At Spoked we take your sleep and recovery very seriously, as we recognise their importance when it comes to training and performance. Therefore, we ask you to tell us about your sleep every day. We currently do this instead of relying on data from sleep trackers, which can be unreliable, but also as most people do not have them. In this blog I’ll be delving into how sleep trackers work, why they aren’t always the best option (although the idea behind them is solid), and why we are currently relying on subjective data.
I will briefly explain what sleep is, the different sleep phases, why it is important for recovery and how much sleep you should aim for. In a future post I will concentrate fully on why sleep is so important, but for this post I will just go into the basics required to explain how sleep trackers work. Then we will examine the different methods of sleep tracking, which range from very scientific sleep labs and brain wave measurements, to most common wrist-based sleep trackers that use movement and heart rate to determine whether you’re asleep or awake, and what sleep phase you are in.
Now the big question: why would you even want to know? Sleep is sleep, right? You can get away with 5 or 6 hours, having a drink makes you fall asleep quicker, and you sleep just fine. Naps and sleeping for long periods are for babies, kids, teenagers, and lazy people! If you think that, you couldn’t be more wrong unfortunately.
Like I alluded to in my previous post, sleep is the most valuable and effective recovery tool you have, and it doesn’t cost a thing (Nedelec et al., 2015). Halson (2019) found that sleep deprivation leads to reduction in performance, mood, cognitive function, memory learning, metabolism, illness, and injury. Things that most of us do or experience, such as caffeine consumption, muscle soreness, injury, jet lag, travel (sleeping in foreign environments) and social media / video games have a profound effect on our sleep. And the brilliantly named Killer et al. (2017) found sleep disturbances during short-term intensified training in trained cyclists. That’s quite a lot that can go wrong with a bit of missed sleep, hence me talking about it all the time.
These anecdotes about people who get by on 3-4hours a night really annoy me, as these are far more the exception than the rule. So how many hours do we really need? Unfortunately, there is no right answer, as this is totally individual. The quick answer is between 7-9 hours, with 8 hours probably being sufficient for most (this is time actually asleep, not just time spent in bed). I can tell you how much sleep people get on average, and that has been shown to be insufficient. In the UK the average time of nightly ‘sleep’ is 7.6hours (I put this in quotation marks, as time in bed and time asleep are not the same). This is for all age ranges, and on average younger people get more sleep. It also includes weekends, so is actually shorter during the week. In the USA the sleep duration during the week was 6.8hours during the week and 7.4hours on weekends. These data do not include sleep latency, so the time asleep will be actually shorter again. This is highlighted in a study by Leeder at al. (2016) where participants spent 8:36hours in bed each night, but only were asleep for 6:55hours of that.
Each day we work out your sleep efficiency (how well you slept) by looking at when you went to bed, how long it took you to fall asleep, how many times you woke up during the night, and when you got up. We pair this with your subjective feedback on how you feel both physically and mentally, as just because you spent a long time in bed does not mean you feel refreshed and strong. Sleep latency (how long it took you to fall asleep) can be quite long, especially if you had a late session or are feeling stressed, so purely spending 9 hours in bed does not equal 9 hours of sleep. This is one of the drawbacks of sleep trackers: they find it difficult to distinguish between not moving / lying in bed (maybe reading) and actually being asleep (Liang et al., 2017).
The validity of sleep trackers has been questioned frequently by sleep experts, and there are of course many different types and models, making analysis of their functionality difficult. There is a big difference between a review by a website or person and a rigorous scientific study, and therefore I will present a range of sources here. Briefly, a scientific study should be the most rigorous, unbiased and trustworthy.
Papers are peer-reviewed (other experts from the area evaluate the paper before it’s published) and in most cases these should be independent from the manufacturer they are investigating. It can get slightly complicated if a certain manufacturer pays for research to be done, as the researchers can try to ‘find’ a more positive angle to show the product in a more positive light. Garmin recently published such a study on their website, and although the methods were no doubt good, it wasn’t peer-reviewed, and as such hasn’t faced any analysis or criticism before they published it - read study here. Unsurprisingly, the results were mainly positive compared to studies I’ve highlighted below.
Anecdotal evidence is the opinion of someone, usually without any scientific data to back up those claims. You’ll find many YouTube videos and online articles about the different sleep trackers, but in most cases, these are based on the experience of a single person with a particular device (there are a few good exceptions, for example - The Quantified Scientist.
Obviously, the information from a scientific study is more what we’re after, so that is what I will concentrate on the most. I am very interested in your experience with sleep and sleep trackers, so please feel free to leave a comment or send me a message
The two main different sleep phases can be split into REM (rapid eye movement) and NREM (non-rapid eye movement) sleep. All NREM sleep is further broken down into deep sleep, light sleep and awake. The time you went to bed, to the time you got up gets analysed and the different phases assigned. It’s important to also include how long it took you to fall asleep (called sleep latency), as the time in bed alone can be misleading.
Deep sleep is seen as the most important sleep phase for recovery for athletes (Halson, 2014). REM sleep is when your body doesn’t move, you have bursts of REM and dreams, i.e., an activated brain and paralysed body. REM sleep is important for brain function in particular. Typically, you spend the most amount of time in light sleep, followed by REM sleep, deep sleep and finally awake (we all wake up several times during the night even if we’re not aware of it).
In deep sleep we release growth hormones that are important for recovery. Optimum conditions for anabolism (building / repair of muscles) occur during deep sleep and restricting the amount of deep sleep has been shown to reduce performance (Halson, 2014). Therefore, if you’re training hard (or even just want to function optimally during the day) then you should definitely try to get as much sleep as possible, and make sure it is good quality as well. Alcohol has shown to reduce our REM sleep and leads to more frequent waking in the later stages of sleep, thus decreasing your sleep quality.
The ‘gold standard’ (scientifically best way to measure something) for measuring sleep is polysomnography (see pic below). Although the best, this is usually only used to assess sleep disorders as it is expensive and difficult. It is quite a complicated set-up and measures brain activity, breathing rate, muscle activity, eye movements, and cardiac activity. However, from this you can get total sleep time, sleep-onset latency, wake after sleep onset, sleep efficiency, sleep fragmentation, number of awakenings, time in each sleep stage, and sleep stage percentages (Halson, 2014). As you can see, measuring sleep correctly is fairly complex, and therefore using less rigorous measuring techniques will lead to problems with accuracy.
Wrist monitors are now very commonly used to analyse our sleep, and there are many different companies offering devices covering a large price range. Most well-known is probably the WHOOP wristband which pairs with your phone and has a substantial monthly fee. Other devices you might have heard of are of course the Apple watch, Fitbit, Oura ring, and various Garmins.
This is not an extensive list, but these all work on similar principles, although use different algorithms. There are also some other methods of sleep data collection, such as headbands and mats to put under the sheets or mattress. The way these work all vary, and some include more, or less data collection points and analysis software.
Most newer sleep wearables provide a four-stage detection of sleep, i.e., wake, light sleep, deep sleep, and REM sleep based mainly on movement detection and heart rate. Some also use heart rate variability, skin conductance and/or skin temperature (Miller et al. 2020). (Caveat: heart rate variability is a good predictor or recovery but needs a separate article to explain why. Briefly, it takes dedicated hardware to measure it correctly).
Evenson et al. (2015) reviewed several sleep trackers and found that they either over-estimate total sleep time and sleep efficiency, and under-estimate wake after onset of sleep, or when the sensitivity of the sensor was changed it under-estimated total sleep time and sleep efficiency and over-estimated wake after sleep onset (in plain English: they got it wrong most of the time).
This meant that some trackers generally over-estimated time asleep by approx. 23 min/day. In a very recent study, Miller et al. (2020) found that WHOOP overestimated sleep time by 8.2±32.9 minutes. Again, that does not sound a lot, but that standard deviation of 32.9 minutes tells us that for some they slept over half an hour more or less than actually reported. A positive step by WHOOP is that bedtime needs to be entered manually. Sargent et al. (2018) found that the average difference in sleep time was 51.5 ± 152.4 min for the Fitbit HR Charge vs polysomnography. Out by one hour on average, but by as much as 3 hours for some!
Although sleep trackers have generally improved, they do rely on movement to assess whether the person is asleep, and as such are prone to error. One thing that is often reported to affect the sleep data is a partner or pet sharing the bed moving.
Some of the algorithms that are supposed to work out your sleep stages are very good, but even different devices from the same company do not always not end up showing the same results (Garmin is one of them). Paying more doesn’t necessarily mean better assessment either. Even sophisticated systems can get sleep phases mixed up. Especially deep sleep and REM sleep are often reported incorrectly (Liang and Martell, 2017). In general time asleep, awake and sleep phases were found to not be accurate (Liang and Nishimura, 2017). Conclusions from several studies were that sleep trackers are reasonable to estimate sleep, but not valid to distinguish between light sleep, deep sleep, REM and wake. And it is precisely the length of these stages that are supposed to inform the apps how good the quality of your sleep was.
One issue that has been raised by doctors and sleep researchers is that an app can claim you slept badly although you don’t feel bad. This might cause anxiety or issues, and that you believe you are tired when in fact you are not. Case studies from the UW Medicine Sleep Centre found that some of the apps were telling people something different from their own personal experience. So even though they did not feel bad because of their sleep, the app told them they hadn’t slept long enough or deeply enough, which caused anxiety.
This is something that has been reported with the WHOOP recovery score and athletes getting over-reliant on them. If you get a bad recovery score despite feeling recovered, then you might question your ability to do a certain session. A lot of this is mental and could cause problems if the scores stay poor despite doing things correctly. Obviously, this can swing the other way too. The app says you slept great and recovered well, but you’re feeling dreadful. What to do? Listen to your body or an algorithm? (answer below)
As we can see above, smartphone apps are mainly poor at determining sleep stages and sleep parameters (with some exceptions). In contrast, a few simple sleep questions are reliable ways of assessing sleep. Minimum needed are bed and wake time, how long it took to fall asleep, waking during the night and ratings of sleepiness and alertness.
What is the take-home message then? As in most cases, common sense should prevail. Firstly, you should know how your sleep tracker actually works and how it determines your sleep score. Once you know that, you’ll be aware that it cannot just be 100% trusted as there are numerous things affecting the results. No one knows your body better than you, so listen to it. Just because your app claims you slept well and / or have a good recovery score does not mean this is actually the case. If you are still feeling tired, sore or flat, then take it easy or have a day off instead of blindly trusting an algorithm. On the flip side, if you had an amazing sleep and feel energised and ready to roll, but your phone is saying ‘uh oh, bad sleep mate. I bet you’re feeling knackered’, then maybe don’t pay too much attention to that and go and smash that PB!
This is why we here at Spoked will always collect your subjective sleep scores, regardless of what your app might say. As you can see, there are still quite a few issues with the data, and even though there might be improvements, it will never know you as well as you know yourself. Although we plan to integrate this kind of data in our app in future, it will always stay linked with the option of a subjective override of the might tech overlords! However, sleep trackers have made you more aware of your sleep, like never before.
Curious in giving Spoked a whirl? You can download it from the Apple and Google app stores for free, and when you're ready, start your free 30-day trial of our Pro plan. Or want to leverage the Spoked technology in your business? Read more here.
Evenson, K.R., Goto, M.M. and Furberg, R.D., 2015. Systematic review of the validity and reliability of consumer-wearable activity trackers. International Journal of Behavioral Nutrition and Physical Activity, 12(1), pp.1-22. Read here.
Halson, S.L., 2014. Sleep in elite athletes and nutritional interventions to enhance sleep. Sports Medicine, 44(1), pp.13-23. Read here.
Halson, S.L., 2019. Sleep monitoring in athletes: motivation, methods, miscalculations and why it matters. Sports Medicine, 49(10), pp.1487-1497. Read here.
Killer, S.C., Svendsen, I.S., Jeukendrup, A.E. and Gleeson, M., 2017. Evidence of disturbed sleep and mood state in well-trained athletes during short-term intensified training with and without a high carbohydrate nutritional intervention. Journal of sports sciences, 35(14), pp.1402-1410. Read here.
Lastella, M., Roach, G.D., Halson, S.L., Martin, D.T., West, N.P. and Sargent, C., 2015. The impact of a simulated grand tour on sleep, mood, and well-being of competitive cyclists. J Sports Med Phys Fitness, 55(12), pp.1555-64. Read here.
Leeder J, Glaister M, Pizzoferro K, Dawson J, Pedlar C. Sleep duration and quality in elite athletes measured using wristwatch actigraphy. J Sports Sci. 2012;30(6):541-5. Read here
Liang, Z., Ploderer, B. and Chapa-Martell, M.A., 2017, May. Is Fitbit fit for sleep-tracking? Sources of measurement errors and proposed countermeasures. In Proceedings of the 11th EAI International Conference on Pervasive Computing Technologies for Healthcare (pp. 476-479). Read here.
Liang, Z. and Martell, M.A.C., 2018. Validity of consumer activity wristbands and wearable EEG for measuring overall sleep parameters and sleep structure in free-living conditions. Journal of Healthcare Informatics Research, 2(1), pp.152-178. Read here.
Z. Liang and T. Nishimura, "Are wearable EEG devices more accurate than fitness wristbands for home sleep Tracking? Comparison of consumer sleep trackers with clinical devices," 2017 IEEE 6th Global Conference on Consumer Electronics (GCCE), 2017, pp. 1-5. Read here.
Miller, D.J., Lastella, M., Scanlan, A.T., Bellenger, C., Halson, S.L., Roach, G.D. and Sargent, C., 2020. A validation study of the WHOOP strap against polysomnography to assess sleep. Journal of Sports Sciences, 38(22), pp.2631-2636.Read here.
Nédélec, M., Halson, S., Delecroix, B., Abaidia, A.E., Ahmaidi, S. and Dupont, G., 2015. Sleep hygiene and recovery strategies in elite soccer players. Sports Medicine, 45(11), pp.1547-1559. Read here.
Sargent, C., Lastella, M., Romyn, G., Versey, N., Miller, D.J. and Roach, G.D., 2018. How well does a commercially available wearable device measure sleep in young athletes?. Chronobiology international, 35(6), pp.754-758. Read here.
