Experimental separation of time of day and homeostatic influences on sleep.

The purpose of the present study was to evaluate the simultaneous effects on sleep of prior time awake (PRW) and time of day (TOD). Eight male subjects spent 13 days in an isolated sleep lab and had three 8-h baseline sleeps and then 18 4-h sleeps, distributed to provide three sleeps starting at 2400, 0400, 0800, 1200, 1600, and 2000. The three sleeps were preceded by 4, 8, and 12 h of PRW, respectively. ANOVA showed that TST and subjective sleepiness increased with PRW and with closeness to the trough of the circadian rhythm of rectal temperature, whereas sleep latency showed the opposite pattern, and rapid eye movement sleep (REM) latency strongly decreased with PRW and with closeness to the trough. Slow-wave sleep (SWS) increased with PRW, whereas SWS latency and final time awake decreased. REM sleep increased with closeness to the circadian trough, and time awake decreased. Multiple-regression analysis showed that REM latency was closely related to increased SWS in the first sleep cycle, reduced SWS latency, and increased PRW [a short PRW before sleep at noon yielded an extremely short (14 min) REM latency]. Sleep latency and final time awake showed almost exactly the same relationship to TOD and PRW. It is concluded that both homeostatic and circadian influences simultaneously affect sleep, that REM latency is very sensitive to the need for SWS, and that the circadian acrophase strongly interferes with sleep. It should be emphasized that the conclusions should not be extrapolated to longer (>12 h) wake spans.