Omni-Sleep: A Sleep Foundation Model via Hierarchical Contrastive Learning of CNS--ANS Dynamic
arXiv:2607.07720v1 Announce Type: new Abstract: Sleep 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 learn
Advances in AI, particularly foundation models and contrastive learning, are increasingly being applied to complex biological data, enabling more sophisticated physiological analyses.
This development indicates meaningful progress in applying AI to complex biological systems, moving towards more physiologically grounded AI models that could revolutionize diagnostics and treatment methods.
The approach to building sleep foundation models shifts from topology-agnostic data fusion to physiologically-informed, architecture-constrained learning, potentially leading to more accurate and interpretable models.
- · AI researchers in biomedicine
- · Sleep diagnostic companies
- · Pharmaceutical companies developing sleep aids
- · Healthcare providers
- · Companies relying on older, less sophisticated sleep analysis methods
Improved accuracy and insights from polysomnography data, leading to better sleep disorder diagnosis.
Development of personalized sleep interventions and therapies based on granular physiological insights.
Extension of this hierarchical contrastive learning approach to other complex biological systems beyond sleep, accelerating AI in precision medicine.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG