Help with Zeo Bedside Unit

Hi,

I'm new to the forum. Also not sure if this is the best place to ask, but there is some activity with Zeo here. Since it's gone defunct I'm not sure where else to ask about such things.

I bought a used Zeo Bedside unit recently that was in good condition. First couple night's readings were suspect, so I found an unopened sensor on ebay and purchased it. Unfortunately I still got the same bogus readings.

Essentially I only get two levels, either REM or light sleep. I know it's faulty because I stare at it while I'm laying there and see if log me as REM sleep while I'm wide awake. Over the course of the night it essentially flips back and forth between these two sleep levels, in a set pattern.

I wanted to use it as a baseline for another device I bought, the ResMed S+. It's a neat device, but there's just no way to know exactly how accurate any of these things are short of a full sleep study (of which I've had several). Either way, Zeo seemed the best equipped to give a good baseline. Unfortunately, that's just not the case as these two are completely not on the same page.

So I have a new headband sensor which seemed to be the key piece that could cause faulty readings. I've reset the bedside until to factory conditions. Obviously fiddled with the headband and tried different positions with the same results. Is there anything else I'm missing here that could cause bad readings, or could I just have a bad bedside unit?

Any info would help. Thanks ahead of time.

Hi Slaphappy,
Review this thread: viewtopic.php?f=1&t=101917&p=955516#p953171
Please be aware that you can not determine the correctness of the bedside viewer display while looking at it. It does not record in real time. There is a least a five minute delay. Additonally, the bedside model does not start recording sleep until it senses that the wearer has entered an uninterrupted sleep stage of at least 10 minutes.
Normal sleep will be comprised mostly of REM and Light with some wake and some deep, in cycles lasting 90-120 minutes. It is not uncommon for those with disturbed sleep to see little or no deep. However, your Zeo should also be registering some wake, at least at the beginning and end of the night. See an example in the thread.
Are you using the monitor/viewer described in the thread above? If so, how about posting a screen shot of an individual night? Note that the Zeo Viewer reports in 30 second epochs while the Bedside display is in 5 minute epochs.

Thanks Jay... No, I'm not using it like the thread you mentioned illustrates.

Maybe I'm relying too much on the bedside display, but I had considered it more of a baseline test before taking the time to export the data for analysis. I can definitely see where there may be more data under the hood than the larger time slices displayed on the unit.

However I had tried to baseline it by wearing during a time I know I was awake, which was about 30 mns. Throughout that entire period it showed me as being in REM sleep on the unit, so I suspected the data was faulty. However there is the possibility that the data is good but not being displayed correctly.

Sounds like it's worth taking the time to try and export the data to look at it in more detail. As I mentioned this is a new process for me, so I've had a total of about 3-4 night's readings. The first 2 or 3 nights was just getting used to the headband. The last night I actually got most of the night with it before having to take it off, so may have some good data to look at.

Yep, you just can't do it while awake looking at the display. Even when you wake during the night and look at the display, because of the delay, it won't necessarily be correct. Once the headband is placed on the dock, resulting in an end of night signal, the data will be written to the card and the display will be fixed for that sleep period.

Yes, do download the viewer, trial the Bedside for a few nights sleeping and I think you'll be able to establish the baseline you're looking for.

As I mentioned earlier, the Bedside won't recognize the data as valid and begin recording until it has detected a sleep stage lasting at least 10 minutes which is uninterrupted by Wake or Undefined.

The 5 minute epochs of the Bedside display are essentially the average of 30 sec epochs contained within it, with the exception of Wake. If one of the 30 sec epochs is wake, the entire 5 minutes will displayed as Wake.

There are no docs with the viewer, but it's pretty easy. It is patterned after the original Zeo.
download and install the program.
Place the Zeo SD card containing sleep data (zeosleep.dat) in the PC card reader
Start the program, click on the SD card button (far top left) or, under File, select Open Dat, navigate to the location of the SD card and open the .DAT file.
The latest 30sec hypnogram will display, the 5 min hypnogram is not generated. Set the desired colors for different sleep stages under preferences.
Multiday displays can be selected from the tool bar.

I have downloaded the viewer and used it to view the data from the bedside unit. I'll post them besides the results from the s+ as a basis for comparison, not necessarily suggesting either is correct but this is what I see.

You can see that the Zeo data suggests that I immediately drop into REM sleep. Whereas the s+ shows me as awake during the same time period. Both change states around the same time dropping into light sleep. Another event around 3am has Zeo showing me in REM in s+ showing me awake for much of that stretch.





The second night was the one I originally referred to where I stayed awake for 30 mns just to watch it on the Zeo. It shows me dropping immediately to deep sleep and then to REM during the time I was awake. After a while I turned on the s+ and went to sleep.



In looking at these one would think they aren't even looking at the same event. None of these give a clear look, but I know the s+ at least shows some of awake event correctly because I noted them. The Zeo never shows a wake event.

So... I'm not sure who to believe at this point

Well, I see your point. Zeo's admitted shortcoming was distinguishing REM from Wake, but in my experience, it over scored wake. In your case, the opposite seems to be true. I don't know why that would be. I suppose it's possible, but probably unlikely, there is some type of electronic interference between the two systems (or anything else you might have in the room). Additionally, I have seen Zeo score Deep when I've worn the headband into the next room (I assume due to a reduced amplitude of signal). Could that be why yours started with Deep?

Just looking at the hypnograms, it looks like the Zeo overscores REM and the S+ underscores it. To me, the most "normal" hypnogram is that of the S+ for Jan 29, but it looks like it could have missed the first REM stage which should have occurred 0300. For 31 Jan, the S+ scores no REM until 0800, again, one would expect to see some around 0300, which the ZEO does indicate.

Your comparison here seems to be incomplete, you have two S+ hypnograms (29th&31st) which include about 8 hours of bedtime, but only one Zeo hypnogram(31st) for only 5 hours. On the 31st, it looks like Zeo starts around 12:30, but the S+ doesn't start until about 1:00. For future comparisons, you might consider cutting and pasting the two hypnograms into a draw program where the timeline can be adjusted to match. It would make comparison easier. See my signature link My Zeo for an example.

Now, I have seen a couple of other Zeo hypnograms where the user was able to "fool" it by lying quietly in bed awake at the beginning of the night watching the display and the Zeo thought the person was in REM. If you think about it, those stages could be pretty close with eye movement while awake mimicking REM.

Anyway, I think you should do some more comparisons over the same periods before deciding which instrument is the more accurate.

Synching the time lines, as I mentioned above, shows some similarity in the two for the 31st.
Deeps flanking some stage around 3. Zeo says REM, S+ says Wake.
Both Say something happened around 0430.
But, other than that...? My concern is the lack of Wake shown by the Zeo.
On the other hand, if you actually went to sleep around 0100, Zeo puts the first REM period about where it should be, about 2 hours in, the 2nd about 90 minutes later, and the 3rd about 60 minutes after that, which are typical periods for REM. The S+ doesn't show any REM until about 0815, 7 hours in, which would not be typical.
As I said above, I think you should continue the trials, try to wear the Zeo thru the night and don't put on the headband until you're ready to go to sleep.

Thanks for the breakdowns...very interesting!

Yes, I only made it halfway through the night with the headband, so that's why on those two nights it stops around 5am. Doesn't sound like there's any other adjusting to be done for the Zeo that may account for some of the weird readings. If electronic interference can cause problems then I may try a night or two where I remove the other device and see if the readings are different.

I've tried adjusting the band, different positions, etc, but the readings were similar no matter where it was on my forehead. Of course the sensors are new, so I can't improve there. I had read in other places where people tried electrodes, so if I can find some may try those just to rule out the interface as much as possible. I've been told I have a thick head so maybe it's causing problems with the sensors

Hopefully after watching over a period of time I can learn to read the tea leaves with the combination of devices giving a fuller picture. I wish there was a way to nail down the different states at least once for a point of reference, but I don't think other than these two devices (short of a sleep study) there's anything I can use at home that could provide that with some assurance. All the other devices seem to be purely based on movement, which I don't trust really. Although some do combine movement with heart rate so maybe could fill in some of the gaps and provide insight.

Slaphappy wrote:Yes, I only made it halfway through the night with the headband, so that's why on those two nights it stops around 5am.

Hi, you do realize the two Zeo postings are for the same night?

I have tried the electrodes, I don't know if you will have any better results with them, I didn't particularly care for them. Another device that could help monitor sleep is a pulse-oximeter. CMS50 models are the most popular here. You can search the forum for info, but probably the two most popular are the CMS50e and CMS50f. The cost is around $100.

I assume that you have not been diagnosed with a sleep disorder and therefore do not have a respiratory breathing machine (CPAP) which could also help in diagnostics? If you do think you have a sleep problem, feeling fatigued, not rested, etc., probably the best thing to do would be to get a lab sleep study (PSG) done after other factors are ruled out such as sleep hygiene, depression, deficiencies, etc.

I agree that the actigraphy devices are not likely to provide the results you're looking for. As far as Zeo goes, I think it is a good tool for establishing trends. It is probably the best consumer product, if you want to know something like, "is deep sleep increasing or decreasing". It's also probably the best at determining sleep stage at any given time as well, but it's not 100%, only something like 80, but over time that accuracy can provide good trend info. See here for links to accuracy assesments viewtopic.php?f=1&t=101917&p=955516#p955516

I suppose a thick head or skull could affect the results, I believe the algorithm does include amplitude of signal as a variable. You might try wearing the headband just over the eyebrows vs just under the hair line, I think those positions provide them greatest difference in amplitude. Unfortunately, there is no fine tuning adjustment available. I have found there seems to be a little break-in period with new sensors, maybe 3 or 4 days where the readings didn't seem to be consistent until settling down. It's possible, I suppose, the sensor cloth could have a little oil or something on it at first. It is permissible, by Zeo, to clean them by wiping the sensor cloth lightly with rubbing alcohol on a soft cloth. Lightly is the operative word.

If you'd like to post a few more nightly results, we might be able to gain a little more insight as to what's happening during your sleep.

I did not realize I had posted the same night twice...Doh! I have updated it and also tried a comparative view similar to what you provided below. It does make sense that the sensors may need some breaking in, so I will keep using them and trying the different positions you suggested.

Was not aware of the pulse-oximeter so will look into them. Sounds like another great data point to have displayed along with the other monitors.

Over the past 15 years I've had 3 separate sleep studies, done by 3 different doctors. All of the studies were similar in that I had a few dozen wake episodes each night. None of them were easily explained by physical problems - no signs of breathing irregularities or restless leg, etc. The first 2 doctors recommended CPAP machines, anyway, which I tried for a few weeks each time, neither of which helped. However I did a lot of work during this time to improve my sleep hygiene to rule all that out. Also had any psychological problems ruled out before starting this process. So since then...just lots of different pills.

One problem I do know of is REM disorder (which I didn't know about until after I was married). However I do find it interesting that Zeo picks up REM where s+ picks up wake and wonder if that may be due to movement while I'm dreaming. That's one of the things I'm trying to track down with all the monitoring. I take something for that, but over time anything that works eventually stops working.

Also, I have chronic migraine which I believe affects it somehow. Basically I have not slept well since I got my first migraine over 20 years ago. I don't think it's a coincidence both happened simultaneously, and somehow the migraines screws with the sleep process.

Either way, I'm hoping to baseline my sleep so that I can better rate the efficiency of the medication and other things during the day (such as when I exercise, etc). So whatever I use for monitoring doesn't have to be lab-study accurate as long as it's consistent.

Here are the corrected JAN29 readings. I'll download some more this weekend and see if there are any changes.




Comparison, lined up on the :45 mark...

The similarities and differences in the Zeo and the S+ are interesting. It does seem a good guess that the S+ is falsely scoring some REM as Wake while the Zeo is scoring some Wake as REM. This could be related to your disorder, but I don't know enough about it to guess. It is possible that Zeo is correctly reporting a shortened REM latency.
I have used the Zeo in the past for just the purpose you describe, monitoring the efficacy of medication. I felt it worked well for me. In my case, however, I was chiefly interested in tracking the trend of Deep.
See the links in my signature for another tool which you might be able to use - IR time lapse video. Many laptops now come with webcam time lapse available so special software might not be needed. Modifying a webcam for IR is easy.
I've got quite a bit scheduled in the next couple of weeks, but I'll be glad to help as I can.
Jay
Edit: Would it be possible to test your Zeo on another person to see if the results are similar?

That's a good suggestion. I could have my wife try it for a night and see if she gets similar readings.

Also like the idea of getting video to see what's going on during those times where the devices disagree, like watching the instant replay. I could re-purpose one of my security cameras since they are designed to work in low light.

Appreciate all the help, btw. The mystery of sleep has been an issue for a while, and trying to get answers can be quite the difficult challenge.

Slaphappy wrote: However I do find it interesting that Zeo picks up REM where s+ picks up wake and wonder if that may be due to movement while I'm dreaming.

Given the symptoms of REM disorder, that sounds plausible, or at least possible.

Check this out... Had my wife try the zeo for a night, and she got great results. Even showed when I came to bed a little before midnight and woke her out of deep sleep.

So something is obviously wrong with my environment, or my head. I'll try tonight to turn off the s+ and not keep my bluetooth phone there. I didn't think any of that would cause interference, but it's possible.

So, at least we know your Zeo will score wake. That is a big improvement. Except for that long period of wake around 0300, your wife's sleep looks pretty good. Was she aware of being awake then?
Electronic interference is certainly possible, though we would think unlikely. However, since ResMed acquired Zeo's assets and the S+ looks as though it was patterned after the original Zeo, it is possible the S+ uses a transmission and encoding method similar to that of the Zeo which might increase the likelihood of interference. The results of your test without the S+ and Bluetooth will be interesting.