Title:
Automatic Detection of Epilepsy from Electroencephalography (EEG) Signals Using Artificial Intelligence (AI)

Authors:
Sri Widodo, Lilik Linggar Rahayuningsih, Indonesia

Abstract:

A central nervous system illness called epilepsy is brought on by aberrant brain activity and can produce convulsions or even unconsciousness in its victims. One way of diagnosing epilepsy is neurological. From a neurological point of view, the underlying cause of epilepsy cases is sometimes unclear. Therefore, the electroencephalography (EEG) technique is used by looking at human brain activity under various conditions. Experts will usually analyse visually the EEG signals to see the brain activity in people with epilepsy. However, with so much data, the visual analysis will take a lot of time and effort. In addition, sometimes there are errors in reading and deciding epilepsy results from EEG visualization results. The goal of the research is to use artificial intelligence (AI) to create an application that uses electroencephalography (EEG) data to identify epilepsy. Convolutional Neural Networks (CNNs) are the AI technique employed in this study. Pre-processing and identification are the two phases of this study approach, which involves categorizing data related to epilepsy (seizures) and no epilepsy (non-seizures). The pre-processing stage is carried out using a toolbox in Python software called EEGLAB, which will produce a feature in the form of the energy spectrum of the EEG signal. The Convolutional Neural Network (CNN) approach will then employ the feature extraction results as input for identification or classification. The Convolutional Neural Network (CNN) approach achieved 90% accuracy in the detection or categorization of epilepsy using EEG pictures.

DOI: http://dx.doi.org/10.51505/ijaemr.2025.1005