Abstracts

Rationale: The creation of personalized brain network models for therapeutic intervention 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 seizure.

Methods: In order to realistically represent seizure transitions, we have developed a physiologically inspired algorithm for activity dependent GABA^­_A depolarization based on chloride accumulation dynamics. The key element of the model is the association of the gain of GABAergic synapses with the pathological accumulation of chloride in pyramidal cells due to high inhibitory input and ion transporter system dysregulation. The model dynamics are autonomous with the exception of the external noise input into the pyramidal cell, i.e., it does not require any time dependent parameter adjustments. 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.

Results: The model successfully reproduced realistic ictal activity in a semi-autonomous manner. Due to fluctuating input, chloride overload into pyramidal cells induced seizure-like activity—including pre-ictal, fast onset and ictal transitions. Crucially, model parameters can be personalized to fit the main features of ictal SEEG recordings in a given patient, in particular in terms of discharge frequency or duration of each seizure phase. We provide some examples of model personalization for clinical cases demonstrating that the patients’ SEEG-recorded seizures can be successfully generated by the personalized model embedded in the subject-specific realistic head model.

Conclusions: By including key physio-pathological mechanisms, the proposed model is able to simulate realistic ictal SEEG electrical activity. The model provides a robust framework for the creation of personalized brain network models of value for the design of therapeutic interventions such as transcranial electrical stimulation (tES) in epilepsy.

Funding: Please list any funding that was received in support of this abstract.: This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement No 855109).