Why We Sleep

by Matthew Walker

  Recent studies by a research team in Japan have now replicated our findings, but they’ve done so by restricting participants’ sleep to five hours for five nights. No matter how you take sleep from the brain—acutely, across an entire night, or chronically, by short sleeping for a handful of nights—the emotional brain consequences are the same.

  When we conducted our original experiments, I was struck by the pendulum-like swings in the mood and emotions of our participants. In a flash, sleep-deprived subjects would go from being irritable and antsy to punch-drunk giddy, only to then swing right back to a state of vicious negativity. They were traversing enormous emotional distances, from negative to neutral to positive, and all the way back again, within a remarkably short period of time. It was clear that I had missed something. I needed to conduct a sister study to the one I described above, but now explore how the sleep-deprived brain responds to increasingly positive and rewarding experiences, such as exciting images of extreme sports, or the chance of winning increasing amounts of money in fulfilling tasks.

  We discovered that different deep emotional centers in the brain just above and behind the amygdala, called the striatum—associated with impulsivity and reward, and bathed by the chemical dopamine—had become hyperactive in sleep-deprived individuals in response to the rewarding, pleasurable experiences. As with the amygdala, the heightened sensitivity of these hedonic regions was linked to a loss of the rational control from the prefrontal cortex.

  Insufficient sleep does not, therefore, push the brain into a negative mood state and hold it there. Rather, the under-slept brain swings excessively to both extremes of emotional valence, positive and negative.

  You may think that the former counter-balances the latter, thereby neutralizing the problem. Sadly, emotions, and their guiding of optimal decision and actions, do not work this way. Extremity is often dangerous. Depression and extreme negative mood can, for example, infuse an individual with a sense of worthlessness, together with ideas of questioning life’s value. There is now clearer evidence of this concern. Studies of adolescents have identified a link between sleep disruption and suicidal thoughts, suicide attempts, and, tragically, suicide completion in the days after. One more reason for society and parents to value plentiful sleep in teens rather than chastise it, especially considering that suicide is the second-leading cause of death in young adults in developed nations after car accidents.

  Insufficient sleep has also been linked to aggression, bullying, and behavioral problems in children across a range of ages. A similar relationship between a lack of sleep and violence has been observed in adult prison populations; places that, I should add, are woefully poor at enabling good sleep that could reduce aggression, violence, psychiatric disturbance, and suicide, which, beyond the humanitarian concern, increases costs to the taxpayer.

  Equally problematic issues arise from extreme swings in positive mood, though the consequences are different. Hypersensitivity to pleasurable experiences can lead to sensation-seeking, risk-taking, and addiction. Sleep disturbance is a recognized hallmark associated with addictive substance use.fn4 Insufficient sleep also determines relapse rates in numerous addiction disorders, associated with reward cravings that are unmetered, lacking control from the rational head office of the brain’s prefrontal cortex.fn5 Relevant from a prevention standpoint, insufficient sleep during childhood significantly predicts early onset of drug and alcohol use in that same child during their later adolescent years, even when controlling for other high-risk traits, such as anxiety, attention deficits, and parental history of drug use.fn6 You can now appreciate why the bidirectional, pendulum-like emotional liability caused by sleep deprivation is so concerning, rather than counter-balancing.

  Our brain scanning experiments in healthy individuals offered reflections on the relationship between sleep and psychiatric illnesses. There is no major psychiatric condition in which sleep is normal. This is true of depression, anxiety, post-traumatic stress disorder (PTSD), schizophrenia, and bipolar disorder (once known as manic depression).

  Psychiatry has long been aware of the coincidence between sleep disturbance and mental illness. However, a prevailing view in psychiatry has been that mental disorders cause sleep disruption—a one-way street of influence. Instead, we have demonstrated that otherwise healthy people can experience a neurological pattern of brain activity similar to that observed in many of these psychiatric conditions simply by having their sleep disrupted or blocked. Indeed, many of the brain regions commonly impacted by psychiatric mood disorders are the same regions that are involved in sleep regulation and impacted by sleep loss. Further, many of the genes that show abnormalities in psychiatric illnesses are the same genes that help control sleep and our circadian rhythms.

  Had psychiatry got the causal direction wrong, and it was sleep disruption instigating mental illness, not the other way around? No, I believe that is equally inaccurate and reductionist to suggest. Instead, I firmly believe that sleep loss and mental illness is best described as a two-way street of interaction, with the flow of traffic being stronger in one direction or the other, depending on the disorder.

  I am not suggesting that all psychiatric conditions are caused by absent sleep. However, I am suggesting that sleep disruption remains a neglected factor contributing to the instigation and/or maintenance of numerous psychiatric illnesses, and has powerful diagnostic and therapeutic potential that we are yet to fully understand or make use of.

  Preliminary (but compelling) evidence is beginning to support this claim. One example involves bipolar disorder, which most people will recognize by the former name of manic depression. Bipolar disorder should not be confused with major depression, in which individuals slide exclusively down into the negative end of the mood spectrum. Instead, patients with bipolar depression vacillate between both ends of the emotion spectrum, experiencing dangerous periods of mania (excessive, reward-driven emotional behavior) and also periods of deep depression (negative moods and emotions). These extremes are often separated by a time when the patients are in a stable emotional state, neither manic nor depressed.

  A research team in Italy examined bipolar patients during the time when they were in this stable, inter-episode phase. Next, under careful clinical supervision, they sleep-deprived these individuals for one night. Almost immediately, a large proportion of the individuals either spiraled into a manic episode or became seriously depressed. I find it to be an ethically difficult experiment to appreciate, but the scientists had importantly demonstrated that a lack of sleep is a causal trigger of a psychiatric episode of mania or depression. The result supports a mechanism in which the sleep disruption—which almost always precedes the shift from a stable to an unstable manic or depressive state in bipolar patients—may well be a (the) trigger in the disorder, and not simply epiphenomenal.

  Thankfully, the opposite is also true. Should you improve sleep quality in patients suffering from several psychiatric conditions using a technique we will discuss later, called cognitive behavioral therapy for insomnia (CBT-I), you can improve symptom severity and remission rates. My colleague at the University of California, Berkeley, Dr. Allison Harvey has been a pioneer in this regard.

  By improving sleep quantity, quality, and regularity, Harvey and her team have systematically demonstrated the healing abilities of sleep for the minds of numerous psychiatric populations. She has intervened with the therapeutic tool of sleep in conditions as diverse as depression, bipolar disorder, anxiety, and suicide, all to great effect. By regularizing and enhancing sleep, Harvey has stepped these patients back from the edge of crippling mental illness. That, in my opinion, is a truly remarkable service to humanity.

  The swings in emotional brain activity that we observed in healthy individuals who were sleep-deprived may also explain a finding that has perplexed psychiatry for decades. Patients suffering from major depression, in which they become exclusively locked into the negative end of the mood spectrum, show what at first appears to be
a counterintuitive response to one night of sleep deprivation. Approximately 30 to 40 percent of these patients will feel better after a night without sleep. Their lack of slumber appears to be an antidepressant.

  The reason sleep deprivation is not a commonly used treatment, however, is twofold. First, as soon as these individuals do sleep, the antidepressant benefit goes away. Second, the 60 to 70 percent of patients who do not respond to the sleep deprivation will actually feel worse, deepening their depression. As a result, sleep deprivation is not a realistic or comprehensive therapy option. Still, it has posed an interesting question: How could sleep deprivation prove helpful for some of these individuals, yet detrimental to others?

  I believe that the explanation resides in the bidirectional changes in emotional brain activity that we observed. Depression is not, as you may think, just about the excess presence of negative feelings. Major depression has as much to do with absence of positive emotions, a feature described as anhedonia: the inability to gain pleasure from normally pleasurable experiences, such as food, socializing, or sex.

  The one-third of depressed individuals who respond to sleep deprivation may therefore be those who experience the greater amplification within reward circuits of the brain that I described earlier, resulting in far stronger sensitivity to, and experiencing of, positive rewarding triggers following sleep deprivation. Their anhedonia is therefore lessened, and now they can begin to experience a greater degree of pleasure from pleasurable life experiences. In contrast, the other two-thirds of depressed patients may suffer the opposite negative emotional consequences of sleep deprivation more dominantly: a worsening, rather than alleviation, of their depression. If we can identify what determines those who will be responders and those who will not, my hope is that we can create better, more tailored sleep-intervention methods for combating depression.

  We will revisit the effects of sleep loss on emotional stability and other brain functions in later chapters when we discuss the real-life consequences of sleep loss in society, education, and the workplace. The findings justify our questioning of whether or not sleep-deprived doctors can make emotionally rational decisions and judgments; under-slept military personnel should have their fingers on the triggers of weaponry; overworked bankers and stock traders can make rational, non-risky financial decisions when investing the public’s hard-earned retirement funds; and if teenagers should be battling against impossibly early start times during a developmental phase of life when they are most vulnerable to developing psychiatric disorders. For now, however, I will summarize this section by offering a discerning quote on the topic of sleep and emotion by the American entrepreneur E. Joseph Cossman: “The best bridge between despair and hope is a good night’s sleep.”fn7

  TIRED AND FORGETFUL?

  Have you ever pulled an “all-nighter,” deliberately staying awake all night? One of my true loves is teaching a large undergraduate class on the science of sleep at the University of California, Berkeley. I taught a similar sleep course while I was at Harvard University. At the start of the course, I conduct a sleep survey, inquiring about my students’ sleep habits, such as the times they go to bed and wake up during the week and weekend, how much sleep they get, if they think their academic performance is related to their sleep.

  Inasmuch as they are telling me the truth (they fill the survey out anonymously online, not in class), the answer I routinely get is saddening. More than 85 percent of them pull all-nighters. Especially concerning is the fact that of those who said “yes” to pulling all-nighters, almost a third will do so monthly, weekly, or even several times a week. As the course continues throughout the semester, I return to the results of their sleep survey and link their own sleep habits with the science we are learning about. In this way, I try to point out the very personal dangers they face to their psychological and physical health due to their insufficient sleep, and the danger they themselves pose to society as a consequence.

  The most common reason my students give for pulling all-nighters is to cram for an exam. In 2006, I decided to conduct an MRI study to investigate whether they were right or wrong to do so. Was pulling an all-nighter a wise idea for learning? We took a large group of individuals and assigned them to either a sleep group or a sleep deprivation group. Both groups remained awake normally across the first day. Across the following night, those in the sleep group obtained a full night of shut-eye, while those in the sleep deprivation group were kept awake all night under the watchful eye of trained staff in my lab. Both groups were then awake across the following morning. Around midday, we placed participants inside an MRI scanner and had them try to learn a list of facts, one at a time, as we took snapshots of their brain activity. Then we tested them to see how effective that learning had been. However, instead of testing them immediately after learning, we waited until they had had two nights of recovery sleep. We did this to make sure that any impairments we observed in the sleep-deprived group were not confounded by them being too sleepy or inattentive to recollect what they may very well have learned. Therefore, the sleep-deprivation manipulation was only in effect during the act of learning, and not during the later act of recall.

  When we compared the effectiveness of learning between the two groups, the result was clear: there was a 40 percent deficit in the ability of the sleep-deprived group to cram new facts into the brain (i.e., to make new memories), relative to the group that obtained a full night of sleep. To put that in context, it would be the difference between acing an exam and failing it miserably!

  What was going wrong within the brain to produce these deficits? We compared the patterns of brain activity during attempted learning between the two groups, and focused our analysis on the brain region that we spoke about in chapter 6, the hippocampus—the information “in-box” of the brain that acquires new facts. There was lots of healthy, learning-related activity in the hippocampus in the participants who had slept the night before. However, when we looked at this same brain structure in the sleep-deprived participants, we could not find any significant learning activity whatsoever. It was as though sleep deprivation had shut down their memory in-box, and any new incoming information was simply being bounced. You don’t even need the blunt force of a whole night of sleep deprivation. Simply disrupting the depth of an individual’s NREM sleep with infrequent sounds, preventing deep sleep and keeping the brain in shallow sleep, without waking the individual up will produce similar brain deficits and learning impairments.

  You may have seen a movie called Memento, in which the lead character suffers brain damage and, from that point forward, can no longer make any new memories. In neurology, he is what we call “densely amnesic.” The part of his brain that was damaged was the hippocampus. It is the very same structure that sleep deprivation will attack, blocking your brain’s capacity for new learning.

  I cannot tell you how many of my students have come up to me at the end of the lecture in which I describe these studies and said, “I know that exact feeling. It seems as though I’m staring at the page of the textbook but nothing is going in. I may be able to hold on to some facts the following day for the exam, but if you were to ask me to take that same test a month later, I think I’d hardly remember a thing.”

  The latter description has scientific backing. Those few memories you are able to learn while sleep-deprived are forgotten far more quickly in the hours and days thereafter. Memories formed without sleep are weaker memories, evaporating rapidly. Studies in rats have found that it is almost impossible to strengthen the synaptic connections between individual neurons that normally forge a new memory circuit in the animals that have been sleep-deprived. Imprinting lasting memories into the architecture of the brain becomes nearly impossible. This is true whether the researchers sleep-deprived the rats for a full twenty-four hours, or just a little, for two or three hours. Even the most elemental units of the learning process—the production of proteins that form the building blocks of memories within these synapses—are stunted by th
e state of sleep loss.

  The very latest work in this area has revealed that sleep deprivation even impacts the DNA and the learning-related genes in the brain cells of the hippocampus itself. A lack of sleep therefore is a deeply penetrating and corrosive force that enfeebles the memory-making apparatus within your brain, preventing you from constructing lasting memory traces. It is rather like building a sand castle too close to the tide line—the consequences are inevitable.

  While at Harvard University, I was invited to write my first op-ed piece for their newspaper, the Crimson. The topic was sleep loss, learning, and memory. It was also the last piece I was invited to write.

  In the article, I described the above studies and their relevance, returning time and again to the pandemic of sleep deprivation that was sweeping through the student body. However, rather than lambaste the students for these practices, I pointed a scolding finger directly at the faculty, myself included. I suggested that if we, as teachers, strive to accomplish just that purpose—to teach—then end-loading exams in the final days of the semester was an asinine decision. It forced a behavior in our students—that of short sleeping or pulling all-nighters leading up to the exam—that was in direct opposition to the goals of nurturing young scholarly minds. I argued that logic, backed by scientific fact, must prevail, and that it was long past the time for us to rethink our evaluation methods, their contra-educational impact, and the unhealthy behavior it coerced from our students.

  To suggest that the reaction from the faculty was icy would be a thermal compliment. “It was the students’ choice,” I was told in adamant response emails. “A lack of planned study by irresponsible undergraduates” was another common rebuttal from faculty and administrators attempting to sidestep responsibility. In truth, I never believed that one op-ed column would trigger a U-turn in poor educational examination methods at that or any other higher institute of learning. As many have said about such stoic institutions: theories, beliefs, and practices die one generation at a time. But the conversation and battle must start somewhere.