Abstracts

Rationale: Determining ongoing seizure risk in children with benign epilepsy with centrotemporal spikes (BECTS) remains a significant challenge. 15% of children will have only one seizure, where others may have over 100 seizures over the course of several years. To date no clinical factors reliably predict disease course. Seizures in BECTS present during a time of maturational changes in cortical physiology, which can be measured from EEG recordings. Increased variability in beta band rhythms have been identified in BECTS. Oscillations in this frequency band are modulated by state of consciousness and thought to reflect intrinsic GABA-mediated inhibitory mechanisms. Here we evaluated whether power in the beta frequency band is associated with seizure risk in children with BECTS. Methods: 22 children with BECTS (15 M, 7 F ages 8 - 15) were recruited for this study. EEG data was collected with a 70-channel electrode cap, sampling rate of 2035 Hz and manually reviewed to identify epochs of wake and sleep (N1/N2). Medication status and most recent seizure from EEG were recorded at the time of visit.EEG data were visually inspected and movement, muscle, and electrode artifacts removed. A minimum of 1.5 minutes (wake: mean 196 sec, range 134 - 200 sec; sleep: mean 189 sec, range 103 - 200 sec) of artifact-free data from each patient was used for analysis of each arousal state. Data were re-referenced to the bipolar montage and power for each state was calculated with non-overlapping 1-second windows using a Hanning taper giving ~1 Hz frequency resolution. Spikes were manually marked by an epileptologist. Time intervals that overlapped with +/- 50 ms around the spikes were not used. The average power values in the beta band (14.9 - 30 Hz) at bipolar channel pairs over the perirolandic areas (C3-C5 and C4-C6) were computed for each available state and averaged together to achieve a single measure of beta band power per subject per state. If one bipolar channel electrode contained artifacts, only one bipolar pair was used for that subject. Average beta power values were log transformed and compared across groups using a 2-tailed student t-test. The relationship between beta power and time from most recent seizure was evaluated using linear regression, controlling for age. Results: We found a significant increase in beta rhythm power in children with BECTS who had had a recent seizure (within 12 months) compared to those who had been seizure free for >12 months during wake (p=0.03) and sleep (p=0.009, Fig 1). We found a relationship between beta power in the perirolandic regions and time from last seizure in wake (p=0.02, n=18) and sleep states (p=0.01, n=18, Fig 2). There was a trend between age and beta power in wake (p=0.09) and sleep (p=0.07). After correcting for age, the relationship between beta power and time to last seizure remained significant in sleep (p=0.05) but not wake (p=0.2). Conclusions: Focal beta power in the perirolandic regions during sleep may provide an easy to compute biomarker for seizure risk in children with BECTS.  Funding: NINDS K23-NS092923