Thursday Tech Tidbit: Garmin Rolls Out Sleep Cycle Details, including REM Sleep

(This is the second half of what should have been ‘Twofer Tuesday Tech Tidbits’ earlier this week, until this tidbit broke free and decided to go live on Thursday instead. Thus…enjoy!)

As of this morning Garmin has rolled out additional sleep analytic data within Garmin Connect, which breaks down your sleep cycles based on leveraging additional data from the wearable to get more details which are in turn translated to specific sleep stages.  These stages are defined as follows (according to Garmin):

1st Stage: 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.

2nd Stage: Deep Sleep: 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 Stage: 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.

Garmin says they’re starting to leverage heart rate variability data from the optical HR sensor to assist in this data, alongside the existing accelerometer data they’d been using to date.  It’s not as well known, but Garmin has actually been leveraging HRV data since early 2017 for non-workout data purposes, such as measuring stress. Early 2017 is when Garmin updated their ELEVATE optical HR sensor to be truly 24×7 at 1-second interval rates.  Whereas previously the interval rate (for something like a Fenix 3HR) was anywhere between a few seconds and tens of minutes.  The new hardware introduced back then opened up this door, alongside others.

To understand the difference between what these data points used to look like, versus today, here’s two nights side by side.  One from a month ago where the new sleep cycle data isn’t enumerated (before/left), and then one from last night where it shows the data (after/right):

2018-06-11 20.35.26 2018-06-11 20.33.57

And for fun, here’s that data within the context of the weekly planner view for the last two weeks. Day’s that are grey/blank are those I didn’t wear the watch, or in the case of this week, remain in the future.

2018-06-11 20.33.21 2018-06-11 20.33.13

And finally, a few more days of my life with the new data details, as well as the secondary view too:

2018-06-11 20.33.57 2018-06-11 20.34.14 2018-06-11 20.34.06

Garmin says that by adding in the HRV data, the side-effect is that they’re also slightly increasing the fall-asleep/wake-up timing as well.  And at least insofar as the last few days of data that I can remember, these fall-asleep/wake-up times are within 1-2 minutes (or less) of my actual times.  The one exception, however, is the 4:30AM wake-up, which didn’t totally capture the cluster-fudge of up and down for the next 3 hours of trying to get two timezone-upset babies to stay asleep for more than 16 minutes at a time.  Not sure I can really blame Garmin there though.

As for compatibility, here’s the watches that are compatible. Basically though it’s any optical-HR sensor laden unit introduced since January 2017.  Anything prior to that lacks the optical HR resolution (frequency) of data to make it work.  These first watches will work as of today (immediately:

  • D2 Charlie
  • Forerunner 645
  • Forerunner 935
  • Vivoactive 3
  • Vivoactive 3 Music
  • Vivosport
  • Vivosmart 3
  • Vivomove HR

This next set of watches will work “soon”. My guess is this is a ‘really darn soon like days or weeks at worst type soon’, and not a ‘next year type soon’. But ya never know.

  • Fenix 5 series
  • Quatix 5
  • Tactix Charlie

And of course, be sure that your Garmin Connect Mobile app is all updated as well.  Oh, and this will all show up in the ‘My Day’ summary tab totaling each day.

2018-06-11 20.33.30 clip_image001

Now I don’t have any meaningful/handy/easy-access way to validate this sleep stage data. I could possibly compare it to other sleep platforms out there, but I didn’t bring any on my current trip.  And for that matter, I haven’t even unpacked them yet and installed them under the bed since moving to Amsterdam. Perhaps someday my unpacking journey will take us there.

In any case – go forth and enjoy more sleep data.

With that, thanks for reading!

Post a Comment

Your email address will not be published. Required fields are marked.
If you would like a profile picture, simply register at Gravatar, which works here on DCR and across the web.

 Notify me of followup comments via e-mail.

You can click here to Subscribe without commenting

Add a picture

*

119 Comments

  1. Dave Lusty

    Hmmm I wonder why the Fenix 5 is delayed. Perhaps related to the Fenix 5 Plus that “nobody” knows about pictured recently on the Garmin website? Considering that 9.10 was released a while ago on the 935 and still in release candidate on the almost identical Fenix this is the only explanation I can think of. It’s more annoying because my Fenix 5 still can’t do the other stuff like trueup that came with this release 🙁

    On another note, do you think they’ll allow setting of watch alarms based on the nearest light sleep to a given time? That would be super cool and a huge step forward, although the plus would need beeps loud enough to hear too…

    • I suspect it’s some minor bug being worked out, based on what I’m seeing. While the FR935 is almost identical, they are worked on by separate teams and the alignment of products in the groupings above roughly shows that.

      I think they’ll eventually get there on watch-based alarms. I feel like Garmin is kinda finally getting to the point (in particular with the Fenix 5 series, but even with CIQ in general) to be able to start turning the screws when it comes to software-based features. They aren’t quite to the development level to be able to charge for those features on a turn-on basis (which is ideally where we’d get to, so we can stop having 98 hardware varieties of things – ala Edge 520 and Edge 520 Plus). I’d like to get to the point where I can buy a base piece of 2-3 hardware choices, and then pay for what I want feature-wise.

      Obviously, that’s a trickier prop to make work from a business standpoint, but certainly very viable if done right.

    • Dave Lusty

      I’d like to see that too, but realistically Garmin need to accept that they sell watches and the software isn’t a value item. Watch straps and other accessories are where the extra profit is, and for that the software just needs to be on all the devices. I get that a cheap watch may have fewer BTLE/Ant channels or lack a baro altimeter, but restricting power is madness when it’s arbitrary. I have 3 edge devices because I wanted 3 sizes, and two watches because one looks nice with a suit and one is nice to run with. Feature differences really frustrate me when they should all work roughly the same.
      As you say though, they seem to be agile enough with their development these days so we’re probably not far off seeing them start to fix some of the more pointless bugs 🙂

    • simon

      seems to be working on the 5x ?

    • Bartokian

      I can confirm it’s working on a 5x…

  2. Chris Watson

    This new update is so boring – just sends me straight off to sleep!

  3. I’m curious about how much of any of this data (1 second hr data, hrv, sleep timing, etc) is downloadable from Garmin? I’m looking to replace a Fitbit Charge HR 2 with a vivosport/active, but am concerned about whether I can get access to my raw data. Thanks!

    • You can download it all via their GDPR stuff these days, though doing something with it is FAR trickier. Same with Fitbit/etc… I started looking at the massive file you can download, but haven’t quite identified where the 24×7 HR data is within it. I’ve found boatloads of other things.

      Perhaps there’s a map/guide for it somewhere (that’d be useful, and I’m sure Garmin has it for internal/support purposes).

    • Interesting. It hadn’t occurred to me that GDPR would apply here, but I’m not seeing an easy link for downloading the data (perhaps because I’m in the US?). I do find a link that allows me to “request data export” (link to garmin.com), but it says it can take up to 30 days to produce. Not particularly useful for day to day management of daily activity / fitness. FitBit has an API that is pretty reasonable, and Garmin must have something similar for individual activities so that sites like tapiriik can move your data around, although that may not apply to all-day data like HR data. Thanks for the info.

    • When I tried it a week or so ago, the link to the Zip file came within a few minutes. Super quick.

      Working on a piece comparing all the GDPR exports and such from different services.

    • Rich Mercer

      I’ll be interested to hear what you find because GDPR has the (unrealistic in my opinion) requirement that any data exported must be portable.

    • The challenge is the word ‘portable’. The data as-is in that export is very much portable. It’s an an easily openable format as generic data. And your .FIT files are also in .FIT files, so that’s all good.

      But as with any standards, it depends on how people use them. There is simply no standard for saving sleep/RHR/step data. Doesn’t exist. Uploaded workout files, yes, a few options there (.FIT/TCX, and to a lesser degree .GPX).

      That’s why when companies tell me they export your workout in .CSV format I want to bang my head against a wall. Sure, it’s open-access, but it’s totally useless. Which is basically the same thing here. I can export that sleep/whatever data from Garmin, but I can’t import it to Fitbit, nor Polar. Albeit interestingly, you can actually import in that data from Fitbit to Garmin (but that was more of a competitive/marketing move).

    • Sam

      I wish I could export detailed information from the other “all-day” health stats to CSV because I really just want to make my own spreadsheet. Right now it just gives me daily averages. Garmin support told me to fill out a feature request form at the link below and when I tried I got the error “This form is currently private and cannot be viewed by the public”. Then the support person stopped responding to me. At least they are working on something over there. link to www8.garmin.com

  4. Marc Simkin

    Ray, do you know the specific version of Garmin Connect Mobile and device software needed to enable this? I’m not seeing any of this data today with the iOS version of Connect and my VivoSport.

    Thanks

    -marc

    • Also not seeing it on my 935, everything up to date according to the app store!

    • simon

      what version of the 935 firmware – working on a 5x with 9.10, latest IOS app

    • Jeff

      It’s working on my 645M. Firmware version 3.30 and Android app version 4.7.

    • Brian Reiter

      I can also confirm that this features is working with my f5x running the 9.10 firmware and the June 1 Garmin Connect app on iOS. Every Sleep Details going back to May 30 shows Deep, Light, REM, and Awake, now.

    • killed and restarted GCM and got a message that I now have Advanced Sleep! No app updates, no watch updates so something must just get triggered at the GC end I’d guess.

    • Marc Simkin

      Thanks Greg.

      I needed to do the same thing.

      I just find it strange, as I did that earlier and it didn’t work. A new version was not downloaded.

      I’m wondering if there was some “switch” that needed to be thrown on the “back end” that took time to propagate through all the accounts in Connect.

      Thanks

      -marc

    • Jase

      I assume that it works on the F5X running the recently released beta or “Public Release Candidate v9.10” firmware? Current latest (public) Fenix 5X firmware that I can see is v8.00 which does not have the latest feature…?

    • Eric Buxton

      I am running 8.0 on my 5x and just checked-the sleep info is there now

  5. Flobble

    “Garmin need to accept that they sell watches and the software isn’t a value item”

    Err, no. Just no.

    I don’t think many people (anyone?) buys a Garmin device primarily because of its ability to tell the time. The value is almost entirely in the software – the tracking, the GPS, the HR measurement, gathering power data, etc. is all made useful by the software. Yes, the underlying hardware provides raw sensor data, but it’s useless without the software layers on top to interpret, reformat, display and transmit meaningful information. And that very definitely includes the associated web services (aka Garmin Connect).

    • Dave Lusty

      Yes Flobble, that’s why you’re buying a Garmin watch rather than a Timex. My point was that Garmin trying to also squeeze a few extra bucks out of us after we buy a £800 watch will not be a good move. I’m happy to pay that for the sapphire Fenix with metal strap because it’s a nice watch. I don’t pay that because it has more features than the VivoActive. Charging for software features will work about as well as car manufacturers charging for updated gps maps – a big fat fail. Ultimately users were disapointed because their car didn’t do what they hoped it would and the same will happen here. History shows again and again that this model just doesn’t work. Happy customers are those that get the best their device can possibly give without feeling like they are being extorted for more money to enable things.
      Apps are a different thing altogether, and there might be a market there but unlikely. I wouldn’t pay for a Triathlon app, but I might pay for a music service which has a free app.

    • Dan G

      “History shows again and again that this model just doesn’t work.” Except all computers (and now phones), ever

    • Mitch W

      Disagree…With a Google Pixel Watch being announced soonish (October?), I am guessing lots of people would buy a triathlon app (in theory). Google will probably not release the watch with fully functioning outdoor multisport tracking capability. However, If they produce a great watch with full featured hardware (waterproof, accurate GPS/HR, etc,), a battery that lasts 3 days (good for Wear OS, ultimately we’d like to see longer), and looks good for (guessing on pricing) $249-399, and there was a multisport app (Polar???) that provided a fully formed athletic software ecosystem, I would throw down some extra cash for the app.

      I only mention Polar because they already have this, except (correct me if wrong) it runs only on m600, not any wearOS device (and artificially limits the ability to do a multisport activity). The real question is what is the value of the software. Would they be willing to sell it as a standalone app and get out of (or at least reduce their role in) the hardware game? What price is worth it to them and that people would be willing to pay?

      It might take a few years, but I think this model will ultimately win out, just like it has for phones (and is ultimately happening to cable and TV). Buy the pretty hardware that you like and has the technical specs you need; purchase the software that you want, and don’t purchase anything you think is fluff.

  6. Ray –

    As always, thanks for the update and insight! Do you have any current buying recommendations based on these new features. I still have a Fenix 3 and am using that with the Scosche Rhythm 24. I was debating about whether to buy a Fenix 5 series, but I don’t want to invest in it and then see Garmin come out with the next version of the Fenix in the coming months.

    • simon

      I can’t see this one specific feature is really worth basing a buying decision on – we don’t even really know if it’s in anyway accurate…or what we can do about it.

      there are many internet rumours of a new 5 (x?) series of watches, the plus, just about to be announced. I’d hang fire a little until we know more.

    • Dave Lusty

      +1 the 5 plus is imminent and there are pics all over the web of them and thier feature list. The 5 is currently discounted though and is a great watch. In fact, I think the sale price ends this week so I’d expect the announcement before this weekend…

    • Brian Reiter

      The enhanced sleep tracking works with the F5 series with the 9.10 (beta) firmware which is in general availability release for the Forerunner 935.

      (Above Ray suggested that Garmin might have discovered some bug in the late beta testing for 9.10 on the F5 series hardware that doesn’t affect the 935 hardware variant, somehow. Or alternatively, it occurs to me that the Fenix team is currently more risk averse about shipping a bug than the Forerunner team.)

      A F5 and F5S Plus might fix the RF issues that are a problem with some accessories. It will probably also add Garmin Pay and possibly music playback. Otherwise, there isn’t enough marketing features to warrant a hardware revision.

      Garmin Pay contactless payment doesn’t work with every bank card. If your bank is not supported then there is no point, although eventually they will probably, mostly all work. And then depending on where you live, contactless payments may or may not be viable at point of sale. link to garmin.com

      Music playback might be a killer feature, too, but will definitely have a major battery life impact if you use it.

    • Patrick Utrecht

      For what it’s worth, the internet rumors list june 20th as a potential announcement day, with the watches in stores early july. Not sure if that is US time/date but we’ll see. And like any rumor, it may be false…

    • Jvr

      Yeah, I think it’s going to be available almost immediately after they announce. They have a link on their “Garmin Pay” page to their shop for the Plus series, but the link points to a 404 (not available). This hints to me that they have it staged and ready to go at the press of a button, so I’m betting announcement next week, and avail either immediately or within a week of announcement.

      I’m getting jittery, I’ve wanted one since I returned my 5X (which I otherwise _loved_) due to poor GPS and pending plus series a few months ago.

    • The Garmin Pay page has since had all references to the Fenix 5 Plus removed. But like another as said the information is pointing towards a 20th June announcement

  7. Goncalo

    This post makes me wanna go to sleep… 🙂

  8. Jay M

    I’ve never noticed issues with wake/sleep times reported by my watch until this week when my F645 reported my wake up time 2 miles into my run. That was a good 30 min after I actually woke up, during which time I spent some time walking around, 5 min on a stationary bike and 5-10 min drive to the trailhead.

  9. the5krunner

    On the occasions when I have delved into sleep tracking and its accuracy and usefulness I have usually initially got very excited and then, ultimately, a little bit disappointed.

    Some of the people who SHOULD know, and that I have exchanged emails with, seem to say that even accurate sleep cycle tracking is not really possible. Even in a lab they probably get the cycles wrong.

    so when I seee comments (above @JayM) like “a good 30 min after I actually woke up” I tend to have a #Sigh moment. NOT AT JAYM…and NOT AT GARMIN …. but rather at the technology in general

    For those wanting background reading, maybe start here: link to en.wikipedia.org

    • Mirko Surf&Run

      Me too I don’t believe very much about accuracy of “deep” and “light” sleep in Garmin Connect. My question is: why is Garmin filling us with wrong data? If they are not sure of what they find, why do they report us? We are having always more data from our devices and from Garmin Connect, but it seems to me alway fewer data are accurate.
      If they want to give us medical data about our health, they should be sure about their results and this analysis should be approved by scientist or doctors.
      I remember that the first days I was very worried because Garmin reported for me very little “deep sleep”, sometimes the whole night was “light sleep”.
      Then I decided not to care about that any more and I feel much better.

    • I suspect like most things that are body-based, it’s never quite as simple as being black and white in terms of ‘working’ or ‘not working’. Just like with optical HR. For some people it works 95% of the time, and for other people it works 5% (or less) of the time.

      Everyone’s body’s are different and in general the tech is quite new – so it’s the same process of slowly sorting through edge cases with each update.

      In my case I generally find the actual timing to be really really good for me (in terms of wake/sleep). I have zero idea on whether or not the sleep cycles are good. Mostly because with little babies causing wake-ups one or more times a night, it’s rather rare that I’m getting a ‘good’ night of sleep. And the times I’m not at home, I’m at a trade-show/etc…which I get even worse sleep. 🙂

      For me, I find it most interesting to just track overall sleep, rather than focusing on all the underlying data.

    • Mitch W

      I have a watch which tracks sleep, and the wake/sleep usually is spot on; sleep cycles..who knows???

      My problem with it though, is ultimately who cares? It’s not actionable. I get messages and warnings, but they are virtually meaningless

      Oh, I had a lack of sleep 5 nights this week… hmmm, I went to bed too late 4 times… Jeez, 6 nights this week with not enough deep sleep!!!

      The problem is I can’t really change much of that. There’s nothing I can do to increase deep sleep (AFAIK). I am not going to stop hanging out with my wife in the evening and start going to bed at 8pm. And I am not going to stop waking up early to workout (best time for my schedule).

      Ultimately, I don’t learn anything I didn’t already know; I don’t get enough sleep. The only answer (sleep more), I am not willing to do.

      Is anyone using sleep data in a more constructive manner than me? I would love to learn what to do with it if I am missing something.

    • Pat

      Agree.

      I went from a Jawbone Up (loved the smart alarm, really helped waking and not feeling groggy), until it stopped charging.

      All the while, I had kept reading about the Basis Peak and how it was one of the more accurate for sleep, until it was discontinued. Looks like Intel is sitting on the patent.

      Then moved to the Fitbit Alta HR, which has had some good info so far. And just recently received the 5X. The total sleep time is similar, but the Fitbit is a lot more sensitive. Last night I was up in the middle of the night for 10 minutes, the Fitbit picked up on this the Garmin did not. Over the course of the night, the Fitbit showed ‘Awake’ for 45 minutes, the Garmin 5 minutes. And given how I felt when I woke up (even before looking at the results) I would say the Fitbit was closer (given some margin of error). Somewhat excited to see this latest update, if even so I do not wear 2 devices to bed (or if I can just wear the Garmin 24/7 and not switch to Fitbit at night). I have found the steps/hr close enough for both, although they both record my bus commute as steps vs phone (samsung s health).

    • Boris

      I’m really surprised, Ray.

      No way to track reliably overall sleep with a device which can’t distinguish between laying on the couch (relaxing, reading something, watching a movie) and really laying in bed (while trying to fall sleep). All current activity trackers are “broken by design” and deliver a lot of bogus metrics – some (maybe) a little less, than others. IIRC there was one device on the market, it’s a while ago, which you strapped on your head – that seemed to work ok’ish, what I’ve heard.

      link to twitter.com

    • I also find the sleep start and end to be fairly accurate and find this useful to look at my overall week, especially when I am travelling to see what sleep I am missing out on ect and the impact the following days when things are more relaxed and back to “normal”

      What I don’t really look at is the inbetween bits as, 1 I can’t validate it, 2 I can tell myself If it’s been a good sleep or not

    • Andrew

      @ Mitch,
      I’m of the same view, sleep data makes for pretty graphs etc. But do I use the data in a meaningful way, no!
      I also thought about how the data could be used against me. Say I crashed a car and killed someone, and somehow lack of sleep came up, then might the data be used as evidence against me. I don’t know the answer, so for the moment, the feature is turned off

    • Bob190

      Bingo!

      Exactly on the money.

    • the5krunner

      guys (& girls)
      the sleep cycle stuff and the going to sleep stuff and the how much sleep i got on the couch or wherever is super nice and super interesting (and probably wrong)

      people want these metrics and are interested in them…and so here they are. garmin meets consumer demand.

      one of the above comments is ‘so what’. which is always the best comment to provoke some thought.

      the answer is that sleeping HRV (and hence maybe sleep cycles) could very well tell you the degree of overnight adaptation you experience. ****ie your response to training.****

      furthermore it could add more meaning and accuracy around the readiness to train metric from a waking hrv test.

      it may or may not link in to the all-day stress metric from firstbeat (i’ve asked)

      so: the prettiness of the charts and cycles is nice.

      BUT the ‘so what’ answer is potentially real and has a potentially actionable impact on your training.

      FWIW: QS/EMFIT has been doing this sort of thing for a few years and non-intrusively.

    • gingerneil

      Bang on the money there Mr 5k.

    • Zoltan

      It is why I bought a Zeo headband, when it was still available.

      I dont believe motion-based sleep tracking.

    • Andrew M

      +1.

      Yes, the watch continues to spew out data points, but is it actionable?

    • wildroamer

      Late to the party, but couldn’t agree more with your take on G’s willingness to leap before looking. Frankly, I find the results of the new analytics, as well as their graphical representations to be puzzling.
      As I understand (and G has said), progression is awake to light to deep, and finally REM. My results are similar to the writer’s, several stretches of light to REM pattern that don’t really follow the norm.
      And why have graphs ordered awake-REM-light-deep?
      Makes little sense, imo.
      Anyway, for me the most telling metric is movement anyway, so whatevs.

  10. Freek

    Will this feature come to Chronos?

  11. Bsquared

    Ray,

    The only sleep tracking accuracy data I’ve seen is from Fitbit’s own research that was published in 2017:
    link to sleepmeeting.org

    Page A26 states 69% accuracy of identifying wake/light/deep/REM in 30-second epochs, and a Cohen’s kappa of 0.52 +/- .14. The Cohen’s scores put the lower end 0.38 (0.52-0.14) as being “poor/fair” and the upper end of 0.66 being “fair-to-good/substantial.”

    Hmm, so Fitbit’s own research on sleep stages shows hit-or-miss accuracy.

    An older 2012 study:
    link to pdfs.semanticscholar.org

    “Fitbit and Actiwatch-64 showed 97.8% and 95.7% of agreement rates with PSG for sleep detection but 19.8% and 38.9% for wake detection, respectively [6]. Finally, both devices were found to consistently misidentify wake as sleep and thus overestimated both total sleep time and sleep efficiency.”

    And while I’d like to think accuracy has improved, Fitbit is strangely quiet on independent research showing the data is reliable and actionable.

    Fun to look at, but Fitbit sleep stages hasn’t proven to be of any value in my training or life.

    Would love to hear differing opinions, or see updated research results into the accuracy and value of having sleep stages data.

  12. So if Garmin can do HRV tracking at night any chance they will add this functionality?
    link to firstbeat.com

    • Seems more actionable to know how recovered you are then the stage of sleep you reached

    • Dan G

      Yeah I wish they’d implement this too. Seems odd that only one, somewhat old device has it.

    • Yeah, I suspect it just comes down to whether or not Garmin wants to pay to license it.

    • gingerneil

      This is fundamental for me. Yes, looking at pretty charts is nice, but it’s the ‘so what’ that counts. Will I go to bed earlier tonight because garmin tells, me I had a, poor sleep last night? Probably not – I’ll do that if I feel tired… But if this is built into the recovery and conditioning advice that the watch is, giving me, I may well take a little more notice.

    • gingerneil

      Not sure why my phone thinks I need all those commas! Maybe it’s over tired!? 😉

    • Eli

      Fitbit adds rem sleep, people tell garmin they want the same functionality as Fitbit. Is that functionality useful? Not really, but people still want it so its worth investing in. They won’t pay for things people don’t ask for. People won’t ask for things they don’t know about.

      Sure when this first came out it was useless as people didn’t want to sleep with a chest strap.

  13. Mark H

    I’m on GCM for Android (May 30?), version 4.7 and FR935 v.9.10 (released 6/4, on my watch slightly later).

    Can we assume this was turned on with the firmware updates, or are they turning it on in GCM remotely?

    I see that GCM has REM going back a week. The tracking was still pretty poor unless you count reading my phone on the toilet as REM. And it’s still counting reading on my phone on the couch as deep sleep and REM.

    I sure hope the tracking is better starting today.

  14. R_Tellis

    When a lot of people who were coming from Fitbit started asking about REM tracking on the Garmin forums I thought it was silly and was glad Garmin wasn’t doing it. REM is measured with and EEG not an EKG.

    After looking at my last few nights and seeing that I apparently don’t have any deep sleep, only light and REM, I stand by my previous assessment.

  15. Juro

    Force-closing and reopening the app on iOS “enabled” this on my standard fenix 5 on 9.10 RC. Thanks fir the tips!

  16. S. Savkar

    I agree with most folks here that tracking deep versus light versus REM modes seems to me useless with the Garmin. Overall amount of sleep has been the only thing I have paid attention to, but as some others have pointed out, it isn’t always that I can do anything about the fact that it just confirms what I knew – I have terrible sleeping habits!

    Anyway, one of these days perhaps I will take it more to heart and actually just say screw it, no more late night TV, off to REM mode for me.

  17. Dan Aponik

    So disappointed that those of us with the 3HR will not be able to take advantage of this update. My watch has not left my wrist in the 19 months that I’ve owned it so that I can monitor my sleep patterns…and now that they are improving the functionality, I’m left in the dark (literally).

    • Yeah, the challenge there is purely hardware based. As you know, the Fenix3 HR doesn’t do true 24×7 HR. It skips around, perhaps missing 10-15 minute (or longer) chunks between readings. Thus, the data for something like this wouldn’t be anywhere near reliable.

      The new optical HR sensors weren’t introduced until January 2017 – a year after the Fenix3 HR came out.

  18. Peter

    All my deep sleep seems to have gone REM, I remain to be convinced…… Vivoactive 3 HR, glad I’m able to BT my Rhythm+ for running, it’s more accurate than Garmin….
    My Rhythm remains the most accurate optical reader during exertion……(possibly bc I can place it above the elbow…..I think Ray has said wrists are not a particularly good place to read HR. I’ve done multiple comparisons with my Wahoo X strap.
    Still this Garmin is an impressive piece of hardware.

  19. Ronny

    Hi. Bin aus Deutschland. Mehr als ein Schlaf am Tag wird trotzdem nicht angezeigt oder? Auch keine Nickerchen?

  20. JTH

    It actually doesn’t surprise me that HRV might play a big role making the data more accurate. As we all know Garmin already use HRV to track your stress reactions throughout the day.

    I actually stopped looking at the sleep tracking details long a go to only roughly how much I get sleep. But what I’m more interested is the stress data. During the day it’s sometimes difficult to say if the stress comes from something you did but generally while sleeping it should always be pretty low. So if it’s on the blue for the whole night I know I’m well rested. After a hard workout it might have a few orange bits or more depending how tired I am. So all in all it is already a pretty good insight to your recovery.

    Another interesting note is how unhealthy habits like weekend drinking affect the stress data while at sleep. It may appear I sleep well (and I usually do, nothing can wake me up after a few beers) BUT, the big difference is in stress data. Sometimes it shocks me how high the stress is, all in orange for the whole night, and if I drink less like up to 3-4 beers it might still go to blue just towards the waking hours.

    It’s kind of common sense that drinking is bad for you, but just proves that the technology works in practice and it’s quite easy to test it when you’re not putting any other stress to your body. Especially for small amount of drinking you may not feel the effects but you can certainly see it in the stress data as your body is doing extra work burning the alcohol.

    Something to consider for the holiday season 😉

  21. Ben

    Great to see Garmin finally giving some attention to sleep tracking again! I am yet to see this on the Fenix 5 with non beta firmware, hopefully it comes through in the next few days.

    Is there any chance of this update or future updates including the ability to track napping? This is a feature that my partner’s Fitbit Blaze can do with seemingly good results and I know this was once possible (albeit cumbersome) with the Fenix 3. In a job where start and finish times vary greatly, naps are a part of life. If a low end Fitbit can do this surely the top end Garmin devices should too.

  22. Chris G

    When I moved from Fitbit to Garmin Fenix 5, this is one of the features I was missing. Is still think that in general Fitbit have a better looking app UI, and is better at visualizing the collected data.

  23. Andy Banks

    And Descent MK1 off the list again @GarminHeath?

    For a £1,300 device which was promised as having the best features of the Fenix 5, it seems to be missing out quite a lot.

    It was released in February and has had a single firmware update in May, is still stuck on a very old CIQ format, which doesn’t support for example Running Power.

  24. Matt S

    Any chance you could do two watches- one on each wrist- to attempt to validate the data. Of course, you would need to load them into separate Connect profiles to have it ‘make sense’ but it would be good to demonstrate.

    • PeterF

      How about a Fitbit on one wrist and a Garmin on the other? Just curious if they at least somewhat agree (with only two you’ll never be sure who is actually correct of course).

  25. Artur

    Ray, do you find any of this sleep data, advance with REM, or non-advanced, of absolutely any use? Given your vast and many years of wearing wearables, do you find that you use this info for any particular purpose? Thanks!

  26. John L2

    Working on my fenix 5x (9.10 beta) as of this morning, and backdated to June 1 (which was before 9.10)

  27. Kathy Ormiston

    How do you get the weekly planner view of the sleep tracker to show up?

    Thanks for the review.

  28. Remco Verdoold

    Do I see the night time feeding of peanut 2?

  29. Jim Andrews

    As a lot of people have said, this is a great demonstration of technology, an amazing feature….but….so what?

    We all know that some things can really affect our sleep quality: work stress / coffee / beer / wine / babies / overtraining. Some of these things we can manage; others we can’t.

    As a few people have noted, a metric which tells us how the sleep we had last night might influence the training we do today would make this “actionable” information. Perhaps that metric is somewhere in the pipeline, and this might just be a step on the way. I hope so.

    However, in the here and now one thing that I wish Garmin would add would be a sleeping equivalent of “MoveIQ” (this being how Connect figures out if we have been for a run / walk / swim without us having to press any buttons). They could call this “SleepIQ”.

    Let me explain. Like Ray, I prefer to sleep with my watch off. I like to leave it to charge by my bed. When my alarm goes off, I get out of bed and put my watch on. If I do that, then Connect doesn’t log my sleep, even though it “knows” that I haven’t had my watch on (no movement or HR input). It shows up as “unmeasurable” time. However this period of time is likely to be within 5-10 minutes of the actual amount of sleep that I had. If Connect recorded it as such, it would be automatically recording my total amount of sleep. For me, that is useful information.

    Although I am not a programmer, it really doesn’t seem like it should be too hard to update either the firmware or Connect so that a period of night-time when my watch isn’t moving and has no HR input is treated as “sleep”. This would record the most actionable metric, which is whether or not we are getting enough sleep.

    So, in the vain hope that someone from Garmin might be reading this, please could you create “SleepIQ”?

    • Paula

      While I agree that a feature along the line of your suggested SleepIQ should be easy enough to implement I don’t think that total sleep is the most actionable sleep related metric, or an actionable sleep metric at all.

      You say it yourself, whenever you get 8 hours sleep, theres still the question of whether that was a very restful sleep or just time that you spent with your eyes closed.

      On the other hand sleeping is the best moment to determine RHR which is a very sensitive and well researched data point that makes for a very actionable metric.

    • Jim Andrews

      Hi Paula

      Thanks for your response. I see from your other responses that you seen to know a great deal about what’s going on behind the scenes with this sort of watch data processing. I think it’s fascinating, and your insights are very helpful.

      On this subject, I get that RHR is a useful metric (I’ve been recording it manually for over 20 years) but that’s not what this is about.

      All I’m saying is that for those of us that don’t like to wear a watch on bed it would be nice to automate the recording of something approximating our sleep time, even if it’s a bit crude.

      As I understand it, my watch measures my RHR in the day anyway.

      My other point is that I don’t really understand what action I will take if I notice an upward or downward trend in my REM, even though it’s totally amazing that my watch can measure it (by proxy).

      Does high REM indicate anything? I know REM rises when we’re learning new motor skills, but I don’t see how it’s going to guide my training.

      Unless, of course, someone knows better in which case I’ll overcome my aversion to sleeping with a watch on!

    • Paula

      I would always give more weight to RHR measured during the night as RHR measured during day. If the two are the same for you (which is possible) then both measurements are viable. Normally however you would expect more factors to skew the picture during the day such as stress, caffeine intake, or the fact that you’re never really completely resting during the day (e.g. writing this actually increases my HR by about 8 bpm right now) many of which do not apply while sleeping.

      While a dramatic reduction of REM sleep would obviously indicate that there may be some serious medical issue (like sleep apnea) or a serious reduction in life circumstances (like babies) I absolutely agree with you that its very much unclear how this kind of info can get actionable when used in normal healthy subjects. You would think one had to establish some sort of baseline for your personal sleep characteristics and then make informed decisions on the basis of the very slight changes that can occur within the range of whats still physiological. What that may be I have no idea.

    • Jim Andrews

      Hi Paula, thanks for the response, very interesting.

      While accepting that the ideal is to monitor HR 24/7, I’d still be keen to see something like ‘SleepIQ* to simply record the fact that I’m probably asleep if it’s night time and my watch isn’t moving or recording HR. As others have noted, there seems to be no way to manually enter hours of sleep in to Connect, so this would be a helpful addition.

      On that note, I’m off to bed. Without my watch! 😉

  30. dan

    It seems like I am always forced into know what my sleep is or have a semi correct resting heart rate displayed.
    Should I wear my Garmin activity tracker all night then I get my sleep metrics but my RHR goes up substantially because I do not seem to be able to sleep for a uninterrupted 4 hour average of time with a RHR as low as during the day when I’m up and sitting at my desk or sitting around my house. Why they cannot just measure my RHR one way all day long is beyond me. Since a higher RHR 7 day average causes me more anxiety than knowing what my sleep is, I’ll continue to leave it lay on the nightstand at night, The moment someone comes up with a tracker that will do both accurately I’m a buyer.

  31. Nathan B

    Hey Ray,
    This is cool and along with True Up, has Garmin going in the right direction SW wise in my opinion.

    One question… You say Garmin have been “leveraging HRV data since 2017”. Do you know why when I select the ‘HRV Stress’ activity on my 935, it still tells me to use my HR strap?

    (then stand still for 3 minutes!)

    Other devices I’ve used recommend doing as waking up (Elite HRV & HRV4Training).

    • Paula

      Nathan, thats because optical HRV Data is too unreliable for usage. So whenever you want to use HRV Stress or use an HRV App on your phone you will not be able to use the optical heart rate sensor. What Firstbeat does is largely ignore that fact and do it anyway and then bake some guesswork into it. Which is why for many people stress levels vary across devices and across firmware versions.

    • More recent Garmin watches can use the HRV data for the Stress app, sans-strap.

      The reality is that since around mid-2017 optical HR sensors can do HRV at rest quite well. Not all companies and not all sensors, but a number of them. It’s doing HRV in a workout that is currently not available to anyone to my knowledge.

      There’s plenty of demo’s that Valencell has done that demonstrate this, stretching back 18-24 months in fact. FirstBeat also showed some of this data at their conference, I believe the presentations are now online too.

    • JimL

      So I noticed this too. On my 935 (Which says it has no new firmware updates via Garmin Connect), when I go to HRV on the device, it says I have to use the strap. However, I can use the calm app and use the stress functions, which use HRV.

      Also, I still don’t have the updated sleep platform. Mine still shows the old data.

  32. Neil Jones

    I notice that on Garmin Connect (not GCM), the key for the different sleep levels (accessed from the question mark at the top right of the sleep data) now shows a colour for manual entries, though I don’t yet see a means of actually doing this.

    However, assuming that your device and therefore GC has your HR/movement data logged, if you manually enter a sleep period, shouldn’t GC be able to retrospectively apply your sleep stages?

    Or is ‘manual’ just for people who take their devices off at night and therefore don’t record any data?

    ———————–

    As an aside, does anyone sometimes find that when they look in the morning, GC has accurately captured their bedtime, but then at some point later in the day it changes its mind and wrongly changes it to the start of their “couch time”?

  33. Shane

    Is there a rollout delay in Asia? Have a 935 with latest firmware and the latest iOS Garmin Connect version

  34. andrew jackson

    How much value is there in this data? I would love to see a post talking about sleep data. I have had an Apple Watch for years which lacks sleep data and never really felt like I was missing out.

    I get that Garmin is now using more advanced metrics like heart rate but I wonder about the accuracy based solely on movement since it is certainly possible to lay still while being perfectly awake.

    • I think it depends. I think everything is a bit of a two steps forward, one step back kinda thing.

      Each step forward (like leveraging HRV data to get better sleep start/stop times) is valuable, and I do find overall sleep tracking valuable. And I think there’s something potentially to be said for REM details once you can start tying that to training load and such.

  35. Kelly

    Hey Ray,
    Curious to see this. As others have pointed out, clinical sleep analysis relies on EEG for sleep stages, but if they can actually provide any validation of what they’re showing (and better yet, any interpretation, e.g. for planning training or other health analysis) this will be really interesting. I know there’s a bunch of sleep clinicians around who quietly roll their eyes about the data that some of these consumer products provide, but it’s clearly the new reality – and who’s to say that a made-up metric from a company might not end up being as useful as a metric that was made up by clinicians a few years earlier.
    If you’re back in Adelaide for TDU next year, and interested to do an actual clinical quality sleep study, let me know, it could possibly be arranged…
    (Also, will be interested to see the post about the data exports from the various platforms that you mention in the comments).

  36. Wim Boonstra

    Now that the Fenix 5 plus have been released and that embargo has lifted for you Ray, has Garmin provided information on when this sleep cycle functionality will be released for existing Fenix 5?

    Or will they not so they can sell more new Fenix 5 plus?

    • You can already use the beta firmware to get the sleep stats. The timing had nothing to do with the Fenix 5 Plus (since they were clear it’s coming shortly). In fact, it was planned to release the same day, but just got held up with some QA stuff.

    • Wim Boonstra

      Thanks, I’ll try that. I hadn’t seen the sleep notes in the firmware update notes, but I guess they don’t tell you everything 🙂

  37. Rui Pedro

    Hi Ray,

    Do you know if this new Advanced Sleep Monitoring feature in Garmin Connect has (or will have) any interaction with other analisys in GC, like Performance Stats? Sleep affects recovery and performance, so…

    Thanks!

  38. Gerry B

    Any idea if the wrist-based HR technology is any different in the Plus series? The wrist-based HR accuracy of my 5S is terrible, much worse than my old Vivosmart HD. To get meaningful HR data while exercising I use my Garmin Tri chest monitor. I was hoping that their wrist-based HR problems would be corrected with the next generation (Plus) watches.

  39. Tariq

    I wonder if they would make this work on Descent MK1, as of time of posting it doesn’t work on it yet.

  40. ice

    R.E.M. identify on FENIX 5 is coming with new firmware 9.20 (actually in roll out)
    Changes made from version 8.00 to 9.20:
    Added support for REM sleep phase detection and improved the accuracy of the existing sleep algorithm.

    • Rene

      Just installed 9.20 on my Fenix 5. TrueUp is working, as it now shows my Edge 820 rides on the Fenix. Nice feature! Been waiting for this from the beginning. I’m curious to the REM sleep. I’ll find out in a few hours 🙂

  41. lynn allen

    I’d love to know how the sleep data compare to the Fitbit Ionic. The Ionic seems strikingly accurate.

  42. Jaques

    Both my Fenix 5 and my Android app have updated. Since the update, my sleep no longer records any “Deep” sleep… only “Light”, “REM”, and “Awake”. I find this difficult to believe that I am only logging REM without any “Deep” sleep.

  43. Atle

    The REM detection is not working with D2 Charlie sw Version 4.20.0.0… yet…

  44. YellowknifeWarrior

    I have to say I’m a bit disappointed in the sleep tanning that Garmin has programmed here. First, the order of Awake, Deep, Light and REM makes no sense, as mentioned before. But more to the point, the watch does not seem to use its internal sensors to figure out what is going on. For instance, this morning I woke up at 0600, and walked around for a bit, sat on the couch for 15 minutes and then moved around some more. The watch decided that even though I was up and moving (as opposed to going for a quick bathroom break and returning to bed) that I was awake for 23 minutes and otherwise in REM sleep. Even the movement tab shows nearly continuous movement for the 52 minutes I was up between 0600 and what the watch perceived to be my waking time (0652).

    I hope Garmin is working on sorting things out. I get very different results on my sleep patterns, especially in terms of when I’m awake or in deep sleep, when I where my wife’s Fitbit Ionic.

  45. vojko

    hi
    …looks like I can’t get sleep data from my fenix 5x plus….and I,m not shure why…I wear the watch every night (for three days now)…but still no sleep data…??
    ….in garmin connect I just get the ” no sleep movement data” notice…and yes, I have been syncing the watch every day…
    …some help would be much apriciated….:)

  46. Am I understanding it right: this “REM” sleep feature will most likely never be available in the fenix 3hr?

    Thanks
    Guglielmo