Abstracts

Rationale: Epileptic spasms are brief seizures typically seen during infancy and young childhood, and often take place in clusters at intervals of several to tens of seconds. Ictal electrocorticography (ECoG) change during each of epileptic spasms is characterized by widespread augmentation of high-frequency oscillations (HFOs) at >80 Hz superimposed on slow activity at < 1 Hz. Here, we conducted a computational ECoG analysis, in an even-by-event manner, to assess the origin, propagation, and termination of spasm-related HFOs. Methods: Extraoperative ECoG signals were acquired using Nihon Kohden Neurofax 1100A Digital EEG Acquisition System. The sampling rate was 1,000 Hz, and the amplifier band-pass filter range was 0.016 - 300 Hz. Epileptic spasms were identified by visual analysis of the ECoG recordings. Quantitative ECoG analysis was performed using a custom code developed in EEGLAB (Swartz Center for Computational Neuroscience, UC San Diego).  At each electrode site, we measured the amplitude of HFO>80 Hz within a 100-ms window sliding every 1 ms. We determined if the quantitative analysis delineated early augmentation of HFOs at the visually-defined seizure onset zone (SOZ) as well as subsequent propagation of ictal HFOs to the surrounding sites. We also determined if termination of HFO augmentation also occurred early at the SOZ and propagated to the surrounding areas. We finally determined if HFOs were suppressed immediately following the offset of spasms. Results: Our analytic software successfully delineated focal augmentation of HFOs at the SOZ determined by visual assessment. Within 500 ms, ictal augmentation of HFOs propagated to widespread regions in the affected hemisphere. Ictal HFO augmentation outside the SOZ often lasted longer than that within the SOZ. Subsets of spasms showed postictal HFO attenuation in the cortical sites involved by ictal HFO augmentation; thereby, the duration of such postictal attenuation was up to twice longer than that of ictal augmentation. Conclusions: This pilot study demonstrated the technical feasibility of time-frequency analysis of ictal and postictal HFOs on an event-by-event basis. Our analysis replicated the "ictal doughnut phenomenon" characterized by disappearance of ictal HFOs earlier in the SOZ compared to that in the surrounding sites. Lingering postictal HFO suppression may reflect neural fatigue of the involved brain region. Further investigation is needed to determine whether such HFO suppression reflects 'latent period' and is related to the event-to-event interval between spasms. Additionally, it remains to be understood if there is an association between postictal suppression of HFOs and the severity of cognitive dysfunction of given patients with epileptic spasms. Funding: None