Multiresolution Spectral Analysis of Epileptic EEG Signals Using Various Wavelet Types

Abstract

In this paper we approach a comparative multi-resolution analysis of EEG signals featuring epileptic activity. The Discrete Wavelet Transform is applied to several sets of EEG recordings, using various wavelet types. For EEG sequences with typical pathological patterns, the multi-level decompositions with various wavelet families, the corresponding spectra, and also the relative energy distribution ratios are calculated and shown. We show that the spectral energy distribution of EEG wave depends on the particular choice of wavelet used in analysis.