Automated seizure detection using multilayer feed forward network trained using scaled conjugate gradient method

Electroencephalogram (EEG) is a tool used in the diagnosis of a common neurological disorder Epilepsy. Analysis of long recordings of EEG by visual inspection for epilepsy is quite a tedious process. In this paper, we present an approach for automated epileptic seizure detection by employing Multi layer Perceptron Neural Network (MLPNN) classifier. Independent Component Analysis (ICA), a statistical tool is used for extraction of features. The ascertained signals are trained under supervision by making use of memory efficient and fast Scaled Conjugate Gradient (SCG) backpropagation algorithm. The data set is taken from a publicly available EEG database. The MLPNN is designed with the tan-sigmoid transfer function in the hidden layer and output layer. The network is analyzed using performance metric like Mean Square Error and confusion matrix. The best classification accuracy is about 100% for the overall dataset. This indicates the proposed method has potential in designing a new intelligent EEG-based assistance diagnosis system for early detection of the electroencephalographic changes.