A novel module based approach for classifying epileptic seizures using EEG signals

Electroencephalogram is one of the most clinically and a scientifically utilized signal recorded from human brain and is powerful source of providing valuable insight of the brain dynamics, playing a prominent role in diagnosis of brain diseases and many cognitive processes. Epilepsy is a persistent, constantly recurring neurological brain disorder characterized by abnormal electrical activity in the brain that acts as a syndrome with divergent symptoms involving spasmodic abnormal electrical activities in the brain. In the present work, seizure classification utilizing EEG signal has been carried out by designing module based CAD system. Design of a CAD system for prognosis of epileptic seizure with less computational complexity and higher accuracy is intended. The focus of this research work is to investigate the effect of module size by dividing the time series into epochs of 256, 512, 1024, 2048 and 4096 samples. With 96.9% overall classification accuracy, and high performance metrics with selected feature set, we infer that the chosen feature set and MLPNN classifier are more promising with the proposed technique.