Abstracts

Rationale: Most seizures have the ability to self-terminate after a few seconds to minutes, suggesting that the brain possesses endogenous mechanisms to curtail excessive neuronal activation. Immediately following termination, myriads of changes are seen electro-clinically which are collectively classified as the POST ICTAL STATE. Most studies have characterized the EEG changes (suppression or delta/theta slowing), but these studies were limited to scalp EEG. Intracranial EEG recorded from cortical and subcortical structures provides rich information about the temporospatial changes in synchrony and spectral evolution at the post-ictal state. We hypothesize that the temporal evolution of synchrony and power spectrum is spatially heterogeneous and seizure onset zone is the last to recover. Furthermore, we posit that the pattern of changes in these two parameters will be different for synchronous ( seizure terminates at the same time in all recorded channels) versus asynchronous (seizure terminates at the different time among recorded channels) seizure offset. Methods: Thirty, post-ictal Stereo-EEG recordings were selected from 15 patients undergoing epilepsy surgical evaluation at the UAB Epilepsy Center. Video-EEG was sampled at 2 KHz, and all recording channels (range= 80-140 electrodes per patient) were included in the analysis. Two minutes of electrocorticogram starting from the earliest time of seizure offset was included for analysis. For synchronous offset, this was same for all channels while in asynchronous offset the EEG was clipped at the earliest offset.  The 2 minute was then divided into 4, 30 seconds long segments, named P1, P2, P3, P4 and also a 30 s long baseline segment (B1) was chosen 5 minutes before the seizure started. The frequency was calculated with power spectrum density (PSD) in each post ictal segments, and the synchrony was defined as the mean phase lag index (PLI) between all channels from 2 s epochs derived from all segments (B1, P1-P4) in 7 frequency bands (delta:1-4 Hz, theta:4-7 Hz, alpha:8-12 Hz, beta:13-30 Hz, low gamma:30-40 Hz, high gamma:45-95 Hz, ripple band:95-150 Hz). PLI within the clinically identified seizure onset zone (SOZ) areas was calculated. Non parametric t-tests, Kruskal-Wallis analysis, multiple comparison test was applied on the mean PLI values. Results: The overall mean PLI values were significantly higher (p < 0.05) compared to the baseline in P1-P4 segments in delta and theta band, and in the P1 segment for the alpha band. Interestingly, P1 had higher mean PLI in the ripple band than P4. The synchronous and asynchronous offset were similar in the delta and theta band, but in the alpha, beta, gamma bands, the asynchronous mean PLI was higher at P1-P3 epochs. The emergence of theta or alpha frequencies were spatially heterogeneous and temporally the last within the SOZ. The PSD in the delta band was significantly (p < 0.05) higher after the synchronous termination compared to the asynchronous. This rate is reversed in the theta, alpha, beta, low gamma band, where the PSD is significantly higher (p < 0.05) in all the time segments. Conclusions: Following seizure termination, patterns of synchronization differentiated between asynchronous and synchronous offset. For synchronous offset, synchrony in the delta and theta band while for asynchronous offset synchrony and frequency power in beta and low gamma bands were higher. These changes in the synchrony and frequency power after the seizure termination could lead us to characterize the level of excitability across the brain areas. Further analysis needed to investigate the connection between the frequency synchronization and the seizure duration, seizure subtype, or post ictal EEG and behavioral state. Funding: Supported by NSF grant OIA 1632891