TIME-FREQUENCY ANALYSIS OF ELECTROENCEPHALOGRAM SERIES .2. GABOR AND WAVELET TRANSFORMS

In this paper we compare two methods, based on the Gabor and wavelet t ransforms, to quantify and visualize the time evolution of frequency c ontents of electroencephalogram (EEG) time series. We found an optimal correlation between EEG visual inspection and the proposed methods in the characterization of the frequency and energy content of character istic activity during an epileptic seizure. The quasimonofrequency beh avior observed in the epileptic EEG series, in a previous work using a Gabor analysis [J. Inst. Electr. Eng. 93, 429 (1946)], is confirmed w ith the analysis using a wavelet. Moreover, the method based on the wa velet transform allows us to build a detector of epileptic events. Bot h methods are exemplified with EEG series obtained with depth electrod es in refractory epileptic patients.