Abstracts

Rationale: Sleep slow wave activity (SWA) is thought to reflect local synaptic strength secondary to synaptic potentiation during wake, and is homeostatically regulated both globally and locally (Tononi and Cirelli, Neuron 2014). A previous high-density electroencephalography (HD-EEG) study in patients with focal epilepsy identified widespread, bilateral increases in non rapid eye movement (NREM) sleep SWA that correlated with seizure frequency during wake; in addition, interictal spike frequency during wake was correlated with focal, unilateral increases in NREM sleep SWA. Here we sought to assess the localizing value of SWA during NREM sleep, as compared to rapid eye movement (REM) sleep, in a patient with reflex epilepsy. Methods: A 61-year-old woman with drug refractory reflex epilepsy was recruited from the University of Wisconsin’s Epilepsy Monitoring Unit. The patient presented with right facial motor seizures triggered by eating. Clinical 10-20 EEG revealed left frontal spikes; several seizures were also captured but were uninterpretable due to movement artifacts. The patient had a normal MRI. An overnight recording with 256 electrode HD-EEG was conducted on the patient and compared to those of 10 healthy volunteers. In each subject, five minute epochs of clean NREM and REM sleep were extracted for preprocessing. Using Matlab software, data was filtered from 1 to 40 Hz, and after rejection of bad channels and noisy epochs, independent component analysis was conducted to remove eye movement, cardiac, and muscle artifacts. Root mean squares of delta (1-4 Hz) and theta (4-8 Hz) band activities were computed to estimate delta and theta power at each channel. After spatial normalization via Z scoring, Statistical Parametric Mapping software was utilized to conduct T tests on consistent differences between the patient and each of the controls, using a random effects approach. Delta and theta power topography differences were assessed within NREM sleep, REM sleep, and between REM and NREM sleep. Results were thresholded at p < 0.05 corrected for multiple comparisons using family-wise error rate. Results: Compared to controls, the patient displayed increased delta and theta power in bilateral centro-temporal regions during NREM sleep, but not during REM sleep. The maximum increase in delta power during NREM sleep presented within the right temporal areas, and for theta power within the left centro-temporal areas. Direct comparisons between NREM and REM sleep revealed focal increases in both delta and theta power in left centro-temporal scalp areas. Conclusions: We observed both widespread and focal increases in NREM sleep SWA in a patient with reflex epilepsy as compared to 10 healthy controls. The increase in bitemporal delta power during NREM sleep confirms our findings from a previous study (Boly et al., Brain 2017), further suggesting that a selective increase in delta power during NREM sleep but not REM sleep may constitute a marker of epileptic disease. The maximum focal increase in theta power during NREM sleep aligned with functional localization of the patient’s ictal semiological findings. This greater localizing value of theta compared to delta power within NREM sleep - likely capturing local, small amplitude slow waves - suggests that analysis of topographical peaks in this frequency band may be of practical clinical value. Direct comparison of NREM versus REM sleep topographies may also provide useful confirmatory findings. Future studies will attempt to further validate this approach by comparing ictal EEG topographies to NREM and REM sleep topographies in a larger cohort of patients with focal epilepsy. Funding: NINDS 1R03NS096379 to M.B.