Artificial Intelligence Powered System for Epilepsy Detection Using EEG Biomarkers

Abstract

One of the most prevalent neurological disorders is epilepsy. Epileptic seizures can occur repeatedly in people with the condition for no recognisable reason. The diagnostic methods dependent on EEG are promising. EEG has uncovered the dynamic functioning of all brain areas throughout time. Its low cost, non-invasiveness, and simply make it crucial for clinical evaluations of brain function. Integrating multiple EEG biomarkers as a compound biomarker could provide a high performance that may accelerate the diagnosis speeds. Artificial intelligence techniques such as machine learning and deep learning provide a significant result in healthcare applications. Logistic Regression (LR), Naive Bayes (NB), and Neural Network (NN) were evaluated using a compound biomarker containing eleven EEG features that were extracted from the Bonn EEG dataset. The aim of this study is to evaluate the feasibility of integrating multiple EEG biomarkers as compound biomarkers for identifying epileptic people. The outcomes showed the performance of all LR, NB, and NN detection models provide a high performance with sensitivity and specificity of greater than 90%.