This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

How does Garmin distinguish between Deep and REM sleep?

Adding my voice to the many reports of sleep discrepancies. VivoActive 3 Music. For weeks now (since I started wearing the watch at night again), I noticed that Garmin Connect presents my sleep as alternating Light and REM sleep, with sometimes only a very brief period of Deep sleep at the start of the night. This does not reflect how I feel I've slept.

This morning, I happened to check my sleep stats immediately after waking. My Connect ap showed over 3hrs of Deep sleep distributed throughout the night, more in the first half than the second, and periods of Light and REM sleep, which looked like a normal sleep pattern to me.

Then, less than an hour later, all but the first half hour of Deep sleep had disappeared, apparently reassigned as light sleep and perhaps a bit of REM.

So now I'm wondering: how does Garmin do this? Why does a plausible sleep registration get re-analyzed, and then change into a much less plausible chart?

  • how does Garmin do this?

    See it explained here: A Sleep Analysis Method Based on Heart Rate Variability 

    At older watch models, the data is being analyzed by the server. Apparently in several passes, with some adaptive or learning algorithm, so it indeed changes with the time as being fine-tuned. At newer models the analysis is done directly on the watch, and I did not notice that it would be changing like it was with the old models.

  • Thanks . That's a very useful and comprehensive answer. It does leave me with the question of the discrepancy of my own assessment of the quality of my night, and the Garmin assessment that I've had little to no deep sleep. That still makes me doubt the validity of the sleep tracking....

  • That still makes me doubt the validity of the sleep tracking....

    I am afraid the only way to tell for sure is having the sleep analyzed by a polysomnography, and comparing it. The subjective evaluation does not have to be necessarily correct. It does not mean the sleep analysis by Garmin is correct either.