Numbers Ain't Everything

Starlette wrote:How are you able to calculate the duration of your events?

Hi Starlette!
I use SleepyHead and they are listed on the Daily tab, on the left side, in a tab called "Events". It give each one's time and duration.

I use ResScan, it counts for me and notes the duration on the event flags. I still love ResScan.

MaxDarkside wrote:Here's my charts from last night, including the brainwaves that you will see above the Zeo sleep stages.

I do not believe those are "brain waves" per se-- rather, a "power spectrogram" of the EEG frequencies.

There's the SOREM again.

TS, given your lack of distinct signs and symptoms of the N. tetrad, I do not believe that to be SOREM. Similarly, given the lack of theta power spectrum correlation of REM scoring, as well as the suggestion from the patent that Zeo uses 2 additional scoring channels, it seems clear to me that Zeo sleep staging does not live on EEG alone.

TS2, it would seem that Zeo would have put some more SWS in there at the end of the night given the delta PS% there.

TS3, that would be an unusual place for SWS (the sleep stage, not the person).

A mystery? Given the apparent accuracy of Zeo in the face of sleep fragmentation, probably not.

I mean, if at least a third of that is ambiguous at best...

deltadave wrote:...that would be an unusual place for SWS (the sleep stage, not the person)...

Although upon further reflection, that would, in fact, be an even MORE unusual place for SWS (the person) to be.

deltadave wrote:

deltadave wrote:...that would be an unusual place for SWS (the sleep stage, not the person)...

Although upon further reflection, that would, in fact, be an even MORE unusual place for SWS (the person) to be.

Not that there's anything wrong with that...

MaxDarkside wrote:

Starlette wrote:How are you able to calculate the duration of your events?

Hi Starlette!
I use SleepyHead and they are listed on the Daily tab, on the left side, in a tab called "Events". It give each one's time and duration.

Good morning.

@ Liz and Max - Going back to my question stated above, I've now attached a copy of last night's data. Confirming I understand you correctly.
The first CA at 9:52pm lasted 21 seconds. Is that correct?

Starlette

Starlette wrote:@ Liz and Max - Going back to my question stated above, I've now attached a copy of last night's data. Confirming I understand you correctly.
The first CA at 9:52pm lasted 21 seconds. Is that correct?

See the (9)... that's the duration. It lasted 9 seconds. The 21 is in reference to the time it occurred... It occurred at 9:52:21 PM. (the 21 is the seconds of the time).

Now I have the CMS 50E pulse oximeter data in our software along with the Zeo data in raw form aligned in time (well, probably close enough). The next step will be to get the ResMed data in there, but I can get a lot of information about what is going on in my sleep (stages, arousals, desats, etc.) with just the pulse-ox and Zeo. (Note; this is not one of my better nights. Woke up a lot, some leaks singing to me, I felt like I didn't need to sleep much, etc. When I woke up at the end, well, I just got up after only about 5 hrs sleep.)

Ok, thanks Max

Starlette

Well IYAM, based on the cardiopulmonary physiologic data and power spectra, all these areas looks like Wake to me:



I mean, there can be a little Heart Rate Variability in REM (if it's Phasic)(vs Tonic) but that's like insane.

Can you upload the Zeo raw data file somewhere (although unless the actual EEG channel is in there this will not be helpful)(unless we can reverse engineer a power spectrogram back into an EEG)?

Do you have your NPSG hypnograms?

OK, I think I got it.

They use the gamma wavelength to create a surrogate EMG to help stage sleep, as shown in this example of an NSPG with hypnogram plotted against a frontal lead (similar to Zeo dry sensor location) filtered to show gamma (everything over ~30Hz):



As you can see, this methodology is efficient at determining Wake (light blue thin line), but of no use in differentiating REM (forest green... errr, "weight-challenged" line)(certainly in this case anyway).

Clearly, they must be creating an eye channel. More fiddlement to follow...

deltadave wrote:Well IYAM, based on the cardiopulmonary physiologic data and power spectra, all these areas looks like Wake to me:

While I do wake from time to time (you can see where I switched off and on the machine during one of those, an obvious indication of being awake) I also do remember dreaming during some of those as well. It may be that my REM is very close to Wake, personally.

I mean, there can be a little Heart Rate Variability in REM (if it's Phasic)(vs Tonic) but that's like insane.

Who says I'm not insane! Bwhahaaaa....

Can you upload the Zeo raw data file somewhere (although unless the actual EEG channel is in there this will not be helpful)(unless we can reverse engineer a power spectrogram back into an EEG)?

I did not record the raw brainwave during Jan 18/19 nor last night, just to save disk space. I can turn on its recording next time, if I like. it's a LOT of data points. However I do have all the alpha, beta, delta, theta, gamma, etc. and pulse oximetry that I could dump to CSV.

Do you have your NPSG hypnograms?

Just on one page, poorly copied, paper, with no easy way to scan at the moment. I can tell you 1) I was in every stage multiple times per hour (it was a train wreck w/o CPAP) and 2) when REM is marked, while I don't have pulse on my copy, my spO2 was wild, mostly desating into the upper 80s and spiked once to 79%. That is to say, during REM, what you see is probably congruent with the NPSG.

Clearly, they must be creating an eye channel. More fiddlement to follow...

Could be. When I simulate REM while awake, closing my eyes and jiggling them all around, the Zeo has no problem picking that up, because of all the muscular activity. I read in someone else's post about how the Zeo should have accelerometers to detect eye movement, but I can tell you they in no way need them. It is very clear when the eyes are moving, particularly left-right, since the sensor is just above them. Up/down shows, but is quite weak.

Here's last night's. I slept better. The Jan 18-19 night was not good for me.
Apologies to Liz for hijacking her thread. Liz, I hope you don't mind but you are right, it's more than just the "numbers" (AHI, being one. You can have a poor AHI and a great night, or the other way around)... so much more. Sleep "stages" are an artificial construct to try to put something continuously valued into discrete categories so it can be "counted" and compared. That is one reason people, machines, clinics, etc. have troubles doing it. It's all about feeling great, OK, maybe "good" at least. That is what I'm after, regardless of how much in one stage or another I spent or what my AHI was.

Zeo seems like it records more data ( sleep stages) than Sleepy Head. Is it free like sleepy head or does it have to be purchased?

cindjo717 wrote:Zeo seems like it records more data ( sleep stages) than Sleepy Head. Is it free like sleepy head or does it have to be purchased?

The Zeo is purchased hardware and software. I got the Zeo Bedside for about $122 on Amazon (price varies I notice, I got a decent price). They have "mobile" and "bedside" units which you can use to create the bar chart at the bottom of my image. I got the Bedside unit because it has a serial port. But they don't provide a cable, so I had to order parts from an electronics supply company and make it myself. It was not hard, but not a trivial task either. Once I had the cable, then I could use ZeoScope software (free like Sleepyhead) to view brainwaves in real-time, streaming out of the Zeo Bedside. To get the chart you see (half-way down my image) we had to write software here to get the data into our software for charting and analysis. We also wrote an importer for *.spor files from the CMS50E pulse oximeter so we could synchronize the data on a time-basis. Next step for me is to maybe get the ResScan data into our software as well. Then we have all the data in an advanced analytical "engine" we have created over the last 20 years, which we typically sell to companies for data analysis, mathematical modeling, prediction / estimation of stuff and finally optimization. My quest is to quantify factors that influence my sleep quality, then optimize them to maximize it.