Correlation dimension changes of the EEG during the wakefulness-sleep cycle.

Abstract

The mathematical tools of chaos theory make the quantification of a time series such as the electroencephalogram possible by calculating its correlation dimensions. In this study, the electroencephalogram was recorded from the vertex in cats during the wakefulness-sleep cycle and its correlation dimension was determined. For the estimation of the correlation dimension a new method, the point-correlation dimension was used. This method is more accurate than others currently used for tracking non-stationarities within the data. The point-correlation dimension was higher during the alert state than in slow wave sleep. It is suggested that during the alert state a different type of pacemaker activity takes over in the generation of the electroencephalogram.