Why We Sleep

by Matthew Walker

  Without exception, every animal species studied to date sleeps, or engages in something remarkably like it. This includes insects, such as flies, bees, cockroaches, and scorpions;fn1 fish, from small perch to the largest sharks;fn2 amphibians, such as frogs; and reptiles, such as turtles, Komodo dragons, and chameleons. All have bona fide sleep. Ascend the evolutionary ladder further and we find that all types of birds and mammals sleep: from shrews to parrots, kangaroos, polar bears, bats, and, of course, we humans. Sleep is universal.

  Even invertebrates, such as primordial mollusks and echinoderms, and even very primitive worms, enjoy periods of slumber. In these phases, affectionately termed “lethargus,” they, like humans, become unresponsive to external stimuli. And just as we fall asleep faster and sleep more soundly when sleep-deprived, so, too, do worms, defined by their degree of insensitivity to prods from experimenters.

  How “old” does this make sleep? Worms emerged during the Cambrian explosion: at least 500 million years ago. That is, worms (and sleep by association) predate all vertebrate life. This includes dinosaurs, which, by inference, are likely to have slept. Imagine diplodocuses and triceratopses all comfortably settling in for a night of full repose!

  Regress evolutionary time still further and we have discovered that the very simplest forms of unicellular organisms that survive for periods exceeding twenty-four hours, such as bacteria, have active and passive phases that correspond to the light-dark cycle of our planet. It is a pattern that we now believe to be the precursor of our own circadian rhythm, and with it, wake and sleep.

  Many of the explanations for why we sleep circle around a common, and perhaps erroneous, idea: sleep is the state we must enter in order to fix that which has been upset by wake. But what if we turned this argument on its head? What if sleep is so useful—so physiologically beneficial to every aspect of our being—that the real question is: Why did life ever bother to wake up? Considering how biologically damaging the state of wakefulness can often be, that is the true evolutionary puzzle here, not sleep. Adopt this perspective, and we can pose a very different theory: sleep was the first state of life on this planet, and it was from sleep that wakefulness emerged. It may be a preposterous hypothesis, and one that nobody is taking seriously or exploring, but personally I do not think it to be entirely unreasonable.

  Whichever of these two theories is true, what we know for certain is that sleep is of ancient origin. It appeared with the very earliest forms of planetary life. Like other rudimentary features, such as DNA, sleep has remained a common bond uniting every creature in the animal kingdom. A long-lasting commonality, yes; however, there are truly remarkable differences in sleep from one species to another. Four such differences, in fact.

  ONE OF THESE THINGS IS NOT LIKE THE OTHER

  Elephants need half as much sleep as humans, requiring just four hours of slumber each day. Tigers and lions devour fifteen hours of daily sleep. The brown bat outperforms all other mammals, being awake for just five hours each day while sleeping nineteen hours. Total amount of time is one of the most conspicuous differences in how organisms sleep.

  You’d imagine the reason for such clear-cut variation in sleep need is obvious. It isn’t. None of the likely contenders—body size, prey/predator status, diurnal/nocturnal—usefully explains the difference in sleep need across species. Surely sleep time is at least similar within any one phylogenetic category, since they share much of their genetic code. It is certainly true for other basic traits within phyla, such as sensory capabilities, methods of reproduction, and even degree of intelligence. Yet sleep violates this reliable pattern. Squirrels and degus are part of the same family group (rodents), yet they could not be more dissimilar in sleep need. The former sleeps twice as long as the latter—15.9 hours for the squirrel versus 7.7 hours for the degu. Conversely, you can find near-identical sleep times in utterly different family groups. The humble guinea pig and the precocious baboon, for example, which are of markedly different phylogenetic orders, not to mention physical sizes, sleep precisely the same amount: 9.4 hours.

  So what does explain the difference in sleep time (and perhaps need) from species to species, or even within a genetically similar order? We’re not entirely sure. The relationship between the size of the nervous system, the complexity of the nervous system, and total body mass appears to be a somewhat meaningful predictor, with increasing brain complexity relative to body size resulting in greater sleep amounts. While weak and not entirely consistent, this relationship suggests that one evolutionary function that demands more sleep is the need to service an increasingly complex nervous system. As millennia unfolded and evolution crowned its (current) accomplishment with the genesis of the brain, the demand for sleep only increased, tending to the needs of this most precious of all physiological apparatus.

  Yet this is not the whole story—not by a good measure. Numerous species deviate wildly from the predictions made by this rule. For example, an opossum, which weighs almost the same as a rat, sleeps 50 percent longer, clocking an average of eighteen hours each day. The opossum is just one hour shy of the animal kingdom record for sleep amount currently held by the brown bat, who, as previously mentioned, racks up a whopping nineteen hours of sleep each day.

  There was a moment in research history when scientists wondered if the measure of choice—total minutes of sleep—was the wrong way of looking at the question of why sleep varies so considerably across species. Instead, they suspected that assessing sleep quality, rather than quantity (time), would shed some light on the mystery. That is, species with superior quality of sleep should be able to accomplish all they need in a shorter time, and vice versa. It was a great idea, with the exception that, if anything, we’ve discovered the opposite relationship: those that sleep more have deeper, “higher”-quality sleep. In truth, the way quality is commonly assessed in these investigations (degree of unresponsiveness to the outside world and the continuity of sleep) is probably a poor index of the real biological measure of sleep quality: one that we cannot yet obtain in all these species. When we can, our understanding of the relationship between sleep quantity and quality across the animal kingdom will likely explain what currently appears to be an incomprehensible map of sleep-time differences.

  For now, our most accurate estimate of why different species need different sleep amounts involves a complex hybrid of factors, such as dietary type (omnivore, herbivore, carnivore), predator/prey balance within a habitat, the presence and nature of a social network, metabolic rate, and nervous system complexity. To me, this speaks to the fact that sleep has likely been shaped by numerous forces along the evolutionary path, and involves a delicate balancing act between meeting the demands of waking survival (e.g., hunting prey/obtaining food in as short a time as possible, minimizing energy expenditure and threat risk), serving the restorative physiological needs of an organism (e.g., a higher metabolic rate requires greater “cleanup” efforts during sleep), and tending to the more general requirements of the organism’s community.

  Nevertheless, even our most sophisticated predictive equations remain unable to explain far-flung outliers in the map of slumber: species that sleep much (e.g., bats) and those that sleep little (e.g., giraffes, which sleep for just four to five hours). Far from being a nuisance, I feel these anomalous species may hold some of the keys to unlocking the puzzle of sleep need. They remain a delightfully frustrating opportunity for those of us trying to crack the code of sleep across the animal kingdom, and within that code, perhaps as yet undiscovered benefits of sleep we never thought possible.

  TO DREAM OR NOT TO DREAM

  Another remarkable difference in sleep across species is composition. Not all species experience all stages of sleep. Every species in which we can measure sleep stages experiences NREM sleep—the non-dreaming stage. However, insects, amphibians, fish, and most reptiles show no clear signs of REM sleep—the type associated with dreaming in humans. Only birds and mammals, which appeared later in the evolutionary timeli
ne of the animal kingdom, have full-blown REM sleep. It suggests that dream (REM) sleep is the new kid on the evolutionary block. REM sleep seems to have emerged to support functions that NREM sleep alone could not accomplish, or that REM sleep was more efficient at accomplishing.

  Yet as with so many things in sleep, there is another anomaly. I said that all mammals have REM sleep, but debate surrounds cetaceans, or aquatic mammals. Certain of these ocean-faring species, such as dolphins and killer whales, buck the REM-sleep trend in mammals. They don’t have any. Although there is one case in 1969 suggesting that a pilot whale was in REM sleep for six minutes, most of our assessments to date have not discovered REM sleep—or at least what many sleep scientists would believe to be true REM sleep—in aquatic mammals. From one perspective, this makes sense: when an organism enters REM sleep, the brain paralyzes the body, turning it limp and immobile. Swimming is vital for aquatic mammals, since they must surface to breathe. If full paralysis was to take hold during sleep, they could not swim and would drown.

  The mystery deepens when we consider pinnipeds (one of my all-time favorite words, from the Latin derivatives: pinna “fin” and pedis “foot”), such as fur seals. Partially aquatic mammals, they split their time between land and sea. When on land, they have both NREM sleep and REM sleep, just like humans and all other terrestrial mammals and birds. But when they enter the ocean, they stop having REM sleep almost entirely. Seals in the ocean will sample but a soupçon of the stuff, racking up just 5 to 10 percent of the REM sleep amounts they would normally enjoy when on land. Up to two weeks of ocean-bound time have been documented without any observable REM sleep in seals, who survive in such times on a snooze diet of NREM sleep.

  These anomalies do not necessarily challenge the usefulness of REM sleep. Without doubt, REM sleep, and even dreaming, appears to be highly useful and adaptive in those species that have it, as we shall see in part 3 of the book. That REM sleep returns when these animals return to land, rather being done away with entirely, affirms this. It is simply that REM sleep does not appear to be feasible or needed by aquatic mammals when in the ocean. During that time, we assume they make do with lowly NREM sleep—which, for dolphins and whales, may always be the case.

  Personally, I don’t believe aquatic mammals, even cetaceans like dolphins and whales, have a total absence of REM sleep (though several of my scientific colleagues will tell you I’m wrong). Instead, I think the form of REM sleep these mammals obtain in the ocean is somewhat different and harder to detect: be it brief in nature, occurring at times when we have not been able to observe it, or expressed in ways or hiding in parts of the brain that we have not yet been able to measure.

  In defense of my contrarian point of view, I note that it was once believed that egg-laying mammals (monotremes), such as the spiny anteater and the duck-billed platypus, did not have REM sleep. It turned out that they do, or at least a version of it. The outer surface of their brain—the cortex—from which most scientists measure sleeping brainwaves, does not exhibit the choppy, chaotic characteristics of REM-sleep activity. But when scientists looked a little deeper, beautiful bursts of REM-sleep electrical brainwave activity were found at the base of the brain—waves that are a perfect match for those seen in all other mammals. If anything, the duck-billed platypus generates more of this kind of electrical REM-sleep activity than any other mammal! So they did have REM sleep after all, or at least a beta version of it, first rolled out in these more evolutionarily ancient mammals. A fully operational, whole-brain version of REM sleep appears to have been introduced in more developed mammals that later evolved. I believe a similar story of atypical, but nevertheless present, REM sleep will ultimately be observed in dolphins and whales and seals when in the ocean. After all, absence of evidence is not evidence of absence.

  More intriguing than the poverty of REM sleep in this aquatic corner of the mammalian kingdom is the fact that birds and mammals evolved separately. REM sleep may therefore have been birthed twice in the course of evolution: once for birds and once for mammals. A common evolutionary pressure may still have created REM sleep in both, in the same way that eyes have evolved separately and independently numerous times across different phyla throughout evolution for the common purpose of visual perception. When a theme repeats in evolution, and independently across unrelated lineages, it often signals a fundamental need.

  However, a very recent report has suggested that a proto form of REM sleep exists in an Australian lizard, which, in terms of the evolutionary timeline, predates the emergence of birds and mammals. If this finding is replicated, it would suggest that the original seed of REM sleep was present at least 100 million years earlier than our original estimates. This common seed in certain reptiles may have then germinated into the full form of REM sleep we now see in birds and mammals, including humans.

  Regardless of when true REM sleep emerged in evolution, we are fast discovering why REM-sleep dreaming came into being, what vital needs it supports in the warm-blooded world of birds and mammals (e.g., cardiovascular health, emotional restoration, memory association, creativity, body-temperature regulation), and whether other species dream. As we will later discuss, it seems they do.

  Setting aside the issue of whether all mammals have REM sleep, an uncontested fact is this: NREM sleep was first to appear in evolution. It is the original form that sleep took when stepping out from behind evolution’s creative curtain—a true pioneer. This seniority leads to another intriguing question, and one that I get asked in almost every public lecture I give: Which type of sleep—NREM or REM sleep—is more important? Which do we really need?

  There are many ways you can define “importance” or “need,” and thus numerous ways of answering the question. But perhaps the simplest recipe is to take an organism that has both sleep types, bird or mammal, and keep it awake all night and throughout the subsequent day. NREM and REM sleep are thus similarly removed, creating the conditions of equivalent hunger for each sleep stage. The question is, which type of sleep will the brain feast on when you offer it the chance to consume both during a recovery night? NREM and REM sleep in equal proportions? Or more of one than the other, suggesting greater importance of the sleep stage that dominates?

  This experiment has now been performed many times on numerous species of birds and mammals, humans included. There are two clear outcomes. First, and of little surprise, sleep duration is far longer on the recovery night (ten or even twelve hours in humans) than during a standard night without prior deprivation (eight hours for us). Responding to the debt, we are essentially trying to “sleep it off,” the technical term for which is a sleep rebound.

  Second, NREM sleep rebounds harder. The brain will consume a far larger portion of deep NREM sleep than of REM sleep on the first night after total sleep deprivation, expressing a lopsided hunger. Despite both sleep types being on offer at the finger buffet of recovery sleep, the brain opts to heap much more deep NREM sleep onto its plate. In the battle of importance, NREM sleep therefore wins. Or does it?

  Not quite. Should you keep recording sleep across a second, third, and even fourth recovery night, there’s a reversal. Now REM sleep becomes the primary dish of choice with each returning visit to the recovery buffet table, with a side of NREM sleep added. Both sleep stages are therefore essential. We try to recover one (NREM) a little sooner than the other (REM), but make no mistake, the brain will attempt to recoup both, trying to salvage some of the losses incurred. It is important to note, however, that regardless of the amount of recovery opportunity, the brain never comes close to getting back all the sleep it has lost. This is true for total sleep time, just as it is for NREM sleep and for REM sleep. That humans (and all other species) can never “sleep back” that which we have previously lost is one of the most important take-homes of this book, the saddening consequences of which I will describe in chapters 7 and 8.

  IF ONLY HUMANS COULD

  A third striking difference in sleep across the animal kingdom is the w
ay in which we all do it. Here, the diversity is remarkable and, in some cases, almost impossible to believe. Take cetaceans, such as dolphins and whales, for example. Their sleep, of which there is only NREM, can be unihemispheric, meaning they will sleep with half a brain at a time! One half of the brain must always stay awake to maintain life-necessary movement in the aquatic environment. But the other half of the brain will, at times, fall into the most beautiful NREM sleep. Deep, powerful, rhythmic, and slow brainwaves will drench the entirety of one cerebral hemisphere, yet the other half of the cerebrum will be bristling with frenetic, fast brainwave activity, fully awake. This despite the fact that both hemispheres are heavily wired together with thick crisscross fibers, and sit mere millimeters apart, as in human brains.

  Of course, both halves of the dolphin brain can be, and frequently are, awake at the very same time, operating in unison. But when it is time for sleep, the two sides of the brain can uncouple and operate independently, one side remaining awake while the other side snoozes away. After this one half of the brain has consumed its fill of sleep, they switch, allowing the previously vigilant half of the brain to enjoy a well-earned period of deep NREM slumber. Even with half of the brain asleep, dolphins can achieve an impressive level of movement and even some vocalized communication.

  The neural engineering and tricky architecture required to accomplish this staggering trick of oppositional “lights-on, lights-off” brain activity is rare. Surely Mother Nature could have found a way to avoid sleep entirely under the extreme pressure of nonstop, 24/7 aquatic movement. Would that not have been easier than masterminding a convoluted split-shift system between brain halves for sleep, while still allowing for a joint operating system where both sides unite when awake? Apparently not. Sleep is of such vital necessity that no matter what the evolutionary demands of an organism, even the unyielding need to swim in perpetuum from birth to death, Mother Nature had no choice. Sleep with both sides of the brain, or sleep with just one side and then switch. Both are possible, but sleep you must. Sleep is non-negotiable.