How Far Can Wearable-Compatible Signals Go? A Controlled Decomposition of Non-EEG Sleep Staging

Consumer wearables increasingly infer sleep stages from signals including heart rate, accelerometry, and photoplethysmography. However, existing studies often report end-to-end performance under a fixed signal setting, making it difficult to determine whether the observed performance comes from genuine physiological decoding, temporal priors, or dataset-specific confounds. The framework is evaluated across a signal-quality ladder spanning Apple Watch Sleep-Accel ($N=31$), the Sleep Heart Health Study ($N=195$, laboratory ECG, respiratory, and SpO$2$ signals), and Sleep-EDF-20 as an EEG+EOG reference, using the same compact Mamba2 model throughout.

• Form factor: wrist

Consumer wearables increasingly infer sleep stages from signals including heart rate, accelerometry, and photoplethysmography. However, existing studies often report end-to-end performance under a fixed signal setting, making it difficult to determine whether the observed performance comes from genuine physiological decoding, temporal priors, or dataset-specific confounds. To address this limitation, we introduce a four-layer controlled decomposition framework for non-EEG sleep staging, covering signal source, physiological representation, temporal prior, and decision layers. The framework is evaluated across a signal-quality ladder spanning Apple Watch Sleep-Accel (N=31), the Sleep Heart Health Study (N=195, laboratory ECG, respiratory, and SpO_2 signals), and Sleep-EDF-20 as an EEG+EOG reference, using the same compact Mamba2 model throughout. Laboratory cardiorespiratory signals reach =0.492, while EEG+EOG reaches =0.796, leaving a residual gap of =+0.304 that reflects missing cortical information rather than temporal modeling alone. Consumer HR/ACC reaches only =0.255, quantifying the additional penalty of derived wearable signals and real-world sensing constraints. Confidence-based abstention provides a calibrated operating mode: removing the 20% lowest-confidence epochs increases from 0.452 to 0.512, while a label-shuffled control collapses to =-0.003. These results support non-EEG sleep staging as coarse, confidence-aware sleep-structure monitoring rather than EEG-equivalent five-class clinical staging.

Verbatim from arXiv (opens in a new tab) · arXiv metadata (abstracts are released under CC0)

No results from this paper on ChatPPG leaderboards yet. Browse the leaderboards.