Sleep. Everyone needs sleep. Some of us might think we can get away with less or some don’t have the luxury of having more or some just have problems getting proper sleep. In any case, not getting enough sleep can be detrimental for our health and well-being in the long run. For athletes, lots of previous research has shown that sleep loss or poor sleep can have a negative impact on their performance.
What happens when an athlete sleeps poorly
Some of the (negative) effects of inadequate sleep include:
- increase in injury rates,
- reduction in accuracy and speed,
- poorer reaction time,
- impaired decision making.
This ultimately has an effect on quality and longevity of athletic career. You could dive deeper into that here: link. So how much or how little sleep does an athlete need? How does an athlete know whether the x number of hours sleep they had last night was good enough or not?
Enter sleep tracking sensors. Sleep tracking sensors (when used) tracks when the user goes to bed and gets up, whether it was in the middle of the night or in the morning. They can identify whether the user was in deep/light/REM sleep, and some even measure metrics like heart rate, breathing rate, oxygen saturation and snoring. It almost sounds like having a polysomnography, which is a sleep study usually done in hospitals, but not quite.
Let’s have a look at what some the different types of sleep tracking sensors are, how they work, and their pros and cons.
What are the different types of sensors?
Wearable Sleep Trackers
The most common wearable sleep trackers are the wrist worn ones. In fact quite a number of “activity-monitoring slash steps-counting plus smart-phone connected plus GPS and heart-rate monitoring” watches (that used to just tell time), now also do sleep tracking. Think Garmin, Fitbit, Suunto and Polar; they all have devices that have sleep tracking capabilities. And let’s not forget the Apple or Samsung Watches that also allow developers to build sleep tracking apps.
They mostly work on the same principle – actigraphy or tracking movement of your wrist (using IMUs) and measuring heart rate (optically) during sleep. Algorithms will analyse the movement and heart rate data and work out whether a person wearing the watch was in deep sleep, light sleep or REM sleep.
There are also a handful of wearable devices that are designed solely for sleep tracking and they measure blood oxygen saturation level (SpO2) similar to pulse oximeters. A couple of examples include EverSleep and Viatom. In terms of data collection, the wearables all connect to their own smartphone app via Bluetooth. Most of them require a login and the data syncs to the cloud but EverSleep only stores the data locally on the smartphone.
Generally, wearable sleep trackers are great for people who travel quite a bit. So it doesn’t matter where they are sleeping, the devices stay on the person and not on the bed or in the room. On the downside, wanting to track your sleep (with wearables) becomes a very intentional action and you need to remember to put it on when you sleep. It also gets tricky if it is a watch you wear during the day because you need to charge the watch at some point.
Sleeping Mats or Under Bed Sensors
Sleeping mats sensors or under bed sensors or mattress sensors are typically sensors that are placed semi-permanently beneath where the person is sleeping. They are usually quite flat and can be placed between the mattress and a bed base or bed frame. Some can be placed underneath the bed sheet or a mattress topper. Although they sound similar, there are some differences between the sensors.
BeautyRest sleep trackers uses piezoelectric sensors to measure pressure/force, and they are designed to be placed under the mattress, located below the pillow or user’s head. Withings’ sleep mat uses an air pressure system and it can also be placed under the mattress near the user’s trunk to measure weight/pressure changes during sleep. Another sleep mat type sensor is the Emfit QS. On first glance, it looks similar in its tracking approach – it measures force/pressure changes. But the unique thing about the Emfit sensor is its sensing material which consists of a quasi-piezoelectric and ferroelectret film. Its high sensitivity allows it to be used for ballistography which is a measure of ballistic forces on the heart. In fact the product’s intent was for athletes to measure heart rate variability (HRV) during sleep and monitor recovery.
These sensors do require a wifi connection so that they can upload the user’s sleep data to the cloud. This is typically established during the initial setup where a calibration process is required. They also need to be plugged in to a wall socket to be powered. The good thing about this type of sensors is that once all that (wifi & power) is setup in the bed, it just tracks sleep passively. There’s no need to turn on or put on anything. Users just go have to bed normally. But it won’t work well for travellers or people who fall asleep on couches more than their beds. Also, some people might not like that fact that they require wifi.
Bedside Sensors
This next type of sensors is also a non-wearable. In fact I would say these are proper non-contact sensors. They are placed next to the bed and next to the sleeping person to be tracked, usually on a bedside table and they be within a certain distance from the person. They look a bit like a speaker or radio or an oversized alarm clock.
Resmed, long known for their continuous positive airway pressure (CPAP) device for treating sleep apnea, developed the S+ which measures movements using ultra low power radio waves. They also refer to it as a novel biomotion sensor. On top of that, the unit uses a microphone to monitor sleep sounds (e.g. snoring) and other environment noises to give an evaluation of each night’s sleep. SleepScore Max is a similar sensor product. It actually uses Resmed’s technology and works almost the exact same way. The company that made it (SleepScore Labs) also came up with an App only solution. It utilises the smart phone’s microphone to detect sounds/noises throughout the night to work out sleep movements and other possible sleep disturbances. Users will need to leave their phones plugged in for it to work.
Placement wise, the above three solutions have to be placed slightly above the top of the mattress level and the sensor/microphone has to point at the user’s chest. The sensor should also be within an arm’s length from the user and they also need to be plugged in to a power socket. But instead of wifi, they connect to their accompanying app via bluetooth.
These sensor solutions do require the user to turn on sleep tracking each night before going to sleep. It is not set and forget like the sleep mat or bed sensors. Although the apps can do push notifications to remind users to sleep on time and to track their sleep. One good thing about this type of sensors is they are a bit more portable than the sleep mat types. One downside is they need to be placed in a specific way and it won’t work very well for people who rolls around in bed and out of the sensor radius or if they like to hug a pillow to sleep and it blocks the radio waves from tracking them.
On Bed Sensors
This last type of sensors are placed on top of the mattress and next to the user’s pillow. Unfortunately, quite a number of such sensors have not stayed in the market for very long. A couple of them include the Sevenhugs Hugone sensor and Samsung’s Sleepsense. These type of sensors mostly rely on accelerometers and gyroscopes (IMUs) to track movement on the bed. So one possible downside of it is: if a person’s mattress can pass the red wine test, (ie a glass of red wine placed on the mattress doesn’t tip when someone jumps on the mattress), chances are their sleep movement won’t get tracked very well by the sensors. I imagine that is a potential reason why the above products have transitioned out.
App/Hybrid Solution
Nevertheless, there are a handful of smart phone apps that do a similar thing using the phone’s in-built IMUs. One example is the Pillow App. Other than tracking movement, it also monitors sleep noise levels using the iPhone’s microphone. For users with an Apple Watch, they can also integrate it to the Pillow App so that sleep motion is tracked with the Watch while the iPhone microphone tracks sounds. In this case, it becomes more of a hybrid on-bed + wearable solution.
What do they track
Sleep duration – Sleep stages – Sleep Score
The most basic thing that all the devices track is when a person goes to bed and when they wake up – that gives sleep duration. They also (attempt to) identify what sleep stage a person is in – either light, deep, REM sleep or awake and the duration of each stage. They also workout a Sleep Score. Different devices might have slightly different variations but they mostly work on a weighted calculation of total sleep time, REM and deep sleep duration, period and number of times awake during the night.
Snoring or Noises
Devices that have a microphone are able to detect a person’s snoring. They could track the number of occurences as well as the duration of the snoring. But other than the noting down those statistics, the sleep devices usually don’t save or record the snoring events which can be a relieve knowing that no one has the opportunity to listen in on our snoring. Other than snoring, a couple of devices also tracks ambient room noises and reports it as disturbances.
Physiological measures
Heart rate and breathing frequency are two commonly tracked metric. That said, not all the devices measure both and the different devices that measure one or the other or both, measure them differently. For heart rate: wearables use optical sensors, sleep mat sensors measures pressure changes and use advanced signal processing to obtain heart rate signal. With breathing frequency: sleep mat sensors has the same approach as above while bedside sensors use sonar technology.
Other interesting features
Apart from the above, there are some unique tracking that certain devices do while others don’t. They include:
- Heart rate variability (HRV) – which is a good metric for assessing the effects of stress on the body
- Blood oxygen level (SpO2) – tracking this during sleep is a good way of monitoring people with sleep apnea
- Breathing disturbances or interruptions in sleep – this is also a way of monitoring sleep apnea
- Body movements or tossing and turning – this is probably a good indicator of restful or unrestful sleep
- Autonomous Nervous System(ANS) balance – this is probably an extension or outcome of HRV tracking
Are they accurate?
Most of the different devices are fairly accurate at tracking when a person has gone to bed or when they wake up, so the sleep duration data is probably reliable. After that, identifying of sleep stages (deep/light/REM) is one that hasn’t been very accurate and many people have had negative feedback (example). With inaccuracy of sleep stage tracking, the sleep score (which relies on that) becomes somewhat pointless.
Snoring is a pretty straightforward tracking where microphones are involved unless there are a lot of ambient room noises. Heart rate tracking with wearables is also a straightforward approach where accuracy has improved over the years with better optical sensors. It is the sleep mat measurements that are the interesting ones. Among them, Emfit has their sleep mat sensing technology tested and validated in a good number of publications – which says something.
Final Thoughts
Pros and Cons: Every type of sleep tracking sensor has their pros and cons. Wearable sensors are portable and good for people who travel often. But if the wearable device is also for daily activity tracking, there will need to be planned “downtime” to charge the device. Generally, the fitness/GPS watches (e.g. Garmin, Fitbit, Polar or Suunto) will automatically detect sleep if the watches are worn. Some do suggest that the watch be kept worn a few hours before sleeping. Smart watches and the sleep specific devices require a few more steps like connecting to the relevant smartphone app and/or turning the app on. Bedside sensors and Under bed sensors are good for people who don’t like to wear a watch to bed. Between the two, Bedside sensors are somewhat more portable because Under bed sensors has setup steps that are bed specific and moving them means going through that setup/calibration again. On the other hand, Under bed sensors work passively once it is setup and simply works in the background – which is great; whereas Bedside sensors require users to activate tracking each night.
Integration with other data: Sleep data by itself it pretty useful but then there is also the opportunity to integrate it with other health or activity data to paint a broader picture. With wearable devices that are also used as activity trackers, they would automatically have sleep and day activities data on the same platform. A few examples are Garmin (Connect), Fitbit (Care), or Polar (Smart Coaching). Withings also has a platform called Health Mate where they integrate data from their smart scales, smart watch and blood pressure monitoring devices. Otherwise lots of devices are able to integrate with platforms such as Apple Health or Google Fit.
For professional sports teams, there are a few athlete management systems (AMS) that can integrate 3rd party sleep sensor data (if the sleep sensors have APIs that allow integration). Fusion Sports and Lumin Sports are a couple of examples that can do that.
Still logging sleep manually/subjectively? So with all these different sleep sensor options available that can suit different needs, there are very little reasons to log sleep data manually. Other than costs of the devices, privacy is probably one other reason I can think of.
If there are other reasons that are holding you back from using a sleep tracking sensor, or if you know of another type of product not listed above, do let us know what they are by leaving a comment below. With that, thanks for reading!
Sports Technology Blog 