Analysis of RR Interval Entropies for Discrimination Between Wake and Sleep States

Abstract

This study use of Approximate Entropy (ApEn) for classification of sleep stages using R-R intervals derived from ECG signals. The results demonstrate that ApEn can effectively differentiate between sleep stages and wakefulness across varying embedding dimensions (m) and time windows. In lower dimensions, higher entropy values correspond to deeper sleep stages, while in higher dimensions, the relationship reverses, associating higher entropy with REM, light sleep, and wakefulness. The analysis also reveals that specific time windows (e.g., 5 minutes) and embedding dimensions (e.g., m = 5) improve the discrimination between sleep and wake states. These findings suggest that ApEn is a valuable tool for non-invasive sleep monitoring and classification. Further studies are recommended to explore its applicability in broader and more diverse populations.