Research On Automatic Detection Of Ictal Epileptic EEG Signal

Epilepsy is a common chronic neurological disease, characterized by a sudden synchronization of electrical activity. It is the second largest neurology-cerebral vascular disease. The waves of epileptic discharge include spikes waves, sharp waves, spike-slow waves, sharp-slow waves etc. So far, electroencephalography(EEG) examination is the most important and most valuable means for the detection of epilepsy by analysis abnormal activity of EEG recordings, it can provide very useful and reliable information for the diagnosis of abnormal brain disease. Doctors can find lesions and employ appropriate treatments through analyzing epileptic EEGs. The most important thing is to detect the existence of spikes or sharp waves in a clinical EEG examination. It is a complicated and time consuming work to check EEG waves and detect the epileptic waves by visual inspection by health care workers. It also needs professionally trained specialists which are seriously lacking in clinical area. Therefore, automatic spike detection in EEG is significant for both detecting the Epileptic waves and reducing the heavy work of doctors.The main research objective of this paper is proposed a new method of epileptic EEG recognition based on existing research, which adopts the method of wavelet analysis combined with nonlinear energy operator(NEO) and approximate entropy (ApEn) to extract feature vector, then employed support vector machines(SVM) to classify these EEG signals on the basis of the Feature vector. Firstly, the EEG signal is decomposed to4levels using discrete wavelet transform. Then compute the nonlinear energy operator of the sub-bands1and sub-bands2’s detailed coefficients. At last, EEG signals are classified by SVM.