Withings Scanwatch Sleep Test (Review)

Hello everyone, my name is Rob and I am a Postdoctoral researcher. In this video I test the sleep tracking accuracy of the Withings Scanwatch.

Instagram:   / quantified_scientist  
Twitter: @QuantifiedRob
Linkedin:   / rob-ter-horst-90756a4a  
Website: www.robterhorst.com

📹Accuracy of the FitBit Charge series:📹
   • FitBit Charge 4 Ultimate Sleep Test  
   • FitBit sleep prediction: scientific t...  
📹Or the accuracy of the Oura ring for sleep prediction:📹
   • Oura Ring Scientific Sleep Test (Review)  

⏱️This comparison is part of a larger project where I spend over 11 hours a week tracking everything in my life. If you'd like to know more:
   • This mad scientist scanned his brain ...  
.⏱️

Timestamps:
00:00 Intro
00:25 Overview
01:17 Video Series Overview
02:10 Test Background
03:25 Results: Example Nights
05:13 Results: Scanwatch Mistakes
05:41 Results: Sleep Stage Percentages
06:18 Results: Sleep Stage Confusion
08:59 Results: Detecting Waking Up
10:14 Limitations
11:14 Conclusions

In my videos I do scientific tests of different tracking devices, I discuss different applications of data analysis, and in the end I hope to use self-tracking to improve my life. If you enjoyed this video or like some of my other content, please consider subscribing to my channel and giving this video a thumbs up, because if you do that the youtube algorithm makes it easier for people to find my video.

So how did I test the Scanwatch? I wore both the Scanwatch and this EEG device called the Hypnodyne Z-max to bed for 8 nights. I
It is important to note here that the Scanwatch does not predict REM sleep, it just predicts light and deep sleep. Therefore I am also interested in seeing if what is supposed to be REM sleep is mostly predicted as light or deep sleep.

This video is the second in a series of 4 videos: In my previous video I did a first test and unboxing of the Scanwatch, in this video I discuss sleep prediction, in the next video I’ll check if the problems I originally had with heart-rate detection after testing the device for just two days persisted in the next few weeks an in the last video I discuss the accuracy of the oxygen saturation measurements and the step counter. As always, since I don’t want to waste your time, timestamps are in the description below and also in the timeline of the video.

The device I’ll be testing the Scanwatch against is the hypnodyne z-max, which is a portable electro-encephalography or EEG device specifically designed for research. This device basically measures your brainwaves and muscle movements in your face, and this is the main way sleep researchers study sleep.
To show you why EEGs are the preferred way of measuring sleep, here you can see the raw EEG data from one of the nights. I will not go into the details, but without knowing anything about EEG, you can already very clearly see the sleep cycles I went through. This shows you the power of using EEG for sleep measurements. I think it’s amazing that your brain goes through this every night, and that most of us are “programmed” in a similar manner to go through all these phases.
To go from these raw brainwaves to sleep stages, I went through each of these recordings manually, and said for each 30 second interval what sleep stage I was in: light sleep, deep sleep, REM sleep or awake. This is similar to how it is done by specialists in sleep studies. This took me about 8 hours in total, which is why I stopped at 8 nights, since I figured this was enough data, and because I basically spent my whole Sunday doing this.

The Scanwatch on the other hand predicts sleep stages based on heart-rate and movement, and maybe blood oxygen saturation in the future. As you can imagine, without brainwave data it is much more difficult to predict sleep stages.

It is important to note there are some limitations to the test I did:
I only tested it on me, and only for eight nights. So I cannot say for sure how this translates to others, and if the data was biased in some way with having only 8 nights.

So to conclude I would say that the Scanwatch is not great yet at predicting sleep. It can often detect when I fall asleep and wake up, though even that is not extremely reliable yet, with 2 out of 8 nights being, so 25%, incorrect.
I hope they will solve the sleep tracking problems in the future, since they will still optimize the software before the actual release. For now I will mostly be relying on my other devices: my Fitbit and my Oura ring, for sleep tracking. And of course the z-max.

However, I did not, and would not, buy the Scanwatch for the sleep prediction.
Thanks for watching! Have a great day, and see you in the next video!