Abstracts

Rationale: Phase amplitude coupling (PAC) between intracranially recorded interictal high frequency oscillations (HFOs) and slow wave activity (SWA) was demonstrated as a potential biomarker of epileptogenicity in a recent study of children with refractory epileptic spasms (ES) (Iimura et al, Epilepsia 2017). Recent studies have demonstrated the detectability of interictal scalp HFOs including ripples (R) and fast ripples (FR). The dynamics of PAC of interictal scalp R and FR have yet to be characterized. We compared the modulation index (MI) quantifying PAC in pre-treatment EEG between children with treatment responsive vs treatment refractory ES. Methods: We retrospectively identified 13 children with diagnosis of epileptic spasms who were found to have HFOs (R and FR) on scalp EEG recorded at 2000 Hz during sleep obtained at initial ES evaluation at our center. Between 10 to 60 minutes of EEG not containing seizure were obtained for each patient. Scalp HFOs were detected using a deep learning based automatic HFO detector as previously reported (Bernardo et al, Clin Neurophysiol 2018). The raw EEG, bandpass filtered EEG (100-600 Hz and 250-600 Hz), and Morlet spectrogram of all putative HFOs were visually inspected to exclude artifactual ‘false ripples’. We calculated the modulation index (MI) quantifying R (100-250 Hz) and FR (250-500 Hz) PAC strength with SWA at the 0.5-1 Hz, 1-2 Hz, 2-3 Hz, 3-4 Hz, and 4-8 Hz passbands using HFO oscillation triggered PAC (Dvorak et al, J Neurosci Methods 2014). The statistical significance of MI was determined via surrogate analysis with phase shuffling of observed samples. We evaluated the HFO occurrence rate (OR) and HFO MI on a per channel basis across all patients. Treatment response was defined as resolution of both ES and hypsarrhythmia with no relapse for at least 3 months. Treatment non-response was defined as persistent ES despite treatment for at least 3 months. Reported p values were adjusted for multiple comparisons with the Holm-Bonferroni method. Results: There were 7 treatment responders and 6 non-responders. Responders and non-responders were similar in respect to age at initial diagnosis (6.7 vs 6.4 mo), reported duration of ES at presentation (35 vs 56 days), and baseline HFO OR (0.11 vs 0.19 HFOs/min/channel). HFOs with PAC with SWA with significance determined by surrogate analysis with p<0.05 was found for both FR and R in 12 out of 13 patients, and in 48 out of 264 channels (18%) across these 12 patients (example of FR and FR PAC shown in Figure 1). FR MI with SWA at 0.5-1 Hz and 4-8 Hz were significantly higher in responders compared to non-responders (p=0.04 and p=0.01, Mann-Whitney U, Figure 2). Conclusions: This study demonstrates the feasibility of measuring PAC from interictal scalp HFOs. Stronger MI between FR and SWA was associated with treatment failure, which is consistent with prior ECoG studies demonstrating HFO PAC as a potential marker of epileptogenicity. Further investigation is warranted to determine the clinical utility of PAC of scalp HFOs. Funding: No funding was received in support of this abstract.