Abstracts

Rationale: A widespread increase in delta activity has been found in patients with focal epilepsy compared to controls during NREM sleep, which was maximum at the seizure focus and correlated with spike frequency during preceding wake (Boly et al., Brain 2017). In both animal and human studies, local increases in slow wave activity (SWA) have also been found during wake after extreme sleep deprivation and/or task over-training. Here we hypothesized that we may also identify some local sleep during wake in patients with focal epilepsy, localized to the area of the seizure focus. As REM sleep is thought to be seizure-protective, we also hypothesized that the seizure focus would not be as consistently identifiable in that state. Methods: Seven patients with temporal lobe epilepsy (TLE, left [n= 6], right [n= 1]) were recruited at the Epilepsy Monitoring Unit of the University of Wisconsin and compared to nine age-matched controls. We also analyzed datasets from two patients with generalized epilepsy. Participants underwent overnight high-density EEG recordings with 256 electrodes. For each dataset, epochs of REM sleep and quiet wake were extracted and filtered from 1 to 25 Hz. Bad channels and noisy epochs were rejected using custom Matlab routines. Independent component analysis was performed to remove muscle and eye movements contaminating the recordings. After bad channel interpolation, SWA (delta (1-4 Hz) and theta (4-8 Hz)) power were computed at each electrode. Power topographies were then spatially normalized using Z scoring and converted to 64x64 2D images. Two sample t-tests were performed using Statistical Parametric Mapping to search for group differences between patients with temporal lobe epilepsy and controls. One sample tests were also performed between each patient with focal or generalized epilepsy and the set of controls to assess for the location of maximum differences in wake and REM, and compared to clinical information about seizure type. All results were corrected for multiple comparisons using family-was error rate and considered significant at cluster-corrected p<0.05 for group analysis, and peak-corrected p<0.05 for individual analyses.   Results: Compared to controls, patients exhibited an increase in delta and theta power in L temporal lobe during wake. In contrast, when comparing wake to REM, focal epilepsy patients exhibited an increase in theta power in R temporal lobe. Local SWA increase in wake and REM in each patient compared to the control group was lateralized to the seizure focus in 6/7 TLE patients (in wake) and 6/7 TLE patients (in REM) using delta power, and 7/7 TLE patients (in wake) and 6/7 TLE patients (in REM) using theta power. Interestingly, patients with generalized epilepsy also showed increases in SWA compared to controls, but these changes were maximum in the midline.  Conclusions: We observed focal increases in SWA in wake in left temporal lobe, the location of the seizure focus for the majority of our TLE patient sample. This increase in power appeared more localized than the increase in power we had previously observed during NREM sleep (unilateral only). In contrast, increase theta power was found when comparing wake to REM on the contralateral side compared to the focus. Best localizing power was obtained for focal theta power during wake (7/7 correct lateralization). Interestingly, our two patients with generalized epilepsy also showed increased SWA during wake but predominantly in the midline. These preliminary results suggest the potential usefulness of considering state-dependent changes in SWA across sleep and wake in patients with focal epilepsy to localize the seizure focus.    Funding: Supported by NINDS 1R03NS096379 and Tiny blue Dot Foundation