Modelling sleep with Gaussian mixture model based on eye movements and delta-activity

The automatic sleep analysis is difficult task. The signals needed to perform the analysis are at least EEG and EOG. In this paper, the analysis is made mainly by using the EOG signal, and only the part of the EEG signal, the delta-band. The Gaussian mixture model (GMM) is calculated from the signals after detection of the eye movements and the processing of the EEG signal. The results show the eye movements and the delta-activity can be used to make discrimination between different sleep stages, although some additional signals, for example EMG, might improve the analysis.