Enhanced pre-surgical evaluation in drug-resistant epilepsy using machine/deep learning models

Surgical removal of the epileptogenic tissue or implantation of a stimulation device represents the best chance for patients with pharmacoresistant epilepsy to be seizure-free. Intracranial EEG acquisition, processing, and machine learning techniques have undergone rapid development and shown promise in the localization of epileptogenic tissue and seizure prediction. However, the transition of these methods into the clinic is slow even though the computational power of computers is sufficient. The project aims to develop a system for automated high-frequency intracranial EEG processing to localize epileptic tissue, predict seizures and surgical outcomes. That entails long-term signal acquisition and storage, application of signal feature extraction algorithms including machine learning and deep learning techniques, and visualization of the results to medical staff. The output of this project will be the software tools for fast and precise localization of epileptogenic foci, seizure, and outcome predictions which could significantly reduce the cost and risk of treatment and increase the wellbeing of patients.