Whether you are an avid biohacker or just take an active interest in your sleep, you may be using, or may be thinking about using a sleep tracker. These trackers come in many shapes and sizes, everything from wrist and headbands to smartphone apps, smart-rings, large thin film sensors for the bed, sound and radio frequency sensors.
Motivated by the rise of big data algorithms and the Quantified Self movement, consumers are increasingly being drawn towards collecting and analyzing their sleep data with these smart devices.
One surprisingly less common question that you should be asking before placing too much stock in the analysis these devices provide is: how accurate are these sleep trackers in tracking the variables they claim to track?
The answer could be very important to you particularly if you intend on using the analysis of your sleep to make decisions.
But the answer depends on the technology used by your device, and how many measurements it integrates to give the best estimate of sleep duration and quality.
In this article, we’re going to take a look at how they work, what they can (and can’t) do, and some of the best ones currently available. First, let’s take a look at how sleep is tracked professionally in a sleep lab. Then, we’ll see how the commercially available sleep trackers stack up in comparison.
How is sleep tracked in the lab?
In an overnight sleep study, also known as a polysomnography, a wealth of information is collected while you sleep to make an accurate assessment of sleep architecture.
This non-invasive study measures brain waves with an electroencephalogram (EEG). An EEG measures the electrical activity of the brain using numerous electrodes placed on the scalp. With the current technology available, EEG brain wave analysis is the most reliable way to measure sleep staging.
The brain wave patterns are created by large groups of nerve cells in the brain as we cycle through the five stages of sleep. For instance, in REM sleep, our brain waves resemble those when we are awake, that is, low amplitude and with no apparent synchrony. EEG electrodes can pick up these signals with a reasonably high signal-to-noise ratio and reliably reflect the phase of sleep we are in.
Other parameters are measured to give a more definite indication of the phase and quality of sleep. This includes tracking eye movement using an electrooculogram, where electrodes are placed above and below the eyes. Eye movements are fundamentally related to the onset of rapid eye movement (REM) sleep, otherwise known as dream sleep.
Body movement and muscle tension are also tracked using electrodes placed on the legs and chest. Sensors attached below the nostrils monitor breathing rates and detect abnormalities in respiration.
Pulse oximetry is used to measure blood oxygen saturation, a non-invasive way to measure oxygen levels, which may be disrupted in sleep apnea patients. Audio and visual recordings are also used in the sleep environment to assess the frequency and intensity of snoring and other parasomnias, which just refers to disruptive sleep disorders in medical terminology.
Based on its comprehensive assessment of multiple sleep measures, this study is most commonly used to diagnose sleep disorders such as sleep apnea and REM sleep behavior disorder.
Sleep tracking variables
While polysomnography can give us a detailed look at the five stages of sleep, consumer sleep tracking devices typically divide sleep into two or three stages: light, deep, and REM. We’ll take a look at the various measures these devices use to evaluate our sleep quality and quantity.
Most sleep trackers indirectly measure how long you’ve been sleeping by analyzing your movements. In this way, inactivity is used as a proxy for a direct sleep measurement.
This is one of the primary data points wearable sleep trackers use to give a best estimate of light, deep, and REM sleep.
The science of using movement to measure sleep/wake behavior is known as actigraphy. Most actigraph devices are worn on the wrist and track movements throughout the night. This data is then processed through specialized algorithms to make an assessment of various sleep parameters such as total sleep time (TST) and wake after sleep onset (WASO).
Movement is registered with an accelerometer, a commonly found MEMS on smartphones and actigraph devices. In addition to estimating time asleep and awake, movement will also be used as an indirect measure of sleep quality. Restful nights are assumed to be less movement-ridden, while restless nights will have frequent periods of movement.
However, it’s possible to be lying still and be awake and/or asleep but moving around and have this wrongly reflected in the device’s morning readout. This is one factor that makes movement-based sleep/wake data limited in their applicability and accuracy.
Additionally, REM is the only sleep stage where the body becomes completely paralysed. During NREM sleep, there is still mild levels of muscle tone present in the body. As a consequence, body movement can still occur, which is why sleepwalking occurs in the deeper stages of NREM sleep.
This is all to say, when you fall asleep may not be accurately captured by a normal sleep tracker because you go from being awake and mostly still to asleep and mostly still — but you still might move. The only time when your body will be entirely still will be 90 minutes after you fall asleep as you enter into REM.
It’s possible that companies measuring this factor in these subtleties into their algorithms but those algorithms are proprietary, and so there’s no way of fully understanding how your sleep data is actually being analysed.
Heart rate and heart rate variability
Heart rate and heart rate variability (HRV) are other metrics used by some modern sleep trackers. While heart rate is measured as the number of heartbeats per minute, HRV refers to the variation in time between each heartbeat. Both are influenced by the state of our autonomic nervous system, which changes throughout sleep.
For instance, heart rate and HRV decrease as we enter into the deeper stages of NREM sleep, but increase as we enter REM sleep. According to this study, NREM sleep stages create increases in the high frequency components of HRV and decreases in the low frequency components, which algorithms can analyze for sleep staging. The opposite trend is found in REM sleep. One study found that heart rate variability was able to predict deep sleep up with up to 87% accuracy.
In this way, these metrics can help clarify the stage of sleep we are in, although not to the same accuracy as is achieved by also incorporating brain wave data.
More recent research has indicated that heart rate variability may be disturbed in individuals with sleep disorders such as sleep apnea and insomnia. Greater HRV is also associated with higher sleep efficiency, and so it may be a useful metric to measure sleep quality as well.
Heart rate is typically measured with optical sensors on wearable devices that infer heart rate based on the amount of reflected light that changes according to blood volume. Other measurement techniques include ballistocardiography, used by the Beddit sleep tracker, which measures the forces associated with heart contractions.
A few sleep trackers such as the SleepScore Max use non-contact bedside sensors that collect radio waves reflected off the body to determine heart rate (as well as respiration rate and movement data). These non-contact sensing methods have been evaluated to have high accuracy and specificity, comparable to top range multi-sensor full contact wrist-worn sleep trackers.
Similar to a sleep study in the lab, some sleep trackers will do a sound analysis using a microphone either on the smartphone or on the wearable. Sleep apps like Sleep Cycle use sound to measure movement indirectly (such as bed sheets moving) with a purported accuracy of about ten inches away.
This removes the need to bring the phone to bed to register movement. Others use the microphone to register noises in the environment, including snoring, which could be impacting sleep quality.
Temperature is closely related to sleep regulation and more broadly, our circadian rhythms. Our body temperature drops a few degrees every night as we prepare for sleep, hitting its lowest in the early morning.
Some wearables such as the Oura Ring continuously track skin temperature and the software plots these trends over the course of days and weeks. This data can give you an understanding of your normal baseline body temperature and how it varies throughout sleep. The sensors may also reflect temperature disruptions from baseline due to things such as sickness and menstruation.
Other sleep trackers like the non-wearable Resmed S+ will track the ambient temperature of the bedroom environment to see how this influences sleep. The recommended room temperature for optimal sleep is between 60-67 degrees Fahrenheit. Higher temperatures may disturb sleep onset, for instance.
Sleep trackers like the Fitbit Surge or Microsoft Band that analyze the sleep environment commonly have ambient light sensors that measure light intensity throughout the night. This information can be used to understand the connection between ambient light levels and sleep quality.
Certain lifestyle factors that influence sleep quality and duration can be tracked in some sleep tracking devices. This includes factors such as the amount of caffeine ingested in the day, stress levels, and eating patterns. In this way, sleep trackers can double as a sleep diary. By logging these lifestyle factors, you can more easily pinpoint which variables influence sleep the most based on the measured data that’s trended over time.
The types of sleep trackers
Wearable sleep trackers
Wearable sleep trackers are largely worn as a wristband and use actigraphy to identify the states of sleep and wakefulness. Other wearables include headbands, finger rings, or pendants that are worn around the neck.
Wristband sleep trackers
Some of the best sleep tracker wristbands currently on the market are the Fitbit sleep tracker and Apple watch sleep tracker.
The Fitbit Versa tracks heart rate, movement, and basic sleep data to give the best estimate of the time spent in light, deep, and REM sleep. It also gives sleep insights for better sleep and tracks daytime activity such as the number of calories burned and the number of steps taken.
The Apple Watch is a multipurpose smartwatch that tracks sleep once the user downloads an app like SleepWatch or Autosleep. Both apps give a trended readout of light, deep, and total sleep time by measuring movements, noises, and heart rate variability.
Headband sleep trackers
Some headband wearables such as the Dreem use EEG sensors to analyze brain activity related to sleep stages. It also outputs biofeedback pink noise tones to aid deep sleep.
The number of sensors and the signal/noise ratio may not be as high as traditional polysomnography, but preliminary studies have shown comparable accuracy. These devices will surely improve over time and be a very accurate longitudinal measure of sleep quality and quantity.
Ring sleep trackers
The Oura Ring is a highly-regarded smart ring that packs many sensors into a small space. It measures body temperature, heart rate, heart rate variability, respiration, and body movements. It also uses pulse oximetry to measure blood oxygen saturation. In addition to tracking sleep, it functions as a fitness tracker, tracking daily activity such as steps and calories.
Best sleep trackers if you don’t want to use a wearable
This is generally the preferred sleep tracking method for those who want to track their sleep but don’t want to attach a device to their body.
These devices record data using infrared technology and thin fabric strips with sensors. They are typically placed either on the mattress or by the bedside.
Sleep monitoring systems
Non-wearables such as the Beautyrest sleep monitor and Beddit sleep monitor integrate with your smartphone and record respiration, heart rate, and body movements using thin sensor systems placed in the mattress.
Using this data, the software makes the best estimate of the time spent in light, REM, and deep sleep and total sleep time. These are then shown as plotted trends over time which can inform better sleep habits.
App-based sleep trackers are the easiest and most cost-efficient way to get started tracking sleep. While virtually hundreds of sleep apps exist for Apple and Android devices, the most popular apps are Sleep Cycle, Sleep Time, Sleep Bot and Sleep Genius.
To use them, the smartphone is placed either on the mattress or on the nightstand near the bed. While the app is on, it records movement data with the microphone. These apps also have a smart alarm feature that claims to wake you up during the lightest sleep.
At Remrise, we suggest not taking your phone to bed with you. Instead, the Remrise app helps you track your sleep through journaling.
The benefits of using a sleep tracker
The currently available sleep trackers don’t precisely reflect your sleep architecture like clinical sleep studies, but they do provide some basic information that can be helpful to understand how well and how long you’ve slept for.
Sleep trackers reinforce consistent sleep/wake times
According to Dr. Stickgold, professor of Psychiatry and director of the Center for Sleep and Cognition at Harvard Medical School, “The greatest value of wearing a sleep tracker device is that it actually keeps a record of when you went to bed and when you got up."
In this way, sleep trackers may be a digital sleep diary that can bring more awareness to our sleep patterns and habits. We know regularity is one of the most important things for having restful sleep. If sleep/wake times are reliably tracked with these devices, and then this knowledge can be used to inform more optimal and consistent sleep schedules and routines.
Additionally, many sleep trackers also have a smart alarm feature that attempts to wake you up during the lightest phase of sleep to minimize sleep inertia. While it may not always be precisely accurate, it may further reinforce sleep predictability, leading to more optimal circadian rhythms and more restful sleep.
Sleep trackers may support beneficial sleep hygiene practices
The data collected can also help you recognize sleep-related patterns and assist in making changes to sleep hygiene-related habits such as exercise times and eating times. If the tracker collects data on noise and light, it may aid in changing certain environmental factors in your sleep environment such as ambient light and noise levels, both of which can disturb sleep.
Sleep tracker anxiety
While sleep trackers have the potential to give us insight into how well we sleep and our sleep patterns, their use can also backfire in some cases and give rise to sleep perfectionism or sleep-related anxiety.
For some, knowing that every movement is tracked continuously or that their sleep duration is being logged can be a stressful experience that can interfere with sleep. For others that decide to track sleep, they may wake up feeling refreshed only to check their app and realize they haven’t slept as well as they had thought, leading them to question their sleep quality (even in spite of their often inaccurate tracker results).
This phenomenon can lead to the opposite of the placebo effect, called the nocebo effect. That is to say, you feel worse because you expect to feel worse based on the demoralizing sleep data from the tracker, leading to an undesirable self-fulfilling prophecy of poor sleep.
A 2017 study published in the Journal of Clinical Sleep Medicine found that sleep trackers can reinforce sleep-related perfectionism or anxiety, a clinical term they refer to as orthosomnia. The authors report three cases of orthosomnia in patients that complained of poor sleep despite normal sleep parameters using polysomnography.
The authors suggest pairing sleep tracker use with cognitive behavioral therapy for insomnia (CBT-I), which can improve sleep habits, reduce sleep-related anxiety, and halt negative thought patterns — effectively ending the obsessive quest for perfect sleep.
Can you get sleep lab-quality data from home?
The question is, if you really wanted to capture multiple variables regarding your sleep, to try to get a closer glimpse of nightly quality of your sleep, could you with these by using multiple consumer sleep trackers?
How to get the most out of the best sleep trackers
Polysomnography is expensive and inconvenient, so it makes sense to wonder if it's possible to leverage commercially-available sleep trackers to create the closest thing to a sleep lab in the home. In this case, the trick is to monitor as many variables as possible that are monitored in a sleep lab environment.
In the spirit of DIY biohacking, the closest way to get towards sleep lab data with the current technology may be to combine wearable and non-wearable sensors to give a more comprehensive picture of sleep quality and duration.
A bedside contactless sensor such as a Resmed S+ would reliably gather information on the ambient environment, such as temperature, light, and noise. Additionally, the movement and respiration data would be supportive of the wearable sensor.
A well-recommended wearable, such as the Oura ring, would capture respiration, body temperature, movement, heart rate, and heart rate variability. When compared against polysomnography in this study, the Oura ring was able to consistently compare against it on many sleep variables, including total sleep time, sleep onset latency, and wake after sleep onset. In terms of sleep staging, the Oura ring was able to agree 65% on light sleep (N1), 51% on deep sleep (N2+N3), and 61% on REM sleep.
Unfortunately, EEG is the most effective way to understand how our minds and bodies move through the different phases of sleep and for now, there are no EEG trackers specifically designed for optimizing sleep yet. However Muse and Dreem have designed EEG products that measure your brain activity. Though, achieving gold standard polysomnography-level precision may still be a few years out.
This is because the sleep lab uses many more EEG sensors placed directly on the skin. For this reason, EEG-based commercial sleep trackers are still very much in their infancy.
With the CDC recently calling sleep a public health epidemic, perhaps a collective movement towards a heightened awareness of sleep is desperately needed. Sleep trackers and more generally, the “quantified self” movement, seem to be playing an important role in this process by giving us a detailed sleep analysis. It’s clear though that sleep is best quantified with polysomnography, based on the wealth and diversity of information it measures.
Nevertheless, commercially-available sleep trackers are convenient, accessible, and relatively inexpensive. They inform users about their sleep habits and how these may influence sleep quality and quantity.
While the sleep reports generated by these sleep trackers may be an important driving force for people to improve their sleep, they should be taken with a healthy dose of skepticism, especially when they are claiming to make accurate, objective measures of specific stages of sleep like REM sleep.
Lastly, if you have a chronic sleep condition or think you might have one, it is best to let the sleep experts analyze your sleep rather than taking matters into your own hands.
Ballistocardiography: Often referred to as BCG, a non-invasive technique for measuring the sudden ejection of blood into the great vessels of the heart with each heartbeat.
Sleep Inertia: A state of impaired cognitive or physical performance that is experienced immediately after awakening. Sleep inertia is more intense after waking from the deeper stages of sleep compared to the light stages.
- de Zambotti M, Baker F, Colrain I. Validation of sleep-tracking technology compared with polysomnography in adolescents. Sleep. 2015;38:1461–1468
- Kamišalić, A., Fister, I., Jr, Turkanović, M., & Karakatič, S. (2018). Sensors and Functionalities of Non-Invasive Wrist-Wearable Devices: A Review. Sensors (Basel, Switzerland), 18(6), 1714. doi:10.3390/s18061714
- Kolla BP, Mansukhani S, and Mansukhani MP, Consumer sleep tracking devices: a review of mechanisms, validity and utility. Expert Rev Med Devices, 2016. 13(5): p. 497–506
- Marino, M., Li, Y., Rueschman, M. N., Winkelman, J. W., Ellenbogen, J. M., Solet, J. M., … Buxton, O. M. (2013). Measuring sleep: accuracy, sensitivity, and specificity of wrist actigraphy compared to polysomnography. Sleep, 36(11), 1747–1755. doi:10.5665/sleep.3142