Reexamination of characteristic of spectral exponent of epileptic EEGs corresponding to levels in wavelet-based fractal analysis

Abstract

Recently, the characteristics of epileptic EEG data have been examined by using the wavelet-based fractal analysis. It was observed that epileptic EEG data may be associated with multiple spectral exponents. In this study, the characteristic of spectral exponents of epileptic EEG data determined from various intervals of levels (or ranges of spectral subbands) is reexamined by applying the wavelet-based fractal analysis to EEG data of epilepsy patients. From the computational results, it is confirmed that the spectral exponent of epileptic EEG data varies corresponding to an interval of levels from which it is determined. Also, it is shown that the spectral exponents of epileptic EEG data obtained during non-seizure period and seizure activity determined from high frequency (10.85–21.70Hz) subband and low frequency (1.36–10.85Hz and 2.71–21.70Hz) subbands are substantially different.