Partial REM-sleep deprivation increases the dream-like quality of mentation from REM sleep and sleep onset.

STUDY OBJECTIVES Sleep onset (SO) is cognitively and physiologically similar to rapid eye movement (REM) sleep, supporting the notion that REM sleep-related processes are 'covertly' active at this time. The objective was to determine if SO mentation is sensitive to REM sleep deprivation. DESIGN Two-group cross-sectional design; sleep recordings for 3 nights. SETTING Standard sleep laboratory with 24-channel polysomnography recording. PARTICIPANTS Fourteen female, 13 male healthy volunteers (18-41 yrs, mean=24.8 +/- 6.07). INTERVENTIONS On Night 2, half were and half were not partially REM sleep-deprived (REMD), recalled REM mentation, and rated it for dream-like quality (DLQ), sleepiness, and sensory attributes. On Night 3, all were awakened from SO substages 4 and 5 for mentation reports and further ratings. REMD measures were derived from scored sleep tracings. MEASUREMENTS AND RESULTS REMD produced increases in DLQ for both REM and SO reports (P < .05); DLQ scores were higher for REM than for SO mentation (P < .001). Covarying sleepiness preserved the (REMD) effect but abolished the REM/SO difference. Whereas 2 sensory attributes (presence of self, visual intensity) tended to distinguish the REM-mentation reports of REMD and control subjects, only 1, self-movement, distinguished their SO mentation reports (P < .06). Multiple regression revealed that increased DLQ of both REM and SO mentation was associated with increased sleepiness and decreased REM sleep time on Night 2. CONCLUSIONS SO mentation responds to REMD much like REM mentation does, a finding consistent with other work supporting the notion of covert REM-sleep processes at SO. DLQ may be mediated by both increases in REM-sleep propensity and a circadian process indexed by sleepiness ratings.