Abstracts

Rationale:
It is now widely accepted that epilepsy is often not localized to a single region, but rather involves a network of regions. As a result, traditional methods of capturing and analyzing seizures have been augmented by techniques that provide insight into underlying brain networks. Single pulse electrical stimulation (SPES) is a novel, quantitative method that is increasingly being applied to understand epileptogenic networks. We combined SPES with graph theoretical methods to investigate the causal influence of mesial temporal regions within and outside seizure networks to provide insight into mechanisms underlying seizure generation and propagation.
Method:
SPES was conducted in 18 patients undergoing intracranial EEG monitoring with stimulation sites including mesial temporal structures: amygdala, hippocampus, and entorhinal cortex (AHE). The amplitudes of the resulting evoked responses were used to quantify the effective connectivity between stimulation and response sites and used to construct weighted, directional networks for graphical analysis. Response amplitude, network density, and graphical centrality measures (degree, hyperlink-induced topic search, and Katz) were compared for nodes inside and outside the mesial temporal structures across three patient groups: focal seizure onset zone in AHE, multifocal ictal onset zone including AHE, and ictal onset zone not including AHE. For patients who underwent bilateral AHE stimulation, metrics were also compared to contralateral MTL not involved in ictal onset. Kruskal-Wallis tests with post-hoc Dunn's tests were used to compare response amplitudes, and linear mixed effects models with post-hoc t-tests were used to compare densities and average centrality.
Results:
The connections within AHE had significantly greater median amplitude (Dunn's tests, p < 0.05, FDR corrected) and density (t-tests, p < 0.05, FDR corrected) compared to all other connection types, regardless of epileptogenicity. However, connections between focal epileptogenic AHE and surrounding non-epileptogenic regions showed increased median amplitude (Dunn's tests, p < 0.05, FDR corrected) and density (t-tests, p < 0.05, FDR corrected) compared to that of multifocal- and non-epileptogenic AHE. Electrodes within focal epileptogenic AHE also had significantly greater (t-tests, p < 0.05, FDR corrected) average network centrality compared to non-epileptogenic regions within patient group (outdegree, hub, authority, Katz-broadcast) and compared to other AHE across patient groups (outdegree, hub).
Conclusion:
Our results demonstrate that the AHE is a distinct highly connected subnetwork that can become hyperexcitable and exert an overall elevated propagative influence over the effective network when it is epileptogenic. Understanding underlying network connectivity and hierarchy may provide insights that eventually lead to improved surgical outcomes.
Funding:
:No funding was received in support of this abstract.