Epilepsy, a condition that causes regular seizures, affects about 65 million people around the globe. Despite treatment advances, epilepsy has long posed a complicated problem to health care providers that, until now, hasn’t been adequately solved: how to know exactly where a seizure starts, which could allow providers to target specific areas of the brain.
The ability to locate the exact location of a seizure would allow surgeons, for example, to use more precise procedures and even avoid surgeries that have high risks and can lead to longer hospital stays for patients. Currently, locating the origin of a seizure can be a time consuming, invasive process. These approaches often involve patients being hooked up to electrodes waiting for a seizure to occur so that a patient’s doctor can develop an accurate map of the brain and pinpoint the exact cause of a seizure for more effective treatment.
Unfortunately, surgery does not have a high success rate due to the difficulty in locating the exact location of a seizure. The ability to identify the exact spot would allow for more precise treatment, which could change the game for epilepsy treatment.
A team of researchers at Johns Hopkins University have developed an artificial intelligence tool that could help overcome some of the limitations to current approaches to locating seizure origin and improve treatment and outcomes for patients. Their work is described in a recent article published in Brain.
Specifically, the AI approach studies how the brain functions as a network of nodes to make predictions about brain activity. The research team looked at brain scans for 65 patients and used the AI tool to predict seizures, where the surgery would originate, and how successful surgery would be. Researchers found the AI could make these predictions with a 79% level of accuracy.
Researchers plan to study their AI in clinical trials in the future.
Sources: Eurekalert!; Brain
