Preictal Network Dynamics Vary Depending on Seizure Generalization: Opportunities for Neuromodulation

Introduction: Seizures in focal epilepsy emerge from seizure onset zones (SOZs) and are relayed by early propagative zones (EPZs)[1]. While the Interictal Suppression Hypothesis (ISH) posits that SOZs and EPZs are suppressed by a widespread network to maintain the interictal period, it remains unclear why seizures nonetheless occur[2]. We hypothesize that seizure onset is the culmination of a progressive increase in ictogenic processes, triggering a compensatory increase in network suppression which is ultimately overcome at seizure onset. Because seizures that generalize recruit EPZs to a greater extent than seizures that remain focal[3], we expected to observe increased preictal suppression of EPZs in seizures that generalize as evidence that network suppression compensates for specific ictogenic processes.

Methods: We studied 44 patients with drug-resistant focal epilepsy undergoing stereo-EEG presurgical evaluation in the Vanderbilt EMU. Regions are designated as SOZs if they initiate seizures and EPZs if they are recruited within 10 seconds of seizure onset. All other regions are designated as non-involved zones (NIZs). Directed connectivity between regions was calculated using partial directed coherence (PDC) across frequency bands, where greater net PDC indicates more inward connectivity. We defined interictal epochs as ‘baseline’ if no seizures occurred within 4 hours. We analyzed dynamic connectivity across the baseline interictal period and the 95 minutes preceding the onset of 53 focal aware seizures (FAS) and 55 focal-to-bilateral tonic-clonic seizures (FBTCS). Statistical significance was evaluated using a paired t-test.

Results: For NIZs, PZs, and SOZs, inward connectivity relative to the baseline interictal period is consistent across the 95 minutes preceding FAS and FBTCS onset (Fig. 1A-B). EPZs receive significantly greater inward connectivity relative to baseline throughout the preictal period of FBTCS onset versus FAS onset (Fig. 2A, p=0.007-0.032), except for the 5 minutes immediately preceding onset (Fig. 2A, p=0.662). Inward connectivity to SOZs does not significantly differ between FAS and FBTCS preictal periods (Fig. 2B, p=0.651-0.898).

Conclusion: Network suppression of EPZs, but not SOZs, is significantly greater during the preictal period of seizures that will generalize compared to seizures that remain focal. Considering that seizures that generalize have been demonstrated to recruit EPZs to a greater extent, this supports the hypothesis that network suppression of seizure zones is a protective mechanism which compensates for underlying ictogenic processes before ultimately being overcome at seizure onset. Augmenting network suppression of EPZs via targeted neuromodulation may reduce secondary generalization risk.