PhD Thesis : Multiscale modeling of epileptic seizure dynamics.
Supervisors : Christophe Bernard & Viktor Jirsa
In this thesis, we use a top-down approach to simulate epileptiform dynamics in-silico. The focus of this work is to unravel mechanisms of seizure genesis and propagation.
The first contribution consists of a literature review of recent concepts in epileptic seizure modeling. The second and main contribution is the development of a mesoscale model successfully reproducing the most important features of seizure dynamics. The third contribution is an extension of the second in the context of seizure propagation between 2 brain regions.
The developed model is based on the Epileptor, a phenomenological dynamical system with 5 state variables reproducing major features of seizure-like events.
A network of bursting and spiking neurons is derived from the Epileptor, with more biologically relevant properties than the former model to assess plausible contributions to seizure from diverse physiological components, such as electro-chemical synaptic couplings and excitability of the neural medium.
Using this paradigm, we successfully reproduce the 4 different phases observed experimentally by induced Status Epilepticus. We find that inhibitory components play a major role in the generation of spike-wave discharges typically observed in-vivo.
We also propose the extension of a companion study modeling seizure propagation between two distinct brain regions, using our more detailed network of neurons with a set of different connectivity patterns. We assess the interplay between epileptogenic and propagation networks in the occurrence of interictal spikes in the latter region.