© ISTOCK.COM/CAREBOTTSome sleep researchers are fond of saying that all animals sleep; that sleep is maladaptive because it takes time away from activities that appear more adaptive, such as mating, seeking food, and looking out for predators; and that no one knows the function of sleep. A good case can be made that each of these statements is false.

To say whether an animal sleeps requires that we define sleep. A generally accepted definition is that sleep is a state of greatly reduced responsiveness and movement that is homeostatically regulated, meaning that when it is prevented for a period of time, the lost time is made up—an effect known as sleep rebound. Unfortunately, the application of this definition is sometimes difficult. Can an animal sleep while it is moving and responsive? How unresponsive does an animal have to be? How much of the lost sleep has to be...

Apart from mammals, birds are the only other animals known to engage in both slow-wave and rapid eye movement (REM) sleep. Slow-wave sleep, also called non-REM sleep, is characterized by slow, high-amplitude waves of electrical activity in the cortex and by slow, regular respiration and heart rate. During REM sleep, animals exhibit a waking-like pattern of cortical activity, as well as physiological changes including jerky eye twitches and increased variability of heart rate and respiration. (See “The A, B, Zzzzs.”) But many more animals, including some insects and fish, engage in behaviors that might be called sleep, such as resting with slow but regular respiration and heart rates and a desensitization to environmental stimuli.

In addition to diversity in the neural and physiological correlates of sleep, species vary tremendously in the intensity, frequency, and duration of sleep. Some animals tend to nap intermittently throughout the day, while others, including humans, tend to consolidate their sleep into a single, long slumber. The big brown bat is the current sleep champion, registering 20 hours per day; giraffes and elephants doze less than four hours daily. What can account for such differences in sleep times? While brain size and brain-body weight ratio do not strongly correlate with the total amount of sleep or with the amount of REM sleep (Nature, 437:1264-71, 2005), persuasive evidence now links food intake to sleep duration. Herbivores such as the elephant and the giraffe must eat and chew constantly because of the low caloric density of their food. Not surprisingly, they have evolved to sleep relatively little. Lions, on the other hand, sleep long—for 14 hours or more—and deeply, especially after they have consumed prey. (See “Slumber Numbers.”)

Breathing also greatly influences sleep time. Because dolphins, whales, and certain other marine mammals need to surface every few minutes to breathe, they cannot have slow waves in both hemispheres of their brain at once. Some marine mammals and birds are able to remain continuously active for weeks with no apparent sleep rebound that would suggest the animals even get sleepy. Although it’s unclear if these animals experience unihemispheric slow waves during these periods of activity, I would argue that, if an animal is swimming, turning to avoid obstacles, and vocalizing, it can’t be asleep.

For animals that do sleep for significant periods of time, one approach to determining the function of sleep is to keep them awake. But because extended sleep deprivation requires repeated awakening, which in humans triggers a surge of cortisol throughout the body, it becomes very difficult to separate the effects due to the loss of sleep from those due to the stress associated with repeated arousal. An additional issue is the effect of interference with task recall resulting from the additional period of learning during the deprivation procedure. A recent, extensive study showed that human sleep parameters were unrelated to memory consolidation on all tasks examined, undermining a vast literature on memory consolidation as a major function of sleep (Sleep, 38:951-59, 2015).

So why do animals sleep? It may be in part simply to conserve energy. At rest, an awake human brain uses 20 percent of the energy consumed by the body, even though it is only about 2 percent of body weight. If a safe sleeping site is found, or if you are a top predator that has no fear of other predators, there is a huge benefit to reducing that 20 percent cost, a reduction that occurs during sleep (Nat Rev Neurosci, 10:747-53, 2009). Recent work by my team with “preindustrial” humans, living under the conditions in which our species evolved, supports this idea that sleep helps organisms save energy, finding that these hunter-gatherers consistently sleep during the coldest portion of the night, rather than their sleep being tightly linked to the light level (Curr Biol, 25:2862-68, 2015). In this way, sleep is on a continuum with hibernation or torpor. By reducing activity and body temperature when the daily temperatures are lowest, substantial energy savings may be achieved.

So it’s fair to say that not all animals sleep in the traditional sense, if they sleep at all (Trends Neurosci, 31:208-13, 2008). And while it is true that we don’t know all the functions of sleep, recent research provides strong support for the hypothesis that energy conservation—a clearly adaptive function—is an important factor. As we continue to study sleep and sleep-like states across the animal kingdom, from insects to mammals, let us keep an open mind about this still-enigmatic phenomenon.

Jerome Siegel is a professor of psychiatry and biobehavioral sciences at the University of California, Los Angeles.


© ISTOCK.COM/DENJA1For the last 15 or so years, scientists around the world have discovered that birds, which are more closely related to reptiles than mammals, exhibit sleeping brain activity similar to that of people (and mice and manatees), engaging in both REM and non-REM sleep. “There must be something about being a bird and being a mammal that causes their brains to need the same kinds of sleep,” says John Lesku, a lecturer at La Trobe University in Melbourne, Australia.

But there are also differences in the brain activity of sleeping mammals and birds, says Niels Rattenborg, who leads the avian sleep group at the Max Planck Institute for Ornithology in Seewiesen, Germany. Unlike mammalian REM sleep, which usually occurs in relatively few episodes that can span several minutes to an hour or more, avian REM sleep typically occurs in hundreds of brief spurts, each lasting just seconds (Curr Biol, 24:R12-R14, 2014).

While the functions of sleep in birds, as in mammals, have been hotly debated, Lesku and Rattenborg say the evidence implicating memory formation and storage is strong. “Several studies suggest that the brain rhythms we see during sleep are playing a role in processing information acquired during the previous day,” says Rattenborg. For example, neurons involved in song production are active in the brains of sleeping juvenile songbirds that learn tunes from adult tutors while awake during the day (J Neurophysiol, 96:794-812, 2006). And researchers have speculated that slow-wave sleep promotes a reduction in synapse strength, weakening less-important memories from the recent past, “so that, at the end of sleep, you’re ready for a new day of learning,” says Lesku.

Like many marine mammals, some birds are capable of unihemispheric sleep, in which one half of the brain remains alert while the other half exhibits electrical signals of slow-wave sleep. The eye connected to the alert half of the brain remains open, while the other typically closes. (REM sleep only occurs in both hemispheres at the same time.) Some bird species must get by on very little sleep altogether. Polygynous male pectoral sandpipers attempt to breed with as many females as possible during a 19-day breeding period, so they stay awake for about 95 percent of this time, Lesku says. “The sleep that they do have is packaged into hundreds of episodes just a few seconds long.”

Other species, such as bar-tailed godwits, will fly nonstop for days or weeks on end to migrate thousands of miles as seasons change. It’s not yet known whether migratory species sleep during flight, however. While no EEG recordings have been performed during flight to date, preliminary work has suggested that slow-wave sleep—even in both hemispheres at the same time—may be possible while on the wing. —Tracy Vence


© ISTOCK.COM/SARAWUTH123In the late 1990s, Joan Hendricks of the University of Pennsylvania School of Veterinary Medicine spent hours sitting in a dark, hot room tapping petri dishes that housed Drosophila. Whenever the flies settled down, Hendricks would disturb their rest, which typically lasted for seven to eight hours each night if left undisturbed. But eventually the flies stopped responding to her taps, to the point that Hendricks was sure they must be dead. But they weren’t dead; they were sleeping (Neuron, 25:129-38, 2000). “It was unbelievably compelling,” she recalls. “I’m bothering these poor flies, they’re clearly at risk by this giant perturbing their universe, and they fell asleep.”

On the other side of the country, Paul Shaw, then a postdoc in Giulio Tononi’s lab at the Neurosciences Institute in San Diego, was undertaking a very similar experiment. And just as Hendricks had observed, his flies exhibited a distinct sleep rebound—following deprivation, they fell asleep more easily and were harder to wake up (Science, 287:1834-37, 2000). “The behavior is just so obvious and so clear,” says Shaw, now an associate professor at Washington University in St. Louis. He, too, concluded that the flies were sleeping.

A lot of things are conserved through­out evolution, so we can use the fly to really understand the genes and molecules that regulate sleep.—Mark Wu,
Johns Hopkins Uni­versity

The Drosophila experiments sent shock waves through the sleep research community, which at that time held that only birds and mammals slept. But Hendricks’s and Shaw’s results were not unprecedented. As early as the 1980s, researchers such as Walter Kaiser and Jana Steiner-Kaiser of the Technische Universität (then Hochschule) Darmstadt in Germany and Irene Tobler of the University of Zurich had documented sleep-like behavior in honeybees, cockroaches, and scorpions. While the Drosophila studies did meet with some resistance, it wasn’t long before the field began to assume that most, if not all, animals sleep, and to recognize that many of the mechanisms regulating sleep-wake cycles are conserved from insects to humans. Even in those first studies, Hendricks’s group found that drugs known to affect sleep in mammals had similar effects in Drosophila, and Shaw and his colleagues observed that several molecular markers known to fluctuate with mammalian sleep-wake states also cycle in flies.

In the last 15 years, research on invertebrate sleep has exploded, with Drosophila now serving as a powerful model system to probe the function of sleep. In insects, as in mammals and birds, most researchers believe sleep serves an important role in memory formation and retention. In 2011, Shaw’s group found that stimulating a group of neurons in the fly brain called the dorsal fan-shaped body could put flies to sleep, and in doing so induce the formation of a long-term memory (Science, 332:1571-76, 2011). And last year Shaw and colleagues were able to reverse memory deficits in a Drosophila  strain that was learning-disabled, as well as in a fly model of Alzheimer’s disease, simply by inducing sleep (Curr Biol, 25:1270-81, 2015). “Somehow sleep is able to fix the brain,” Shaw says.

Research has also revealed the importance of sleep for learning and memory in honeybees, which must remember the locations of multiple food sources and communicate that information to their hive mates. Last year, Randolf Menzel of the Free University of Berlin and his colleagues presented sleeping honeybees with an odor cue they had experienced during awake learning, and this improved their memory performance after they woke up (Curr Biol, 25:2869-74, 2015). “It’s a brilliant study,” says Shaw. “Just like humans and rats—it’s incredible.”

Invertebrate sleep researchers are now probing the brain circuitry that regulates sleep-wake cycles: How do wake- and sleep-promoting neurons interact, and what causes an animal to fall asleep or awaken? Other avenues of research focus on how sleep affects brain plasticity and functions beyond memory, and on the molecular mechanisms at play in sleep regulation.

“We’re screening thousands of [Drosophila] lines,” says Amita Sehgal of the University of Pennsylvania Perelman School of Medicine. “[It’s] a totally unbiased approach toward [identifying] what genes or proteins are required to maintain daily levels of sleep.” And so far, the results have pointed to the same genes in flies and in people. “In the limited amount of human work that’s been done, the same molecules are coming up,” says Sehgal.

“A lot of things are conserved throughout evolution, so we can use the fly to really understand the genes and molecules that regulate sleep,” agrees Johns Hopkins University’s Mark Wu, who in 2014 characterized the function of WIDE AWAKE in flies and found that the gene has a mouse homolog whose expression is enriched in the suprachiasmatic nucleus, the master circadian pacemaker in the mammalian brain (Neuron, 82:151-66, 2014). “Flies do lots of things that are very similar to humans,” Wu says. “Neurotransmitters are similar [and] do similar things. The neural circuits, although they look different at first glance, often do similar things.”

And Shaw points out at least one way that insects may even be a better system for studying human sleep than rodents. “The mouse is like a narcoleptic human; they can’t stay asleep,” he says. “In some ways the fly looks more like us than the mouse does.”—Jef Akst


© ISTOCK.COM/PAULBULL Despite anecdotes from scuba divers, fishermen, and aquarium owners, science knows very little about fish sleep. What is known, from less than a decade of work on zebrafish (Danio rerio), is that the fish sleep—from a behavioral perspective and, it’s increasingly understood, on genetic and neural levels—in ways that are remarkably similar to mammals, birds, and other animals.

In 2001, for example, Boston University researcher Irina Zhdanova (then at MIT) and her colleagues found that zebra-fish exhibit characteristic behaviors such as motionlessness, decreased sensitivity to sensory stimulation, and a sleep rebound response following rest deprivation (Brain Research, 903:263-68). Then, in 2010, Caltech neuroscientist David Prober (then at Harvard) and colleagues screened about 6,000 small molecules and found that the vast majority of those that were known to affect mammalian sleep had similar impacts on sleep in larval zebrafish (Science, 327:348-51). More recently, Prober and his labmates also found that melatonin is required for the circadian control of sleep in zebrafish, just as it is in humans and other mammals (Neuron, 85:1193-99, 2015).

“There are huge similarities in terms of gene networks, neural transmitter systems, ontogeny, and development of sleep,” says Karl Karlsson, a Reykjavik University neuroscientist who studies sleep in zebrafish.

Scientists are now working to nail down more-definitive markers of sleep in the species, says Prober, but zebrafish are already proving to be excellent models of sleep in humans. “It doesn’t even matter if it’s ‘sleep,’” he says. “The question is: ‘Can we discover anything using the fish that’s useful for humans?’ We think the answer to that is yes.”

One advantage of studying zebrafish is that, as with Drosophila, mice, and rats, their genome has been completely characterized. Zebrafish are also vertebrates (unlike fruit flies), diurnal (unlike mice and rats), and, as larvae, transparent. “It’s a luxury to be able to look into a whole vertebrate,” says Philippe Mourrain of Stanford University. “You can look at the vasculature. You can look at the cell migration. You can look at the creation and disappearance of synapses. You can look at the entire brain. It’s a very powerful system for neuroscience.”

Taking advantage of this transparency, Mourrain and his colleagues pioneered the documentation of sleep-dependent impacts on the accumulation and pruning of synapses in the zebrafish’s brain. “We could look through different sleep/wake cycles directly at synapses and see if they emerge or disappear during those wake and sleep phases.” Based on this work, Mourrain’s group provided evidence that zebrafish sleep serves to prune away less-important synaptic connections forged during waking hours, a process that likely strengthens overall cognition (Neuron, 68:87-98, 2010). “We were the first ones to demonstrate that in zebrafish you have this kind of decrease during sleep of the synapses,” he says.

And zebrafish are likely to continue providing clues about the largely mysterious phenomenon of sleep, Prober says. “It’s still a relatively small field,” he says. “There are many more people who work on rodents and flies than on fish, but the fish community is growing very rapidly as people are appreciating some of the advantages that it brings.”—Bob Grant


© ISTOCK.COM/LASZIO SZIRTESIMost marine mammals, unlike their terrestrial cousins, can enter a state of being literally half awake: while one hemisphere of the brain remains alert, the other produces the slow brain waves akin to those seen in terrestrial mammals during non-REM sleep.

John Lilly, director of the Communication Research Institute in the Virgin Islands in the 1960s, first suggested that dolphins sleep unihemispherically when he observed them closing one eye at a time while at rest either at the surface or on the bottom of their pools. In 1972, EEG recordings from a female pilot whale revealed that neural activity characteristic of slow-wave sleep alternated between the whale’s brain hemispheres.

A few years later, researchers from the USSR observed that dolphins experienced wakefulness and intermediate states, hovering between waking and sleeping, in one or both hemispheres, but slow-wave sleep in only one hemisphere at a time. The animals could also swim slowly and surface to breathe without waking the sleeping hemisphere. When the researchers recorded dolphins’ EEG patterns over 24-hour periods, they discovered that the average total sleep time for each hemisphere was approximately 4.5 hours per day, though individual sleep sessions lasted only about 40 minutes (Neurophysiology, 20:398-403, 1988).

Unihemispheric sleep may allow these marine mammals to stay alert for long periods of time without detrimental effects. In 2012, a team led by researchers from the National Marine Mammal Foundation in San Diego found that bottlenose dolphins can remain responsive continuously for periods of at least 15 days, swimming, echolocating, and reacting to stimuli at random intervals without any noticeable cognitive deterioration (PLOS ONE, 7:e47478, 2012). “It makes you wonder whether the unihemispheric slow wave that you see in dolphins really should be called ‘sleep’ at all,” says Jerome Siegel of the University of California, Los Angeles, and the author of the introduction to this feature. But dolphins do show evidence of sleep rebound within each hemisphere when tracked with implanted electrodes: if the dolphin is periodically disturbed so as to consistently wake up one hemisphere, the deprived half of the brain will attempt to fall asleep more often and stay asleep longer (J Sleep Res, 1:40-44, 1992).

For other marine mammals, researchers have turned up behavioral evidence suggesting they, too, sleep half a brain at a time. Siegel’s group has observed captive killer whales and a captive gray whale calf at Sea World, both of which exhibited resting behaviors that resembled those of dolphins. “[The gray whale] seemed to open one eye at a time, so that is consistent with what we’ve seen in the killer whale and the dolphin,” says Siegel. “In all likelihood, they have unihemispheric sleep, but nobody’s actually measured.”

In the wild, tagging studies of sperm whales offer further clues. These animals take resting dives in which they remain still and silent under the water for up to 30 minutes, says Saana Isojunno, a fellow in the sea mammal research unit at Scotland’s University of St Andrews. Sperm whales in a resting cycle surface to breathe for several minutes and then dive again, repeating the process as many as 10 times in a row unless they are disturbed by sounds such as killer whale vocalizations. But the whales take these dives only about 5 percent of the time during tagging studies, says Isojunno. “Resting behavior is rare.”

Some semiaquatic marine mammals such as fur seals also display unihemispheric slow-wave sleep, but with a twist. While on land, fur seals experience REM sleep like other terrestrial mammals, but in the water, they transition to a unihemispheric sleep pattern in which they swim on their sides with the “awake” flipper paddling in the water to keep them afloat. And, like dolphins, fur seals don’t seem to experience any negative effects from missing out on REM sleep.

“You bring them back up on land, and they just go back to the normal mode of land sleep,” says Siegel. “They kind of break all the rules as to what we think of as normal in terms of sleep regulation.”
—Karen Zusi


Nonanimal organisms such as plants, fungi, and bacteria do not have central nervous systems, which is how sleep manifests in animals. And many animals will continue to sleep, albeit irregularly, even if their circadian clocks are disrupted. For these reasons, most researchers would agree that plants and other nonanimals do not sleep in the traditional sense, but they do stick to a tight daily schedule, dictated by molecular clocks akin to those in animals.

As early as the 18th century, French astronomer Jean-Jacques d’Ortous de Mairan noted that mimosa plants continued to open their leaves each day and close them each night even when kept in total darkness. That was the first inkling that plants (and other organisms) keep time internally to anticipate environmental changes, rather than simply reacting to light in the environment.

Many activities fluctuate with a plant’s circadian rhythms. Most species time their photosynthesis to peak at noon with the brightest sunlight, and undergo growth spurts at night, elongating their stems using the stores of starch accumulated during the day. Just before dawn, Arabidopsis thaliana mounts immune defenses, anticipating the arrival of downy mildew spores with the sunrise. Later in the day, the plant’s immunity is dialed down, so it doesn’t stunt overnight growth. (See “Holding Their Ground,” The Scientist, February 2016.)

“We’ve got this beautiful data now in plants to show how clock proteins are involved in every major decision a plant has to make,” says Steve Kay, a molecular geneticist at the Scripps Research Institute in La Jolla, California.

Examples of circadian rhythms exist in every kingdom of life. The green alga Chlamydomonas reinhardtii swims up toward the light during the day and toward nitrogen sources at night. Some fungi release spores in a daily cycle as temperature and humidity fluctuate. The salt-loving archaeon Halobacterium salinarum shows daily rhythms in gene expression, most likely to adjust its metabolism to the levels of oxygen dissolved in its environment. Some bacteria even maintain daily routines in metabolism (PLOS ONE, 4:e5485, 2009). Circadian biologists once thought that bacteria would never keep a 24-hour clock, as a bacterium may live less than a day. This assumption was incorrect, says Susan Golden, who studies circadian rhythms in cyanobacteria at the University of California, San Diego. “The great-grandchildren know what time great-grandma thought it was,” and daily cycles emerge when observing populations of bacteria.

Importantly, all of these species maintain their circadian rhythms when subjected to constant darkness or light, suggesting that the clock is internally driven, just like sleep in animals. Indeed, sleep is simply an output of circadian timekeeping that animals evolved alongside other clock-driven states, such as body temperature and expression of metabolic genes.

“The circadian clock in general has evolved in biological organisms to allow organisms to deal with the perpetual changes in our environment,” says Janet Braam, a plant biologist at Rice University. “Sleep is an output [of the clock] more specific to certain organisms, like certain animals.”—Kate Yandell


Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?