Abstracts

This is a Late Breaking abstract

Rationale: Understanding seizure propagation, and in particular the mechanisms which limit or detain seizure spread is of high therapeutic interest. In animal models, an inhibitory response in the cortex surrounding the propagating ictal wavefront has been demonstrated, and evidence of this has been seen in humans. In this work we sought to explore whether a clear inhibitory component can be seen in human seizures, not only on a submillimeter scale as previously shown, but rather throughout the epileptic network. We used two measures to infer neuronal inhibition: high gamma power and the 1/f slope and offset. First, we postulated that a decrease in high gamma power may imply a decrease in neuronal firing, based on the established correlation between the two. We then extracted the aparametric fit of the 1/f curve, the slope of which has been shown to reflect excitation/ inhibition balance, to examine the dynamic between excitatory and inhibitory tone in relation to the seizure onset zone (SOZ).

Methods: A total of 71 seizures recorded from 15 patients with medically refractory epilepsy who underwent presurgical evaluation with intracranial electrodes were analyzed. All channels were scrutinized for decreases of high gamma band power (70 – 130 Hz) compared to baseline. High gamma decreases were then characterized morphologically and by their distance from the SOZ. In a subset of 49 seizures from 14 patients, the FOOOF algorithm was used to parameterize neural power spectra (range 1-40 Hz) in a sliding window which allowed analysis of the 1/f slope as it changed throughout the seizure. _x000D_
Results: Decreases in high gamma power compared to baseline were seen in in at least one electrode in all seizures. These high gamma decreases either persisted throughout the seizure or were transient and were at times heralded by a brief increase in high gamma. They were most frequent approximately 5cm from the SOZ though they were seen at varying distances from the SOZ, and their amplitude was unaffected by this distance. They were significantly more likely to occur in contacts outside the SOZ.  The 1/f fitted alpha slope was extracted for a subset of 49 seizures and was shown to evolve compared to baseline as the seizure progressed. This dynamic differed depending on proximity to the electrode of seizure onset, with a decrease in alpha slope in electrodes proximal to the SOZ and an increase in electrodes distant from it.

Conclusions: The identification of decreases in high gamma band power during a seizure is highly suggestive of inhibition, seen primarily in regions outside the SOZ. The observed increases in the alpha slope of the 1/f curve further support this notion. Identification of high gamma decreases at various distances from the SOZ suggests that these inhibitory responses to the seizure go beyond contiguous spread and are rather a network phenomenon. How and whether this inhibition shapes seizure dynamics and propagation remains an open question, but a deepened understanding of ictal inhibition could pave the way for novel, spread directed interventions in epilepsy.

Funding: None