Smartphone sleep apps

By Karyn O'Keeffe 06/05/2013 4


For a few months now, I’ve been playing around with a sleep app to get a better idea of how easy they are to use, what data I could collect, and whether I’d remember to input my sleep data every day.  When I first started using it, I was mainly interested in my sleep duration and sleep timing.  Recently though, I’ve noticed the developers for my particular app have been adding more and more bells and whistles.  There was always the option to indicate lights on and off times for your sleep period, with additional options to limit phone functions, such as receiving calls and emails overnight.  However, my app now claims that it can use the accelerometer in my smartphone to accurately detect movement overnight, and interpret movement as sleep stages.

Many apps now include this feature, and provide users with a range of outputs from basic sleep structure to specific recommendations about how to improve sleep alongside their behaviour and sleep patterns.  What is concerning though, is that some of these recommendations are not scientifically-based or in line with best practise.  Most of these apps also claim to be extraordinarily accurate in their movement detection and sleep stage interpretation.  These claims are somewhat surprising given that sleep researchers themselves have been using movement as a surrogate for sleep and wake for some time, and not one of us thinks we can accurately interpret sleep stages from that data.  (Believe me, if it was possible, we would use it in a second.)

Activity monitors (actigraphs) are fairly reliable at detecting total sleep time (sleep duration).  Although not quite as accurate, they also provide an acceptable measure of sleep onset latency (the time from lights out to falling asleep) and sleep fragmentation (broken sleep).  Actigraphs are less reliable in clinical populations than in healthy adults.  It is not known whether this reduced accuracy extends to those with sub-clinical sleep difficulties.

Actigraphs cannot reliably measure different sleep stages, and I think this is highly evident in the graphical representations of sleep that many of these apps produce.  Often their sample diagrams show vastly abnormal sleep structure, and slow wave sleep shortly before waking.  If you’re getting a healthy amount of sleep each night, slow wave sleep just prior to waking is extremely unlikely.

Many sleep apps now include a feature to optimise our time of waking, so that we wake from light sleep. This is purported to reduce sleep inertia, which is a sensation of grogginess, confusion and decreased functioning, experienced immediately on waking.  Waking from slow wave sleep is thought to contribute to sleep inertia; however, sleep structure, sleep routine and time of day are also likely to be involved.  However, if we are to assume that sleep stage detection was reliable and given that regularly waking from slow wave sleep would suggest restricted sleep duration, and/or poor sleep routine or timing, I wonder if the focus should be on improving overall sleep health, rather than fixation on alarm clock times.

That said, a few apps provide great advice on healthy sleep, with sensible advice on how to achieve it.  Additionally, many people report benefit from using these apps (for example, waking feeling more refreshed) but I’m inclined to think that this results from a focus on achieving better sleep, and perhaps the effect of receiving personalised feedback on your own sleep, than the performance of the app itself.


4 Responses to “Smartphone sleep apps”

  • I’ve heard about these apps (but not cared enough to investigate how they work) how do they pick up the movement data? I can’t imagine they can detect much from a bedside table but I don’t like the idea of sleeping with a phone strapped to my body. Are they meant to pick up vibrations when placed on the mattress?

    That would presumably depend greatly on the sort of mattress.

    so, what’s the deal?

    • There seem to be a range of apps that use movement (note, there are also apps/devices that use other physiological measures). Some are similar to the ones used in research and clinical practice, in that they use a device worn on the wrist that can transfer data to an app on the smartphone. Most, however, seem to use the accelerometer in the smartphones themselves. The phone needs to be placed on the mattress each night. But as I’m sure you can imagine, it is likely the quality of and night-night variability in the data would greatly depend on a large number of factors: where you placed your smartphone in relation to your body, what position you slept in that night, how much you moved around the bed that night, how close your partner was to you…

  • …if you’ve got one of those magic mattresses from the ads where you can have a glass of wine on one side and can jump up and down on the other….

    yeah, about what I figured. Thanks.

  • Very interesting and congruous with what we expect the good doctor would say.

    Rather than diagnose, smartphone sleep apps should simply collect accelerometer data (the above-noted variability makes this a dubious approach), prompt the subject to easily input sleep-related study information, and relay said data package to a professional for analysis and recommendations.