Variability measures in neural systems: bases for an automatic sleep staging independent of spectral analysis

Abstract

In this work we introduce some nonlinear neural dynamics techniques and analyze the performance of two variability measures related to the system uncertainty and predictability in the context of rats sleep study. We propose the bases of a novel automatic sleep staging procedure that is independent of the spectral analysis and can be extended to human sleep EEG data with no need of simultaneous recordings as fMRI, EOG, EMG or polysomnograms.