Automatic identification of an epileptic spike pattern in an EEG signals using ANN

Abstract

This work emphasizes to automatically detect the pattern called ‘epileptic spike’ from electroencephalogram (EEG) signal using multilayer perceptron (MLP). The analysis work is carried out through using the receiver operating characteristics (ROC). Electroencephalograph is used to record the electrical activity of the brain. Classification of the (EEG) signal plays a vital role in the diagnosis of epilepsy. The verification of epileptic seizure requires longterm EEG monitoring of 24 h or more. The signal is a huge collection of data and unfortunately, medicos uses the traditional method of visually interpreting EEG signal through personal experience to identify the transient event of epilepsy. This method of visual interpretation is tedious and time-consuming and, may result in an erroneous judgment. Hence, efficient EEG signal analysis is required for the diagnosis of epilepsy. This model is evaluated on the basis of sensitivity and selectivity and experimental result highlights the good precision of the model. The overall accuracy of the model is computed as 99.11%.