Millions of people wear digital sleep trackers to bed every night, waking to see the pattern of their slumber. But, to date, the understanding of these night graphs is relatively simplistic.
It’s easy enough to tell how long a person spends in various modes of rest and how often they wake up. But the bigger picture, how individual nights and even weeks of sleep tie together, remains an open question, one that a team of researchers at UC San Diego, working with colleagues at UC San Francisco and the City University of New York, is beginning to explore by analysing millions of nights of sleep across thousands of people.
The key, as detailed in a paper recently published in the online scientific journal npj Digital Medicine, is analyzing longer periods of sleep for each individual, creating a landscape of a person’s rest that stretches for months.
Building on work by a team of researchers at the University of Tokyo, the team, co-led by UCSD bioengineer and data scientist Ben Smarr and integrative clinical psychologist Ashley Mason at UCSF, analyzed 5 million hours of sleep logged by more than 33,000 people who used the Oura Ring, a sleep-sensing digital device made by a private company in Finland.
Researchers followed methods used by the Tokyo research team using the Oura data to construct three- to six-day sleep “phenotypes,” or sleep patterns, characterized by the ring’s detection of periods when each person was awake and asleep, and the duration of those factors, during a nearly 10-month period from January 1 through October 24, 2020.
Using computers to cluster these sleep fingerprints together based on the similarities observed in their patterns of slumber, five central phenotypes emerged. The most common was one in which people got about eight hours of sleep for at least six days in a row.
Other patterns were more interrupted, ranging from people who slept continuously for about half of the nights measured, but slept for shorter periods of fewer than three hours the remainder of the time. Another found that some sleep continuously most of the time, but occasionally have interrupted sleep about once per week. The rarest pattern of all was for those who slept for short periods of time every night, waking regularly.
While other studies have been limited in the amount of time that data is examined — the Tokyo team processed only a few days — the UCSD effort examined how individuals moved from one sleep pattern to another, seeing transitions from full-night sleeping to more interrupted patterns.
They found that it is rare for a person to stay in the full night’s sleep mode continuously. Occasional switches to other modes were relatively common. And mode-switching was still more common for those with chronic conditions, such as sleep apnea and diabetes, or illnesses such as COVID-19 and influenza.
“I love this metaphor of landscapes where you’re moving through sleep space, and that is actually telling you a lot about your health journey,” Smarr said.
Lead author Varun Viswanath, an electrical and computer engineering graduate student at the UCSD Jacobs School of Engineering, added that understanding patterns of sleep over longer periods of time, and how people typically move between them, has the potential to make the data collected by sleep trackers more valuable to those who wear them night after night.
“Often we think about our sleep as what it was like last night, but, over a week, your sleep is dynamic,” Viswanath said. “You might have a set pattern of sleeping later on weekends and getting up earlier on weekdays and capturing that dynamic, what that variability looks like within a longer period of time, gets closer to understanding.
“If you have a pattern of going through a set of different sleep characteristics, and that pattern changes, that might be a bigger indicator than just having one or two bad nights of sleep.”
After reviewing the paper, Dr. Shalini Paruthi, a sleep medicine doctor and member of the American Academy of Sleep Medicine’s Emerging Technology Committee, said the study’s availability to track so many individuals for nearly a whole year is unique.
“The big question is, now that we’ve got millions of people wearing sleep trackers, can we turn that into something that’s really useful for society?” Paruthi said.
She said that she could see how knowledge of typical landscapes of sleep could help when trying to help patients diagnose whether or not what they are experiencing is normal or not.
“When people come in with complaints about their sleep, some are actually very fixated on, last month, you know, there was one night I just slept terribly, and I couldn’t figure out if it was something I ate or what,” Paruthi said. “With something like this, it allows you to say, ‘you know, if you only had one or two bad nights last month, that’s actually pretty normal, you were spot on, you’re actually doing great.’”
Better understanding of longer-period sleep tracker data and how it may correlate to medical conditions, she added, can only help in getting people diagnosed and treated sooner.
“Hopefully, another group of researchers reads this study and says, ‘oh, that’s really interesting, what if we added this or tried tweaking that,’ because that’s going to push use closer and closer to the answers that can improve quality of life,” Parthi said.
Most Popular
Comments / 0