Estimation of epileptic seizure by using Lyapunov exponent, wavelet entropy and Artificial Neural Networks

Brain signals are widely used for diagnosing epilepsy disease. The objective of this study is to design an automated system for differentiating epileptic EEG signals from non epileptic ones. The EEG signals used in the study comprise both healthy and epileptic signals which have been taken from patients during seizure. The signals were analyzed in phase space by means of Lyapunov exponent and wavelet entropy. Some features were identified from this phase space data and automatically classified by an adapted Artificial Neural Networks (ANN).