Detection of Epileptic Seizure EEG Signal Using Multiscale Entropies and Complete Ensemble Empirical Mode Decomposition

Epilepsy is a severe neurological disease which is diagnosed by analyzing Electroencephalogram. The epileptic seizure detection technique based on multiscale entropies and complete ensemble empirical mode decomposition (CEEMD) is proposed in this paper. CEEMD is used for the estimation of sub-bands and two multiscale entropies; multiscale dispersion entropy (MDE) and refined composite MDE are extracted from the sub-bands. The feature selection method, configured by hybridizing the filter based and wrapper based method, is used to select relevant multiscale entropies. The hybrid method has not only reduced features but also improved classification performance. An artificial neural network is trained with relevant features and performance is measured using classification accuracy, sensitivity and specificity. Five clinically relevant classification problems are used to assess the proposed technique. The performance is also compared with the state of the art techniques. The proposed technique has shown an improvement in detection of seizures and can be used to build the clinical system for epileptic seizure detection.