In hindsight, starting a sleep diary two weeks before travel plans promised to cause major disruptions wasn't my wisest idea. Then again, it has afforded me the chance to observe how well technology can keep up with my shifting body clock, not to mention the jet lag.
For the purpose of this week's entry I've spent seven days tracking my sleep using two wrist-worn devices, the Misfit Ray and the Fitbit Blaze, as well as the Beddit 3, a third gen sleep device that sits between the mattress and sheet.
Last week I tested a handful of wristworn devices just to see how accurate I found the tracking to be, and whether I could glean anything about my own sleep quality from them. Some inaccuracies and inconsistencies surfaced as expected, but the data did confirm a few things: I needed to sleep earlier, I needed to not oversleep; and that my sleep quality was being disturbed by outside noise, as well as possible other factors that hadn't yet surfaced. Above all, I realised that sifting through endless numbers wasn't going to improve matters but likely make them worse, as I noticed my sleep anxiety increasing the further down the data hole I traveled.
The Fitbit Blaze and Misfit Ray track sleep automatically, and for sleeping/waking up times I've found their results to be very close. What they detect between those points has been a different matter, with two notable disagreements over time restless/awake this week I want to flag. I knew jet lag would mean I'd be waking up at bizarre times, and on Sunday night Fitbit recorded 8h 49mins of total sleep, an hour of which was 'restless'. On the other hand Misfit said I had managed to get 9h 46min of shut-eye. One problem seems to be that Fitbit removes 'restless' sleep from your total, along with awake time, when in fact I may have simply been in a lighter phase. Interestingly, Beddit counted 9h 51min, much closer to Misfit's reading, detecting 34 minutes of awake.
Left: Fitbit Middle: Misfit Right: Beddit
Then things were reversed on the Saturday, with Misfit tracking 9h 53min of sleep, Fitbit recording 11hr 7min, and Beddit totalling 10 hour 52min. Those are some big fluctuations, and my theory is that it's down to how sensitively the different sensors have been tuned, as Stanford's Jamie M. Zeitzer mentioned in our first chat.
Other than a blip on Sunday night where it failed to pick up me waking up to check the time on my Fitbit (yes, the Blaze did detect it, I'd be concerned if it didn't) I've found the Beddit 3 to be quite a good device. It's barely noticeable under the sheet, and through a combination of sensors can measure heart rate variability, breathing, snoring and, of course, movement. Beddit's CEO tells me the HVR measure gives it the edge on most devices which check periodic heart rate, not variability.
Not only does it feel like a more thorough way of tracking sleep, at least so far, it feeds back bullet points on your night's sleep along with tips for improving it. I snore a lot, and the app has kindly told me this already (ok, ok, some people may have told me this before), and I now know that it's something I should keep an eye on.
Unfortunately, without any way of tracking the brain, the device is never going to be as accurate as an EEG, but when I had a chat with Beddit CEO Lasse Leppäkorpi this week he said he believes Beddit is significantly better than the gamut of wrist-worn devices. He puts an EEG at about 90% accuracy for tracking sleep, wrist trackers at around 30-40% accuracy, and Beddit at about 70/80%. It might seem relatively low by Beddit's own claims, but that's the nature of sleep tracking - even the best method doesn't get it right 100% of the time, something Zeitzer also told me.
If technology could score our ability to sleep on planes, I wouldn't even make bottom of the class. I can count on one hand the number of times I've successfully slept for an hour or longer flight, and this week - made of two 10 hour flights to attend the inaugural Wareable Tech Awards in London - was no exception.
However, I did manage to get one burst of an hour on my first flight, and the Blaze actually detected this, while the Ray didn't. I wasn't sure whether the conditions would be right for a tracker to detect a nap - it was late afternoon, the plane was probably moving me around a bit, and being upright meant I wasn't in an ideal position. Napping is healthy for a number of reasons, which is why I was glad to see the Fitbit detect it. It's also worth pointing out that I wouldn't be able to use the Beddit here due to the way it's designed, so that's one minor flaw. Of course most of these devices are really focused on the "main meal", but I still feel like tracking naps has its benefits.
So, bit of a weird week for sleeping overall, but some interesting and positive results from using the Beddit 3. My plan from here is to try some similar devices that don't need to be attached to the body to work, and see how I find them by comparison. Eventually, I want to move from tracking my sleep and using the data to guide me, to actually using technology that works more actively to improve sleep.
But before then, I need to make sure I'm making the improvements I've already set out for myself, and that I'm using what I believe to be the most accurate tech I can find. I also want to find more devices that offer useful guidance, and not just numbers.
I've already made some small improvements by going to bed earlier and trying to ensure I'm getting as many hours as possible. While I'm sticking to my goal of not sleeping too much on weekends so as to not disrupt my pattern, I'm not counting Saturday, due to jet lag disrupting those patterns mercilessly.
Right, no more planes. I'm back in my bed, and I'm staying here.