Semantic channel aware model improves polysomnography sleep analysis
SemPSG: A Semantic Channel-Aware Foundation Model for Polysomnography Analysis
Machine Learning
Summary
Polysomnography (PSG) records many body signals during sleep to understand how people sleep and detect sleep problems. Different hospitals use different setups, making it hard to use one computer model everywhere. The authors created SemPSG, a special model that understands what each signal means and adapts to different setups. This helps it learn better from diverse data and improves tasks like identifying sleep stages and predicting health issues from sleep data.
What this means in practice
- •For hospital data teams: Analyze sleep study data from different machines using one adaptable model that improves accuracy across varying channel setups.
- •For wearable device designers: Build sleep monitoring features that handle various sensor configurations by incorporating semantic understanding of signal channels.
Authors
Junyu Chen, Chenxi Liu, Shiqin Tang, Hao Miao, Wanyun Ling, Ziyue Li, Hongbin Liu, Gaofeng Meng
Abstract
Polysomnography (PSG) integrates multiple physiological signals to provide a comprehensive characterization of human sleep, yet its heterogeneous channel configurations across centers pose substantial challenges for transferable representation learning. Existing foundation models mainly focus on physiological modeling or temporal learning, while channel identity is often treated as a fixed structural index, overlooking the physiological semantics encoded by signal modality and reference configuration. To this end, we propose SemPSG, a Semantic channel-aware foundation model for heterogeneous PSG analysis. SemPSG explicitly represents the physiological semantics of channel identity and incorporates them into both signal representation learning and channel aggregation, enabling flexible modeling across diverse data configurations. Specifically, a semantic-conditioned time-series encoder captures signal-specific temporal dynamics and cross-signal interactions, while a multi-view image encoder extracts complementary time-frequency and morphological patterns from the same physiological recordings. We evaluate SemPSG on sleep and health-related tasks, including sleep staging, sleep-disorder breathing analysis, disease prediction, cognition and emotion recognition, and demographic estimation. Extensive experiments demonstrate consistent improvements over both general-purpose time series foundation models and PSG-specific foundation models, together with generalization across heterogeneous datasets across diverse channel configurations.