Time-frequency analysis of epileptic EEG patterns by means of empirical modes and wavelets

Abstract

In this paper we perform a time-frequency analysis of epileptic EEG patterns based on two approaches for characterizing nonstationary multi-frequency signals, namely, the continuous wavelet transform (CWT) and the empirical mode decomposition (EMD). Possibilities and limitations of both these techniques are considered, and a combined approach for automatic pattern detection is proposed.