Automatic processing of EEG signals for seizure detection using soft computing techniques

Epileptic seizures, a crucial neurological disorder, reflect the excessive and hyper-synchronous activity of neurons in the brain. Human knowledge of functioning of the brain is still insufficient to understand the neurophysiology of suddenly occurring epileptic seizures. But the detection of the disorder and recognition of the affected brain area is essential for the clinical diagnosis and treatment of epileptic patients. Epilepsy is not only a disorder, but rather acts as a syndrome with divergent symptoms involving episodic abnormal electrical activities in the brain. EEG is the most economical and effective tool with high temporal resolution for understanding the complex dynamical behavior and studying physiological states of the brain. The research presented in this paper, aims to develop a computer aided diagnostic system utilizing EEG data to diagnose whether the person is epileptic or not. We present here various methodologies that could be implemented in hardware for monitoring an epileptic patient. Statistical features depicting morphology of EEG signals are extracted, selected and utilized to classify the signals by Artificial Neural Network, Radial Basis Function, Naive Bayes Classifier, K means classifier, Support vector machine. Efficacy of technique is evaluated on the basis of performance measures, sensitivity, specificity and accuracy. It has been observed that artificial neural network and support vector machine with radial basis function kernel are more successful as compared to other soft computing paradigms.