What is better than the drowsiness that eases you into sleep? It’s such a guiltless, hedonic experience: muscles relax, eyelids flit then shut, consciousness slips. Soon you’re in the early stages of dreamless, non-rapid eye movement sleep, when slow-moving delta waves of electrical activity wash across your brain.
That’s good drowsiness.
Then there’s the bad kind, when your neurobiology is insisting that it’s time to sleep, but circumstances simply can’t—or shouldn’t—allow it. You might be driving a car or operating on a patient or banging away on a keyboard.
That drowsiness starts a tug-of-war between wake and sleep that makes it difficult to maintain attention—especially in monotonous situations—and slows down reaction times. “Wake state instability” is David Dinges’s term for it, and few people know more about sleep deprivation and the drowsy brain than Dinges, chief of the division of sleep and chronobiology at the University of Pennsylvania. “You get to the point where you’re fine one second. The next second, you’re lapsing. You’re fine again, then you lapse again,” he explains.
More people are spending more time in this herky-jerky state of mind than ever before. Survey after survey indicates that people are sleeping less and less. Last year, for example, the Centers for Disease Control reported that 30% of employed Americans said they sleep six hours a night or less. Fifty years ago, those people were a tiny minority. While some people are true “short sleepers,” genetically endowed so that they hold up well when sleep deprived, hundreds of sleep lab studies and 40 years of epidemiological data show that six hours is not enough for most people, and that bad things—depression, metabolic maladies, heart disease, premature mortality—start to happen outside the range of seven to eight hours of sleep per night.
Researchers like Dinges are slowly piecing together the puzzle of what happens inside a drowsy brain. PET imaging has shown a general decrease in brain activity after sleep deprivation, with larger reductions in some areas like the thalamus, which helps to regulate sleep. Functional MRI (fMRI) has painted a more complicated and nuanced picture. While lack of sleep results in less activity in the fronto-parietal parts of the cortex—those needed for task-specific attention—increased activity in the thalamus, anterior cingulate, and other arousal structures can get busy in attempt to compensate.
Work hours are only part of the equation.
In the last several decades, the drowsiness problem has swelled to epidemic proportions, but collectively, we’re only dimly aware of its depth and extent. The fight against it has been hampered by both societal misunderstandings and a lack of scientific knowledge. Drowsy driving legislation is on the books or has been proposed in about a dozen states, but because there isn’t an objective measure of drowsiness, it’s difficult to draft and enforce stricter laws. Work rules and regulations are another stumbling block because they’re rooted in the mistaken notion that fatigue from working long hours is where the danger lies. In reality, Dinges says, work hours are only part of the equation. Commute times and personal habits outside of work are just as important. “You’ll find people who work 80 hours a week and get plenty of sleep. You’ll also find people who work 30 hours a week and don’t,” Dinges says. “The whole culture is using a set of rules that don’t apply.”
Technology might ride to the rescue. Automated cars could turn drivers into just another passenger, removing drowsiness and its risks from the equation entirely. In the meantime, car companies are loading up their vehicles with safety features that, for the most part, monitor the car and correct its course or warn the driver. Monitoring the driver has taken a back seat, which some safety experts say is a missed opportunity.
Dinges’s ambitions run higher than mere monitoring. He wants an all-fronts attack that would predict, prevent, and detect drowsiness and then intervene if necessary—not just in cars but in any circumstance where inattention can cause significant harm. But first, he and his colleagues have to understand what’s happening in our brains during the flickering moments between wake and sleep.
Paying Attention to Attention
In the early 1920s, the founding father of sleep research Nathaniel Kleitman discovered as a young researcher that it wasn’t easy keeping six University of Chicago male students awake for three straight nights. Observers had to keep them moving all night long or else they’d conk out. Some would try to slip the watchful eye of their minders, pretending to go for a walk in a corridor so they could “walk over to some corner, sit down, and fall asleep almost immediately.”
Sleep researchers have come up with other means for keeping their subjects awake, but Kleitman’s 1923 account of his classic “experimental insomnia” study includes observations about the effects of sleep deprivation that still ring true. During the day, for example, circadian rhythms kicked in, and the overwhelming desire to sleep eased up. But Kleitman’s study volunteers couldn’t listen to a lecture or take notes without words turning into unintelligible scribbling, and on the day after the third sleepless night, taking notes was “entirely impossible.” Yet their ability to name letters and do mental arithmetic wasn’t affected.
Research has since confirmed that the effects of sleep deprivation and subsequent drowsiness are uneven. A few areas of high-level thinking, such as logical reasoning, is left pretty much intact when people stay up all night, while multitasking and flexible thinking end up in shambles. Some researchers pushed the idea that sleep deprivation has a particularly pronounced effect on working memory and executive function—all the processing power needed for planning, plotting, and coordinating. But others—including Kleitman, back in the day—have noted that it is monotony and inherently less-engaging tasks that are the real snoozers for the sleep deprived. Boredom makes rested people annoyed and frustrated, Dinges says. But it puts tired people right to sleep.
By dint of hundreds of experimental results and review articles, Dinges has advanced the notion that lack of sleep most reliably and fundamentally affects attention and vigilance—our ability to focus in a stable way on one aspect of the external world swirling around us. If economics is the science of scarcity, then attention is brain economics: “It allocates processing resources to the stimulus of interest,” explains Michael Chee, a sleep researcher at the Duke-National University of Singapore Medical Graduate School. The things that don’t have our attention receive fewer cognitive resources.
While learning, executive function, and short-term memory are all affected by sleep deprivation, Dinges argues that attention is “foundational,” the base layer of cognition upon which those and other types of thinking are built. If attention is shaky, then they’ll wobble, too. Several years ago, Dinges reviewed papers published on 70 short-term (less than 48 hours) sleep deprivation studies that included a total of 1,533 study subjects. He crunched the results of the individual studies and concluded that, of the cognitive domains, sleep deprivation affects simple, sustained attention the most.
Lapses—errors of omission—are one way we experience this flickering attention. But Dinges’s own experiments, where participants pressed a button in response to a cue, suggest it may also cause us to overcompensate. “You know you’re supposed to not be lapsing, so you compensate by pressing the button too soon. So now you’re committing errors of commission, too,” he says.
Sleep deprivation amps up risk taking, too. Chee says that in two of the gambling studies he conducted, he discovered what casino owners probably figured out long ago—brains short on sleep are more likely to hope for big gains and undervalue losses. Functional MRI of sleep-deprived brains showed that the ventral striatum, a part of the brain that serves to anticipate gain, was more active, while the insula, where many emotions are processed, was less active. Because other types of decisions fraught with risk don’t produce these effects, Chee is now testing various types of gambling games to tease out more precisely which aspects are affected by sleep deprivation.
“The brain can’t tell”
Of course, people can and do recognize when they’re nodding off. Often, though, people push themselves to stay awake. Those motivational thoughts can dominate their self-awareness. That’s part of why drowsiness can be so dangerous. If people are driving drowsy, the desire to stay awake until they reach their destination may push them to drive faster so they can get there quicker, Dinges says. Now they’re driving drowsy and speeding. Not a good combination.
In Dinges’s pioneering studies, which restricted sleep to four or six hours per night over several weeks, he found that people are actually pretty good at judging their impairment during the first few days of sleep restriction. But as the number of short-sleep nights piles up, people grow more impaired and their judgment worsens. At that point, they think they’ve adapted to getting a few hours of sleep a night, but they haven’t. “Humans are abysmally poor judges of their true incapacitation from sleep loss,” Dinges says. “Our subjective sense of ourselves is very much dependent upon acute, immediate changes, not chronic conditions,” he adds. “At some point, a brain can’t tell.”
Yet Dinges has also noticed that people respond to sleep deprivation differently. We generally fall into three categories, he says. Type 1s keep it together and continue to perform well on various cognitive tests, though like others, they do notice that they’re sleepy. Most people are type 2s: sleep deprivation causes deficits, but circadian rhythms perks them up. Type 3s have little tolerance for sleeplessness. After about 16 hours of being awake, they experience cognitive deficits, too, but don’t benefit from “circadian rescue.”
These divisions may have a genetic underpinning. Researchers have identified “sleep genes” and polymorphisms associated with differences in how people respond to sleep deprivation, though there are probably not just one or two. At least in the experimental setting, the type 1, 2, and 3 groupings follow a normal distribution, suggesting that a number of genes are involved in susceptibility to sleep deprivation.
Under the Hood
Though the genetic picture remains hazy, our understanding of what’s happening inside the brain is growing sharper. Imaging technology is allowing researchers to see what’s going on in drowsy brains. Sleep researchers have had the most luck recently with fMRI, which measures neuronal activity indirectly by measuring oxygenated blood flow. No radiation is involved, and it can be repeated under many different conditions.
Drowsiness studies using fMRI have also revised our understanding of how the brain focuses our attention. Attention has been viewed as very much a top-down, frontal lobe–directed phenomenon—essentially, that we must actively boss it around. But Chee’s fMRI-enabled research of visual processes has shown that the parietal lobe, which processes sensory information, is consistently affected by sleep deprivation. Research by Govinda Poudel and colleagues at the New Zealand Brain Research Institute has shown that the task-oriented parts of the mildly sleep-deprived brains start out okay, but as they work neuronal activity drops off. Meanwhile, activity is high in the arousal network, including the thalamus and anterior cingulate, reflecting the drive to stay awake.
Sleep is a dynamic, bottom-up process, analogous to a flock of birds changing direction.
Other fMRI-enabled research has shown the consequences of sleep deprivation at the brain’s network level, particularly the default mode network, which comprises areas of the brain that are active during inward cognitive experiences (like daydreaming and remembering) but go off-line during external tasks. If you prevent a sleep-deprived brain from following its natural instinct to sleep, the lively “talk” among the various parts of the default mode network grows sparse, as does the communication between the default mode network and its counterpart, which activates during externally oriented tasks.
Findings like these, which identify very local spots of decreased neuronal activity, support an explanation of sleep put forward by James M. Krueger, Giulio Tononi, and others. Sleep, Krueger explains, is not orchestrated by one particular part of the brain but is instead a dynamic, bottom-up process, analogous to a flock of birds changing direction. Small networks of neurons, called cortical columns, tip into a sleep-like state—Krueger is careful to avoid using the word sleep —depending on how active the neurons have been. As more and more cortical columns go into this sleep-like state, sleep occurs as an emergent property.
The mental fumbling that occurs when you haven’t had enough sleep is a matter of not having enough “awake” cortical columns, Krueger says. “When 5,000 of the 10,000 columns go into a sleeplike state, your performance will drop just because you don’t have as many columns involved in fulfilling the performance.”
Living With Less
The thing about sleep deprivation and drowsiness is that we know it’s bad for us but we cheat on sleep anyway. Dinges thinks our brains are a house divided, wired so that the higher parts flout what our mid- and lower-brain biology is trying to tell us. “Our brains elect to use time as a commodity and violate this neurobiology that regulates the timing of wake and sleep in all other species.”
Occasionally, ignoring those signals can have catastrophic consequences. The Exxon Valdez oil spill, where a drowsy captain ran his supertanker aground, is one example, the Three Mile Island nuclear disaster another. But more often, the consequences are smaller in scale and underplayed. For example, for years government reports and independent research estimated that 3-4% of fatal car crashes involved drowsy drivers. But when Brian Tefft, a researcher at the AAA Foundation for Traffic Safety, filled in the gaps in the data using statistical techniques, he calculated that 16.5% of fatal car wrecks in the United States—resulting in about 4,000 deaths—involved a drowsy driver. Clearly it’s more than a minor problem.
Drowsy driving legislation is on the books or has been introduced in roughly a dozen states. Many of these states have laws that are on the innocuous end of the spectrum, emphasizing awareness and education. At the other end is New Jersey, which has one of the toughest. “Maggie’s Law,” named for a college student who was killed by a driver who hadn’t slept for 30 hours, allows prosecutors to charge drivers with vehicular homicide if they cause a fatal crash after not sleeping for 24 hours. Still, prosecutions under such statues are rare and difficult. Drunk driving laws might be a model for driving while drowsy laws, but without an easy, objective test for drowsiness like blood alcohol content for drunk driving, it’s hard to imagine how they’d work.
Mitigation is a more practical—and consequential—strategy than criminalization. Some like rumble strips and guard rails, are already in place. Others, like that being developed by Azim Eskandarian, are on the near horizon. Eskandarian, a professor of engineering and applied science at George Washington University and director of the school’s Center for Intelligent Systems Research, has created a system for detecting the small, moment-to-moment corrections drivers make to stay in their lane. When drivers are alert, those “lane keeping” corrections are continual, he explains, but when they’re drowsy, they occur less often and are more drastic.
While you can’t buy Eskandarian’s system in a car today, manufacturers currently offer other collision-warning and automatic braking systems. Such monitoring and warning systems are a bridge, keeping us safe until driverless cars becomes a reality. For now, Dinges says, “you’ll still have the joy of driving, but there will be increasing protections against the catastrophic event.”
Not all drowsiness mitigation needs to be engineered: A good nap can work wonders. A study published in the Journal of the American Medical Association last year showed that a “protected sleep” policy for medical residents that involving taking away cell phones for a five-hour sleep period not only increased the amount of time they actually slept, it also improved their alertness when they woke.
Despite such measures, we shouldn’t expect the problem of sleep deprivation and drowsiness to go away, Dinges says. Our modern, stimulated brains do not go gentle to bed. “I think we want more time awake and alert and able to do things. We love it,” Dinges says. “We’ll sit up at night, whether it’s playing video games or shopping on the computer or doing work off hours. We love cognitive flexibility.”
Lim J, Tan JC, Parimal S, Dinges DF, Chee MWL. 2010. "Sleep Deprivation Impairs Object-Selective Attention: A View from the Ventral Visual Cortex." PLoS ONE 5(2): e9087. doi: 10.1371/journal.pone.0009087