Why We Sleep

by Matthew Walker

  Alcohol consumed by a mother readily crosses the placental barrier, and therefore readily infuses her developing fetus. Knowing this, scientists first examined the extreme scenario: mothers who were alcoholics or heavy drinkers during pregnancy. Soon after birth, the sleep of these neonates was assessed using electrodes gently placed on the head. The newborns of heavy-drinking mothers spent far less time in the active state of REM sleep compared with infants of similar age but who were born of mothers who did not drink during pregnancy.

  The recording electrodes went on to point out an even more concerning physiological story. Newborns of heavy-drinking mothers did not have the same electrical quality of REM sleep. You will remember from chapter 3 that REM sleep is exemplified by delightfully chaotic—or desynchronized—brainwaves: a vivacious and healthy form of electrical activity. However, the infants of heavy-drinking mothers showed a 200 percent reduction in this measure of vibrant electrical activity relative to the infants born of non-alcohol-consuming mothers. Instead, the infants of heavy-drinking mothers emitted a brainwave pattern that was far more sedentary in this regard.fn6 If you are now wondering whether or not epidemiological studies have linked alcohol use during pregnancy and an increased likelihood of neuropsychiatric illness in the mother’s child, including autism, the answer is yes.fn7

  Fortunately, most mothers these days do not drink heavily during pregnancy. But what about the more common situation of an expectant mom having an occasional glass or two of wine during pregnancy? Using noninvasive tracking of heart rate, together with ultrasound measures of body, eye, and breathing movement, we are now able to determine the basic stages of NREM sleep and REM sleep of a fetus when it is in the womb. Equipped with these methods, a group of researchers studied the sleep of babies who were just weeks away from being born. Their mothers were assessed on two successive days. On one of those days, the mothers drank non-alcoholic fluids. On the other day, they drank approximately two glasses of wine (the absolute amount was controlled on the basis of their body weight). Alcohol significantly reduced the amount of time that the unborn babies spent in REM sleep, relative to the non-alcohol condition.

  That alcohol also dampened the intensity of REM sleep experienced by the fetus, defined by the standard measure of how many darting rapid eye movements adorn the REM-sleep cycle. Furthermore, these unborn infants suffered a marked depression in breathing during REM sleep, with breath rates dropping from a normal rate of 381 per hour during natural sleep to just 4 per hour when the fetus was awash with alcohol.fn8

  Beyond alcohol abstinence during pregnancy, the time window of nursing also warrants mention. Almost half of all lactating women in Western countries consume alcohol in the months during breastfeeding. Alcohol is readily absorbed in a mother’s milk. Concentrations of alcohol in breast milk closely resemble those in a mother’s bloodstream: a 0.08 blood alcohol level in a mother will result in approximately a 0.08 alcohol level in breast milk.fn9 Recently we have discovered what alcohol in breast milk does to the sleep of an infant.

  Newborns will normally transition straight into REM sleep after a feeding. Many mothers already know this: almost as soon as suckling stops, and sometimes even before, the infant’s eyelids will close, and underneath, the eyes will begin darting left-right, indicating that their baby is now being nourished by REM sleep. A once-common myth was that babies sleep better if the mother has had an alcoholic drink before a feeding—beer was the suggested choice of beverage in this old tale. For those of you who are beer lovers, unfortunately, it is just that—a myth. Several studies have fed infants breast milk containing either a non-alcoholic flavor, such as vanilla, or a controlled amount of alcohol (the equivalent of a mother having a drink or two). When babies consume alcohol-laced milk, their sleep is more fragmented, they spend more time awake, and they suffer a 20 to 30 percent suppression of REM sleep soon after.fn10 Often, the babies will even try to get back some of that missing REM sleep once they have cleared it from their bloodstream, though it is not easy for their fledgling systems to do so.

  What emerges from all of these studies is that REM sleep is not optional during early human life, but obligatory. Every hour of REM sleep appears to count, as evidenced by the desperate attempt by a fetus or newborn to regain any REM sleep when it is lost.fn11 Sadly, we do not yet fully understand what the long-term effects are of fetal or neonate REM-sleep disruption, alcohol-triggered or otherwise. Only that blocking or reducing REM sleep in newborn animals hinders and distorts brain development, leading to an adult that is socially abnormal.

  CHILDHOOD SLEEP

  Perhaps the most obvious and tormenting (for new parents) difference between the sleep of infants and young children and that of adults is the number of slumber phases. In contrast to the single, monophasic sleep pattern observed in adults of industrialized nations, infants and young kids display polyphasic sleep: many short snippets of sleep through the day and night, punctuated by numerous awakenings, often vocal.

  There is no better or more humorous affirmation of this fact than the short book of lullabies, written by Adam Mansbach, entitled Go the F**k to Sleep. Obviously, it’s an adult book. At the time of writing, Mansbach was a new father. And like many a new parent, he was run ragged by the constant awakenings of his child: the polyphasic profile of infant sleep. The incessant need to attend to his young daughter, helping her fall back to sleep time and time and time again, night after night after night, left him utterly exasperated. It got to the point where Mansbach just had to vent all the loving rage he had pent up. What came spilling out onto the page was a comedic splash of rhymes he would fictitiously read to his daughter, the themes of which will immediately resonate with many new parents. “I’ll read you one very last book if you swear,/You’ll go the fuck to sleep.” (I implore you to listen to the audiobook version of the work, narrated to perfection by the sensational actor Samuel L. Jackson.)

  Fortunately, for all new parents (Mansbach included), the older a child gets, the fewer, longer, and more stable their sleep bouts become.fn12 Explaining this change is the circadian rhythm. While the brain areas that generate sleep are molded in place well before birth, the master twenty-four-hour clock that controls the circadian rhythm—the suprachiasmatic nucleus—takes considerable time to develop. Not until age three or four months will a newborn show modest signs of being governed by a daily rhythm. Slowly, the suprachiasmatic nucleus begins to latch on to repeating signals, such as daylight, temperature change, and feedings (so long as those feedings are highly structured), establishing a stronger twenty-four-hour rhythm.

  By the one-year milestone of development, the suprachiasmatic nucleus clock of an infant has gripped the steering reins of the circadian rhythm. This means that the child now spends more of the day awake, interspersed with several naps and, mercifully, more of the night asleep. Mostly gone are the indiscriminate bouts of sleep and wake that once peppered the day and night. By four years of age, the circadian rhythm is in dominant command of a child’s sleep behavior, with a lengthy slab of nighttime sleep, usually supplemented by just a single daytime nap. At this stage, the child has transitioned from a polyphasic sleep pattern to a biphasic sleep pattern. Come late childhood, the modern, monophasic pattern of sleep is finally made real.

  What this progressive establishment of stable rhythmicity hides, however, is a much more tumultuous power struggle between NREM and REM sleep. Although the amount of total sleep gradually declines from birth onwards, all the while becoming more stable and consolidated, the ratio of time spent in NREM sleep and REM sleep does not decline in a similarly stable manner.

  During the fourteen hours of total shut-eye per day that a six-month-old infant obtains, there is a 50/50 timeshare between NREM and REM sleep. A five-year-old, however, will have a 70/30 split between NREM and REM sleep across the eleven hours of total daily slumber. In other words, the proportion of REM sleep decreases in early childhood while the proportion of NREM sleep actually increases, even though total sleep time d
ecreases. The downgrading of the REM-sleep portion, and the upswing in NREM-sleep dominance, continues, throughout early and midchildhood. That balance will finally stabilize to an 80/20 NREM/REM sleep split by the late teen years, and remain so throughout early and midadulthood.

  SLEEP AND ADOLESCENCE

  Why do we spend so much time in REM sleep in the womb and early in life, yet switch to a heavier dominance of deep NREM sleep in late childhood and early adolescence? If we quantify the intensity of the deep-sleep brainwaves, we see the very same pattern: a decline in REM-sleep intensity in the first year of life, yet an exponential rise in deep NREM sleep intensity in mid- and late childhood, hitting a peak just before puberty, and then damping back down. What’s so special about this type of deep sleep at this transitional time of life?

  Prior to birth, and soon after, the challenge for development was to build and add vast numbers of neural highways and interconnections that become a fledgling brain. As we have discussed, REM sleep plays an essential role in this proliferation process, helping to populate brain neighborhoods with neural connectivity, and then activate those pathways with a healthy dose of informational bandwidth.

  But since this first round of brain wiring is purposefully overzealous, a second round of remodeling must take place. It does so during late childhood and adolescence. Here, the architectural goal is not to scale up, but to scale back for the goal of efficiency and effectiveness. The time of adding brain connections with the help of REM sleep is over. Instead, pruning of connections becomes the order of the day or, should I say, night. Enter the sculpting hand of deep NREM sleep.

  Our analogy of the Internet service provider is a helpful one to return to. When first setting up the network, each home in the newly built neighborhood was given an equal amount of connectivity bandwidth and thus potential for use. However, that’s an inefficient solution for the long term, since some of these homes will become heavy bandwidth users over time, while other homes will consume very little. Some homes may even remain vacant and never use any bandwidth. To reliably estimate what pattern of demand exists, the Internet service provider needs time to gather usage statistics. Only after a period of experience can the provider make an informed decision on how to refine the original network structure it put in place, dialing back connectivity to low-use homes, while increasing connectivity to other homes with high bandwidth demand. It is not a complete redo of the network, and much of the original structure will remain in place. After all, the Internet service provider has done this many times before, and they have a reasonable estimate of how to build a first pass of the network. But a use-dependent reshaping and downsizing must still occur if maximum network efficiency is to be achieved.

  The human brain undergoes a similar, use-determined transformation during late childhood and adolescence. Much of the original structure laid down early in life will persist, since Mother Nature has, by now, learned to create a quite accurate first-pass wiring of a brain after billions of attempts over many thousands of years of evolution. But she wisely leaves something on the table in her generic brain sculpture, that of individualized refinement. The unique experiences of a child during their formative years translate to a set of personal usage statistics. Those experiences, or those statistics, provide the bespoke blueprint for a last round of brain refinement,fn13 capitalizing on the opportunity left open by nature. A (somewhat) generic brain becomes ever more individualized, based on the personalized use of the owner.

  To help with the job of refinement and downscaling of connectivity, the brain employs the services of deep NREM sleep. Of the many functions carried out by deep NREM sleep—the full roster of which we will discuss in the next chapter—it is that of synaptic pruning that features prominently during adolescence. In a remarkable series of experiments, the pioneering sleep researcher Irwin Feinberg discovered something fascinating about how this operation of downscaling takes place within the adolescent brain. His findings help justify an opinion you may also hold: adolescents have a less rational version of an adult brain, one that takes more risks and has relatively poor decision-making skills.

  Using electrodes placed all over the head—front and back, left side and right, Feinberg began recording the sleep of a large group of kids starting at age six to eight years old. Every six to twelve months, he would bring these individuals back to his laboratory and perform another sleep measurement. He didn’t stop for ten years. He amassed more than 3,500 all-night assessments: a scarcely believable 320,000 hours of sleep recordings! From these, Feinberg created a series of snapshots, depicting how deep-sleep intensity changed with the stages of brain development as the children made their often awkward transition through adolescence into adulthood. It was the neuroscience equivalent of time-lapse photography in nature: taking repeat pictures of a tree as it first comes into bud in the spring (babyhood), then bursts into leaf during the summer (late childhood), then matures in color come the fall (early adolescence), and finally sheds its leaves in the winter (late adolescence and early adulthood).

  During mid- and late childhood, Feinberg observed moderate deep-sleep amounts as the last neural growth spurts inside the brain were being completed, analogous to late spring and early summer. Then Feinberg began seeing a sharp rise in deep-sleep intensity in his electrical recordings, right at the time when the developmental needs of brain connectivity switch from growing connections to shedding them; the tree’s equivalent of fall. Just as maturational fall was about to turn to winter, and the shedding was nearly complete, Feinberg’s recordings showed a clear ramping back down in deep NREM-sleep intensity to lower intensity once more. The life cycle of childhood was over, and as the last leaves dropped, the onward neural passage of these teenagers had been secured. Deep NREM sleep had aided their transition into early adulthood.

  Feinberg proposed that the rise and fall of deep-sleep intensity were helping lead the maturational journey through the precarious heights of adolescence, followed by safe onward passage into adulthood. Recent findings have supported his theory. As deep NREM sleep performs its final overhaul and refinement of the brain during adolescence, cognitive skills, reasoning, and critical thinking start to improve, and do so in a proportional manner with that NREM sleep change. Taking a closer look at the timing of this relationship, you see something even more interesting. The changes in deep NREM sleep always precede the cognitive and developmental milestones within the brain by several weeks or months, implying a direction of influence: deep sleep may be a driving force of brain maturation, not the other way around.

  Feinberg made a second seminal discovery. When he examined the timeline of changing deep-sleep intensity at each different electrode spot on the head, it was not the same. Instead, the rise-and-fall pattern of maturation always began at the back of the brain, which performs the functions of visual and spatial perception, and then progressed steadily forward as adolescence progressed. Most striking, the very last stop on the maturational journey was the tip of the frontal lobe, which enables rational thinking and critical decision-making. Therefore, the back of the brain of an adolescent was more adult-like, while the front of the brain remained more child-like at any one moment during this developmental window of time.fn14

  His findings helped explain why rationality is one of the last things to flourish in teenagers, as it is the last brain territory to receive sleep’s maturational treatment. Certainly sleep is not the only factor in the ripening of the brain, but it appears to be a significant one that paves the way to mature thinking and reasoning ability. Feinberg’s study reminds me of a billboard advertisement I once saw from a large insurance firm, which read: “Why do most 16-year-olds drive like they’re missing part of their brain? Because they are.” It takes deep sleep, and developmental time, to accomplish the neural maturation that plugs this brain “gap” within the frontal lobe. When your children finally reach their mid-twenties and your car insurance premium drops, you can thank sleep for the savings.

  The relationship betwe
en deep-sleep intensity and brain maturation that Feinberg described has now been observed in many different populations of children and adolescents around the world. But how can we be sure that deep sleep truly offers a neural pruning service necessary for brain maturation? Perhaps changes in sleep and brain maturation simply occur at roughly the same time but are independent of each other?

  The answer is found in studies of juvenile rats and cats at the equivalent stage to human adolescence. Scientists deprived these animals of deep sleep. In doing so, they halted the maturational refinement of brain connectivity, demonstrating a causal role for deep NREM sleep in propelling the brain into healthy adulthood.fn15 Of concern is that administering caffeine to juvenile rats will also disrupt deep NREM sleep and, as a consequence, delay numerous measures of brain maturation and the development of social activity, independent grooming, and the exploration of the environment—measures of self-motivated learning.fn16

  Recognizing the importance of deep NREM sleep in teenagers has been instrumental to our understanding of healthy development, but it has also offered clues as to what happens when things go wrong in the context of abnormal development. Many of the major psychiatric disorders, such as schizophrenia, bipolar disorder, major depression, and ADHD are now considered disorders of abnormal development, since they commonly emerge during childhood and adolescence.