Canonical correlation analysis and neural network (CCA-NN) based method to detect epileptic seizures from EEG signals

Abstract

In this research, a novel method based on Canonical Correlation Analysis (CCA) and Artificial Neural Network (ANN) to detect epileptic seizures from EEG signals is proposed. CCA was applied on EEG signals and feature vectors corresponding to Eigen values were extracted. These Eigen values were fed as input to Artificial Neural Network (ANN)‘s widely explored model Multilayer Perceptron Neural Networks (MLPNNs) for classification between occurrence of non-epileptic seizures and epileptic seizures. The extracted Eigen values using CCA proved to be a better epileptic seizures detector and provide average classification accuracy, sensitivity and specificity as 92.583%, 93.25% and 91% respectively.