A Systematic Review of Machine Learning-Based Methods for Detecting Epileptic Episodes in Advance

Abstract

Epilepsy is a chronic neurological disorder which is characterized by recurring seizures and affects around 50 million people globally, making it problematic to the healthcare systems and diminishing patients' quality of life. Traditional seizure detection methods, primarily based on electroencephalogram (EEG) analysis and clinical observation, are constrained by their limited accuracy and real-time monitoring capabilities. Nevertheless, there has been the availability of the machine learning (ML) which opened new doors for early detection and prediction epileptic events. There are ML algorithms such as support vector machines (SVMs), neural networks, deep learning models which comprises of the convolutional neural networks (CNNs) and recurrent neural networks (RNNs) that have demonstrated significant potential in identifying precursor patterns and biomarkers from EEG data. The early detection of epileptic events is crucial for the well-being of  people with epilepsy because it enables timely interventions that would ease the management of this condition. Machine learning (ML) is a recent tool used to predict epileptic seizure by analyzing large datasets with intricate algorithms in order to detect particular patterns indicative of forthcoming activities. Based on the secondary data this systematic review explores prominent research contributions in various machine learning-based approaches for the early detection of epileptic events, highlighting the methodologies, data types, performance metrics, and key findings from recent studies. By synthesizing and discreetly analysing the current state of research, this review aims to delve deeply into the most effective strategies and identify areas for future investigation.