Proving the Business Case for the Internet of Things

Researchers use wearables to study sleep patterns

Steve Rogerson
January 4, 2018

Chronobiologists Till Roenneberg and Eva Winnebeck from Ludwig Maximilian University in Munich have developed a way to record the dynamics of sleep under natural conditions.
The study of human sleep has so far been possible only under controlled conditions in dedicated sleep laboratories. Thanks to the development of an analytical technique, experimental subjects can sleep in their own beds while contributing to the scientific understanding of what remains a mysterious biological phenomenon.
For a simple wrist-worn device now suffices to record the essential characteristics of natural sleep bouts. The technique is validated in a paper published in the journal Current Biology, which reports work carried out under the direction of Roenneberg, who heads the Human Chronobiology Group at the Institute of Medical Psychology at LMU, and Winnebeck, who is in charge of its sleep laboratory.
Sleep is not a resting state in which people remain immobile. Instead it is characterised by phases of involuntary movement, which correspond to distinct physiological states.
In the third phase of their long-term human sleep project, the chronobiologists at LMU took advantage of this to study the dynamics of sleep phases in more than 16,000 sleep bouts from 593 experimental subjects between the ages of eight and 92. By focusing on phases of immobility and contrasting them with phases of movement during natural sleep, they were able to detect rhythms in movement that recurred at intervals of approximately 90 minutes.
The study confirms that the alternation of bursts of movement and bouts of immobility essentially correspond to the known cycles of REM (rapid eye movement) and non-REM (deep) sleep. These phases are associated with the onset and cessation of involuntary REMs that occur several times every night. However, while these sleep cycles of REM and non-REM sleep have so far been measured by means of electroencephalography in sleep laboratories, the pattern of locomotor inactivity during sleep (LIDS) can now be easily detected with wrist-worn monitors.
“Our new method makes it possible to collect objective information on the sleep characteristics and phases exhibited by individuals outside the usual laboratory setting, using a simple recording and analysis technique,” said Roenneberg. “With this approach the whole world becomes a sleep laboratory.”
As a result, the scientists have already found that men move more during their sleep than women. Furthermore, shift workers, who frequently have to alter their sleeping times, also move more than day workers with regular sleeping times. And the younger one is, the more pronounced are the movement rhythms and the more one increases one’s movement towards the end of the sleep period.
The wristwatch-like device used is an actimeter, which can provide a continuous record of wrist locomotor activity over periods of up to several months. It works on the same principle as commercially available smart watches that detect changes in position and acceleration, which also try to monitor sleep phases.
“Our method is, however, simple, transparent and designed specifically for long-term measurements,” said Roenneberg.
At all events, the validation of the new tool’s efficacy should make it easier for researchers to investigate the factors that affect the quality of natural sleep, which has hitherto been difficult to measure objectively. It may therefore someday be able to help people who suffer from insomnia and other sleep disturbances.
“We also hope that it will make people more aware of the importance of sleep for our general health and wellbeing,” said Roenneberg.
To measure a person's sleep, researchers have always relied on costly and time-consuming approaches that could only be used in a sleep lab.
"There has been practically no possibility of getting detailed sleep structures in a normal life setting over a long period of time," said Roenneberg. "You can't easily give somebody an EEG to take home and have next to the bed. You can't do this over six weeks or six months. We are going to see things nobody has seen before."
The wrist-worn research gadget can be purchased for $150. It is akin to commercially available self-trackers used by consumers. The gadgets, called actimeters, record data on wrist movement from which one can obtain activity patterns for up to three months. The researchers used the actimeters to assess rest and activity cycles not just over the course of the waking day, but also during sleep itself.
The findings are the latest in a larger, ongoing human sleep project, designed to learn more about sleep and its essential role in people’s lives by collecting sleep data on thousands of people in the real world. The team had been collecting information on sleep duration and quality via questionnaires. The next step was to find a way to collect objective measurements of sleep characteristics on similarly large numbers of people.
In the new study, the team looked to actimeter data collected from the subjects. But the patterns of activity during sleep collected using the devices appeared rather messy. It was hard to discern the cyclical sleep patterns normally seen with other, more complicated devices in the lab.
Then, they noticed something: by focussing on periods of inactivity during the night, a much clearer cyclical pattern began to emerge. The researchers used a simple conversion to measure inactivity as opposed to activity on a scale of near zero to 100, with 100 representing total inactivity.
"It was flabbergasting how it clarified the structures," Roenneberg said.
The LIDS measures showed that movement patterns reflect sleep cycles and replicate the dynamics seen in the lab. The data showed no sex differences in LIDS-derived sleep dynamics, although men move more than women do. They did observe large differences among individuals based on their age and work schedules.
Roenneberg said that it wasn't clear at first how the inactivity cycles matched up to the patterns of REM and non-REM sleep typically measured in the lab. Further study revealed that periods of least activity reflected deeper sleep. Those of greater activity corresponded to light and REM sleep. That's because, during REM sleep, the extremities frequently twitch and those twitches are detected by the actimeters.
As the researchers collect data in this way on many more people, they hope to come up with new and much more objective ways to measure not just sleep but also sleep quality. Such measures are essential for evaluating whether interventions to improve sleep actually work.
"Right now, we're not able to judge the outcome of interventions," Roenneberg said. "If, for example, we change school times, is sleep quality changed? What about shift work times or indoor lighting? All interventions necessary to improve sleep today are only judged by sleep duration and by asking people how they feel they have slept. There's no objective way to measure sleep quality, and we need this desperately."
Roenneberg said they were now poised to measure and compare sleep of people living in different cultures, climates, latitudes and lifestyles. They ultimately plan to build online infrastructures to allow anyone to upload actimetry recordings and receive meaningful feedback on their sleep.
"Many devices have tried to use activity to assess sleep structures, but our method is simple, transparent, and works especially in long-term recordings," Roenneberg said. "This will help many who have sleep problems and will hopefully increase the appreciation for the importance of sleep for our health and well-being."