[Detection of spikes in epileptic EEG based on Multiresolution Tsallis' entropy]

In this paper, the detection of the spikes in the epileptic EEG signal was studied based on Multiresolution Tsallis' entropy (MRET). First, EEG signals were decomposed into wavelet series, and then, at every scale, the abnormal spikes were distinguished from normal background EEG activities by using MRET. The analysis of 6 patients' EEG data showed that the abnormal epileptiform spikes in EEG can be accurately detected with this method, which opens up the perspectives of building up automatic detection devices for spikes in EEG. Compared with the Shannon entropy, the MRET provides one with more detailed information.