Abstracts

Rationale: It is currently theorized that the ictal wavefront, a slow-moving wave of tonic neuronal firing, is responsible for the recruitment of cortical areas into a seizure by producing low frequency synaptic barrages that synchronize neurons ahead of and behind it [1-2]. Although easily identified in multielectrode recordings, this wavefront is difficult to characterize in raw electrocorticography (ECoG) traces of seizures; however, it may be possible to indirectly observe the wavefront’s passing in ECoG via its low frequency effects on neuronal populations. Methods: Here we used a short-time Fourier transform (STFT) spectral analysis of ECoG recordings from patients undergoing intracranial seizure monitoring to identify the low-frequency (0.5 – 16 Hz) power changes that occur during focal to bilateral tonic-clonic seizures (F2BTCS). Using an automated contour detection algorithm, we then extracted the dominant power contour from the channel-averaged spectrograms of F2BTCSs, allowing us to characterize low frequency power changes throughout a seizure’s evolution. Results: In 10 seizures from 6 patients, the average raw power spectrogram during an F2BTCS consistently showed a maximum power contour initiating in the 4-8 Hz range, approximately at the same time as the ramp in voltage observed in the channel-averaged voltage trace. On the whole, this power contour stabilized in frequency initially, but slowly and consistently decreased in frequency during post-recruitment, reaching the 1-2 Hz range at termination. Extracting these dominant power contours from the channel-averaged spectrograms, we observed that power localizes in frequency along this decreasing contour initially during the voltage ramp at recruitment, de-localizes in frequency during post-recruitment, and re-localizes along this contour as termination approaches. Individual channels demonstrated the same power localization trends observed on average during seizure initiation and approaching termination; however, in the post-recruitment epoch, channels demonstrated variability in which time points displayed power localization and de-localization along the power contour. Conclusions: This work shows that at seizure initiation, low frequency power is concentrated in the same frequency range as the spike rate of synchronized neurons prior to and following ictal wavefront passage [1-2]; this suggests that the timing of ictal wavefront passage may be able to be defined in ECoG channels spectrally. Further, this work supports previous research regarding synchrony in seizures, suggesting that spatially distributed cortex during F2BTCSs on average share a dominant frequency of oscillation per time point, especially at seizure initiation and as termination approaches [3]. Conversely, it also shows desynchronization of individual electrodes from this dominant frequency at various time points in post-recruitment [3]. Further work will be done to corroborate these findings in a complex partial seizure cohort and to use low frequency power contour features of individual channels to identify important seizure landmarks, allowing us to gain insights into how seizures spread in cortical and subcortical networks. References: [1] Weiss SA, et al. (2013). High frequency oscillations distinguish two types of seizure territories in humans. Brain 136: 3796–3808. [2] Smith EH, et al. (2016). The ictal wavefront is the spatiotemporal source of discharges during spontaneous human seizures. Nat Comm 7:11098. [3] Jiruska P, et al. (2013). Synchronization and desynchronization in epilepsy: controversies and hypotheses.  J Physiol 591(4): 787–79. Funding: NIH Medical Research Scholars Program, NIH Intramural Research Program