An individualized Neural Mass Model of ictal activity based on GABA-A pathology for personalization of brain stimulation protocols in epilepsy

Abstract

Abstract The creation of personalized brain network models for therapeutic intervention with brain stimulation is a promising direction of research in epilepsy. Early work in the last decades has shown that neural mass models can be used to represent realistic epileptic seizure transitions. We provide here a novel semi-autonomous neural mass model that includes the dynamics of chloride accumulation and is capable of realistically reproducing the electrical activity recorded by an SEEG electrode in the epileptogenic zone during a seizure. The neural mass model includes a physiologically inspired algorithm that relates activity-dependent GABA-A depolarization to the pathological accumulation of chloride in pyramidal cells. Due to fluctuating input, chloride overload into pyramidal cells induced seizure-like activity—including pre-ictal, fast onset, and ictal transitions. The fact that the model is capable of generating spontaneous realistic seizures constitutes a novelty in the field of computational modeling with neural masses in epilepsy. With the aim of generating SEEG data to compare with intracranial recordings from epileptic patients, the neural mass model is first embedded in a layered model of the cortex, and then in a realistic head model. The model also accounts for the effects of electric fields on the membrane of cell populations. We provide some examples of model personalization for clinical cases demonstrating that the patients’ SEEG-recorded seizures can be successfully reproduced by the personalized model embedded in the subject-specific realistic head model and that they can be controlled with a DC field. By including key physio-pathological and physical mechanisms, our model simulates realistic ictal SEEG electrical activity and reflects the effects of brain stimulation. This provides a robust framework for the creation of personalized brain network models for the design of therapeutic interventions such as transcranial electrical stimulation (tES) in epilepsy. Keywords: epilepsy, neural mass model, GABA depolarization, personalized models