Researchers have developed an advanced AI system capable of predicting epileptic seizures with an impressive 99.6% accuracy—up to one hour before they occur. This breakthrough, reported by IEEE Spectrum, provides patients valuable time to prepare and take medication before a seizure strikes.
Currently, approximately 50 million people worldwide live with epilepsy, as noted by the World Health Organization, with 70% able to manage their seizures through medication. The new AI system, created by Hisham Daoud and Magdy Bayoumi at the University of Louisiana at Lafayette, significantly enhances existing prediction techniques.
Traditional methods rely on EEG (electroencephalogram) tests to analyze brain activity and then apply predictive models. In contrast, this innovative approach simultaneously records brain activity and predicts the electrical channels that activate during a seizure using a sophisticated deep learning algorithm.
Although widespread availability is still some time away as the research team develops a custom chip for processing, this technology promises to be a life-changing advancement for people living with epilepsy.