Non Stochastic Propagation of Interictal Spikes as a Measure of Human Brain Connectivity

As a means to understand human neocortical brain connectivity, we analyzed long term intracranial recordings from patients undergoing epilepsy surgery monitoring with subdural grid electrode arrays. These recordings detect seizures, but more frequently they detect interictal spikes that like seizures, can spread across the cortex. While spike propagation can be inferred by visual inspection of temporal relationships of spikes on different electrodes, there is inherent variation that makes a manual approach subject to possible false positives and negatives. In order to develop an unbiased method to overcome this obstacle, we measured the propagation patterns of IISs using a multivariate auto regressive algorithm called discrete short time direct Directed Transfer Function (dsdDTF), used previously to measure seizure propagation, as a means to identify common directionalities of interictal spike propagation using different frequency bands. EEG data was superimposed on 3D brain reconstructions as a means to relate spike propagation to neocortical anatomy. We specifically examined the propagation patterns of interictal spikes on 10 patients who had long term electrode arrays overlying their central sulcus that separates the sensory and motor cortices from each other. We found the dsdDTF method was able to see all of the spike propagations visualized manually, but with much greater detail. Our algorithm identified patterns of directed IIS propagation that were reproducible on separate epochs for each of the 10 patients. We also found that interictal spikes preferentially propagate to adjacent cortical regions more frequently than distant areas, suggesting greater functional connectivity at these shorter distances. Finally, we found that the complex gyral pattern of the brain influences interictal spike propagation with very low rates of interictal spike propagation across the central sulcus compared to other adjacent brain regions. This study provides the first application of dsdDTF to neocortical interictal spiking and demonstrates clear spatial relationships on how interictal spikes propagate that could reflect the underlying functional connectivity of the human neocortex. This novel, quantitative technique also has the potential to help identify epileptic brain regions and improve outcome in patients undergoing epilepsy surgery.