AI learned to predict 99% of epileptic seizures

Scientists from the University of Louisiana have introduced an artificial intelligence system that can predict epileptic seizures with an accuracy of 99.6%. The device can do this an hour before the first signs of a seizure appear.

According to IEEE Spectrum, the system will give people enough time to prepare for a seizure and take the right medicine. According to the World Health Organization, about 50 million people worldwide have been diagnosed with epilepsy, and 70% of these patients can control their seizures with medication.

Although this is not a complete solution to the problem, the new artificial intelligence system developed by Hisham Dowd and Magdie Byoumi of the University of Louisiana is an important step forward in comparison with existing methods of forecasting.

Current seizure detection methods analyze brain activity using an EEG test (electroencephalogram) and then apply a predictive model. The new system does both at the same time using the deep learning algorithm, which displays brain activity.

The researchers added: in order to introduce the system for widespread use, it takes some more time. Now they are working on a custom chip that can help in processing the necessary algorithms.

hightech.fm