Improving epileptogenic zone estimation using Bayesian inference on neural field models

Epilepsy remains a significant medical challenge, particularly in drug-resistant cases where surgical intervention may be the only viable treatment option. Identifying the epileptogenic zone, the brain region responsible for seizure initiation, is a critical step in surgical planning. Combining dynamical system models and the neuroimaging data of epileptic patients in a Bayesian framework has previously been shown to be a promising approach to identify the epileptogenic zone. However, previous studies employed coupled neural mass models to describe the whole brain seizure dynamics and hence could only provide a highly coarse spatially estimate of the epileptogenic zone. In this study we propose an extension of the Bayesian virtual epileptic patient framework to a neural field model which could improve the spatial resolution by several orders. Performing model inversion using neural field models is a challenging task since: (i) it is computationally expensive to compute gradients over a neural field simulation and (ii) parameter space can be very high dimensional. We demonstrate that using pseudo-spectral methods and spherical harmonic transforms it is feasible to perform Bayesian model inversion on a neural field extension of the reduced Epileptor model. We found that the neural field extension not only improves the spatial resolution but also significantly reduces the number of false positives.