Sleep Staging from the EEG Signal Using Multifractal Detrended Fluctuation Analysis

The quality of sleep is closely related to human health. Accurate monitoring of sleep quality can play an effective supervisory role in helping people improve the quality of sleep. The analysis of the electroencephalogram (EEG) can yield much useful information about sleep stage. In the present study, the MF-DFA algorithm was applied to stage the different sleep states. The two key parameters in MF-DFA algorithm the segmented length s and the order q of fluctuation function were determined by the sleep EEG data in MIT-BIH polysomnography database, and verified by the experiment. The results demonstrated that the h(q) with q=0,1,2 and s=10~100 can distinguish the wake, shallow sleep and deep sleep states in MIT-BIH database accurately, and can reflect the process of sleep state better.