Interpretable Framework for Sleep Monitoring: Applying Statistical Control Charts to Physiological Data Streams

Area of Science:

• Physiological monitoring
• Statistical process control
• Sleep science

Background:

• Polysomnography (PSG) generates complex, non-linear physiological time-series data, posing interpretability challenges.
• Analyzing sleep-related physiological variability requires transparent and explainable methods.

Purpose of the Study:

• To explore the feasibility of applying statistical process control (SPC) charts to cardio-respiratory signals from PSG.
• To assess if SPC control charts can provide transparent and interpretable analysis of sleep-related physiological variability.

Main Methods:

• Cardio-respiratory signals from a public PSG dataset were preprocessed and analyzed using univariate control charts.
• Sleep stage annotations were used to contextualize physiological variability across wake and non-REM sleep.
• Control chart rule violations were quantitatively examined relative to sleep-state transitions.

Main Results:

• Control chart rule violations were more frequent during wakefulness and wake-non-REM sleep transitions.
• Variability flags from SPC control charts showed structural correspondence with annotated sleep stage dynamics.
• Control charts remained relatively stable during sustained non-REM sleep.

Conclusions:

• SPC control charts offer a transparent and interpretable framework for analyzing physiological variability in sleep data.
• This method supports future research in sleep-state analysis and explainable data-driven approaches.
• The study demonstrates feasibility, not diagnostic performance or accuracy.