Fourier and wavelet analyses of normal and epileptic electroencephalogram (EEG)

Human normal and epileptic electroencephalogram (EEG) signals have been analyzed using Fourier transform (FT) and continuous wavelet transform (CWT) as well as discrete wavelet transform (DWT). The power spectrum densities of both normal and epileptic EEG showed 1/ f^a fluctuations as a function of frequency f. The alpha rhythms that appeared in the normal EEG were not observed in the epileptic EEG. FT could not reveal the localized spikes and complexes that are typical among epileptic seizures. On the contrary, both CWT and DWT analyses could clearly detect localized low-frequency epileptic components in the EEG records. Our final goal of the study is the automatic detection of the epileptic disorders in the EEG in order to support the diagnosis and care of the epileptic syndromes and related seizure disorders.