Classifying Detection of Epileptic EEG Based on Approximate Entropy in Wavelet Domain

In the analysis of epileptic EEG data, the typical presence of epileptic activity includes spike wave, sharp wave, spike-and-slow complex wave and sharp-and-slow complex wave. Each of these epileptic EEG has different time-frequency characteristics. If they are detected by identical detection rule, it is impossible to obtain optimal detection result. In this paper, we present a classifying detection method to automatically detect different kinds of epileptic EEG data using the discrete wavelet transform (DWT) combined with approximate entropy (ApEn). Spike wave, spike-and-slow complex wave and sharp-and-slow complex wave are detected by this method and the optimal detection rules are achieved. And it assures a higher detection rate with a lower false detection rate.