Leveraging Classical Methods for Accurate Epilepsy Detection with EEG Signals

Epilepsy is a neurological disorder characterized by abnormal electrical activity in the brain and its diagnosis is usually based on the analysis of electroencephalogram (EEG) signals. However, due to the inherent nonlinearity and chaotic nature of EEG signals, visual inspection is a time-consuming, labor-intensive, and experience-demanding process. Therefore, the main objective of this study is to present an efficient computer-aided approach for the analysis of EEG signals aiming at epilepsy detection. Here, 23 features were extracted from EEG signals obtained from 121 healthy and epileptic participants. These features were then fed individually to three different machine learning classifiers to evaluate their discriminative abilities. Among the extracted features, the theta band feature extracted from the F3C3 electrode showed a high performance with the decision tree classifier, achieving an F1 score of 0.774 and an accuracy of 0.813.