REM sleep occurs in cycles of about 90-120 minutes throughout the night, and it accounts for up to 20-25% of total sleep time in adult humans, although the proportion decreases with age (a newborn baby may spend 80% of total sleep time in the REM stage). In particular, REM sleep dominates the latter half of the sleep period, especially the hours before waking, and the REM component of each sleep cycle typically increases as the night goes on.
As the name suggests, it is associated with rapid (and apparently random) side-to-side movements of the closed eyes, a phenomenon which can be monitored and measured by a technique called electrooculography (EOG). This eye motion is not constant (tonic) but intermittent (phasic). It is still not known exactly what purpose it serves, but it is believed that the eye movements may relate to the internal visual images of the dreams that occur during REM sleep, especially as they are associated with brain wave spikes in the regions of the brain involved with vision (as well as elsewhere in the cerebral cortex).
Brain activity during REM sleep is largely characterized by low-amplitude mixed-frequency brain waves, quite similar to those experienced during the waking state – theta waves, alpha waves and even the high frequency beta waves more typical of high-level active concentration and thinking. These show up as the typical “saw-tooth” brain wave pattern on an electroencephalogram (EEG) and, because of these similarities with the waking state, REM sleep has earned the moniker “paradoxical sleep”. The brain’s oxygen consumption, reflecting its energy consumption, is also very high during this period, in fact often higher than when awake and working on a complex problem.
Breathing becomes more rapid and irregular during REM sleep than during non-REM sleep, and the heart rate and blood pressure also increase to near waking levels. Core temperature is not well regulated during this time and tends towards the ambient temperature, in much the same way as reptiles and other cold-blooded animals. Sexual arousal is also common during REM sleep and the male penis and female clitoris become aroused and erect for substantial periods during this sleep stage, regardless of whether or not any dreams in progress are of an erotic nature.
Although the muscles become more relaxed during non-REM sleep, they become completely paralyzed and unresponsive during REM sleep. This virtual absence of muscle tone and skeletal muscle activity is known as atonia, and it occurs because the brain impulses that control muscle movement are completely suppressed (other than those controlling the eye movements and one or two other essential functions, like the heart, diaphragm, etc, that allow us to breathe and remain alive). The source of these inhibitory signals (which utilize the neurotransmitter norepinephrine) is in a specific part of the pons region of brainstem called the locus coeruleus.
The majority of dreams – certainly the most memorable and vivid dreams – occur during REM sleep, and it is thought that the muscular atonia that accompanies it may be a built-in measure to protect us from self-damage which could occur while physically acting out these vivid REM dreams. This hypothesis is borne out by Michel Jouvet’s early experiments on cats in which the muscle inhibition nerves were severed, leading these cats to physically stalk invisible prey and run and jump around wildly during the dreams of REM sleep.
Neurologically, REM sleep is activated by secretion of the neurotransmitter acetylcholine and inhibited by the neurotransmitter serotonin, and this effect is principally generated in the pons region of the brainstem. In experiments on animals, it has been shown that the surgical destruction of a particular group of nerve cells in the pons can eliminate REM sleep completely, suggesting that the active function of these cells (rather than merely the deactivation of wakefulness mechanisms) is necessary for REM sleep.
Although lack of REM sleep leads to surprisingly few negative effects on behaviour, it has been shown to impair the ability to learn complex tasks, suggesting that REM sleep is a vital component of our sleep patterns, particularly during early childhood development, when REM sleep makes up a much larger percentage of total sleep. This is backed up by the fact that, if REM sleep is repeatedly interrupted or shortened, then longer REM “rebound sleep” tends to occur at the next opportunity in compensation (instead of slowly moving through the various stages of non-REM sleep first, the sleeper slips quickly into REM sleep, and stays there longer than usual).
Some memory consolidation, particularly of procedural and spatial memory, also takes place during this stage, although perhaps not to the same extent as during the deeper, later stages of non-REM sleep. It has been noted that people tend to spend more time than usual in REM sleep following days when they have been in unusual situations requiring them to learn a lot of new tasks.
Although most people do not tend to wake after each cycle of REM sleep, as some animals do, we are more likely to wake from REM sleep than from non-REM sleep. Usually, these “micro-awakenings” are of a few seconds only, and the sleeper does not normally remember them. If over-stimulated, though, a person may wake up fully, and it may take the length of an entire sleep cycle (1.5 – 2 hours) to get back to sleep.
Although it is usually assumed that REM sleep (and the dreams that go with it) is a physiological necessity, recent findings have muddied the waters somewhat. For example, in cases of REM sleep deprivation, individuals tend to compensate by dreaming more during non-REM sleep. Animals deprived of REM sleep for as long as two months seem to be able to continue with very little perceptible behavioural or physiological injury, and humans taking certain antidepressant medications that result in little or no REM sleep also show few apparent negative consequences.