RESUMEN
Sleep is a prominent behaviour in the lives of animals, but the unresponsiveness that characterizes sleep makes it dangerous. Mammalian sleep is composed of two neurophysiological states: slow wave sleep (SWS) and rapid-eye-movement (REM) sleep. Given that the intensity of stimuli required to induce an arousal to wakefulness is highest during deep SWS or REM sleep, mammals may be most vulnerable during these states. If true, then animals should selectively reduce deep SWS and REM sleep following an increase in the risk of predation. To test this prediction, we simulated a predatory encounter with 10 wild-caught Norway rats (Rattus norvegicus), which are perhaps more likely to exhibit natural anti-predator responses than laboratory strains. Immediately following the encounter, rats spent more time awake and less time in SWS and REM sleep. The reduction of SWS was due to the shorter duration of SWS episodes, whereas the reduction of REM sleep was due to a lower number of REM sleep episodes. The onset of SWS and REM sleep was delayed post-encounter by about 20 and 100 min, respectively. The reduction of REM sleep was disproportionately large during the first quarter of the sleep phase, and slow wave activity (SWA) (0.5-4.5 Hz power density) was lower during the first 10 min of SWS post-encounter. An increase in SWA and REM sleep was observed later in the sleep phase, which may reflect sleep homeostasis. These results suggest that aspects of sleep architecture can be adjusted to the prevailing risk of predation.
