Abstracts

Rationale: It is well demonstrated that interictal epileptic discharges (IEDs) in focal epilepsy are influenced by sleep, and in particular that rapid eye movement (REM) sleep has a suppressing effect on epileptic activity. The decrease in epileptic activity is assumed to be due to EEG desynchronization which is mainly mediated by cholinergic neurotransmission. Based on these considerations we investigated the occurrence of IEDs and high frequency oscillations – a novel biomarker of the epileptogenic zone – across both phasic and tonic REM sleep. We hypothesized that interictal events are even more suppressed during phasic REM sleep because of additional intermittent cholinergic drive.Methods: Twelve patients underwent polysomnography during long term intracerebral EEG recording. After sleep staging, segments of phasic REM sleep were identified and marked. An equal number of segments of matching duration was marked randomly during tonic REM sleep in the same sleep cycle. We computed the EEG power in frequencies up to 30 Hz and from 30 to 500 Hz in the tonic and phasic segments. In the same segments we marked IEDs and high frequency oscillations: ripples in the 80–250 Hz band and fast ripples in the 250–500 Hz band. We grouped the intracerebral channels into channels in the seizure onset zone (SOZ), channels in the irritative zone but outside the SOZ, and channels without any epileptic activity, presumably recording only physiologic activity. We compared the EEG power and the proportion of events in phasic and tonic segments, and tested whether the rate of ripples was significantly different in the three channel groups, as ripples might be of pathologic or physiologic origin.Results: Power in frequencies below 30 Hz was lower during phasic than tonic REM sleep (phasic to tonic power ratio lower than one, p<0.001, Bonferroni corrected), possibly reflecting increased desynchronization. All types of interictal events were significantly less frequent during phasic REM sleep compared to tonic REM sleep (39% of 2767 spikes, 35% of 3162 ripples, and 18% of 708 fast ripples (Fig. 1). In contrast to ripples in channels with epileptic activity, presumably physiologic ripples were more abundant during phasic REM (73% of 336 ripples in normal channels, 30% of 816 in channels in the irritative zone but outside the SOZ, 28% of 2010 in the SOZ – in all cases the differences are highly significant, Bonferroni corrected p<0.001). During tonic REM sleep the mean ripple rate was significantly different among the three groups of channels (Bonferroni corrected p<0.001). During phasic REM sleep the difference between mean ripple rate in normal channels and channels in the irritative zone outside the SOZ was not statistically significant (Fig. 2).Conclusions: Phasic REM sleep has an enhanced suppressive effect on epileptic activity compared with tonic REM sleep, corroborating the role of EEG desynchronization in the suppression of epileptic activity. In contrast, physiological ripples were increased during phasic REM sleep, possibly reflecting REM-related memory consolidation and dreaming.