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

Sleep Tracking Issues

Thought I would just add my experiences with the first couple of weeks of sleep tracking with my 245.  I see lots of threads complaining about the Garmin sleep tracking accuracy and capabilities so I am getting the impression that it is not really very accurate.  But just in case anyone has some suggestions as to how I might improve my sleep tracking, here are my observations and issues:

It seems the device is depending almost entirely on movement and these first few points/questions have to do with that.  I guess it makes sense that movement would be the only way for Garmin to determine sleep as it is only strapped to your wrist but that lead to accuracy issues for other people too?  The 245 it seems to start saying I am asleep when I am reading in bed for an hour or so before actually going to sleep.  Also in the morning it always shows many, sometimes over 100, steps, even if I have not left the bed all night.  I do tend to sleep on my stomach with my arms up under my pillow so the watch may move around a bit, even when I am sleeping.  It does however seem to accurately show if I actually awake during the night.  This I know from remembering that I woke up once or twice, and then seeing that reflected accurately on the timeline.

Another observation is that I do not ever wake unrefreshed and I usually sleep 7-9 hours.  I am never tired during the day.  I have never been told that I am a restless sleeper or that I snore or breathe loudly during sleep.  Never been told I sleep anything other than peacefully with even breathing.  However, the watch almost never shows deep sleep and when it does it is only a couple or a few minutes.  The deep sleep it might report also comes several hours into the night, rather than near the beginning of the night where it should normally appear.  The device shows mainly light sleep, interspersed by many (8 to 14) shorter periods of REM.

Lastly the Pulse Ox and respiration is all over the place.  Pulse Ox averages in the low 90% range but at times drops briefly to low 80s.  Once it single drop to 78%.  Should I discount that monitoring as inaccurate?  

The Pulse Ox readings may be related to respiration?  Generally my sleep respiration is almost the same as my waking resting respiration, 12 to 14 breaths per minute.  However during sleep the 245 registers several deep down spikes to as low as 7 breaths per minute.

How does all this compare to what others are seeing?  If my arms do move a bit under my pillow or the weight of my body and head compresses my wrist and the watch, might that be why I register almost no deep sleep and so many distinct, short periods of REM?

Edit:  Talked to Garmin and because I sleep with arms/wrists, under my pillow and head, the added pressure to the watch is very likely making the Pulse Ox inaccurate as that sensor is very affected by pressure of skin contact. This could also be impacting accuracy of respiration and heart rate while sleeping as the added pressure between the watch and wrist can impact the HR sensor also and that in turn affects how respiration is calculated.  The Garmin rep also stated that the phases of sleep could be accurate or inaccurate depending on wrist location and how much movement when sleeping.  He felt that me seeing almost no deep sleep could certainly be incorrect given that I am never tired during the day and my sleep is reported as quiet and sound.  Basically if you sleep on your side or back with your wrist free and clear and do not move your arm much during sleep, it may be fairly accurate.  Otherwise, probably not so much.

So the sleep tracking features seem really to be more of a curiosity and fun thing to monitor rather than anything really that accurate at all times and for all people.  If I had any perceived sleep issues of my own, I would certainly not depend on one of these watches for accurate sleep information.  However for ongoing heart rate, averages, maximums, etc. the watch seems very good.  And I love the other tracking features, activity recording, and notifications.

  • I looked at my daily summary of respiration and it is off even during the day when awake and moving around.  There were a few times today when I was working in the yard when my connect data says my respiration rate was at 16 brpm, then dropped for a couple minutes to 7 brpm, then went back to 17 brpm.  I know for sure that was not the case so I think it is not accurate during sleep either.  I mainly purchased this watch for activity tracking, heart rate monitoring and the Connect features and those do seem to work well. At first I was excited to analyse sleep but now I don't think that function is all that accurate or helpful.

  • timely article:  ft.com article on sleep monitoring

    a few excerpts:

    But there is a more fundamental problem raised by the notion of tracking your sleep. Sleep is not like other areas of our health.  You might say that it is the quantum physics of our lives: the simple act of measurement alters what it is we are measuring.

    Focus on your sleep too much, and you will find it more difficult.

    If the “gold standard” technique (EEG) of measuring sleep has these limitations,
    what hope is there for devices that simply measure movement?

    do I think that you should invest in a sleep tracker? The answer is a resounding no.


    Guy Leschziner is a consultant neurologist and author of ‘The Secret World of Sleep: Journeys through the Nocturnal Mind’ (Simon & Schuster)

  • Thanks for the excerpt.  Would be an interesting article to read I'm sure. Would need to be a serious news junkie to spring for a digital only $360.00 per year subscription though.  Was tempted to get the $1.00 four week trial, but sometimes those are difficult to stop from converting to a regular subscription!

    I do agree that movement measuring has a large chance of error.  I am a good example of that because I sleep on my stomach and consequently move my arms around quite a bit even though my body pretty much stays put.  I am sure my need to occasionally reposition my arms because they are under the weight of my head and pillow is why when I wake the watch says I have taken 15 to 50 or so steps during the night. (Yes, I am sure I don't sleep walk!) 

  • Garmin sleep tracking is useless. (former owner of a Fenix 5S Plus and Venu). #worstsleeptrackinginthebusiness

  • Maybe not quite useless but not very good.  Since the original post I have become convinced that movement is pretty much all the watch is detecting to tell if you are sleeping or not.  

    First it takes note of the time you have entered for your normal sleep times.  If you go to bed 2 hours later than what you indicated is normal, then during that two hours all sorts of things can happen.  For example if you are watching TV and not moving, then it will indicate you are asleep. Then get up to take the dog out and you are shown as having woken up.  I find for this the best thing to do is the next morning, edit your go to bed time to cut out the 2 hours you stayed up later.

    The same thing happens if you are in bed reading after your recorded normal go to bed time.  It will show you asleep even though you are wide awake reading.  

    For me, I think I move my arm quite a bit while sleeping because the watch shows me awake many times during the night when I am not.  

    I also tried a few tests.  A couple of times when I did wake up during the night, I got up and moved a bit, then quietly lay in bed awake for 10 minutes or so then got up again for a moment.  In the morning Garmin showed me awake for a couple of minutes, but then asleep for several minutes, then awake again.

    Not sure how Garmin is trying to detect deep and REM sleep but there is no pattern to it that I can see. It does not seem to parallel respiration, heartbeat, movement or the pulse ox readings.  I do think that often the REM detection is correct though if it correlates with dreaming.  Often I will see REM sleep indicated for a while before I wake up in the morning and I know I was dreaming just before I woke up.

    It is all very interesting but I am pretty sure it is not that accurate and should be taken as entertainment not science.  I know if I felt I had sleep issues, I would not depend on this watch to give me any valuable information.  I love a lot about the rest of the watch though!

  • The current state of Garmins sleep tracking is so useless they might as well just turn the feature off to save battery. Garmin seemed to think walking nearby the bedroom after 11pm meant "he's probably asleep now". Often the analysis said I was in REM sleep while watching tv or playing some game on my pc. Got tired starting every day editing the sleep data.

    On weekends when I stay up later than usual I tend to get around 14h of sleep according to Garmin...YEY! Too bad the truth was around 8-9h so the watch/Garmin connect only got it wrong by 5h.

    Fitbit, Samsung, Huawei, Suunto, Polar etc have WAY WAY WAY better sleep tracking than Garmin. I now have the Polar Vantage V after testing most of the flagships.

    Like you said I also liked a lot about the Garmin watches, but the ridiculous sleep tracking was too annoying. Also the strength workout where it supposedly auto count your reps and recognise what workout your doing is just a mess and not working. 

  • I think this may be mentioned somewhere above, but maybe it was in another thread....  Garmin seems to use some software or a third party to tweak the sleep information every morning.  My information syncs through Connect upon awakening every day.  If I go to Connect and look at it within 30 to 60 minutes or awakening, the information is often crazy inaccurate.  Other times it less off, but still not good.  Always within an hour or so, the information is significantly changed and better reflects what sleeping I think I actually did.  So somehow Garmin or a 3rd party is analyzing the sleep data, making changes and updating Connect.  Maybe this daily revision process will evolve to even more accurate data?  Or maybe not as the data does seem to be mainly limited to movement and people move their wrists in many different ways when sleeping.  

    As I stated before, my watch often shows 100 or more steps when I awake.  So my wrist movement must be significant for the watch to think I took 100 steps when I didn't even get out of bed. Makes sense that would also affect Garmin's assessment of how much I was awake and how little deep sleep I got during the night.

  • Coming from a Fitbit Charge 3, the Garmin sleep tracking is a joke in comparison. I sleep on the side with my arm free (i.e. not under the pillow) and I have the same issues. I get virtually zero deep sleep registered, whereas the Fitbit registered at least about an hour every night. I don't know how they manage to get such poor analysis in 2020, it makes me question whether I can trust any of their other data, such as aerobic training effect etc. It's surprising to me that a serious company like Garmin would overlook sleep tracking, especially given their investments in Body Battery, Stress levels etc - all of which depends so much on the quality of your sleep. 

  • Couldn't have said it better myself. Was shocking to experience the useless sleep tracking Garmin offers. The most expensive watches in the industry and they can't even get sleep tracking and sleep analysis right. Makes you question if all the other metrics is pure bs too.. 

  • I agree. It's terrible.