Abstracts

Rationale: Is seizure activity determined by short-range (mesoscale) connections, or are ictal discharges governed by long-range (macroscale) effects? And at these different ranges, does feedforward inhibition play a role? While these questions have previously been investigated using primarily single-scale modalities, the answers may include cross-scale network interactions. Therefore, we investigate and model the relationship between the macroscopic local field potentials (LFP) and the mesoscopic multiunit action potentials (spike trains) from human seizure activity. Methods: Using spike trains from microelectrode arrays (MEA) and LFPs from electrocorticography (ECoG) in patients with focal epilepsy (n=4), we employ spike-triggered averages to quantify the spike-LFP relationship during ictal and interictal states. Furthermore, we model the cross-scale interactions in two parts: 1) a spatiotemporal neural field model with local interactions for the mesoscopic, propagating spike activity; and, 2) a single neural mass with an excitatory and inhibitory population to represent the surrounding macroscopic volume of tissue. Results: Spike-triggered averages show differing short-range (at the MEA) and long-range (across the ECoG) effects during ictal time periods, which can be attributed to varying roles of feedforward inhibition. On the short-range, the averages are dominated by excitation, as is evidenced by a distinct downward deflection in the averaged LFP. This result is in agreement with previous findings of our group that suggest that failure of inhibition allows the ictal wavefront to propagate. Contrastingly, for a range of distances up to 10 cm from the MEA, spike-triggered averages show significant effects in the theta band (4-8 Hz) that plateau ~3cm from the spiking territory and include both excitation and inhibition. Applying bifurcation analysis to our neural mass model, we find that intact feedforward inhibition at the macroscale allows for a distinct route between the ictal to inter-ictal state, enabling a post-ictal process. Conclusions: Sustained seizure activity is the result of interactions across network scales, with mesoscopic spike activity showing macroscopic effects centimeters away. Feedforward inhibition plays a dual role in seizures. Failed feedforward inhibition at the mesoscale allows for propagation of multiunit action potentials, but effective feed forward inhibition at the macroscale creates a distinct transition path from the ictal to inter-ictal state, i.e. seizure termination. Funding: This work is funded by NIH R01 NS095368.