Sleep in non-human animals refers to how the behavioral and physiological state of sleep, mainly characterized by reversible unconsciousness, non-responsiveness to external stimuli, and motor passivity, appears in different categories of animals.
Rats kept from sleeping die within a couple of weeks, but the exact function of sleep is still unknown.
Sleep can follow a physiological or behavioral definition. In the physiological sense, sleep is a state characterized by a reversible unconsciousness, special brainwave patterns, sporadic eye movement, loss of muscle tone (possibly with some exceptions; see below regarding the sleep of birds and of aquatic mammals), and a compensatory increase following deprivation of the state. In the behavioral sense, sleep is characterized by non-responsiveness to external stimuli, the adoption of a typical posture, and the occupation of a sheltered site, all of which is usually repeated on a 24 hour basis. The physiological definition applies well to birds and mammals, but in other animals (whose brain is not as complex), the behavioral definition is more often used. In very simple animals, behavioral definitions of sleep are the only ones possible, and even then the behavioral repertoire of the animal may not be extensive enough to allow distinction between sleep and wakefulness.
Sleep as a phenomenon appears to have very old evolutionary roots. The nematode C. elegans is the most primitive organism in which sleep-like states has been observed. Here, a lethargus phase occurs in short periods preceding each moult, a fact which may indicate that sleep primitively is connected to developmental processes. Raizen et al.'s results furthermore suggest that sleep is necessary for changes in the neural system.
The electrophysiological study of sleep in small invertebrates is complicated. However, even such simple animals as fruit flies appear to sleep, and systematic disturbance of that state leads to cognitive disabilities. There are several methods of measuring cognitive functions in fruit flies. A common method is to let the flies choose whether they want to fly through a tunnel that leads to a light source, or through a dark tunnel. Normally, flies are attracted to light. But if sugar is placed in the end of the dark tunnel, and something the flies dislike is placed in the end of the light tunnel, the flies will eventually learn to fly towards darkness rather than light. Flies deprived of sleep require a longer time to learn this and also forget it more quickly. If an arthropod is experimentally kept awake longer than it is used to, then its coming rest period will be prolonged. In cockroaches that rest period is characterized by the antennae being folded down and by a decreased sensitivity to external stimuli. Sleep has been described in crayfish, too, characterized both by passivity and increased thresholds for sensory stimuli as well as changes in the EEG pattern, markedly differing from the patterns found in crayfish when they are awake.
Sleep in fish is not extensively studied. Some species that always live in shoals or that swim continuously (because of a need for ram ventilation of the gills, for example) are suspected never to sleep. There is also doubt about certain blind species that live in caves. Other fishes seem to sleep, however. For example, zebrafish, tilapia, tench, brown bullhead, and swell shark  become motionless and unresponsive at night (or by day, in the case of the swell shark); Spanish hogfish and blue-headed wrasse can even be lifted by hand all the way to the surface without evoking a response. A 1961 observational study of approximately 200 species in European public aquaria reported many cases of apparent sleep. On the other hand, sleep patterns are easily disrupted and may even disappear during periods of migration, spawning, and parental care.
Reptiles have been subjected to electrophysiological studies of sleep. That is to say that electrical activity in the brain has been registered when the animals have been asleep. However, the EEG pattern in reptilian sleep differs from what is seen in mammals and other higher animals. In reptiles, sleep time increases following sleep deprivation, and stronger stimuli are needed to awaken the animals when they have been deprived of sleep as compared to when they have slept normally. This suggests that the sleep which follows deprivation is compensatorily deeper.
There are significant similarities between sleep in birds and sleep in mammals, which is one of the reasons for the idea that sleep in higher animals with its division into REM and non-REM sleep has evolved together with warm-bloodedness. Birds compensate for sleep loss in a manner similar to mammals, by deeper or more intense SWS (Slow-wave sleep).
Birds have both REM and non-REM sleep, and the EEG patterns of both have similarities to those of mammals. Different birds sleep different amounts, but the associations seen in mammals between sleep and variables such as body mass, brain mass, relative brain mass, basal metabolism and other factors (see below) are not found in birds. The only clear explanatory factor for the variations in sleep amounts for birds of different species is that birds who sleep in environments where they are exposed to predators have less deep sleep than birds sleeping in more protected environments.
A peculiarity that birds share with aquatic mammals, and possibly also with certain species of lizards (opinions differ about that last point), is the ability for unihemispheric sleep. That is the ability to sleep with one cerebral hemisphere at a time, while the other hemisphere is awake. When only one hemisphere is sleeping, only the contralateral eye will be shut; that is, when the right hemisphere is asleep the left eye will be shut, and vice versa. The distribution of sleep between the two hemispheres and the amount of unihemispheric sleep are determined both by which part of the brain has been the most active during the previous period of wake—that part will sleep the deepest—and it is also determined by the risk of attacks from predators. Ducks near the perimeter of the flock are likely to be the ones that first will detect predator attacks. These ducks have significantly more unihemispheric sleep than those who sleep in the middle of the flock, and they react to threatening stimuli seen by the open eye.
Opinions partly differ about sleep in migratory birds. The controversy is mainly about whether they can sleep while flying or not. Theoretically, certain types of sleep could be possible while flying, but technical difficulties preclude the recording of brain activity in birds while they are flying.
Different animals sleep different amounts. Some animals, such as bats, sleep 18–20 hours per day, while others, including giraffes, sleep only 3–4 hours per day. There can be big differences even between closely related animals. There can also be differences between laboratory and field studies: for example, researchers in 1983 reported that captive sloths slept nearly 16 h a day, but in 2008, when miniature neurophysiological recorders were developed that could be affixed to wild animals, sloths in nature were found to sleep only 9.6 h a day.
As for birds, the main rule for mammals (with certain exceptions, see below) is that they have two essentially different stages of sleep: REM and non-REM sleep (see above). An animal's feeding habits are associated with its sleep length. The daily need for sleep is highest in carnivores, lower in omnivores and lowest in herbivores. Humans do not sleep unusually much or unusually little compared to other animals, but we sleep less than many other omnivores. Many herbivores, like Ruminantia (such as cattle), spend much of their wake time in a state of drowsiness, which perhaps could partly explain their relatively low need for sleep. In herbivores, a direct correlation is apparent between body mass and sleep length; big animals sleep more than smaller ones. This correlation is thought to explain about 25% of the difference in sleep amount between different animals. Also, the length of a particular sleep cycle is associated with the size of the animal; on average, bigger animals will have sleep cycles of longer durations than smaller animals. Sleep amount is also coupled to factors like basal metabolism, brain mass and relative brain mass.
Mammals born with well-developed regulatory systems, such as the horse and giraffe, tend to have less REM-sleep than the species which are less developed at birth, such as cats and rats. This appears to echo the greater need for REM-sleep among newborns than among adults in most mammal species.
Since monotremes, egg-laying mammals, are considered to represent one of the evolutionarily oldest groups of mammals, they have been subject to special interest in the study of mammalian sleep. As early studies of these animals could not find clear evidence for REM sleep, it was initially assumed that such sleep did not exist in monotremes but developed after the monotremes left the rest of the mammals and became a separate, distinct group. However, EEG registrations of the brain stem in monotremes show a firing pattern that is quite similar to the patterns seen in REM sleep in higher mammals. In fact, the largest amount of REM sleep known in any animal is found in the platypus.
Among others, seals and whales belong to the aquatic mammals. Seals are grouped in earless seals and eared seals, which have solved the problem of sleeping in water differently. Eared seals, like whales, show unihemispheric sleep. The sleeping half of the brain does not awaken when they surface to breathe. When one half of a seal's brain shows slow wave sleep, the flippers and whiskers on its opposite side are immobile. While in the water, these seals have almost no REM sleep and may go a week or two without it. As soon as they move onto land they switch to bilateral REM and NREM sleep comparable to land mammals, surprising researchers with their lack of "recovery sleep" after missing so much REM.
Earless seals sleep bihemispherically like most mammals, under water, hanging at the water surface or on land. They hold their breath while sleeping under water, and wake up regularly to surface and breathe. They can also hang with their nostrils above water and in that position have REM sleep, but they do not have REM sleep underwater.
REM sleep has been observed in the pilot whale, a species of dolphin. Whales do not seem to have REM sleep, nor do they seem to have any problems because of this. One reason REM sleep might be difficult in marine settings is the fact that REM sleep causes muscular atony; that is to say, a functional paralysis of skeletal muscles that can be difficult to combine with the need to breathe regularly.
Unihemispheric sleep refers to sleeping with only a single cerebral hemisphere. The phenomenon has been observed in birds and aquatic mammals, as well as in several reptilian species (the latter being disputed: many reptiles behave in a way which could be construed as unihemispheric sleeping, but EEG studies have given contradictory results). Reasons for the development of unihemispheric sleep are likely that it enables the sleeping animal to receive stimuli, threats, for instance, from its environment, and that it enables the animal to fly or periodically surface to breathe when immersed in water. Only non-REM sleep exists unihemispherically, and there seems to exist a continuum in unihemispheric sleep regarding the differences in the hemispheres: in animals exhibiting unihemispheric sleep, conditions range from one hemisphere being in deep sleep with the other hemisphere being awake to one hemisphere sleeping lightly with the other hemisphere being awake. If one hemisphere is selectively deprived of sleep in an animal exhibiting unihemispheric sleep (one hemisphere is allowed to sleep freely but the other is awoken whenever it falls asleep), the amount of deep sleep will selectively increase in the hemisphere that was deprived of sleep when both hemispheres are allowed to sleep freely.
The neurobiological background for unihemispheric sleep is still unclear. In experiments on cats, where the connection between the left and the right halves of the brain stem is severed, the brain hemispheres show a desynchronized EEG where the two hemispheres can sleep independently of each other. In these cats, the state where one hemisphere slept non-REM and the other was awake, as well as one hemisphere sleeping non-REM with the other state sleeping REM were observed. Interestingly, the cats were never seen to sleep REM sleep with one hemisphere while the other hemisphere was awake. This is in accordance with the fact that REM sleep, as far as is currently known, does not occur unihemispherically. (See also Unihemispheric slow-wave sleep.)
Animals that hibernate are in a state of torpor, differing from sleep. Hibernation markedly reduces the need for sleep, but does not remove it. Hibernating animals end their hibernation a couple of times during the winter so that they can sleep.