Development of EEG-based epileptic detection using artificial neural network

Epilepsy is one of the most common neurological disorders causing from repeating brain seizures that are the result of the temporal and sudden electrical disturbance of the brain. Electroencephalogram (EEG) signal plays an important role in the diagnosis of epilepsy. This project proposed to develop a system that can detect epilepsy based on EEG signal using artificial neural network. Discrete Wavelet Transform (DWT) and Fast Fourier Transform (FFT) were applied as feature extraction methods. These features then set as input to the feedforward neural network with backpropagation training algorithm to get the classification accuracy. The accuracy of DWT with 10000 epochs is 97% while accuracy of FFT method gives 53.889% accuracy. The combination of DWT and FFT extracted features give the highest accuracy, which is 98.889%. The classification accuracy depends on the number of epoch and the features from the feature extraction. Increased number of epoch gives long response time to train the network.