[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82012-en":3,"doc-seo-82012-105":29,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},82012,7971461740886,"Theodore","https://ap-avatar.wpscdn.com/davatar_3d24733baf745e90a7e4bdd5f77d97b2",8,"Research & Report","Omni-Sleep: A Sleep Foundation Model via Hierarchical Contrastive Learning of CNS–ANS Dynamics","Omni-Sleep is introduced as a sleep foundation model that leverages the physiological CNS/ANS partition to guide topology-constrained representation learning from multimodal polysomnography signals (EEG, EOG, EMG, ECG, respiration). The model learns structured representations through intra-system consistency, inter-system synchronization aligning brain–body dynamics, and latent-space masked temporal modeling for long-horizon sleep dynamics. Pretraining on 100,000+ hours of multi-center multimodal PSG supports sleep staging and multi-disease classification, outperforming baselines in label efficiency, cross-dataset generalization, and missing-modality robustness.","arXiv :2607 .07720v2 [ cs .LG] 10 Jul 2026  \n2026-7-13  \nOmni-Sleep: A Sleep Foundation Model via Hierarchical Contrastive Learning of CNS–ANS Dynamics  \nZhoujie Hou* 1,3 , Song Wang* 1,3 , Kexin Lou* 1,3 , Mo Wang1,3 , Chen Wei1,3 , Quanying LiuB 1,2,3  \n1 Department of Biomedical Engineering, Southern University of Science and Technology, Shenzhen, China, 2 Shenzhen Loop Area Institute, Shenzhen, China, 3 Omni-Intelligence, Shenzhen, China  \nSleep physiology arises from the coordinated dynamics of the central nervous system (CNS) and autonomic nervous system (ANS), as reflected by multimodal polysomnography signals including EEG, EOG, EMG, ECG, and respiration. However, existing sleep foundation models often fuse heterogeneous biosignals in a topology-agnostic manner, overlooking their physiological organization. We introduce Omni-Sleep, a sleep foundation model that uses the CNS/ANS partition as a physiological prior for topology-constrained representation learning. Omni-Sleep learns structured representations through three objectives: intra-system consistency, which captures shared subsystem-level factors within neural and cardio-respiratory signals; inter-system synchronization, which aligns subsystem trajectories to model brain–body dynamics; and latent-space masked temporal modeling, which captures longhorizon sleep dynamics. Pre-trained on over 100,000 hours of multi-center multimodal PSG data, Omni-Sleep is evaluated on sleep staging and multi-disease classification. Across datasets and modalityablation settings, Omni-Sleep outperforms strong foundation-model baselines, showing improved label efficiency, cross-dataset generalization, and robustness to missing modalities. These results highlight the value of physiological hierarchy for generalizable sleep representation learning. Code is available at [https://github.com/AutoBrain-sleep/OmniSleep](https://github.com/AutoBrain-sleep/OmniSleep).  \nKeywords: Sleep; self-supervised learning; multimodal PSG; foundation model  \nCode: [https: // github. com/ AutoBrain-sleep/ OmniSleep](https: // github. com/ AutoBrain-sleep/ OmniSleep)  \n1. Introduction  \nSleep is a dynamic physiological process governed by coordinated interactions between central neural activity and autonomic cardio-respiratory regulation [3, 26] . In clinical practice, this brain–body coupling is captured by polysomnography (PSG), which records heterogeneous signals spanning EEG, EOG, EMG, ECG and respiratory channels. These signals are informative not only for sleep staging, but also for assessing multi-system health risks that manifest through altered sleep physiology [1, 8, 11, 18, 24] . Recent self-supervised foundation models for neural data, have shown that large-scale pretraining can learn transferable brain representations and reduce reliance on costly expert labels [5, 13, 27–31] . This paradigm has also been adopted for sleep analysis through largescale pretraining on polysomnographic recordings [9, 24] . However, existing multimodal methods often align heterogeneous physiological signals in a unified representation space, overlooking their physiological organization [17, 20, 23, 24] . This flat fusion neglects a key property of PSG: CNS-derived signals and ANS-related signals follow distinct physiological manifolds, yet exhibit stage-dependent and time-varying synchronization [6, 32] . In real-world deployments, these limitations are amplified under domain shifts and missing modalities. [5] .  \n* Equal contribution. B Correspondence to: [liuqy@sustech.edu.cn](liuqy@sustech.edu.cn).  \nClinical PSG with Multimodal Signals Real-world Deployment Challenges Downstream Tasks  \nFigure 1 | Omni-Sleep targets real-world PSG deployment.  \nTo bridge this gap, we propose Omni-Sleep, a sleep foundation model that uses the CNS/ANS partition as a physiological prior for topology-constrained sleep representation learning. Pretrained on over 100,000 hours of multi-center PSG data, Omni-Sleep combines hierarchical ","cbCaisiGroqwTogI","https://ap.wps.com/l/cbCaisiGroqwTogI","pdf",1612146,1,12,"English","en",105,"# Introduction\n# Related Work\n## CNS-ANS Coupling in Sleep\n## Multimodal PSG Representation Learning","[{\"question\":\"What physiological prior does Omni-Sleep use for learning sleep representations?\",\"answer\":\"It uses the CNS/ANS partition as a physiological prior, constraining representation learning to respect the physiological topology.\"},{\"question\":\"How does Omni-Sleep learn structured representations?\",\"answer\":\"It combines three objectives: intra-system consistency, inter-system synchronization, and latent-space masked temporal modeling for long-horizon dynamics.\"},{\"question\":\"What tasks and evaluation results are reported for Omni-Sleep?\",\"answer\":\"Omni-Sleep is evaluated on sleep staging and multi-disease classification, showing improved label efficiency, cross-dataset generalization, and robustness under missing-modality settings.\"}]",1784177567,30,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"omni-sleep-a-sleep-foundation-model-via-hierarchical-contrastive-learning-of-cnsans-dynamics","",{"@graph":35,"@context":84},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/omni-sleep-a-sleep-foundation-model-via-hierarchical-contrastive-learning-of-cnsans-dynamics/82012/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What physiological prior does Omni-Sleep use for learning sleep representations?","Question",{"text":74,"@type":75},"It uses the CNS/ANS partition as a physiological prior, constraining representation learning to respect the physiological topology.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does Omni-Sleep learn structured representations?",{"text":79,"@type":75},"It combines three objectives: intra-system consistency, inter-system synchronization, and latent-space masked temporal modeling for long-horizon dynamics.",{"name":81,"@type":72,"acceptedAnswer":82},"What tasks and evaluation results are reported for Omni-Sleep?",{"text":83,"@type":75},"Omni-Sleep is evaluated on sleep staging and multi-disease classification, showing improved label efficiency, cross-dataset generalization, and robustness under missing-modality settings.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,121,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":28,"slug":120},"research-report",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":124,"slug":125},9,"Religion & Spirituality",20,"religion-spirituality",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":124,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]