Analysis of the Local Dynamics of Interictal Discharge Propagation Using a Traveling Wave Model

Epilepsy is one of the most common neurological diseases in the world, with about 30% of cases not amenable to pharmacological intervention and potentially requiring surgical intervention. The process of localizing the epileptogenic zone – the area associated with seizure initiation in patients with focal epilepsy – involves examining different areas of the brain for the presence of interictal discharges. In this paper, we propose a new methodology for noninvasive investigation of the fi ne spatiotemporal structure of interictal discharges observed on the magnetoencephalogram (MEG). We applied a traveling wave model to regularize the inverse MEG problem. The algorithm represents the neural activity generating an interictal discharge as a superimposition of local waves propagating along radial paths and generated by a single point source. The LASSO method with positive coeffi cients was applied to determine the combination of waves generated with different parameters giving the best match with the recorded MEG for each discharge. The properties of the algorithm were analyzed using realistic simulations of MEG data. The method was then applied to analysis of MEG data from three patients with drug-resistant multifocal epilepsy. Some of the discharges yielded wave-like patterns with clear propagation dynamics, while for others, the observed activity could not be explained by the wave superimposition model. Moreover, discharges with clear propagation dynamics showed marked spatial clustering correlating with the epileptogenic zones described in the case histories of two of the three patients.