A sleep stage estimation algorithm based on cardiorespiratory signals derived from a suprasternal pressure sensor

Luca Cerina

The automatic estimation of sleep structure is central in translating sleep research from clinical laboratories to people’s homes. Oftentimes, portable sensors (like smartwatches) compromise capabilities in favor of usability, resulting in surrogate measures of sleep stages and other physiological signals, such as respiration. We propose the use of cardiorespiratory signals obtained by a Suprasternal Pressure sensor as an alternative. The sensor is already used in sleep-disordered breathing research, and now we extended its usefulness by separating its respiratory and cardiac components automatically. The signals are then fed to a neural network to estimate sleep stages, with promising results.