ー 8 8 MATTHEW WALKER 711 genes had been abnormally revved up in their expression by the loss ofsleep, while the other halfhad been diminished in their expression, or shut down entirely. The genes that were increased included those linked tO chronic inflammation, cellular stress, and various factors that cause cardiovascular disease. ong those turned down were genes that help maintain stable metabolism and optimal immune responses. Subse- quent studies have found that short sleep duration ⅷⅡ so disrupt the activity ofgenes regulating cholesterol. ln particular, a lack 0f sleep ⅶⅡ cause a drop in high-density lipoproteins (HDLs)— a directional profile that has consistently been linked t0 cardiovascular disease. lnsufficient sleep does more than alter the activity and readout Of your genes; it attacks the very physical structure 0fyour genetic mate- ⅱ itself. The spiral strands of DNA ⅲ your cells float around ⅲ the nucleus, but are tightly wound together intO structures called chromo- somes, rather like weavmg individual threads together to make a sturdy shoelace. ld just like a shoelace, the ends ofyour chromosomes need t0 be protected by a cap or binding tip. For chromosomes, that protec- tive cap is called a telomere. lfthe telomeres at the end Ofyour chromo- somes become damaged, your DNA spirals become exposed and your now vulnerable genetic code cannot operate properly, like a fraying shoelace without a tip. The less sleep an individual obtains, or the worse the quality of sleep, the more damaged the capstone telomeres of that individual's chromosomes. These are the findings 0f a collection 0f studies that have recently been reported in thousands Of adults in their forties, fifties, and sixties by numerous independent research teams around the world. t Whether this association is causal remains to be determined. But *Beyond a simple lack Of sleep, Dijk's research team has further shown that inappropri- ately timed sleep, such as that imposed by jet lag or shift work, can have equally large effects on the expression of human genes as inadequate sleep. By pushing forward an indi- vidual's sleep-wake cycle by a few hours each day for three days, Dijk disrupted a massive one-third of the transcribing activity of the genes ⅲ a group ofyoung, healthy adults. Once again, the genes that were lmpacted controlled elemental life processes, such as the tim- ing of metabolic, thermoregulatory, and immune activity, as well as cardiac health. tThe significant relationship between short sleep and short or damaged telomeres is observed even when accounting for other factors that are known tO harm telomeres, such as age, weight, depression, and smoking.

ー 3 6 MATTHEW WALKER the repeating beep 仕 om a heart monitor in a hospital: beep, beep, beep. NOW picture the dramatic sound effect you hear in emergency room television dramas when a patient starts to slip away as doctors franti- cally try to save their life. At first, the heartbeats are constant—beep, beep, beep—as are your responses on the visual attention task when you are well rested: stable, regular. Switch t0 your performance when sleep-deprived, and it is the aural equivalent 0fthe patient ⅲ the hospi- tal going int0 cardiac arrest: beep, beep, beep, beeeeeeeeeeeeeep. Your performance has flatlined. NO conscious response, no 1 れ Otor response. A microsleep. And then the heartbeat comes back, as ⅥⅡ your perfor- mance—beep, beep, beep—but only for a short while. Soon, you have another arrest: beep, beep, beeeeeeeeeeeeeep. 、 microsleeps. Comparing the number of lapses, or microsleeps, day after day across the four different experimental groups gave Dinges a second key finding. Tho s e individuals who slept eight hours every night main- tained a stable, near-perfect performance across the ⅲ weeks. Those ⅲ the three-night total sleep deprivation group suffered catastrophic imp airment, which was no real surprise. After the first night Of no sleep at all, their lapses in concentration (missed responses) increased by over 400 percent. The surprise was that these impairments continued t0 escalate at the same ballistic rate after a second and third night of tOtal sleep deprivation, as ifthey would continue tO escalate in severity if more nights Of sleep were lost, showlng no signs offlattening out. But it was the ⅲ , 0 partial sleep deprivation groups that brought the most concerning message ofall. After four hours of sleep for six nights, participants' performance was just as bad as those who had not slept for twenty-four hours straight—that is, a 400 percent increase in the number of microsleeps. By day 11 on this diet of four hours of sleep a night, participants' performance had degraded even further, matching that of someone who had pulled ℃ back-to-back all-nighters, going without sleep for forty-eight hours. MOSt worrying 伝 om a societal perspective were the individuals in the group wh0 obtained six hours ofsleep a night—something that may sound familiar to many of you. Ten days of six hours of sleep a night was it tOOk tO become as impaired in performance as going without sleep for twenty-four hours straight. And like the total sleep deprivation

W HY W E S L E E P 2 2 3 nor provide useful responses—the typical way that cognitive scientists assess the workings 0f the brain. Short 0f lucid dreaming, which we ⅶⅡ address at the end 0f this chapter, sleep scientists have been left wanting ⅲ this regard. We have frequently been resigned t0 passively 0b serving brain activity during sleep, without ever being able to have participants perform tests while they are sleeping. Rather, we measure waking performance before and after sleep and determine if the sleep stages or dreaming that occurred in between explains any observed benefit the next day. I and my colleague at Harvard Medical School Robert Stickgold designed a solution t0 this problem, albeit an indirect and imperfect one. ln chapter 7 1 described the phenomenon ofsleep inertia—the car- ryover Ofthe prior sleeping brain state intO wakefulness ⅲ the minutes after waking up. We wondered whether we could turn this briefwindow ofsleep inertia t0 our experimental advantage—not by waking subjects up ⅲ the mormng and testing them, but rather by waking individuals up 仕 om different stages 0fNREM sleep and REM sleep throughout the night. The dramatic alterations in brain activity during NREM and REM sleep, and their tidal shifts in neurochemical concentrations, dO not reverse instantaneously when you awaken. lnstead, the neural and chemical prop ertie s 0f that particular sleep stage ⅷⅡ linger, creating the inertia periOd that separates true wakefulness 仕 om sleep, and last some minutes. Upon enforced awakening, the brain's neurophysiology starts out far more sleep-like than wake-like and, with each passing minute, the concentration Of the prior sleep stage 仕 om which an indi- vidual has been woken ⅵⅡ gradually fade from the brain as true wake- fulness rises to the surface. By restricting the length 0f whatever cognitive test we performed to just ninety seconds, we felt we could wake individuals up and very quickly test them in this transitional sleep phase. ⅲ doing so, we could perhaps capture some 0f the functional properties of the sleep stage 仕 om which the participant was woken, like capturing the vapors of an evaporating substance and analyzing those vapors tO draw conclusions about the properties ofthe substance itself. lt worked. We developed an anagram task in which the letters of

WHY WE SLEEP 3 2 ー the lnstitute ofMedicine, part 0f the US National Academy of Sciences, issued a report with a clear statement: working for more than sixteen consecutive hours without sleep is hazardous for bOth the patient and resident physician. You may have noticed my specific wording in the above paragraph: first-year residents. This is because the revised rule ()t the time ofwrit- ing this b00 has only been applied t0 those in their first year of train- ing, and not tO those in later years Ofa medical residency. Why? Because the Accreditation CounciI for Graduate MedicaI Education—the elite board ofhigh-powered physicians that dictates the American residency training structure—stated that data proving the dangers ofinsufficient sleep had only been gathered ⅲ residents in their first year 0f the pro- gram. As a result, they felt there was no evidence t0 justify a change for residents ⅲ years MO tO five—as ifgetting past the twelve-month point ⅲ a medical residency program magically confers immunity against the bi010 cal and psychological effects of sleep deprivation—effects that these same individuals had previously been so provably vulnerable to just months before. This entrenched pomposity, prevalent in so many senior-driven, dogmatic institutional hierarchies, has no place in medical practice my opimon as a scientist intimate with the research data. Those boards must disabuse themselves of the we-suffered-through-sleep- deprivation-and-you-should-too mentality when it comes tO traming, teaching, and practicing medicine. Of course, medical institutions put forward other arguments to justify the old-school way 0f sleep abuse. The most common harkens back to a William Halsted-like mind-set: without working exhaustive shifts, it will take far t00 long to train residents, and they ⅶⅡ not learn as effectively. Why, then, can several western European countries train their young doctors within the same time frame when th ey are lim- ited tO working no more than forty-eight hours in one week, without continuous long periods Of sleeplessness? Perhaps they are just not as well trained? ThiS, t00 , is erroneous, since many ofthose western Euro- pean medical programs, such as ⅲ the UK and Sweden, rank among the top ten countries for most medical practice health outcomes, while the m 引 0 ⅱ Of US institutes rank somewhere between eighteenth and

ー 4 0 MATHEW WALKER The heady cocktail of sleep loss and alcohol was not ad 〃レ e , but instead 襯記ゆ〃 c 砒ル e. They magnified each other, like two drugs whose effects are harmful by themselves but, when taken together, interact tO produce truly dire consequences. After thirty years Of intensive research, we can now ans 、 many of the questions posed earlier. The recycle rate 0f a human being is around sixteen hours. After sixteen hours Of being awake, the brain begins t0 fail. Humans need more than seven hours 0f sleep each night tO maintaln cognitive performance. After ten days Ofjust seven hours of sleep, the brain is as dysfunctional as it would be after going with- out sleep for twenty-four hours. Three んⅡ nights ofrecovery sleep (). e. , more nights than a weekend) are insufficient tO restore performance back to normal levels after a week 0f short sleeping. Finally, the human mind cannot accurately sense hOW sleep-deprived it is when sleep- deprived. Ⅵ宅 shall return tO the ramifications Of these results in the remain- ing chapters, but the real-life consequences 0f drowsy driving deserve special mention. This coming week, more than 2 million people in the US ⅷⅡ fall asleep while driving their motor vehicle. That's more than 250 , 000 every day, with more such events during the week than week- ends for obvious reasons. More than 56 million Americans admit tO struggling t0 stay awake at the wheel 0f a car each month. As a result, 1.2 million accidents are caused by sleepiness each year ⅲ the United States. Said another way: for every thirty seconds youve been reading this b00 に there has been a car accident somewhere in the US caused by sleeplessness. lt is more than probable that someone has lost their life in a fatigue-related car accident during the time you have been reading this chapter. You may find it surprising t0 learn that vehicle accidents caused by drowsy driving exceed those caused by alcohol and drugs co 襯わ加 e Drowsy driving alone is worse than driving drunk That may seem like a controversial or irresponsible thing tO say, and I dO not wish tO trivi- alize the lamentable act 0f drunk driving by any means. Yet my state- ment is true for the following simple reason: drunk drivers are Often late in braking, and late in making evasive maneuvers. But when you fall asleep, or have a microsleep, 0 レ立 0 臾 r け加 g 記 ge 既 A person wh0

2 6 8 MATTHEW WALKER this b00k at eleven p. m. ⅲ New York City, having been surrounded by electric light all evening. Your bedside clock may be registering eleven p. m. , but the ommpresence 0f artificial light has paused the internal tick-tocking 0f time by hindering the release of melatonin. Biologically speaking, you've been dragged westward across the continent tO the internal eqmvalent Of Chicago time (ten p. m. ), or even San Francisco time (eight p ・ m ・ ) ・ Artificial evening and nighttime light can therefore masquerade as sleep-onset insomnia—the inability t0 begin sleeping soon after getting into bed. By delaying the release of melatonin, artificial evening light makes it considerably less likely that you'll be able to fall asleep at a reasonable time. When you do finally turn out the bedside light, hop- ing that sleep ⅷⅡ come qtuckly is made 心 the more difficult. lt ⅶⅡ be some time before the rising tide of melatonin is able tO submerge your brain and bOdy in peak concentrations, instructed by the darkness that only now has begun—in other words, before you are biologically capable oforgamzing the onset ofrobust, stable sleep. What ofa petite bedside lamp? How much can that really influence your suprachiasmatic nucleus ? A lOt, it turns out. Even a hint Of dim light—8 to 10 lux—has been shown to delay the release of nighttime melatonin in humans. The feeblest 0f bedside lamps pumps out twice as much: anywhere 仕 om 20 to 80 lux. A subtly lit living room, where most people reside in the hours before bed, ⅶⅡ hum at around 200 lux. Despite being just 1 to 2 percent of the strength of daylight, this ambient level ofincandescent home lighting can have 50 percent ofthe melatonin-suppressing influence within the brain. Just when things looked as bad as they could get for the suprachias- matic nucleus With incandescent lamps, a new invention in 1997 made the situation far worse: blue light—emitting diodes, or blue LEDs. For this invention, Shuji Nakamura, lsamu Akasaki, and Hiroshi Amano won the NobeI Prize in physics ⅲ 2014. lt was a remarkable achieve- ment. Blue LED lights offer considerable advantages over incandescent lamps in terms oflower energy demands and, for the lights themselves, longer life spans. But they may be inadvertently shortening our own. The light receptors in the eye that communicate "daytime" to the suprachiasmatic nucleus are most sensitive tO short-wavelength light

WHY WE SLEEP to breathe. If んⅡ paralysis was t0 take hold during sleep, they could not swim and would drown. The mystery deepens when we consider pinnipeds (one 0f my - time favorite words, 仕 om the Latin derivatives: 2 加〃 2 "fin" and e 市 & 00t " ) , 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 心 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 soupqon ofthe stuff, racking up just 5 t0 10 percent 0f the REM sleep amounts they would normally enjoy when on land. Up t0 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 ofNREM sleep. These anomalies dO not necessarily challenge the usefulness Of REM sIeep.Without doubt, REM sleep, and even dreaming, appears t0 be highly useful and adaptive in those species that have it, as we shall see in part 3 0f the b00k. That REM sleep returns when these animals return t0 land, rather being done away with entirely, affirms this. lt is simply that REM sleep does not appear t0 be feasible or needed by aquatic mammals When ocean. During time, 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 dOl- phins and whales, have a total absence 0f REM sleep (though several of my scientific colleagues Ⅱ tell you l'm wrong). lnstead, 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 hid- ing in parts Ofthe brain that we have not yet been able tO measure. ln 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. lt 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 sleep- ing bramwaves, does not exhibit the choppy, chaotic characteristics Of REM-sleep activity. But when scientists looked a little deeper, beautiful

40 MAITHEW WALKER its gate, and which are not. Should they gain privileged passage, they are sent up tO the cortex at the top Of your brain, where they are con- sciously perceived. By locking its gates shut at the onset 0f healthy sleep, the thalamus imposes a sensory blackout ⅲ the brain, prevent- ing onward travel 0f those signals up t0 the cortex. As a result, you are no longer consciously aware Ofthe information broadcasts being trans- mitted 仕 om your outer sense organs. At this moment, your brain has lost waking contact with the outside world that surrounds you. Said another way, you are now asleep. The second feature that instructs your own, self-determined judg- ment ofsleep is a sense Oftime distortion experlenced ⅲ杁 MO contradic- tory ways. At the most obvious level, you lose your conscious sense Of time when you sleep, tantamount tO a chronometric void. Consider the last time you Ⅱ asleep on an airplane.When you woke up, you proba- bly checked a clock to see how long you had been asleep. Why? Because your explicit tracking 0f time was ostensibly lost while you slept. lt is this feeling 0f a time cavity that, in waking retrospect, makes you confi- dent you've been asleep. But while your co 〃 & c / 0 ″ & mapping of time is lost during sleep, at a 〃 0 ル CO 〃 sc / 0 ″ & level, time continues tO be cataloged by the brain with incredible precision. l'm sure you have had the experience ofneeding to wake up the next morning at a specific time. Perhaps you had to catch an early-morning flight. Before bed, you diligently set your alarm for 6 : 00 a. m. MiracuIously, however, you woke up at 5 : 58 a. m. , unassisted, right before the alarm. Your brain, it seems, is still capable of logging time with quite remarkable precision while asleep. Like so many other operations occurring within the brain, you simply don't have explicit access t0 this accurate time knowledge during sleep. lt all flies below the radar Of consciousness, surfacing When needed. One last temporal distortion deserves mention here—that oftime dila- tion ⅲ dreams, beyond sleep itself. Time isn't quite time within dreams. lt is mo st Often elongated. Consider the last time you hit the snooze but- ton on your 記 m , having been woken 仕 om a dream. Mercifully, you are gvmg yourself another delicious five minutes 0f sleep. You go right back to dreaming. After the 0 ed five minutes, your alarm clock faithfully sounds again, yet that's not what it felt like t0 you. During those five min-

CHAPTER 5 Changes in Sleep Across the Life Span S LE E P B E FO R E B ー RTH Through speech or song, expecting parents will often thrill at their abil- ityto elicit small kicks and movements 仕 om their in utero child. Though you should never tell them this, the baby is most likely fast asleep. prior to birth, a human infant ⅶⅡ spend almost all of its time in a sleep-like state, much ofwhich resembles the REM-sleep state. The sleeping fetus iS therefore unaware OfitS parents' performative machinations. Any CO- occurring arm flicks and leg bops that the mother feels 仕 om her baby are most likely tO be the consequence ofrandom bursts ofbrain actiV1ty that 与甲 i 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 ⅲ utero, the immature fetus's brain has yet t0 construct the REM-sleep muscle-inhibiting system adults have in place. Other deep centers 0f the fetus brain have, however, already been glued in place, including those that generate sleep. lndeed, by the end ofthe second trimester of development (approximately week 23 of pregnancy), the vast majority of the neural dials and switches required t0 produce NREM and REM sleep have been sculpted out and wired up. As a result 0f this mismatch, the fetus brain still generates formi- dable motor commands during REM sleep, except there is no paralysis tO hOld them back. Without restraint, those commands are freely trans- lated 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 approxi-

W HY W E S L E E P Figure 6 : The Urge to Sleep Greatest urge tO sleep (Greatest distance between Process-C and Process-S) (Sleep drive)¯、、 Process-S 3 3 Process-C Circadian ()a ke d rive) / am sleep llpm 70E sleep 1 ユ pm 70m What happens to all of the accumulated adenosine once you do fall asleep? During sleep, a mass evacuation gets under way as the brain has the chance to degrade and remove the day's adenosine. Across the night, sleep lifts the heavy weight 0f sleep pressure, lightening the adenosine load. After approximately eight hours of healthy sleep ⅲ an adult, the adenosine purge is complete. Just as this process is ending, the marching band 0f your circadian activity rhythm has fortuitously returned, and its energwing influence starts tO approach. When these れ V() processes trade places in the morning hours, wherein adenosine has been removed and the rousing volume Of the circadian rhythm is becoming louder (indicated by the meeting ofthe れ lines ⅲ figure 6 ) , we naturallywake up (seven a. m. on day , 0 , in the figure example). FO ト lowing that Ⅱ night ofsleep, you are now ready tO face another sixteen hours ofwakefulness with physical vigor and sharp brain function. INDEPENDENCE DAY, AND NIGHT Have you ever pulled an "all-nighter"—forgoing sleep and remaming awake throughout the following day? If you have, and can remember much Of anything about it, you may recall that there were times when you felt truly miserable and sleepy, yet there were other moments when, despite having been awake for longer, you paradoxically felt 襯 0 尾 alert. Why? I don't advise anyone to conduct this self-experiment, but assess- ing a person's alertness across twenty-four hours Of tOtal sleep depri- vation is one way that scientists can demonstrate that the い VO forces determining when you want t0 be awake and asleep—the twenty-four-