Modeling the Stationarity and Gaussianity of Spontaneous Electroencephalographic Activity

Abstract

Considerable motivation exists for the development of an adequate statistical model for spontaneous electroencephalographic (EEG) activity. At present, almost all methods of time-domain and frequency-domain EEG analysis are based on implicit assumptions regarding the statistical characteristics of the underlying random process, particularly with respect to the extent of stationarity and Gaussianity of the process. However, the actual characteristics of specific EEG ensembles have not been extensively investigated. In this paper, a technique is proposed for estimating the degree of wide-sense stationarity and the degree of Gaussianity of an ensemble of EEG records. Results which have been obtained by applying this technique to three relatively large ensembles of multichannel EEG data are also described. In addition, the comparative advantages of employing alternate methods of EEG analysis are discussed in relation to the estimated degree of stationarity and Gaussianity of the particular EEG ensembles under consideration.