Epileptic seizure classification from intracranial EEG signals: A comparative study EEG-based seizure classification

Epilepsy is a neurological disorder of the brain, characterized by recurrent seizures, i.e., unusual behaviors. Electroencephalogram (EEG) is a test that measures and records the electrical activity of the brain, and is widely used in the detection and analysis of epileptic seizures. However, it is often difficult to identify critical changes in the EEG waveform by visual inspection, thus opening up a vast research challenges for developing automated methods for quantifying such subtle changes. Moreover, manual interpretation becomes complex due to nonlinear and non-stationary nature of the EEG signals. Hence, it is necessary to develop a quantitative method to automatically identify the normal and epileptic brain activities. In this study, we propose a supervised machine learning method to classify non-ictal (normal or preictal or interictal) and ictal (during seizure) brain activities from EEG signals. Several time-domain and time-frequency domain features are extracted to characterize EEG signals, which are then applied to linear discriminant analysis (LDA) or nonlinear support vector machine (SVM) classifier in order classify brain epilepsy. A preliminary experiment with two sets of EEG signals from `focal' and `nonfocal' channels (50 EEG signals per set, each of 10 second duration) of five subjects shows 79.20% accuracy of the epilepsy classification using a combination of features from the time frequency domain.