Epilepsy is increasingly understood as a disorder of distributed brain networks rather than a condition arising from a single isolated focus. The aim of this thesis was to investigate how epileptic activity emerges, propagates, and becomes self-sustaining across multiple biological scales, from large-scale brain circuits in vivo to patient-derived neuronal networks in vitro.
To address this question, three complementary studies were conducted.
The first study examined the role of interhemispheric hippocampal communication in seizure generation in a mouse model of temporal lobe epilepsy induced by intrahippocampal kainate. It showed that long-lasting ictal events, but not isolated interictal spikes, depended on the activity of the contralateral hippocampus. Seizure onset was consistently associated with coordinated firing and low-frequency coupling between both hippocampi, indicating that ictogenesis relies on pathological inter-hippocampal interactions. These findings identify bilateral hippocampal coupling as a potential biomarker and therapeutic target for seizure prevention.
The second study investigated how a focal hippocampal insult progressively evolves into a distributed epileptic network. Longitudinal multi-region recordings revealed that, despite largely unilateral structural damage, epileptiform activity rapidly engaged both hemispheres. Early network maturation was better captured by hemispheric asymmetry, particularly in fast-ripple activity, than by absolute spike rates, and this asymmetry predicted later seizure burden. Chemogenetic manipulations further demonstrated that different network nodes play distinct and dynamic roles during epileptogenesis. While silencing the epileptogenic focus early after status epilepticus did not prevent disease progression, inhibition of the contralateral hippocampus unexpectedly worsened pathological activity. These results show that the contralateral hippocampus initially exerts a compensatory influence before later contributing to seizure generation, highlighting the resilience and adaptability of the epileptic network.
The third study extended this network-based perspective to a human model of epilepsy using brain organoids derived from induced pluripotent stem cells of a patient with Glucose Transporter Type 1 Deficiency Syndrome (GLUT1-DS). Under normal glucose conditions, patient and control organoids developed broadly similar neuronal activity. However, under reduced glucose availability, patient-derived organoids displayed marked hyperexcitability, reproducing the metabolic sensitivity observed clinically. This work establishes patient-derived brain organoids as a relevant platform for modeling epilepsy mechanisms and testing personalized therapeutic approaches.
Together, these studies support a multi-scale and network-centered view of epilepsy. They show that temporal lobe epilepsy rapidly becomes a bilateral and resilient network disorder, that remote brain regions can either restrain or promote epileptogenesis depending on disease stage, and that human patient-derived models provide valuable translational tools. Overall, this thesis contributes to a better understanding of epilepsy as a dynamic network disease and supports the development of future network-guided therapeutic strategies.
