Grant: Project Grant

Drug-resistant epilepsy affects many patients whose seizures cannot be controlled with medication. These seizures are unpredictable and can cause injury, loss of independence, and increased risk of premature death. Neurostimulation devices, which deliver electrical pulses to the brain, offer an alternative treatment for some patients. However, current devices stimulate either continuously or only after a seizure has begun. Although these approaches can reduce seizure frequency, only about 10% of patients achieve long-term seizure freedom. This project will develop a new intelligent neurostimulation system that predicts seizures before they occur and delivers targeted stimulation to prevent them. The system combines artificial intelligence with adaptive optimization of stimulation patterns, allowing therapy to be tailored to each patient's brain activity. Over time, the system will learn which stimulation patterns are most effective for preventing seizures in each individual. Our team has already demonstrated key elements of this approach. In animal studies, prediction-triggered stimulation reduced seizures by 90%, with over 80% of animals becoming seizure-free. We have also demonstrated the feasibility of AI-guided neurostimulation in patients undergoing short-term monitoring. In this project, we will evaluate this technology in patients undergoing clinical monitoring for epilepsy. Using a wearable system connected to implanted electrodes, we will test whether seizure prediction and adaptive stimulation can reduce seizures during several weeks of monitoring. The system will also be designed to operate within the strict size and power limits required for future implantable devices. By demonstrating seizure prevention in freely moving patients, this research will help bridge the gap between early research and future implantable therapies. Ultimately, this work could lead to more effective, personalized treatments for people living with drug-resistant epilepsy.