Garmin Advanced Sleep Monitoring

One of the emerging trends in consumer-based sports and activity technology has been the spread of heart rate variability analysis (HRV).

What is HRV? Let’s say you and I both have a HR of 60bpm. We could both have quite different HRV (RR). My heart might beat exactly once every second and you think that would be good…but you’d be wrong. Taking an example, your heart rate might vary between be 1/10th of a second faster than mine to 1/10th slower. You have greater variability but the EXACT same number of beats per minute.

Whilst a low RESTING heart rate is generally good; a HIGH HRV is generally good. A high HRV generally means there is less stress in your body and the HRV data which shows that is derived from the nervous system(s) in your body (via your heartbeat).

Stress is not necessarily bad. Exercise creates stress and that’s generally good but you can also get stress from work and, let’s say, that’s generally bad. HRV CANNOT DISTINGUISH between the sources of stress. Merely that there is stress.

That’s HRV 101.
Let’s come back to what you should now be able to see in Garmin Connect as it has now introduced SLEEP STAGES and measures them, in part, with HRV.

Sleep Stages

In a normal night of sleep, you will cycle through the different sleep stages: light sleep, deep sleep and REM (rapid eye movement) sleep.

  • The first stage is light sleep. Your eye movements and muscle activity start to slow down as your body prepares for deep sleep. This stage of sleep helps with relaxation.

  • The next stage is deep sleep, in which your eye and muscle movements stop completely, and your heart rate and breathing slow down. Your body goes into restoration mode, helping with recovery, building bone and muscle, and boosting your immune system.

  • Last in the sleep cycle is REM sleep, which cycles from short bursts to longer stretches as you leave deep sleep. This is the dreaming stage, when your brain is almost as active as when you’re awake. REM sleep is important for forming memories and processing information.

Source: Garmin

As well as seeing your night’s sleep cycles you can see trends over time… in the ‘usual’ Garmin Connect way.

Here is a night of sleep where I would say I slept well but, look, virtually no deep sleep. At least according to Garmin

In the past Garmin were making partly informed guesses based on movement data from your watch’s accelerometer eg lots of movement might indicate light sleep. Now Garmin are claiming to use additional data such as HRV. Garmin specifically say “such as HRV” and that’s probably because they will also use other data like your baseline HR, temperature, motion and maybe other things too.

Their informed guesses will likely be as good as Fitbit’s, Apple’s or any other vendor you have heard of that makes wearables. All vendors will not derive accurate sleep stages ( see wikipedia: Polysomnography). Actually, maybe Garmin will even be a little bit better than the mass-market competition as their ELEVATE optical HR sensor, in its latest iteration, is perhaps the most accurate from the main watch vendors.

Currently Garmin advanced sleep monitoring works with: vívoactive® 3, Forerunner® 645, Forerunner® 935, vívosport™, vívosmart® 3, vívomove® HR, vivoactive 3 Music and Forerunner 645 Music. Fenix 5 and other wearables coming soon. It is likely that will involve the new Fenix 5 plus series.

So What?

I appreciate I am sounding a little cynical so far. And I have been. Apologies. As you’ve probably guessed…there’s a BUT coming. A big one.

The problem here is that Garmin (and all the other vendors) are responding to consumer needs and wants. People seem to want to know their 24×7 HR; they want to know their steps; and they want to know their sleep cycles. Fair enough. That can’t hurt.

As Douglas Adams once classified the Earth, It’s ‘Mostly Harmless

Chest straps have ‘always’ been able to get HRV data. It was just that no one would wear them 247. Then came optical HR and everyone (well, me) got very excited. The initial reality of the inaccuracy of optical HR still exists BUT now the newer optical HR sensors only really have problems at high HR levels and in non-ideal circumstances (ie when exercising!). At sleeping and resting levels of HR they are generally pretty good to the point now where the vendors claim to be able to get HRV from optical sensors at resting levels.

Let’s think about exercising again for a minute and say you’ve done 2x20mins at 90%FTP or maybe you’ve just run 3x 5 minute miles. You did exactly as specified but so what? You don’t know the impact on your body, you’re measuring inputs. Measuring your HR during exercise might go some way to saying how your body is really being stressed (I know, I know, HR can be wrong!). You can in fact look at HRV as you exercise to look at the stress as it happens but that is an article for another day.

BUT you can also look at your HRV later that day as you sleep. you can use HRV to assess changes in stress. And, my understanding is, that scientists believe that adaptation to exercise mostly happens when you sleep. So REALLY what you want to measure is not the input of watts or minute-miles but rather the ADAPTATION to them.

Remember, exercise is REALLY BAD FOR YOU. If you do it non-stop you WILL die. But rather it’s the ADAPTATION to the stimulus of exercise that is good for you. #Sleep.

But HOW MUCH are you sleeping and HOW MUCH ARE YOU ADAPTING? #QualitySleep.

So it’s not simply measuring how long you sleep, although that’s a great start, but also the QUALITY of your sleep. Maybe also the quality of your sleep cycles. I also don’t think it’s simply a case of athletes JUST having quality DEEP SLEEP (see above for definition) to measure their adaptation to exercise. I think it’s more involved than that.

The BIG ‘SO WHAT’ moment for consumer-athletes is probably close. And that is where we can really see and quantify our adaptation perhaps looking at how HRV gets higher throughout the night and integrating that to get a measure of adaptation ( Eight Sleep and others already do that). To an extent, Garmin/Firstbeat’s ‘All day stress’ measure probably already goes a long way towards achieving just that. I’ve asked a few questions to people who should know about this in detail and may update this post further with their responses.

I recently wrote a speculative piece on the Fenix 6 (six) for 2019. The area of sleep and adaptation was one area where I cited the scope for improvements for athletes. But with the anticipated features for the Fenix 6 of Galileo, Garmin Pay, TOPO maps and music ALL coming to the imminent Fenix 5 plus then maybe I was wrong. Maybe the clever stuff will start to appear on our watches VERY soon. Maybe some of that clever stuff is really quantifying ADAPTATION as you sleep?

Now all YOU have to do is train hard to make the clever stuff happen 🙂 It was ever thus.

Factettes – from my HRV ‘training’

  • Evening alcohol – delayed or zero adaptation
  • Sleeping tablets – delayed or zero adaptation
  • Ironman training – progressive and almost imperceptible week-on-week decline in HRV until a breaking point was reached. I broke.
  • HRV – tells you that you are about to be ill. Really it does!!…sometimes.
  • HRV – tells you sometimes you’re fine to train hard on the bike the day after a hard run session. Sometimes not. And your plan may well say take a rest day as the default action after a hard day #WastedValuableTrainingTime
  • You sleep longer you usually get more adaptation

Source: Don’t forget that pinch of salt.

Garmin Fenix 6 – view on the future – opinion, including Forerunner 945


Reader-Powered Content

This content is not sponsored. It’s mostly me behind the labour of love which is this site and I appreciate everyone who follows, subscribes or Buys Me A Coffee ❤️ Alternatively please buy the reviewed product from my partners. Thank you! FTC: Affiliate Disclosure: Links pay commission. As an Amazon Associate, I earn from qualifying purchases.

4 thoughts on “Garmin Advanced Sleep Monitoring

  1. Good/solid post, T5KR. 👏👍

    „As a final experiment, we tried to add a little movement to the measurement. Nothing crazy, just computer work, a little muscle contraction, etc. … you have certainly noticed how it is key not to move while using the camera based measurement, as movement introduces a lot of artifacts. … as it is simply not possible to use optical technologies during movement, when looking at HRV analysis … However, always pay attention to data quality and artifacts, as movement highly affects optical measurements.“

    „​By looking at baseline correlations over periods of months, you might find factors that consistently have an effect on your physiology, and make adjustments. For example, the correlation between HRV & sleep quality (even subjectively annotated), shows up consistently across a broad range of individuals. It’s easy to understand that poor sleep gets you physiologically tired, and if you see this relation showing up in your data, maybe it’s time to try a few sleep hacks (cooler temperature, less screens, etc.). Nonetheless, beware of spurious correlations.“

    1. Schsche have a ‘data quality’ flag attached to their readings. I don’t know if Garmin is the same. but if they do that then this POTENTIALLY gets around some of the issues you highlight from hrv4training

      1. I’m going to comment from the lay point-of-view here…

        But just on the surface, all it looks like Garmin has done here is simply change how the watch interprets the levels of Deep Sleep and added the REM layer to the mix.

        Just from my history since they have implemented it I get lots of Light Sleep (I am actually genetically predisposed to being a light sleeper), have a very small window of Deep Sleep (since the change) and get a larger window of REM sleep (now with the change).

        If we are putting these into percentiles, My sleep measurement would be Approx 70% light–30% Deep with a negligible amount of movement, before the change. Right now it looks more like 60% light–5% Deep–35% REM. Again without looking at the math…all they did was change the deep sleep levels to incorporate REM into the equation.

        And again coming from a lay person’s point-of-view on the subject, that just seems like a gimmick to me and less science–which, when you really start looking at Garmin’s current approach to their devices (and current lineup barring the next generation of devices), seems to be more gimmick. I would think (and expect) there would be a need for a specialized sensor/device to really be able to get accurate data (not just a reliance on an external sensor like the chest strap). Throwing these options late into the lifespan of a device generation seems not good (Like the recent running POWER support for the HRM-TRI).

        Perhaps Garmin is just “testing” this option now and will have a proper way to measure it later (with some “NEW” sensor we haven’t heard about) with the next generation I don’t know but for now I really don’t think it means anything worth basing training/effort/life around.

      2. i think ALL the tech and algorithms are probably there now.
        yes i’d say you are right in that garmin are just fiddling around.
        it will be firstbeat that do the deed.

Comments are closed.