Abstracts

Rationale: Subtraction ictal SPECT coregistered with MRI (SISCOM) is a valuable presurgical tool in patients with medically refractory epilepsy. The complex perfusion patterns observed in SISCOM is suggested to reflect the ictal onset and propagation pathways. While the utility of SISCOM perfusion changes in localizing the ictal onset areas has been widely discussed, very few studies have investigated the electrophysiological correlate of the ictal perfusion patterns [1]. The aim of this study is to understand the neural correlate of ictal perfusion changes in a consecutive series of 95 patients who underwent ictal and interictal SPECT as part of their presurgical evaluation and later underwent stereo-EEG (SEEG) evaluation between 01/2016 and 12/2016. Methods: To evaluate the relationship between SISCOM and SEEG seizures, semiological changes during ictal SPECT seizures were compared with semiological changes during SEEG seizures. For every patient, a two minutes baseline segment of interictal SEEG, at least two hours preceding a seizure, was identified. Patients were excluded from the analysis if (1) SPECT injection was undertaken during the postictal phase, (2) poor correlation between semiology during ictal SPECT and SEEG seizures, (3) non-availability of raw SEEG baseline and ictal episodes and (4) poor quality of SEEG data. SEEG seizure and interictal baseline data were downloaded and analyzed using Matlab. The SISCOM map was co-registered to the SEEG implantation map and the SISCOM z-score for each SEEG contact was extracted. Spectral decomposition for the baseline and ictal SEEG data was performed using Welch periodogram. For every contact, we estimated excitation and suppression power in six EEG frequency bands as illustrated in Fig 1A [1]. Basically, excitation and suppression can be visualized as the increase and decrease, respectively, in total power across different frequency bands for the entire seizure compared to baseline.  For every patient, excitation and suppression power across channels were compared with SISCOM z-score using cross-correlation analysis. Statistical significance of the cross-correlation coefficient was accessed using students t-test. False discovery rate was used to correct for multiple comparisons. Results: A total of 70 patients met the specific inclusion criteria (Table 1). Fig. 1B shows the results of the cross-correlation analysis for the six frequency bands. SISCOM z-score was positively and statistically significantly correlated (p < 0.05) with the excitation power for the alpha to high gamma frequency band and with the suppression power in the delta frequency band (Fig 1B). Similar results were observed when thresholded SISCOM maps were used (See Fig. 1C), i.e. hyperperfused areas (z≥1.5) had a higher excitation power in the alpha to high gamma frequency bands when compared to hypoperfused areas (z≤-1.5). Conclusions: One of the principal conclusions of this study is that hyperperfused brain areas have increased excitation of activities in alpha to high gamma frequency range when compared to hypoperfused areas. Further, hyperperfused areas have increased suppression of activities in delta frequency range. This denotes that hyper and hypoperfused areas have different EEG power distribution in higher frequency bands during seizures. The study demonstrates that SISCOM perfusion patterns correlate with electrophysiological changes as observed via SEEG and therefore contain useful information for generating an implantation hypothesis.[1] Tousseyn S, Krishnan B, Wang ZI, Wongwiangjunt S, Nayak CS, Mosher JC, et al. Connectivity in ictal single photon emission computed tomography perfusion: a cortico-cortical evoked potential study. Brain 2017;42:282–6. Funding: None