Sleep On It

Sleep might just be the most important part of daily health—and the biggest new target for biomedical engineering.

Every night, around the world, 7 billion people lie down to sleep. Their eyes close, their bodies relax, and their brain waves begin to smooth from the chaos of wakefulness into slower, synchronized waves. As their thoughts begin to lose coherence, a part of the brain called the ventrolateral preoptic nucleus, located deep behind the eyes, begins to blanket the nervous system in quieting chemicals like gamma-aminobutyric acid and galanin, shutting down the networks of wakefulness and turning off the body’s awareness of the world around it. It is the most important part of these people’s days.

Over the course of the night, these sleepers, if they’re typical, will cycle through four stages of sleep—three nonrapid eye movement (NREM 1–3) stages and rapid eye movement (REM). The deeper the NREM stage, the slower and deeper their brain waves will be. In NREM 2, this slowing is interrupted by short flurries of activity and high-amplitude spikes—sleep spindles and K-complexes—that burr irregularly through the brain, though by NREM 3, these will have blurred into the deep and steady rhythms of slow-wave sleep.

During this phase, waves of hyper- and depolarization methodically sweep through the neurons of the cortex and thalamus in cycles of up to 20 per minute, like a chorus of crickets that build to crescendo over and over again. The body’s breathing is deep and regular. Blood pressure, insulin sensitivity, and the stress hormone cortisol drop, and the tissue-repairing growth hormone levels surge. Eventually, this rhythmicity stops: the brain becomes active, and acetylcholine floods in at concentrations twice that of wakefulness. Inhibitory chemicals paralyze the body, except for the eyes, which begin to move rapidly. This is REM sleep—dreaming sleep, or paradoxical sleep, because in an electroencephalogram (EEG), the brain truly does look almost like that of someone who’s awake. Then, eventually, the REM stage will end and transition into light sleep, NREM 1, and the cycle will begin all over again.

Why we sleep is still up for debate, but sleep we must. The act of sleeping and the biological rhythms that drive it are tied in to almost every major organ system, every major chemical communications network—even the very patterns of our protein transcriptions. Failure to sleep enough—and very few of us do sleep enough—can increase a person’s risk of obesity, diabetes, even certain cancers.

“Sleep is this profoundly important behavior,” explains Russell Foster, a circadian neuroscientist and head of the Nuffield Laboratory of Ophthalmology at the University of Oxford, United Kingdom. “It is not an indulgent behavior where the brain is shut down, it’s a critical aspect of our biology—and, finally, that is being appreciated.”

Nature’s Soft Nurse

Before scientists recognized sleep as the active process that it is, it was conceived largely as an exercise in passivity. It was the absence of wake—not quite death, but not very interesting either. Not until the arrival of the EEG did that view begin to shift, first in 1937 with the describing of NREM sleep stages by Alfred Loomis in his New York laboratory, and then with the discovery of REM sleep and its connection to dreaming in the 1950s by Chicago researchers Nathaniel Kleitman, Eugene Aserinsky, and William Dement, which kicked the field into high gear.

Since then, scientists have established the basics of a complex two-process system believed to drive sleep: sleep homeostasis and the circadian rhythm. The first is the most mysterious—that’s the need for sleep that builds the longer a person stays awake. Researchers suspect that something must accumulate during wake that acts to promote sleep, which then dissipates upon actually sleeping. Currently, the best candidate for this “something” is adenosine, a chemical known to inhibit transmitter systems associated with wake—thereby promoting sleep—and which, in fact, does build up over the day as a by-product of the biochemical reactions that power cells. By no coincidence, it’s also the very molecule that caffeine blocks. But since multiple chemicals from all sorts of different physiological systems also meet at least some of these same criteria, it’s most likely that more than one chemical drives the pressure to sleep.

The circadian rhythm is much better understood—this is the body’s biological clock, the heart of which lies in a 20,000-cell part of the brain called the suprachiasmatic nucleus (SCN), which is wired directly to the eyeballs and “set” by exposure to light. Its roughly 24-hour oscillation comes from an interplay of genes and proteins that operate in a cycle of transcription, binding, inhibition, and degradation that takes about a day to complete. Via projections through the brain, the SCN directly promotes both wake and sleep. But more than that, it also helps orchestrate the daily rhythms of the billions of peripheral clocks located in essentially every organ of the body: levels of inflammatory cytokines from the immune system rise and fall, and fat cells break down and produce lipids at different rates according to the time of day. The intestinal clock works to promote greater motility and absorption during the day than at night, and many metabolic hormones including the hunger hormone ghrelin, the fat-cell produced satiety hormone leptin, and pancreatic insulin all vary in concert to aid our daily feeding and fasting.

“Our hormones are actually set up to process food better during the daylight hours,” explains Michelle Miller, who co-leads the Sleep, Health, and Society research program at Warwick Medical School, Coventry, United Kingdom. But, as she explains, the flip side of that is that when we eat and sleep against the clock or skimp on sleep, our bodies backfire. To illustrate, she points to research showing that dieters who ate their main meal earlier in the day lost more weight more quickly than those who ate later in the day—even when there were no real differences between the groups’ caloric intake. Similar evidence also suggests that people who sleep around five hours a day may tend to eat more at night and gain weight, and conversely, eat fewer fats and carbohydrates, and lose weight when they’re better rested.

As a matter of fact, over the last ten years, shift work and disordered sleep have been linked with every major modern ailment: obesity, cancer, heart disease, mood disorders, and neurodegenerative diseases like Parkinson’s and Alzheimer’s. Those who consistently sleep five hours a night or less are 28% more likely to have diabetes, 15% more likely to have had a stroke, and 48% more likely to have had heart disease, compared to their better-rested counterparts. Their reactions to influenza and hepatitis vaccines are blunted. Six hours of sleep per night or less ups the risk of a premature death by 12%, which, as Miller and her colleagues point out in a 2010 paper, “if causally related, would equate to over 6.3 million attributable deaths in the United ­Kingdom in people over 16 years of age and over 25 million attributable deaths in the United States in people over the age of 20 years.”

At least some of these effects may be mediated by epigenetic mechanisms—indeed, researchers believe that between 3 and 20% of genes operate in a circadian fashion. But these patterns are all complicated by the fact that the SCN is not the only way to set these peripheral clocks—an enormous number of internal signals and external inputs can reset circadian gene expressions. Food and fasting are particularly potent circadian stimulants: up to 15% of the expressed gene transcripts in the mouse liver show circadian rhythms, but when mice are fasted, more than 80% of these transcripts stop cold. Shift workers in particular, face high odds of obesity, metabolic syndrome, and fatty liver disease relative to their nine-to-five colleagues, and researchers suspect that at least part of this might be due to a food-induced decoupling of the digestive clocks from the master SCN.

Such effects don’t stop at the body either. Poor sleep and sleep deprivation slash attention, memory, perseverance, and judgment. In children, it can manifest as restlessness and moodiness, leading to misdiagnoses of attention-deficit-hyperactivity disorder. Some research has shown that just 17–20 hours without sleep is on par with a blood alcohol content between 0.05 and 0.1%, and the economic costs of sleep-related car accidents in the United States have been estimated at US$43–56 billion. Cumulatively, the effects of sleep deprivation can sum to staggering amounts. In Australia alone, the combined total costs of sleep disorders have been estimated at almost US$7.5 billion.

“We’re going back to this question of ‘what is the role of sleep in health?’ and it looks like pretty much wherever you look, you see something,” says Michael Grandner, a sleep researcher at the Perelman School of Medicine at the University of Pennsylvania, in Philadelphia. “It makes sense. Just like your diet affects everything somehow or other, either weakly or strongly—sleep is the same way.”

Sleepless in Society

The tragedy of this research is that at the very moment science is uncovering all these deadly effects of sleep loss, the modern world has come to sleep less and less. A third of adults in the United States sleep six or fewer hours a night these days—two hours less than they slept on average half a century ago, and the U.K. Sleep Council estimates that similarly impoverished sleep affects nearly half of Britons.

There are plenty of reasons for this trend—there’s our globalized 24-hour commerce and communications, our workday cultures, late-night social lives, and ubiquitous technologies. “We’re all carrying these mobile devices, checking our e-mail late into the night—for some people in the middle of the night—it didn’t used to be like that,” says Steven Shea, director of the Oregon Health and Science University’s Center for Research on Occupational and Environmental Toxicology. “As a society, we’re shooting ourselves in the foot.”

But sleep disorders also comprise a good part of the problem. The U.S. Centers for Disease Control and Prevention report that at least 50–70 million adults in the United States have a sleep or wakefulness disorder; Australian experts have estimated that chronic primary sleep disorders affect at least 6% of the population, and probably more.

The newly revised third edition of International Classification of Sleep Disorders (ICSD-3) lists over 70 different sleep disorders and variants. Some of these are comparatively uncommon—narcolepsy, the collapse-where-you-are-and-sleep disorder that’s caused by a lack of the hormone orexin, affects just 0.02–0.18% of the populace. Parasomnias like sleepwalking and sleep talking are similarly rare and affect children more than adults—though one exception to this is REM sleep behavior disorder, where the brain fails to paralyze the body during sleep. It often emerges in older men as a result of early neurodegenerative disorders like Parkinson’s disease, which can cause damage to parts of the brain’s sleep centers.

On the other hand, there are disorders like chronic insomnia, which afflicts an estimated 10% of the population, and transient insomnia, which affects up to a third. Researchers still don’t know for certain the root cause of this particular issue, but there is some evidence to suggest that insomniacs might be more physiologically activated than their better-rested peers: studies have shown that insomniacs seem to have a higher body temperature around bedtime, a higher metabolic rate across the 24-hour period than noninsomniacs, and higher frequencies in their EEGs during sleep, and they may also release higher levels of stimulatory compounds like cortisol and adrenocorticotropic hormone.

Easily the defining sleep disorder of the modern age, however, is the breathing disorder obstructive sleep apnea (OSA). The gist of this condition is self-suffocation: during sleep, the body’s muscles relax, and the weight of the neck collapses down to block the upper airway. Oxygen falls, blood pressure drops, then spikes, and the body’s fight-or-flight system surges to life to wake the person up and restart breathing. This can happen ten, 30, or 100 times an hour in a given night, and, most of the time, the patient doesn’t even know it’s happening—they just know that they’re exhausted. Unsurprisingly, perhaps, the long-term effects of these nightly episodes can be deadly: a sevenfold increased risk for road accidents; higher rates of hypertension, diabetes, and coronary artery disease; a two- to threefold increase in the risk of dying from heart attack or stroke; and memory problems.

Because OSA is both a result of excess weight and a contributor to it (by way of all the metabolic disruption that comes with poor sleep), it has spread hand-in-hand with the obesity epidemic. Today, common statistics for the prevalence of OSA typically range from 2 to 7% in women and men, but numbers as high as 24% of men and 9% of women have also appeared in the literature, and the disease is believed to still be underdiagnosed by up to 85%—which makes for a great many people at risk for a great many problems, and they may have no idea.

Smarter Sleeping

The combination of medicine’s growing respect for sleep with the broad prevalence of inadequate sleep has begun to open new technology opportunities. For instance, the gold standard for any sleep diagnosis has traditionally been polysomnography (PSG), a multisensor overnight test conducted in a sleep clinic. But now, in the face of growing financial and clinical pressures, PSG is now giving way to home sleep apnea test devices (See Grifantini’s article “How’s My Sleep?” in this issue of IEEE Pulse).

Likewise, a host of creative technologies for apnea and other sleep issues are under development worldwide: dual-heat-flux probes embedded into neck pillows from Seoul National University; high-resolution pressure-sensitive bedsheets from the University of California, Los Angeles; a bedside unit from the Colorado-based SleepImage that can monitor movement, body position, and cardiac information through a simple chest electrode; and smart mattress pads, shirts, and pillowcases. Some of these are for sleep apnea, but many of them are intended to help researchers begin to study sleep in a larger, longitudinal fashion.

“Overnight sleep studies are great for measuring one night,” Grandner says, “but they don’t capture what your sleep is like in the real world.” This shift toward portable home technologies, he says, “will help us focus sleep away from being a purely diagnostic discipline—which it largely was—to more of a chronic disease management discipline, which is what it should be, because sleep disorders are chronic disease.”

Back in the lab, meanwhile, other researchers are making new forays into the field to try to get at the many unknowns that remain—like why we need to sleep. A widely hailed paper by a team at the University of Rochester, New York, reported last year that in mice, the influx of cerebrospinal fluid into the brain increases by 60% while they sleep, flushing out cellular trash like -amyloid proteins, like a sort of nightly cerebral dishwasher. Similarly, at the University of Wisconsin in Madison, a group headed by Giulio Tononi has published a series of studies exploring the parallel hypothesis that the brain may be pruning back synapses wired over the course of a day’s experience, so as to sharpen up the necessary memories while maintaining a manageable signal-to-noise ratio. And at Massachusetts General Hospital, Sleep Division Director Matt Bianchi has begun re-examining a decade’s worth of patient PSG data to see what can be found.

“If you take the standard PSG and you just analyze it in a different way,” he says, “it almost doesn’t matter which signal you’re talking about: you can get something more out of a deeper analysis than what the standard scoring brings you.” Imagine a world, he says, where a clinician could not only tell a patient whether they have apnea, but start to substratify them by type. “Maybe there’s actually 20 kinds of apnea,” he speculates. “Or maybe it could be: your REM sleep looks okay to the naked eye, but the spectral analysis shows us something else—or any number of combinatorial things that you can ask, taking advantage of the rich data.”

Slowly but surely, sleep’s importance is beginning to pervade society too. Schools in the United States have begun to push back their start times to accommodate adolescent sleep patterns. Over the past few years, companies like Google, Cisco Systems, and Proctor and Gamble have installed nap pods in their workplaces. The entire town of Bad Kissingen, Germany, last year announced its intention to reorient itself around people’s circadian rhythms and sleep. An entire industry of sleep monitors has emerged in force for the health-conscious consumer, from the simple—the Jawbone UP and Fitbit Flex actigraphy bracelets—to the more sophisticated, like the Beddit mattress probe and the multisensor Basis Science band (see “Regular Sleep” and refer to the article on page 14 in this issue by Kristina Grifantini, “How’s My Sleep?”).

Regular Sleep

How are regulators and device makers dealing with the wild west of consumer-aimed mobile sleep health technology?

By some estimates, 97,000 mobile health apps exist today, and nearly three quarters of them are aimed at the consumer. What’s more, this same consumer is expected to collectively purchase more than 17 million wearable tracker bands this year. In three years’ time, 3.4 billion people will own a smartphone, and at least half of them will be using those phones to run mobile health apps. It is, in short, the dawn of the age of mobile consumer health.

Thanks to medicine’s growing appreciation for the importance of good sleep to health and well-being, sleep tracking will almost certainly play a role in this expansion of consumer-driven tech (see Kristina Grifantini’s article “How’s My Sleep?” on page 14 in this issue). Popular devices like the Jawbone UP and the Fitbit Flex band already provide some sleep-tracking capabilities, offering wearers the chance to monitor how long and how well they slept through the night, often with a rough play-by-play of their stages of sleep, and a new generation of more sophisticated technologies like the Beddit mattress sensor will only build upon this further.

Technologies like these have the potential to do genuine good just by making individuals more aware of their sleep and the factors that influence it for better or worse. But they aren’t wholly without their risks, and the current range in the quality and quantity of such tools pose a daunting challenge to regulators, who have only just begun to turn their attention to this area.

All mobile health devices have to face certain issues—data protection and privacy, transparency, ease of use. But at the moment, when it comes to mobile sleep health, possibly the biggest question mark is accuracy. Right now, the majority of consumer sleep-tracking devices rely on actigraphy, or motion sensing, to derive their data. By default though, actigraphy doesn’t measure “sleep”—only neural sensors like the EEG can do that—it just measures how much a person’s body moves during the night. The technology works well enough for normal healthy sleepers, but it has a high margin of error, tends to over- and underestimate sleep—by well over an hour sometimes—and flat out isn’t reliable for disordered sleepers. This poses two problems: first, inaccuracies could potentially give a person with an undiagnosed sleep disorder a cleaner bill of healthy sleep than warranted, discouraging them from seeking needed treatment. Second, it could also lead to false alarms in healthy people. “In turn,” write a team of engineers from the University of Oxford in a 2013 review of sleep-screening apps and issues, “this may lead to large-scale inappropriate resource allocation or may even overwhelm the health care system, even if the app is 99% accurate.”

The regulatory community has been slow to act compared to the speed of the mobile marketplace, but that’s beginning to change. In the fall of 2013, the U.S. Food and Drug Administration (FDA) released its final guidelines for medical apps, updating a draft version that had been under review since 2011. In essence, the administration is picking its battles: it’s flat out not going to regulate innocuous apps, like those that just provide education or information—medical dictionaries or interactive diagrams—but it will definitely regulate mobile health tech that could cause harm to a patient if they don’t work properly—targeted disease diagnostic apps, apps that calculate insulin diagnosis, etc. If it acts like existing technology that is already regulated, chances are the mobile version will also be subject to regulation.

However, sleep and fitness trackers fall squarely into a middle ground, where the FDA “intends to exercise enforcement discretion.” It might regulate them, or not, as it sees fit depending on their potential risk. Naturally, this creates something of a gray zone for tech creators. Preparing for FDA registration or approval means extra steps, extra costs, and long times before a product goes to market, but what if something in this category goes to market without regulation, causes significant harm, and lands the device’s makers in court? What’s more, while these are final guidelines, they are by no means written in stone. The FDA has already made clear that they will keep tabs on the situation and change their stance if needed, and several pieces of legislation have been proposed to the Congress and Senate that would diminish the administration’s health information technology authority.

Meanwhile, Europe is in an even greater state of flux. Right now, the European Union (EU) doesn’t formally distinguish between general wellness and tracking apps and established medical and diagnostic devices. There are directives that help manufacturers decide whether their device or software falls under medical device or in vitro diagnostic devices, and if a software or mobile technology does fall under either of these categories, they are regulated as such. But if something doesn’t really fit under either of those categories, it’s not clear what rules should apply. Right now, the EU simply doesn’t have anything like the FDA’s flexible “discretionary” power. This is likely to change within the next couple of years. Earlier this year, the European Commission published a green paper on mobile health, inviting comment from the larger community onto a series of different questions specific to consumer mobile health tech, ranging from whether current EU legislation is sufficient to account for emerging lifestyle and well-being apps, to how to safeguard user data and privacy, and the best ways to ensure app safety and transparency. Feedback on these points will be collected until the beginning of July 2014 to be integrated, the Commission hopes, into updated policy decisions at some point over the following year.

How are companies coping with the uncertainties? One of the most longstanding specimens in the consumer health tracking group, the BodyMedia armband, is registered with the FDA as a class II, 510(k)-exempt “isokinectic testing and evaluation system”—but that categorization is both partially related to the way the company has heavily involved itself in the clinical research realm for obesity and diabetes, an extreme rarity in this area. In fact, most other such devices actively steer themselves in the opposite direction.

“We degraded the service in terms of medical accuracy, and we don’t make any medical claims, but we do it on purpose,” says Uli Gal-Oz, chief executive officer of the California and Israel-based company SleepRate, which combines heart rate variability data from a chest sensor with smartphone recordings of environmental noise to analyze sleep and give targeted cognitive-behavioral tips. The company’s walking a fine line, because it does pinpoint insomnia specifically in the way it promotes itself; however, it’s also low enough in risk—its tips include things like waking up at the same time every day—that it’s unlikely to raise serious red flags. “Basically,” he explains, “it’s all about reaching the consumers and helping them as early as possible—before they develop a medical condition.” It’s the difference between taking the time to submit a medical-grade product to the FDA or giving a simpler “wellness” version to the consumer right now.

“They don’t want to cross the barrier into a diagnostic device, because that crosses a whole bunch of different systems,” says Douglas Kirsch, a neurologist and sleep specialist with Harvard Medical School and Brigham and Women’s Hospital in Boston, Massachusetts, speaking of consumer sleep tech makers. “The fact is that if they can do it on a consumer device, it can probably be more readily adapted than the clinical device.”

Marco Della Torre, the vice president of product science at the California-based company Basis Science, agrees. His company has developed a sensor-packed wristwatch-style fitness tracker that combines actigraphy with measures for skin temperature, perspiration, and blood flow to derive both standard fitness variables, like calories burned, and sleep stage analyses. It’s a little more sophisticated than something like the Jawbone, but it’s likewise unregistered with the FDA. “There is so much we can learn about ourselves, and improve for everyday health and fitness. You don’t need to be a medical device to get a lot of benefit from a system like this,” Della Torre says. Regulation, he points out, isn’t the only way that mobile tech like his can try to establish its quality. Basis has recently released a white paper study reporting that sleep stage data derived by the tracker correlated with PSG by a coefficient of 0.92 (where perfect correlation is 1.0). The study has not been submitted to a journal for peer review, but it’s more than many competitors have done. Alternatively, voluntary health app certification and review sites are beginning to spring up outside of formal regulation, such as the Health Apps Library in the United Kingdom, which only lists apps that have been vetted by the National Health Service to ensure their compliance with data protection laws, reliability of information, and minimal potential harm.

Sleep monitoring and fitness trackers provide so many potential benefits to people now, he says. Regulating that more stringently could make that more difficult to access, Della Torre worries. “All those consumer benefits—I would hate to see any change in people’s ability to access that kind of information,” he says. But he’s optimistic. “It’s one of those things we’re all watching very closely, and it’s tough to speculate where the FDA will go,” he says. “But I don’t think that’s going to happen.”

“It’s definitely a time of rapid change, but it’s an exciting time, too,” says Ronald Chervin, director of the Sleep Disorders Center at the University of Michigan Health System in Ann Arbor. “We have enormous potential going forward to improve the hardware, and even given the current hardware, to use signal analysis to really make these tools more useful, with the goal being inexpensive, easy, widely available sleep screens or assessments. There’s nobody who doesn’t sleep. We spend a third of our lives asleep, and, for most of medicine’s history, that third has been ignored—to the detriment, unfortunately, of our waking daytime health.” Sleep health, he says, “has a lot to offer to human health and well being, and I think we’re going to see that.”

Shannon Fischer is a freelance science writer living in Boston, Massachusetts.