Predicting Epileptic Seizures with Artificial Intelligence

Description

Epileptic seizures are caused by uncontrolled, simultaneous, electrical discharges of millions of nerve cells in the brain. The majority of those affected can be helped with medication. However, some patients remain resistant to therapy, i.e., the seizures cannot be controlled. In these cases of drug-resistant epilepsy, patients often remain isolated and are restricted in their everyday lives because the risk of a sudden seizure is too high. In our research project, we have attempted to develop an early warning system for epileptic seizures by creating a forecast based on changes in biomarkers. To this purpose, we have developed a mobile sensor network of wearables that can be used to monitor certain bodily functions around the clock. We have analyzed the
recorded data using cutting-edge deep learning methods in order to discover patterns in multivariate time series data that indicate preictal states and can therefore signal an impending seizure. Initial results demonstrate the functionality of our method.

Period	10 Apr 2025
Event title	Forschungsinteraktion TNF - MED 2025
Event type	Conference
Location	Linz, AustriaShow on map
Degree of Recognition	National