Neuronal silence as a predictive biomarker and target for epileptic seizures suppression.

Epilepsy is a prevalent neurological disorder marked by abnormal synchronized neuronal firing, which can often lead to long-term cognitive and physical impairments. In this work, we introduce a reliable biomarker for seizure prediction. Through simulations of a conductance-based neuronal network model that reproduces spontaneous seizure-like events, we identify that slow potassium channels play an important role in seizure generation. Our key finding is the consistent presence of a prolonged period of neuronal silence that precedes the seizure onset, establishing it as a physiologically relevant biomarker for seizure prediction. Notably, this silence is also identified in human electrophysiological data, confirming its physiological and clinical relevance. Based on this biomarker, we develop a targeted suppression strategy that, in our simulations, significantly shortens long seizure duration by up to 93%. Our results establish the network silence as a predictive and clinically translatable biomarker for seizure dynamics, opening new avenues for improved forecasting and personalized neuromodulation therapies in epilepsy.