Abstracts

Rationale: Interictal EEG epileptiform spikes are a well-recognized hallmark of epilepsy as they are present in nearly all cases, independent of etiology. Defined as brief (<250 ms) electrical discharges, spikes are the most useful biomarker to predict epilepsy after a first seizure. Although spikes are useful for diagnosing epilepsy and may have a role in increasing propensity for seizures, the relationship between spikes and epilepsy resolution is unclear. Further, the essential characteristics required to define a spike are controversial, and inter-rater agreement in identifying spikes is low. Benign epilepsy with centrotemporal spikes (BECTS) is a stereotyped self-limited electroclinical epilepsy syndrome characterized by distinct bilateral, independent, sleep activated spikes in the centrotemporal region on EEG. All patients with BECTS ultimately enter disease resolution, the likelihood of which correlates strongly with duration seizure free. We used objective techniques to characterize spike frequency and morphological features in a prospective, cross sectional cohort of children with BECTS to help clarify the relationship between spike rate and spike morphological characteristics over the course of disease resolution. If spike characteristics are found to change over time, longitudinal EEGs can be used to predict resolution. Further, a better understanding of the relationship between spike morphology and disease could provide insights into the critical characteristics that contribute to their pathology. Methods: Fifteen children diagnosed with BECTS (ages 8-14.9 y) were included. EEG were recorded with a 70 channel cap based on the 10-10 electrode placement system at a 2035 Hz sampling rate. Non-REM sleep was selected for analysis and all data were referenced to the average reference. To identify spikes, we applied the Persyst 13 spike detector (Persyst Development Corporation, San Diego). Detections were visualized and detections that coincided with clear artifacts or normal sleep architecture were removed. All other detections in any of the central or temporal electrode derivations were included in the analysis. Spike rate and three morphological measurements (mean duration, mean height, and mean angle) were evaluated. We also evaluated whether spike morphological features became more stereotyped over disease resolution (e.g. the intra-individual variance in each of the morphological measures). We compared these measures to duration seizure-free, which we have previously shown to be a reliable proxy for likelihood of entering terminal resolution in BECTS, using a regression model, including age as a covariate. Results: Spike frequency (p-value of 0.011, aR2=0.32) and spike height (p=0.026, aR2=0.49) negatively correlated with the duration seizure-free. Spike duration (p=0.062, aR2=0.11) trended to correlate with duration-seizure free. There was no relationship between spike angle and duration seizure free (p=0.92, mean=19.95, range=9.7-35.7 degrees). There was no relationship between the variance of spike height (p=0.11), spike duration (p=0.12), or spike angle (p=0.23) and duration seizure free. Conclusions: We evaluated a cohort of children with BECTS, a common, transient developmental epilepsy syndrome characterized by epileptiform discharges in the centrotemporal regions. We found that spike frequency, duration, and height each decrease as subjects enter epilepsy resolution and longitudinal changes in these measures may help predict resolution. Overall, these data suggest that spikes are more frequent, taller, and last longer during the active disease in BECTS. Overall, these data also demonstrate that spikes evolve over a continuous morphological spectrum with variable height and duration within and across subjects, which likely contributes to the poor inter-rater agreement in their detection. The consistently sharp angle (< 36 degrees) observed in spikes over the course of disease suggests this feature may be a reliable characteristic for spike detection. Funding: K23-NS092923