Abstract

Info

Title:: Detection of epileptic spikes in EEG based on wavelet packet transform

Author(s):: ZHU Ningning1; LI Hao1; DENG Xiaoqiao1; YU Ming2; LI Xiaolong1; 1. School of Electronics and Information, Jiangsu University of Science and Technology, Zhenjiang 212000, China 2. Department of Neurology, Affiliated Hospital of Jiangsu University, Zhenjiang 212000, China

Keywords:: Keywords: epileptic spike detection wavelet packet transform signal reconstruction missed detection rate false detection rate

Abstract:: Abstract: An algorithm combining the physical characteristics (amplitude, frequency) of spikes and wavelet packet transform is proposed for better the spike detection in the electroencephalogram (EEG) signals of epilepsy patients. Firstly, wavelet packet transform is used to decompose the epilepsy EEG signals by wavelet packet decomposition the EEG frequency (0-30 Hz) into 3 layers. Secondly, the third layer node frequencies of S(3, 0) (0-10.85 Hz), S(3,1) (10.85-21.70 Hz), S(3,2) (21.70-32.55 Hz) are reconstructed according to the frequency range of EEG signals. Finally, the spike amplitude is taken as the detection threshold to extract the epileptic spikes in healthy period and the intermittent period between epilepsy attacks, and during epilepsy attacks. The experimental results show that when the sampling frequency of the data is 173.61 Hz and the signal length is 23.6 s, the algorithm can extract the spike signals of different epilepsy patients in different periods, and that the rates of false detection and missed detection of the proposed algorithm are 12.02% and 11.70%. The proposed algorithm has a good performance in the detection of epileptic spikes.