuracy for NB and IE, but lower recall for minority classes (EC and AO). GAN augmentation improved overall accuracy (from 94.6% to 97.8%) and recall for minority classes (EC: 76.9% → 93.2%; AO: 74.6% → 91.8%). In binary classification (asynchrony vs NB), sensitivity improved from 94.4% to 98.0%, reducing missed detections. Conclusion: Combining GAN-based augmentation with hybrid deep learning improved recognition of PVA, particularly for underrepresented categories. This strategy addresses dataset imbalance and enhances the clinical applicability of automated monitoring in ICU settings. Patient-ventilator asynchrony Mechanical ventilation Generative adversarial networks Deep learning Full Text Supplementary Files strobechecklist.pdf supplymentalmaterials.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-8145560","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":548285777,"identity":"2b6abd81-231d-4078-92a5-965e28fb3293","order_by":0,"name":"Shuai Ren","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0ElEQVRIiWNgGAWjYJACZiDm4ZdgbICyidUiOYOxsYEkLQwGNxgYidNi3t57+HNhm52M8e3m9gcMFdaJDexnD+DVInPmXJr0zLZkHrM7B4EOO5Oe2MCTl4BXi4REjhkzbxszj9mNxMYGxrbDiQ0SPAb4tci/Mf7M21bPYzwDpOUfMVqACqR52w7zGEiAtDQQo4Unx0ya59xxHgmgX2YkHEs3buPJIaCF/YzxZ56yanv+2e0PPnyosZbtZz+DXwsqSABiNhLUj4JRMApGwSjAAQC/FT9Z4hGlRwAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0003-4407-0265","institution":"Beijing Institute of Technology","correspondingAuthor":true,"prefix":"","firstName":"Shuai","middleName":"","lastName":"Ren","suffix":""},{"id":548285778,"identity":"c5a6536a-91ca-478a-ba1e-1424438ec0d9","order_by":1,"name":"Rongheng Zhao","email":"","orcid":"","institution":"Beijing Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Rongheng","middleName":"","lastName":"Zhao","suffix":""},{"id":548285779,"identity":"eaac4b16-d6d6-48da-a46e-461cd8994f4e","order_by":2,"name":"Xiaohan Wang","email":"","orcid":"","institution":"Beijing Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Xiaohan","middleName":"","lastName":"Wang","suffix":""},{"id":548285780,"identity":"15c40093-2d63-43c3-9c76-357d767ad2d5","order_by":3,"name":"Xiaoqian Shi","email":"","orcid":"","institution":"Beijing Chao-Yang Hospital Capital Medical University: Beijing Chaoyang Hospital","correspondingAuthor":false,"prefix":"","firstName":"Xiaoqian","middleName":"","lastName":"Shi","suffix":""},{"id":548285781,"identity":"79cb0b58-c9ff-46d3-8ad4-e5a7d86d8557","order_by":4,"name":"Tao Wang","email":"","orcid":"","institution":"Beijing Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Tao","middleName":"","lastName":"Wang","suffix":""},{"id":548285782,"identity":"0357d82b-7016-4709-86c2-1284d0b99c02","order_by":5,"name":"Haichao Li","email":"","orcid":"","institution":"Beijing Chao-Yang Hospital Capital Medical University: Beijing Chaoyang Hospital","correspondingAuthor":false,"prefix":"","firstName":"Haichao","middleName":"","lastName":"Li","suffix":""},{"id":548285783,"identity":"60f39b98-cc89-4f83-8e9c-dceb3fb9c4e7","order_by":6,"name":"Zujin Luo","email":"","orcid":"","institution":"Beijing Chao-Yang Hospital Capital Medical University: Beijing Chaoyang Hospital","correspondingAuthor":false,"prefix":"","firstName":"Zujin","middleName":"","lastName":"Luo","suffix":""}],"badges":[],"createdAt":"2025-11-18 12:43:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8145560/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8145560/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":96926833,"identity":"7265035a-e644-4755-ab96-e8b0d83f79e2","added_by":"auto","created_at":"2025-11-27 14:26:05","extension":"xml","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11680,"visible":true,"origin":"","legend":"","description":"","filename":"icmeICMED2502162.xml","url":"https://assets-eu.researchsquare.com/files/rs-8145560/v1/e6b28369d883adaf5a61115d.xml"},{"id":96926860,"identity":"9842b2b7-cae1-47c6-9b20-92db0f88c8a4","added_by":"auto","created_at":"2025-11-27 14:26:12","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1019,"visible":true,"origin":"","legend":"","description":"","filename":"ICMED250216230207.go.xml","url":"https://assets-eu.researchsquare.com/files/rs-8145560/v1/a565b13ed9b0a4b6f083bce9.xml"},{"id":96926841,"identity":"97683d2e-c675-48ea-8bda-075028be2319","added_by":"auto","created_at":"2025-11-27 14:26:08","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":821,"visible":true,"origin":"","legend":"","description":"","filename":"ICMED2502162Import.xml","url":"https://assets-eu.researchsquare.com/files/rs-8145560/v1/eb311ddfd48f5fe19d97d8cf.xml"},{"id":97900645,"identity":"4a4c676c-ebf7-4399-b157-66b3f7ef14a3","added_by":"auto","created_at":"2025-12-10 15:45:42","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2219843,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript1118.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8145560/v1_covered_4e2fb8d2-7671-4542-9b2c-8729dc65449d.pdf"},{"id":96926825,"identity":"2f35a388-8aa1-4810-aa22-7ca66eaf92c8","added_by":"auto","created_at":"2025-11-27 14:26:02","extension":"pdf","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":156818,"visible":true,"origin":"","legend":"","description":"","filename":"strobechecklist.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8145560/v1/5b4e095d2aab456dbff1b827.pdf"},{"id":96926830,"identity":"6de2b30f-cb16-4885-8837-4ee496883a21","added_by":"auto","created_at":"2025-11-27 14:26:04","extension":"pdf","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":735481,"visible":true,"origin":"","legend":"","description":"","filename":"supplymentalmaterials.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8145560/v1/b4ab0ba5ac6d5457a1b9871c.pdf"}],"financialInterests":"","formattedTitle":"Automatic Recognition and Classification of Patient–Ventilator Asynchrony Using Deep Learning with Generative Adversarial Networks","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":"Patient-ventilator asynchrony, Mechanical ventilation, Generative adversarial networks, Deep learning","lastPublishedDoi":"10.21203/rs.3.rs-8145560/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8145560/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003ePurpose: \u0026nbsp;\u003c/strong\u003ePatient–ventilator asynchrony (PVA) is frequent in mechanically ventilated patients and is associated with poor outcomes. Automated recognition has the potential to improve clinical care, but existing methods are limited by data scarcity and class imbalance. We aimed to develop and validate a deep learning model enhanced by synthetic data generation to improve the accuracy of PVA detection and classification.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e We developed a two-step framework: (1) a self-attention CNN–LSTM generative adversarial network (SACL-GAN) to generate realistic respiratory cycles, and (2) a hybrid multi-head attention CNN–BiLSTM classifier (MHACBL) model trained on both real and synthetic data. 4,907 respiratory cycles from 32 ICU patients were annotated into normal breathing (NB) and four asynchrony categories: ineffective effort (IE), early cycling (EC), double triggering (DT), and airway oscillation (AO). Model performance was assessed using accuracy, precision, recall, and F1-score and other metrics.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eSACL-GAN successfully generated high-quality synthetic data with low mean squared error (MSE: 0.58-2.84 for pressure, 1.68-8.07 for flow) and dynamic time warping (DTW: 3.03-6.47 for pressure, 9.61-12.24 for flow). Without augmentation, the MHA-CBL classifier achieved high accuracy for NB and IE, but lower recall for minority classes (EC and AO). GAN augmentation improved overall accuracy (from 94.6% to 97.8%) and recall for minority classes (EC: 76.9% → 93.2%; AO: 74.6% → 91.8%). In binary classification (asynchrony vs NB), sensitivity improved from 94.4% to 98.0%, reducing missed detections.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion: \u003c/strong\u003eCombining GAN-based augmentation with hybrid deep learning improved recognition of PVA, particularly for underrepresented categories. This strategy addresses dataset imbalance and enhances the clinical applicability of automated monitoring in ICU settings.\u003c/p\u003e","manuscriptTitle":"Automatic Recognition and Classification of Patient–Ventilator Asynchrony Using Deep Learning with Generative Adversarial Networks","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-27 14:23:00","doi":"10.21203/rs.3.rs-8145560/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":"427207db-2068-48f5-af73-7498ba4d3046","owner":[],"postedDate":"November 27th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-12-10T09:21:22+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-27 14:23:00","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8145560","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8145560","identity":"rs-8145560","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}