Apnea Burden-Guided Framework: Enhancing Out-of-Distribution Generalization in PPG-Based Sleep Apnea Characterization

Summary

Sleep apnea carries substantial cardiovascular and metabolic risk, but polysomnography is too complex and costly for long-term preventive monitoring at home. In wearable systems, blood oxygen saturation is usually the dominant predictor while the morphological content of the PPG pulse wave is left underexploited. This paper proposes an apnea-burden prediction framework and uses it to test how much PPG morphology adds, with a specific focus on out-of-distribution generalization rather than in-distribution accuracy alone.