Abstracts

Rationale:
  Synaptic potentiation during wake translates into an increase in slow wave activity (SWA) during sleep, while sleep SWA power exponentially declines overnight as a reflection of sleep-dependent synaptic downscaling. In a previous HDEEG study, we found increased SWA at locations matching the seizure focus; increased HDEEG sleep SWA was also correlated with seizure frequency. To further investigate causal links between sleep SWA and epileptic activity, we analyzed ictal recordings and interictal sleep of patients implanted with intracranial macro-electrodes. In particular, we wished to investigate relationships between ictal phase-locked high gamma (PLHG) - a proxy for neuronal firing - and increased sleep SWA.      
Method:
A total of 6 patients were included in the analysis: 3 patients with depth electrodes (59, 24, and 16 electrodes in cingulate or hippocampal areas) and 3 patients with a combination of grid and strip intracranial subdural electrodes (55, 94, and 80 electrodes; 1 with poorly localized seizures, and 1 without seizures – fronto-temporal and parietal areas). For each patient, a whole night of recordings without seizure was selected and manually sleep scored (using sleep identified by increased delta/beta ratio and presence of high amplitude slow waves and spindles). Interictal spikes were identified by an epileptologist. After bad channel and epoch rejection and rejection of 1-second epoch centered around each spike, we computed the power for delta and beta bands and averaged delta power and delta/beta ratio over the first and last hour of sleep in each electrode. Delta power and delta/beta ratio values within the first and last hour and their difference overnight were Spearman-correlated to ictal PLHG.         Results  Visual inspection of power data distribution revealed that the highest delta power was most consistently found in the clinically identified seizure onset zone. The average correlation between power during sleep and ictal PLHG was positive in each of the 5 patients with seizures (reaching significance at p< 0.05 at the individual level in the 3 patients with the most frequent seizures). Delta power and delta/beta ratio showed an average decline across all channels in 4/5 patients (the patient with the net overnight increase was also the patient with the most frequent seizures). In the 4 patients where seizure focus could be localized, a positive correlation was observed between ictal PLHG and a paradoxical increase in delta/beta ratio overnight (reaching intra-individual significance in the 2 patients with most frequent seizures). In contrast, overnight change in delta/beta ratio was negatively correlated with ictal PLHG in the patient where the focus could not be identified.