Features extraction and classification for Ictal and Interictal EEG signals using EMD and DCT

Electroencephalogram (EEG) is a record of electricalsignal to represent the human brain activity. Many researchersare working on human brain as they are fascinated by the idea ofsecret, thought and feeling from the external and/or internalstimuli. Feature extraction, analysis, and classification of EEGsignals are still challenging issues for researchers due to thevariations of the brain signals. Different features are used toidentify epilepsy, coma, encephalopathies, and brain death, etc.However, we have observed that features extracted from samekinds of signal transformations are not effective to differentiatethe epilepsy periods including Ictal (active seizure period) andInterictal (interval between seizures) of EEG signals. In thispaper we present a new approach for feature extraction usinghigh frequency components from DCT transformation. Then, wecombine the new feature with the existing feature (amplitude andfrequency modulation bandwidth) extracted from the empiricalmode decomposition. These features are then used as an input toleast squares support vector machine (LS-SVM) to classify Ictaland Interictal period of epileptic EEG signals from TemporalLobe brain location. Experimental results show that the proposedmethod outperforms the existing state-of-the-art method forbetter classification of Ictal and Interictal period of epilepsyusing benchmark dataset.