Towards automatic sleep staging via Cross-Recurrence Rate of EEG and ECG activity

Abstract

This paper investigates the non-linear dynamic relationship between electroencephalogram (EEG) and electrocardiogram (ECG) signals during sleep. These relationships were studied with Cross-Recurrence Rate (CRR), a non-linear measure that studies the recurrence of the phase space trajectories of dynamical systems. Data from 10 subjects during sleep were obtained from the MIT-BIH Polysomnographic database and the CRR between ECG and EEG signals was estimated. The investigations revealed strong coupling relationships between ECG and EEG that varied according to the underlying sleep stage. From a physiological perspective, the findings indicate an increase in EEG and ECG during deep sleep, while also indicating the feasibility of potential CRR application for automatic sleep staging.