Seizure prediction model based on method of common spatial patterns and support vector machine

Records of brain electrical activity from intracranial and scalp EEG of seven patients with different types of epilepsy are analyzed to predict the epileptic seizure onset. A method based on the CSP and SVM is introduced. This is an efficient method to predict epileptic seizures: from 52 pre-seizure signals, the seizure onsets in 23 of those are predicted. Through this method, we propose a seizure prediction model which gets an accuracy rate represented by predictions/seizures of 5/20-5/5 and a pseudo-prediction rate of 1.6-10.9 per hour.