AI system can detect epileptic seizures early and warn the patient

avatar

Researchers at the University of Louisiana have an AI system developed which may detect epileptic seizures very early. This technology could significantly improve the lives of many patients in the future.

Epileptic seizures are difficult to predict

Worldwide, about 50 million people suffer from epilepsy. This disease is caused by the exchange of electrical signals between cells, and this then leads to chaos in the brain. This disease then manifests itself through epileptic seizures. Nearly 70 percent of the patients are using medication to remain under control. The great problem and suffering of the patients are mainly caused by the randomness and unpredictability of the seizures. Epileptic seizures usually break without clear signs, often in the most adverse situations. This anxiety accompanies those affected on a daily basis, which is why researchers at the University of Louisiana have been working on a solution to this problem.

AI System detects seizures 99% early

For years researchers have been working on a solution to predict epileptic seizures, significantly reducing patient suffering. To do this, the researchers used an electroencephalogram (EEG) to make predictions. In a 2-step process, brain activity had to be recorded manually to make predictions. The difficulty, however, is the individual characteristics of the brain. The researchers from Louisiana have now packaged these two steps into an AI system, which gives faster and more accurate information. Different electrodes placed in several places record the spatial and temporal activity of the brain. For the system to work, the system must be set up individually with the patient. This can be done by means of portable EEG electrodes outside a hospital also. The model has 99% accuracy. Now researchers have to work on suitable and everyday hardware. This will record the activities in the form of a chip and then accompany the patient constantly and warn them at an early stage.



0
0
0.000
0 comments