Automatic Epileptic Seizure Detection Based on EEG Signals Using Deep Learning
Keywords:Electroencephalogram (EEG), epileptic seizure, time and frequency domain features, Convolution Neural Network (CNN), deep learning
About one percent of the world's population suffers from epilepsy. A patient with epilepsy must be diagnosed early and accurately if they are to have any chance of being treated successfully. One method for diagnosing epilepsy is by carefully analyzing the Electroencephalogram (EEG) signal. We propose a method for signal processing EEG signals that detect epilepsy based on time-frequency features extracted from the signal and used as input for a neural network classifier. With the help of a convolutional neural network with deep learning, better and more efficient features were obtained and an accurate diagnosis was provided. It resulted in a significant difference between the two individuals upon analysis of the EEG signal. As compared with the previous method, the proposed technique distinguishes between healthy and epileptic signals with specificity 98 ± 2%, sensitivity 99 ± 0.7%, accuracy 98 ± 0.6%, and F-score 98 ± 0.5%. It is possible to use EEG signal analysis to detect the onset of seizures, especially in infants, as an effective tool to diagnose cases of suspected clinical signs of seizure onset.