Computerized method for scoring of polygraphic sleep recordings.

Abstract

A new method for automatic analysis of polygraphic sleep recordings is described. The analysis is to be carried out on a small laboratory computer in two phases. The first phase is a real-time signal processing during which the EEG signal is being divided into quasi-stationary segments of uneven length and homogeneous in their spectral content, ands the EOG and EMG signal analysed. The second phase is a very short off-line operation, comprised of sequential clustering of the segments, using fuzzy subset theory, and optimal fuzzy partition. The output available is the hypnogram, describing the sequence of sleep stages during the night, graphs describing the relative power changes of the main EEG activities (delta, theta, alpha and sigma), and 3-dimensional plots of the spectral changes during the night. In addition, a few quantitative measures related to the relative amount of the various sleep stages and to the number of rapid eye movements are calculated.