The Science Behind Sleep Tracking In Fitness Apps

Sleep-tracking apps and fitness trackers have become increasingly popular, but how do they work? Sleep-tracking technology has advanced significantly in recent years, with apps now able to monitor sleep patterns, heart rate, body movement, and even the different stages of sleep. This is achieved through various methods, including in-built accelerometers and gyroscopes, which track movement and allow sensors to estimate when the user is active or inactive. This process is known as actigraphy. Some apps also use sonar technology to assess movement and breathing rate, while others use photoplethysmography (PPG) to track heart rate and movement. While the accuracy of sleep-tracking apps has been questioned, with some studies showing accuracy levels of only 78% when identifying sleep versus wakefulness, they can still provide valuable insights into sleep patterns and duration.

Accelerometers and gyroscopes

Gyroscopes are another motion-sensing device that, when used in conjunction with accelerometers, can provide more accurate sleep tracking. This process is known as actigraphy, a non-invasive technique used to assess cycles of activity and rest. Actigraphy can be used to estimate sleep duration, but it is not very good at detecting different sleep stages. For example, it may think that a person has fallen asleep when they are just not moving very much.

To improve accuracy, sleep tracking typically pairs movement tracking with other technologies, such as optical heart-rate monitoring and sound detection. Some trackers also record environmental factors like the amount of light or temperature in the bedroom. For example, the Muse S Gen 2 Headband uses an accelerometer and gyroscope alongside a PPG sensor to measure heart rate and blood circulation. The Google Pixel Watch 2 combines an accelerometer with a gyroscope and a heart rate sensor to track sleep stages.

In addition to wearables, some phone apps also use the phone's accelerometer and gyroscope to track movement and sound data to monitor sleep. For example, the Pillow app uses the phone's accelerometer and gyroscope, along with sound-sensing technology and Apple Health data, to track sleep.

Photoplethysmography (PPG)

The most common clinical application of PPG is the modern pulse oximeter, which was developed in 1991. PPG signal processing has evolved to include added inputs and artificial intelligence/machine learning/deep learning (AI/ML/DL) algorithms. These advances have enabled wearables to incorporate additional data outputs such as sleep tracking since 2013.

The blood underneath the skin reflects the red light emitted by a smartwatch and absorbs the green light. Each heartbeat allows for more absorption of green light, and in between heartbeats, less green light can be absorbed. This enables the device to track the user's heart rate and movement data to determine when they are asleep.

PPG has been used to train deep learning models to perform sleep staging on PPG signals from polysomnograms. Differences in signal characteristics, including amplitude, frequency, and variability, can be observed between signals obtained during different stages of sleep.

The limitations of PPG include oxymetry signal failure, motion artefacts, and signal processing. However, there are technical solutions to overcome these issues. The potential medical applications of PPG are numerous, including home-based detection of sleep-disordered breathing (SDB) and long-term monitoring of insomnia and circadian rhythm sleep disorders.

Sonar technology

The Sleep.com app, powered by SleepScore, also uses sonar to track sleep. The app sends silent signals into the sleep environment, and when these sound waves are reflected into the phone's microphone, an advanced algorithm interprets their shape and movement, measuring breathing rate and body movement. The app's algorithms then interpret these signals to track the user's journey through the various sleep stages. The frequencies used by the app are above the level of normal human hearing and are considered safe by medical professionals. The app's sonar technology is also accurate enough to measure the sleep of one person, even if they share a bed with a partner.

The Sleep as Android app also offers users the option of using sonar, although when tested by the New York Times, the app rarely counted REM sleep. The app also insisted that the tester had spent 90% of his sleep in deep sleep, which is unlikely.

Polysomnography (PSG)

PSG provides an accurate assessment of sleep quality and can aid in diagnosing sleep disorders such as sleep apnea and narcolepsy. It is also useful for detecting periodic limb movements and parasomnias. The procedure typically costs between $500 and $7000, depending on the lab and location.

PSG is often compared to consumer sleep tracking devices and fitness trackers to evaluate their accuracy. Several studies have compared PSG with popular fitness trackers and apps, including Fitbit, Apple Watch, and SleepScore. While fitness trackers offer convenience and a wealth of health data, they may not match the accuracy of PSG.

PSG data is interpreted by a healthcare professional, whereas fitness trackers use proprietary algorithms. It is important to note that fitness trackers do not directly measure sleep but rather estimate it through metrics like inactivity and movement. They can, however, provide valuable insights into sleep patterns and trends, helping users optimise their sleep habits and routines.

Wearables vs phone apps

Sleep tracking wearables and phone apps are designed to help you understand your sleep patterns and improve your sleep quality. While both types of technology rely on similar methods to track sleep, there are some key differences between the two.

Wearables

Wearable sleep trackers come in the form of smartwatches, rings, or fitness trackers that are worn on the wrist. These devices collect data through direct skin contact, measuring heart rate, heart rate variability, movement, breaths per minute, and skin temperature. Some popular wearables include the Apple Watch, Google Pixel Watch, Fitbit, and Oura Ring. These devices can provide a wealth of data, including time spent in each sleep stage and sleep cycles. They are convenient for those who want concrete guidance on improving their sleep and maintaining peak performance throughout the day. However, they may be less accurate in detecting certain sleep stages and can be uncomfortable or easy to misplace.

Phone Apps

Phone apps, such as Sleep.com, use sonar technology to track sleep without requiring a wearable device. By sending silent signals into your sleep environment, the app can measure your breathing rate and body movement to provide an accurate picture of your sleep patterns. Phone apps are a good option for those who want a non-contact sleep tracking method that is discreet and convenient. However, they may not offer the same level of detailed data as wearables and may have limited functionality compared to dedicated sleep tracking devices.

Both wearables and phone apps have their advantages and disadvantages when it comes to sleep tracking. Wearables provide more detailed data and guidance but may be less comfortable or easy to lose. Phone apps offer a contactless and discreet option but may have more limited functionality. Ultimately, the best choice depends on your personal preferences and how much data and guidance you want to receive about your sleep.

Frequently asked questions