Abstracts

Rationale: A large body of evidence suggests that pathologic high-frequency activity is a hallmark of ictogenesis. Recently, exaggerated cross-frequency phase-amplitude coupling (PAC) between the phase of slow rhythms and the amplitude of fast rhythms has been associated with the seizure onset zone (SOZ), specifically during ictal states. Given the well-documented role of low-frequency rhythms in coordinating activity across disparate brain regions, PAC is well positioned to facilitate the dissemination of focal onset ictal activity. Therefore, we hypothesize that ictal activity within the SOZ should precede exaggerated increases in PAC, which occur only when the seizure generalizes. Correspondingly, local increases in PAC within regions of secondary seizure generalization should either coincide with or temporally precede the spread of overt ictal activity to that location.Methods: We conducted preliminary analysis on one artifact-free seizure recorded with intracranial EEG in a patient with intractable epilepsy. The seizure start time and SOZ was identified by a board-certified epileptologist. The signal from each electrode was decomposed into phase (2-30 Hz) and amplitude (15-350 Hz) using a complex-Morlet wavelet. PAC was computed separately for the 20 seconds immediately preceding seizure onset (preictal) and the first 20 seconds of the seizure (ictal). Significant PAC was determined using Monte Carlo simulations (1000 surrogates) and a p=0.05 (two-tailed, Bonferroni corrected) with the added condition that at least one adjacent phase and amplitude frequency layer was also significant. For each electrode, both the line-length (LL, a surrogate measure for ictal activity) of the 1-400 Hz voltage time-series and the PAC for all frequency pairs was computed using a 4-second moving average. These measures were then cross-correlated to find the relative time lag when PAC and LL were most strongly correlated. These time lags were then averaged across all ictally significant PAC frequency pairs for each electrode.Results: From the preictal to the ictal window, the number of electrodes displaying significant PAC increased (4/46 and 14/46, respectively), and the overall magnitude of significant PAC increased 6.6-fold. The 2 electrodes identified as the SOZ were included in the 14 electrodes with ictally significant PAC with the remaining 12 displaying ictal activity shortly after seizure onset. Cross-correlation analysis revealed that increases in ictal activity dramatically preceded increases in PAC by 305 msec in one of the SOZ electrodes. In the remaining electrodes, increased PAC and LL either co-occurred with 4 msec (n=5) or PAC preceded LL by anywhere from 20-742 msec (n=8).Conclusions: These data point to a focal region within the SOZ where ictal activity precedes exaggerated PAC, in contrast to all secondarily involved ictal regions. This finding is consistent with our hypothesis that PAC facilitates the spread of focal onset ictal activity. We intend to assess the robustness of this finding with a larger cohort.