The Problem with Tracking Sleep Data

The 2022 Tour de France Femmes was decided in the Vosges mountains, during a brutal seventh stage with three category-one climbs. Dutch rider Annemiek van Vleuten attacked on the second climb, then opened up a four-minute gap on the final push of the day, a grueling 3,163-foot ascent of the Grand Ballon. It was the hardest day of the Tour, and with another mountain stage coming the next day, recovery was crucial. But with their legs fried, their cortisol levels soaring, and their nervous systems cranked in fight-or-flight mode, would the riders actually be able to sleep properly?

Surprisingly, the answer was yes—or at least, mostly. Nine of the Women’s Tour riders were wearing Whoop bands on their wrists; their data, which was published earlier this year in Sports Medicine—Open, showed that the riders got an average of 7.6 hours of sleep that night, compared with an overall average of 7.7 hours both before and after the Tour. They did, however, spend a little more time than usual in light sleep and less in restorative REM sleep. Whether that matters in any practical sense is the fundamental question confronting athletes, coaches, and sports scientists as they enter a new era of sleep tracking. The technology is better than ever; we just have to figure out what to do with it.

Tracking Sleep Stages Is Still a Challenge

Sleep is hardly a new biohack, but it has been a hot topic in performance circles ever since neuroscientist Matthew Walker’s 2017 book Why We Sleep. The problem with first-generation sleep trackers, though, was that they relied on accelerometers and basically assumed that if you weren’t moving, you were asleep. The latest generation of devices is more sophisticated, adding heart-rate measurements and other physiological cues like breathing rate and skin temperature to refine their algorithms, and able to tell the differences between distinct sleep stages. As a result, says Charli Sargent, a sleep scientist at Central Queensland University in Australia and lead author of the Tour de France study, “The whole world is becoming a sleep laboratory.”

Companies like Apple, Garmin, Oura, Polar, and Whoop have gotten very good at detecting sleep. Compared with sleep-lab studies, where subjects are wired up to record brain and muscle activity, the latest consumer wearables were typically 86 to 89 percent accurate at determining whether a wearer was asleep or awake, Sargent and her colleagues found. Detecting individual sleep stages, on the other hand, is still a work in progress: the wearables only got it right 50 to 61 percent of the time.

The picture for athletes is more complex. Many of the new sleep-stage algorithms rely on heart-rate variability, or HRV, the subtle fluctuations in timing from one beat to the next. HRV changes with sleep stage, but it’s also influenced by vigorous exercise. Indeed, Sargent found that HRV was systematically lower after mountain stages in male Tour de France riders. Another new study, led by Marc Poulin of the University of Calgary, had a group of healthy volunteers do a hard interval workout in the early evening, then tracked their sleep with an HRV-based Polar watch as well as collecting gold-standard sleep-lab data. The good news: the accuracy of the sleep tracker was undiminished by the workout.

What Can Athletes Do with the Data?

Overall, then, wearable sleep trackers are already pretty good, and they will likely continue to improve. The next question—the really hard one—is what we should do with the data. If cyclists are getting less REM sleep after mountain stages, what should they do differently? “Ride easier” isn’t useful advice; and it hardly seems like we need a fancy algorithm to give us the usual sleep-hygiene advice about bedtimes, alcohol, and electronics before bed.

For some people, simply having objective data about when to hit the hay and when to wake up might function as a useful reminder to cover these bases, in the same way a step tracker spurs you to get your 10,000 steps. Athletes might also be interested in seeing how their sleep changes at altitude, as an indicator of whether they’ve acclimatized and are ready for hard workouts. And there may eventually be subtler insights: for example, preliminary data from Poulin’s lab in older adults suggests that those who don’t get enough deep sleep are more likely to develop cognitive problems years later. For now, the best approach is to establish a baseline and then look for changes, Sargent says. If you usually get 15 to 20 percent deep sleep and that changes to 10 to 15 percent, you should probably figure out why.

Against these putative benefits, you have to weigh the risks. Poor sleep is not always a problem that can be solved by trying harder and worrying more about it—or by collecting sleep-tracking data. “Anxiety related to sleep can be both a symptom and a cause of some types of sleep problems,” Sargent acknowledges. The study that sticks in my mind, from Oxford University in 2018, involved giving subjects bogus feedback about whether they’d slept well or poorly. Those who were told that they’d slept poorly the night before reported feeling scattered, fatigued, and cranky. A little bit of data can be a dangerous thing, especially if its accuracy is questionable.

As for the mystery behind the surprising finding that Tour cyclists sleep just fine, thank you very much, even after the physiological disruption of brutal mountain stages, Sargent and her colleagues propose a disarmingly simple explanation. The cyclists prioritized sleep: they went to bed early and consistently, and gave themselves plenty of time there; ergo, they slept well. Earlier studies found that super-intense endurance exercise, especially when repeated day after day, led to diminished sleep—but the new generation of athletes are on top of it. There will be plenty to learn in years to come from the new sleep-measurement techniques, combined with robust analytical approaches like machine learning and AI. “I consider sleep to be the next frontier in physiology,” Poulin says. But none of it matters if you’re not putting in your time in the sack.

For more Sweat Science, join me on Threads and Facebook, sign up for the email newsletter, and check out my book Endure: Mind, Body, and the Curiously Elastic Limits of Human Performance.