Author(s) :

Kenneth Yu.

Volume/Issue :

Abstract :

A seizure is a sudden, uncontrolled surge of electrical activity in the brain that affects millions of people leading to serious impairments. Seizures can be caused by many factors commonly attributed to epilepsy; however, other factors, such as brain tumors and drugs, can induce seizures as well. Such seizures, especially in early childhood, can have a devastating impact on a child’s development, or induce physical harm due to the immediate loss of consciousness and other severe side effects. Currently, seizure treatments are limited, with medications and surgery being the two main points. However, models based on artificial intelligence and machine learning have emerged as newer prevention methods for seizure treatments, allowing for forecasting ahead of time to apply treatments in case a seizure can happen. This systematic review paper investigates recent literature, from 2022 to 2025, and explores the application of artificial intelligence and machine learning to seizure forecasting and detection. The sources are screened through the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flowchart. Additionally, the review addresses the limitations, reliability, accuracy, and barriers to widespread use indicating that seizure care is shifting from reaction to prediction. Through artificial intelligence and machine learning, seizures can now be detected before they occur, allowing earlier intervention and improved patient safety. This review establishes the state of the art in models, performance, and challenges, and will help researchers continue to advance in the realm of seizure forecasting, thereby improving the quality of life of many people around the globe.