Devices/programs to help evaluate sleep status

So here's another shot of the ResMed S+ Hypnogram compared to changes in Respiratory and Flow Rate graphics from SleepyHead. The Hypnogram correctly identifies a Wake period around 02:20 when I got up for a couple of minutes. Hypnogram REM periods coincide very nicely with Respiratory Rate changes we might think are REM when looking at the SleepyHead Graphic. Hypnogram Deep Periods generally occur during periods of calmer Respiratory Rates as expected. I do think the Hypnogram misses a period of REM around 23:30 where there is a fluctuation in Respiratory Rate. This could be due to the parameters of the S+ algorithm. The period might be too short, or too soon after sleep onset, etc. I have yet to find any significant detail regarding the algorithm and data display.


For a fairly decent overview of the state of consumer sleep monitors see:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4640400/
Nat Sci Sleep. 2015; 7: 147–157.
Published online 2015 Nov 5. doi:  10.2147/NSS.S94182
PMCID: PMC4640400

Consumer sleep monitors: is there a baby in the bathwater?

The rapid expansion of consumer sleep devices is outpacing the validation data necessary to assess the potential use of these devices in clinical and research settings...
We suggest that multistage validation efforts are crucially needed, from the level of sensor data and algorithm output, to extrapolations beyond healthy adults and into other populations and real-world environments.

Soothest Sleep wrote:Has anyone experience with the Sleepace RestOn device? Like the Withings Aura, there is a band that slips under the bedsheet; reports are generated through a phone/device app. A local dealer has them on sale, and I am tempted to purchase. I've read a mix of online reviews, but was wondering if anyone in this community has tried it.

Thanks,
Jean

Being a gadget freak, I have tried just about every one you can think of. The Reston quit working after about 2 weeks, connections broke in the sensor strap. I got another one, and I was VERY careful, but it quit also. Just got a Sleepscore Max put out by ResMed (replacement for the S+, I think). They claim to have done it in partnership with Dr. Oz. So fap\r, not impressed. It seems like a repackaged S+.

jsielke wrote:Being a gadget freak, I have tried just about every one you can think of. . .

Thank you for your comments about the Sleepace RestOn; they've helped me make my decision.

Jean

Recently I got a copy of the Costco magazine that included some information about two products related to this thread:
1) BodiMetrics O2 Vibe (retail about $200, Costco $145) https://www.costco.com/BodiMetrics-O2-V ... 77334.html Manufacturer's information https://bodimetrics.com/product/o2-vibe/
2) BodiMetrics Performance Monitor (retail $300 Costco $220) https://www.costco.com/BodiMetrics-Perf ... 77319.html Manufacturer's info
https://bodimetrics.com/product/bodimet ... e-monitor/

Both are wearable devices that monitor multiple data points for the user.
Vibe: A. Heart Activity, B. Systolic Blood Pressure, C. Blood Oxygenation, D. Temperature, E. RPP Trending Chart, G. Relaxation Index, plus a pedometer mode
PM: monitors the same set of data points and pedometer function as Vibe for up to 4 users (3 configurable and 1 guest). PM CAVEAT: DO NOT USE IF YOU HAVE AN IMPLANTED CARDIAC DEVICE, SUCH AS PACEMAKER AND/OR IMPLANTED CARDIO-DEFIBRILLATORS (ICDS).

Both are BlueTooth 4.0 enabled and readout through iOS and Adroid devices (compatiblity chart at website)
The most serious downside of the Vibe for me is the lack of Windows support - I don't have and don't particularly want a smartphone, iPad, or other iOS or Android device, aging Ludite that I am.

Has anyone here any experience with either of these?

Quick performance update about Fitbit and last night.
The darn cat sat on the machine at 4 AM...so it got turned off and woke me up and I had to swat the cat.
FitBit totally missed the known awake time at 4 AM.

My observation on FitBit Charge 2 from 2 months of daily usage.

- Fitbit has two modes - normal and sensitivity. Below feedback is based on 'normal' as I understand 'sensitive' mode throws charts off bit and I didn't dare to try
- I have compared and correlated the Fitbit recorded data with actual HD video recording, SH data and oximeter (pulse/heart rate in this context) readings for accuracy, so some attention to detail is involved

Positives
- Pretty much all rolling over and repositioning in the bed has been recorded as arousals.
- Sleep stages are estimated and may not be accurate but in my opinion still dependable rather than know nothing
- Sleep and wake up times are fairly accurate if not perfect and so are the sleep cycles
- Heart Rate is almost accurate and in-line with the pulse rate reading from oximeter

Negatives
- Sometimes if not most of the time, recorded data lags actual time by 1-2 minutes in recording. No big deal I guess
- Two movements within a 10 min (approx) interval is considered as awake time of the entire duration rather between those two movements rather than two separate awake events. Ex: you rollover at 11:31 once and another at 11:38 though you are asleep in between. Rather than two awake events one at 11:31 and 11:38 the awake duration is recorded as 7 minutes.
- Arousals rounded to 30 seconds so if you just rolled around for 4 seconds it is rounded to 30 and since these events tend to happen few times a hour the total awake time can falsely appear greater in %. I chose to ignore percent and just look at the whole number in this case
- I noticed when slept with arm (one that has fitbit on) on the chest, the chest (and consequently hand) movement caused by the breathing entirety is marked as awake even though one is clearly sleeping. So this could be few to tens of minutes. This has potential to throw the charts in terms of overall percentages of individual sleep cycles off limits. For me this has been more a rare situation fortunately. Not sure if this kind of recording is an anomaly or common occurrence but something to take into account especially the device appears to use solely on two parameters (HR and movement)
- Smart alarm feature is really silent vibration and couple repeats if we don't acknowledge the alarm. Many times than not I missed the alarm which I was happy about (assuming I was having good sleep ). So the feature helps but I consider that dumb given the name for the feature. There is some outcry in the Fitbit Charge 2 community forums on the absence of feature to wake one up when in light sleep stage as apparently Pebble and some other devices (which unfortunately have been discontinued) have feature to wake up you from light sleep instead of deep/REM cycles. I have been actively tracking that forum and for some reason FitBit company chose to not include that even though it's acquisition of Pebble had it.

Good to be aware of
- Doesn't correlate between N1 and N2 sleep stages and are combined into light sleep.
- If using Fitbit only for sleep purposes ensure to wear it for few mins before and after sleep so that FitBit's algorithm records it properly

Algorithm (consider a novice opinion pls)
I have to imagine that Fitbit has 2 parameters to rely on - movement and heart rate and I believe that's what they advertize as. I noticed (at least in my case) that deep sleep is marked by lack of body movement (no arousals) and a combination of regular heart bate. REM determined by the extreme irregularities of heart beat especially when the period is marked with higher than normal/average pulse rate events. Awake % is explained in the negative sections. Everything else is pretty much light sleep which correlates to smaller irregularities combined with or without arousals.

Overall with the quirks mentioned in the negatives above, I think it has been accurate to me in terms of sleep and wake times as I used to correlate that everyday with my SH data. It also helps me have a fair understanding on how I was able to sleep overall in a snapshot which is a whole lot better than nothing/guessing. It's not perfect but it does add value when combined with other data depending on what you use it for sleep tracking and especially true when one is trying to fine track or tune their therapy.

Last but not least and funny -Once I forgot the fitbit at my cousin's place (5 hrs away) and had it shipped to me. Over the 2 days UPS had it in transit (there is no option to shut the device off except for draining the battery), it recorded 9 hrs of sleep in the night and 3 hrs of sleep in the morning. Go figure!!

fog.apnea wrote:Fitbit has two modes - normal and sensitivity. Below feedback is based on 'normal' as I understand 'sensitive' mode throws charts off bit and I didn't dare to try

I didn't/haven't tried "sensitive" mode. In fact I forgot about it until you mentioned it again. Maybe I will try it if I don't forget again.

For the most part it is accurate enough to suit my needs. It's not perfect. Sometimes it misses known awake times which is the only real asleep vs awake that I can know for sure about. Not often though. Most of the time it seems to be fairly accurate. The sleep stages I take with a grain of salt since I can't prove or disprove any of that. Based on the percentages though...probably fairly accurate if I am getting the normal sleep stages which I expect I am.

Last night it showed 2 awake times between going to sleep and getting up...the first one I maybe remember but I thought it was later in the night but I didn't check the clock. The cat woke me up briefly and the only way I know the time is to check the clock on the FitBit but I didn't check the clock for that wake up. Second time was at 6:06 AM when I woke up and considered getting up and I did check the FitBit clock and decided to stay in bed and I must have fallen back to sleep because the next time I woke up it was daylight and 7:30 AM and I got up.
So for last night it does match my memory. Now of course we all know that what we might remember isn't necessarily all that happened but I don't really care to dig deeper. Not going to do the video or pulse ox thing because I just don't want or need to.

I do think that some (not necessarily all) of the "arousals" it reports are very brief wake ups that I do have sort of a memory of because I have some FitBit reports with no awake times reported and I know I remember a brief awakening or 2 or 3. I will almost always have a handful of those for any number of reasons.
Some are probably normal cycle awakenings and some are from the cat or dog and some from pain and turning over in bed.
So I have a vague memory of some awakenings but the FitBit might not show them except as the faint pink of the arousal thing.
I haven't had a night with no remembered brief awakenings at all in so long I can't recall when.
It's normal for me and as long as I don't remember a huge number of them then I count it as a good night.

All in all still happy with the FitBit...had a problem last week where it was showing going to sleep an hour later than I was (and I know for sure I didn't lay in bed for an hour before I went to sleep) but the awake time was correct. So the total time asleep was an hour short. Had a hard time figuring out how to fix it but a reset of the time zone seems to have fixed it.
The syncing issues I had...seems to be related to Windows and the pairing. If I don't let Windows step in and get involved it syncs up quite nicely but as soon as I let Windows "Pair" I run into trouble and can't even manually sync it. Today and yesterday it synced just fine and I didn't even have to do it manually and I still tell Windows to stay out of it when I get the pop up window asking me if I want Windows to Pair with the device. It's a bit annoying to have to close the box each time it syncs but a lot easier than having to into the settings and unpair every thing and all that.
Near as I can figure out it is just a problem with some Windows versions and I drew the short straw.

It's not perfect but close enough to give me a fairly accurate picture of how the night went.