Why We Sleep

by Matthew Walker

  Fire was no perfect solution, however, and ground sleeping would have remained risky. An evolutionary pressure to become qualitatively more efficient in how we sleep therefore developed. Any Homo erectus capable of accomplishing more efficient sleep would likely have been favored in survival and selection. Evolution saw to it that our ancient form of sleep became somewhat shorter in duration, yet increased in intensity, especially by enriching the amount of REM sleep we packed into the night.

  In fact, as is so often the case with Mother Nature’s brilliance, the problem became part of the solution. In other words, the act of sleeping on solid ground, and not on a precarious tree branch, was the impetus for the enriched and enhanced amounts of REM sleep that developed, while the amount of time spent asleep was able to modestly decrease. When sleeping on the ground, there’s no more risk of falling. For the first time in our evolution, hominids could consume all the body-immobilized REM-sleep dreaming they wanted, and not worry about the lasso of gravity whipping them down from treetops. Our sleep therefore became “concentrated”: shorter and more consolidated in duration, packed aplenty with high-quality sleep. And not just any type of sleep, but REM sleep that bathed a brain rapidly accelerating in complexity and connectivity. There are species that have more total REM time than hominids, but there are none who power up and lavish such vast proportions of REM sleep onto such a complex, richly interconnected brain as we Homo sapiens do.

  From these clues, I offer a theorem: the tree-to-ground reengineering of sleep was a key trigger that rocketed Homo sapiens to the top of evolution’s lofty pyramid. At least two features define human beings relative to other primates. I posit that both have been beneficially and causally shaped by the hand of sleep, and specifically our intense degree of REM sleep relative to all other mammals: (1) our degree of sociocultural complexity, and (2) our cognitive intelligence. REM sleep, and the act of dreaming itself, lubricates both of these human traits.

  To the first of these points, we have discovered that REM sleep exquisitely recalibrates and fine-tunes the emotional circuits of the human brain (discussed in detail in part 3 of the book). In this capacity, REM sleep may very well have accelerated the richness and rational control of our initially primitive emotions, a shift that I propose critically contributed to the rapid rise of Homo sapiens to dominance over all other species in key ways.

  We know, for example, that REM sleep increases our ability to recognize and therefore successfully navigate the kaleidoscope of socioemotional signals that are abundant in human culture, such as overt and covert facial expressions, major and minor bodily gestures, and even mass group behavior. One only needs to consider disorders such as autism to see how challenging and different a social existence can be without these emotional navigation abilities being fully intact.

  Related, the REM-sleep gift of facilitating accurate recognition and comprehension allows us to make more intelligent decisions and actions as a consequence. More specifically, the coolheaded ability to regulate our emotions each day—a key to what we call emotional IQ—depends on getting sufficient REM sleep night after night. (If your mind immediately jumped to particular colleagues, friends, and public figures who lack these traits, you may well wonder about how much sleep, especially late-morning REM-rich sleep, they are getting.)

  Second, and more critical, if you multiply these individual benefits within and across groups and tribes, all of which are experiencing an ever-increasing intensity and richness of REM sleep over millennia, we can start to see how this nightly REM-sleep recalibration of our emotional brains could have scaled rapidly and exponentially. From this REM-sleep-enhanced emotional IQ emerged a new and far more sophisticated form of hominid socioecology across vast collectives, one that helped enable the creation of large, emotionally astute, stable, highly bonded, and intensely social communities of humans.

  I will go a step further and suggest that this is the most influential function of REM sleep in mammals, perhaps the most influential function of all types of sleep in all mammals, and even the most eminent advantage ever gifted by sleep in the annals of all planetary life. The adaptive benefits conferred by complex emotional processing are truly monumental, and so often overlooked. We humans can instantiate vast numbers of emotions in our embodied brains, and thereafter, deeply experience and even regulate those emotions. Moreover, we can recognize and help shape the emotions of others. Through both of these intra- and interpersonal processes, we can forge the types of cooperative alliances that are necessary to establish large social groups, and beyond groups, entire societies brimming with powerful structures and ideologies. What may at first blush have seemed like a modest asset awarded by REM sleep to a single individual is, I believe, one of the most valuable commodities ensuring the survival and dominance of our species as a collective.

  The second evolutionary contribution that the REM-sleep dreaming state fuels is creativity. NREM sleep helps transfer and make safe newly learned information into long-term storage sites of the brain. But it is REM sleep that takes these freshly minted memories and begins colliding them with the entire back catalog of your life’s autobiography. These mnemonic collisions during REM sleep spark new creative insights as novel links are forged between unrelated pieces of information. Sleep cycle by sleep cycle, REM sleep helps construct vast associative networks of information within the brain. REM sleep can even take a step back, so to speak, and divine overarching insights and gist: something akin to general knowledge—that is, what a collection of information means as a whole, not just an inert back catalogue of facts. We can awake the next morning with new solutions to previously intractable problems or even be infused with radically new and original ideas.

  Adding, then, to the opulent and domineering socioemotional fabric that REM sleep helps weave across the masses came this second, creativity benefit of dream sleep. We should (cautiously) revere how superior our hominid ingenuity is relative to that of any of our closest rivals, primate or other. The chimpanzees—our nearest living primate relatives—have been around approximately 5 million years longer than we have; some of the great apes preceded us by at least 10 million years. Despite aeons of opportunity time, neither species has visited the moon, created computers, or developed vaccines. Humbly, we humans have. Sleep, especially REM sleep and the act of dreaming, is a tenable, yet underappreciated, factor underlying many elements that form our unique human ingenuity and accomplishments, just as much as language or tool use (indeed, there is even evidence that sleep causally shapes both these latter traits as well).

  Nevertheless, the superior emotional brain gifts that REM sleep affords should be considered more influential in defining our hominid success than the second benefit, of inspiring creativity. Creativity is an evolutionarily powerful tool, yes. But it is largely limited to an individual. Unless creative, ingenious solutions can be shared between individuals through the emotionally rich, pro-social bonds and cooperative relationships that REM sleep fosters—then creativity is far more likely to remain fixed within an individual, rather than spread to the masses.

  Now we can appreciate what I believe to be a classic, self-fulfilling positive cycle of evolution. Our shift from tree to ground sleeping instigated an ever more bountiful amount of relative REM sleep compared with other primates, and from this bounty emerged a steep increase in cognitive creativity, emotional intelligence, and thus social complexity. This, alongside our increasingly dense, interconnected brains, led to improved daily (and nightly) survival strategies. In turn, the harder we worked those increasingly developed emotional and creative circuits of the brain during the day, the greater was our need to service and recalibrate these ever-demanding neural systems at night with more REM sleep.

  As this positive feedback loop took hold in exponential fashion, we formed, organized, maintained, and deliberatively shaped ever larger social groups. The rapidly increasing creative abilities could thus be spread more efficiently and rapidly, and even improved by that ever-increasing amount of
hominid REM-sleep that enhances emotional and social sophistication. REM-sleep dreaming therefore represents a tenable new contributing factor, among others, that led to our astonishingly rapid evolutionary rise to power, for better and worse—a new (sleep-fueled), globally dominant social superclass.

  Chapter 5

  Changes in Sleep Across the Life Span

  SLEEP BEFORE BIRTH

  Through speech or song, expecting parents will often thrill at their ability to elicit small kicks and movements from their in utero child. Though you should never tell them this, the baby is most likely fast asleep. Prior to birth, a human infant will spend almost all of its time in a sleep-like state, much of which resembles the REM-sleep state. The sleeping fetus is therefore unaware of its parents’ performative machinations. Any co-occurring arm flicks and leg bops that the mother feels from her baby are most likely to be the consequence of random bursts of brain activity that typify REM sleep.

  Adults do not—or at least should not—throw out similar nighttime kicks and movements, since they are held back by the body-paralyzing mechanism of REM sleep. But in utero, the immature fetus’s brain has yet to construct the REM-sleep muscle-inhibiting system adults have in place. Other deep centers of the fetus brain have, however, already been glued in place, including those that generate sleep. Indeed, by the end of the second trimester of development (approximately week 23 of pregnancy), the vast majority of the neural dials and switches required to produce NREM and REM sleep have been sculpted out and wired up. As a result of this mismatch, the fetus brain still generates formidable motor commands during REM sleep, except there is no paralysis to hold them back. Without restraint, those commands are freely translated into frenetic body movements, felt by the mother as acrobatic kicks and featherweight punches.

  At this stage of in utero development, most of the time is spent in sleep. The twenty-four-hour period contains a mishmash of approximately six hours of NREM sleep, six hours of REM sleep, and twelve hours of an intermediary sleep state that we cannot confidently say is REM or NREM sleep, but certainly is not full wakefulness. It is only when the fetus enters the final trimester that the glimmers of real wakefulness emerge. Far less than you would probably imagine, though—just two to three hours of each day are spent awake in the womb.

  Even though total sleep time decreases in the last trimester, a paradoxical and quite ballistic increase in REM-sleep time occurs. In the last two weeks of pregnancy, the fetus will ramp up its consumption of REM sleep to almost nine hours a day. In the last week before birth, REM-sleep amount hits a lifetime high of twelve hours a day. With near insatiable appetite, the human fetus therefore doubles its hunger for REM sleep just before entering the world. There will be no other moment during the life of that individual—pre-natal, early post-natal, adolescence, adulthood, or old age—when they will undergo such a dramatic change in REM-sleep need, or feast so richly on the stuff.

  Is the fetus actually dreaming when in REM sleep? Probably not in the way most of us conceptualize dreams. But we do know that REM sleep is vital for promoting brain maturation. The construction of a human being in the womb occurs in distinct, interdependent stages, a little bit like building a house. You cannot crown a house with a roof before there are supporting wall frames to rest it on, and you cannot put up walls without a foundation to seat them in. The brain, like the roof of a house, is one of the last items to be constructed during development. And like a roof, there are sub-stages to that process—you need a roof frame before you can start adding roof tiles, for instance.

  Detailed creation of the brain and its component parts occurs at a rapid pace during the second and third trimesters of human development—precisely the time window when REM-sleep amounts skyrocket. This is no coincidence. REM sleep acts as an electrical fertilizer during this critical phase of early life. Dazzling bursts of electrical activity during REM sleep stimulate the lush growth of neural pathways all over the developing brain, and then furnish each with a healthy bouquet of connecting ends, or synaptic terminals. Think of REM sleep like an Internet service provider that populates new neighborhoods of the brain with vast networks of fiber-optic cables. Using these inaugural bolts of electricity, REM sleep then activates their high-speed functioning.

  This phase of development, which infuses the brain with masses of neural connections, is called synaptogenesis, as it involves the creation of millions of wiring links, or synapses, between neurons. By deliberate design, it is an overenthusiastic first pass at setting up the mainframe of a brain. There is a great deal of redundancy, offering many, many possible circuit configurations to emerge within the infant’s brain once born. From the perspective of the Internet service provider analogy, all homes, across all neighborhoods, throughout all territories of the brain have been gifted a high degree of connectivity and bandwidth in this first phase of life.

  Charged with such a herculean task of neuro-architecture—establishing the neural highways and side streets that will engender thoughts, memories, feelings, decisions, and actions—it’s no wonder REM sleep must dominate most, if not all, of early developmental life. In fact, this is true for all other mammals:fn1 the time of life when REM sleep is greatest is the same stage when the brain is undergoing the greatest construction.

  Worryingly, if you disturb or impair the REM sleep of a developing infant brain, pre- or early post-term, there are consequences. In the 1990s, researchers began studying newly born rat pups. Simply by blocking REM sleep, their gestational progress was retarded, despite chronological time marching on. The two should, of course, progress in unison. Depriving the infant rats of REM sleep stalled construction of their neural rooftop—the cerebral cortex of the brain. Without REM sleep, assembly work on the brain ground to a halt, frozen in time by the experimental wedge of a lack of REM sleep. Day after day, the half-finished roofline of the sleep-starved cerebral cortex shows no growth change.

  The very same effect has now been demonstrated in numerous other mammalian species, suggesting that the effect is probably common across mammals. When the infant rat pups were finally allowed to get some REM sleep, assembly of the cerebral rooftop did restart, but it didn’t accelerate, nor did it ever fully get back on track. An infant brain without sleep will be a brain ever underconstructed.

  A more recent link with deficient REM sleep concerns autism spectrum disorder (ASD) (not to be confused with attention deficit hyperactivity disorder [ADHD], which we will discuss later in the book). Autism, of which there are several forms, is a neurological condition that emerges early in development, usually around two or three years of age. The core symptom of autism is a lack of social interaction. Individuals with autism do not communicate or engage with other people easily, or typically.

  Our current understanding of what causes autism is incomplete, but central to the condition appears to be an inappropriate wiring up of the brain during early developmental life, specifically in the formation and number of synapses—that is, abnormal synaptogenesis. Imbalances in synaptic connections are common in autistic individuals: excess amounts of connectivity in some parts of the brain, deficiencies in others.

  Realizing this, scientists have begun to examine whether the sleep of individuals with autism is atypical. It is. Infants and young children who show signs of autism, or who are diagnosed with autism, do not have normal sleep patterns or amounts. The circadian rhythms of autistic children are also weaker than their non-autistic counterparts, showing a flatter profile of melatonin across the twenty-four-hour period rather than a powerful rise in concentration at night and rapid fall throughout the day.fn2 Biologically, it is as if the day and night are far less light and dark, respectively, for autistic individuals. As a consequence, there is a weaker signal for when stable wake and solid sleep should take place. Additionally, and perhaps related, the total amount of sleep that autistic children can generate is less than that of non-autistic children.

  Most notable, however, is the significant shortage of REM sleep. Autistic in
dividuals show a 30 to 50 percent deficit in the amount of REM sleep they obtain, relative to children without autism.fn3 Considering the role of REM sleep in establishing the balanced mass of synaptic connections within the developing brain, there is now keen interest in discovering whether or not REM-sleep deficiency is a contributing factor to autism.

  Existing evidence in humans is simply correlational, however. Just because autism and REM-sleep abnormalities go hand in hand does not mean that one causes the other. Nor does this association tell you the direction of causality even if it does exist: Is deficient REM sleep causing autism, or is it the other way around? It is curious to note, however, that selectively depriving an infant rat of REM sleep leads to aberrant patterns of neural connectivity, or synaptogenesis, in the brain.fn4 Moreover, rats deprived of REM sleep during infancy go on to become socially withdrawn and isolated as adolescents and adults.fn5 Irrespective of causality issues, tracking sleep abnormalities represents a new diagnostic hope for the early detection of autism.

  Of course, no expecting mother has to worry about scientists disrupting the REM sleep of their developing fetus. But alcohol can inflict that same selective removal of REM sleep. Alcohol is one of the most powerful suppressors of REM sleep that we know of. We will discuss the reason that alcohol blocks REM-sleep generation, and the consequences of that sleep disruption in adults, in later chapters. For now, however, we’ll focus on the impact of alcohol on the sleep of a developing fetus and newborn.