Caractérisation et localisation de réseaux épileptiques en utilisant de signaux EEG de surface et de profondeur

Epilepsy is a neurological disorder affecting 0.6% to 0.7% of the world’s population, characterized by recurrent seizures due to abnormal brain activity. For certain epileptic patients who do not respond well to drug treatment, surgery may be necessary, requiring a good knowledge of the epileptogenic zone responsible for triggering the seizures. Our work focuses on the inference of effective brain connectivity from depth and surface EEG signals, in order to better identify epileptic networks. Initially, we explored dictionary-based approaches and proposed a new approach which, when applied to intracerebral signals, enabled us to correctly identify propagation graphs during the ictal phase. We then developed an innovative “all-in-one” approach combining simultaneous source localization and connectivity estimation from surface signals. This proved to be both efficient and robust on expert-simulated patterns.