Zeo Mobile or Zeo Bedside?

[davelikesbeer]

I'm considering getting Zeo. It seems like a cool toy especially for those with sleeping issues.

For those using what is now called Zeo Bedside, do you think you would have preferred Zeo Mobile had it been available when you bought Zeo?

The difference seems to be that for Mobile, the data is transmitted from the headband to your Android or iPhone and then can be uploaded directly from the phone to their site (no computer is required). I'm not sure if this is automatic.

I'm also not sure what happens to the data once it is uploaded.

I think I want the mobile version, but I'm concerned about the size of the app and the data. I'm already finding myself having to delete some apps in order to install others. However, the convenience of not having another device and not having to boot a computer to upload data sounds good. Not to mention the mobile version is $50 cheaper.

[Uncle_Bob]

From the Zeo web site i see they mention that the mobile version comes with a charge dock.
I'm confused, is the charge dock for your phone or for the headband?
I assume the communication between the headband and phone is via bluetooth, in which case i wonder if you could use the ipad?

At the moment I think i would go with the bedside version as its a proven product. The mobile one has not even been released yet.
Also i don't know if I'm always going to have my smartphone. If i break or lose a smartphone i normally punish myself and replace it with a bare bones old style phone for the duration of my contract.

[Lizistired]

Hi Dave, You've summed up all the issues. I guess it depends on how mobile you are. I look at my ResScan data on the computer anyway and usually overlay the Zeo data with it so I'm already on the computer. If you travel alot you would be able to use it on the road. But I don't use it every night anyway. I did at first, but now once or twice a week. I'm trying to limit the things that could interfere with my sleep.
Mobile... the price is right. But if I want, I can view the graph and statistics on the console without uploading to the website. It also has a smart alarm, designed to wake you in light sleep. I haven't had much luck with it since something else always wakes me up before I want to get up, but it seems to have potential.

Uncle Bob, the charging dock is for the Zeo. I don't know how much the app would drain your gadget. I can't really call mine a phone, since it sucks at that. I need a bigger nightstand!
I'm sure Jay will come on and give you his opinion.

[Jay Aitchsee]

Thanks for the intro, Liz.

Dave, I'm afraid you're going to have to call Zeo to get the answers to all your questions. The smartphone app looks cool, and for everyday use, I'm sure it would be fine. Probably better than the bedside unit for displays, etc. But, it's not clear to me if you can access the data by PC once it's collected by the smartphone. If not, it would be a deal breaker for me because I want to be able to "sync" that data with other sources like ResScan, Pulse-Ox, and video. Another shortcoming may be the inability to get at the data for your own use. With the Bedside unit, the data is recorded to a generic SD card and uploaded. You can download the data from site to get more detailed analysis then is provided by the site, there is alpha software available which allows you to read the card directly, and it's possible to obtain "raw" data directly from a serial port on the bedside unit to monitor brainwaves in real time (also requires alpha or beta software).
Now, I'm not so obsessed (yet) that the last two things would be a deal breaker for me, but as I said, not being able to easily integrate the data with my other sources would be.
So if you're a hacker, you probably want the bedside unit. If not, the phone will probably be more than adequate. Give 'em a call. They're nice folks.

Jay

[davelikesbeer]

I am a hacker, and I would like to have access to the data. It appears they already have python libraries to read the data from the SD card, http://zeorawdata.sourceforge.net/ . Which is cool.

I imagine a company that does this is also one that would use the same write format on their mobile app. But you are correct, the only way to know for sure is to call them.

Jay, can you tell me approximately how much data is written for one night?

IIRC, your goal is to be able to build a home sleep lab for under $1000. Assuming you design a whole system that works, it's really sad that it can only be used for hobby purposes. It does seem you could piece together all the parts for someone to do a self home study that would do a good job diagnosing...however, what could one do with that information?

[RandyJ]

I'm intrigued by the Zeo but I'm not convinced that an inexpensive device like the Zeo can correctly discern sleep phases etc. Has anyone read any independent analysis (not Zeo's own literature) that attests to its accuracy?

If so, could you post a link to that material here?

Thanks a lot!

[Jay Aitchsee]

Dave, I don't know how much data is written each night (I've not tried to measure it), but I think I know this: their sample rate is every 10 seconds, their write rate is every 30 seconds, so for an 8 hour night there are only 8*60*2=960 data points. My spread sheet with 77 days data is about 35 KB. Zeo infers the 2GB SD card might slow from starting to fill after 3 months or so. So for the amount of data written per night in MB, I think it's pretty small. One more thing to ask them when you call.

Randy, Zeo doesn't claim to be 100% accurate. In absolute terms, they claim a PSG is about 85% accurate. In those terms, they claim to be 75% accurate or, as they put it, 88% as accurate as a PSG. I think I have stated their claims correctly. If you have a Zeo, you can catch it in a lie. It might say you are asleep when you know you are awake, etc. Some of that comes simply from the way the data is presented. For example, the primary graph is in 5 minute intervals which is made up of 30 second data points, which were produced by 10 sec samplings. This information is smoothed and averaged in the final presentation. So, you could have been asleep 3 minutes ago, been awake for a minute or so and the graph might report you as asleep when you look at it. Now, if you need to know what you were doing every 30 seconds, you can download the data and look at it. In my opinion, it is accurate enough to provide a reasonable picture of nightly sleep stages. I think the value lies not in a single night's accuracy, but in looking at trends over several nights. For example, while taking a drug for sleep, I feel worse and looking at Zeo over time I see increased awakenings. Or, looking at when my apneas occur with ResScan compared to Zeo indicates that my apneas rarely occur during REM. That kind of thing. I will bet, though, that the results of 5 nights at home on a Zeo will produce more accurate information about a persons average sleep than 1 night at a sleep lab. That said, my Zeo reports look very similar (almost identical) to my PSG's, but I don't have the independent link you're looking for, I'm not sure who would fund it - a competitor? Boy, do I sound like I'm selling these things or what? Randy, all I can tell you is that I was very skeptical too, but I've found the Zeo to be a useful tool.

With my sleep Lab for under a $1000, I can work with my sleep doctor to find the reason(s) for my non-restorative sleep (I hope).

Jay

[RandyJ]

Thanks Jay.

I guess I was wondering if any independent groups (like Consumer Reports) had reviewed/tested the Zeo to see if it can do what it says, ie can it really read brain waves and interpret them like a PSG. When it says you were in REM, can you really trust it?

For example, I was talking about a wristwatch that measures pulse via EKG with a cardiologist and he told me that he had tested one model against the readings that a real EKG machine was giving in the hospital, and it was spot on accurate. I was wondering if any testing groups had tried out the Zeo in a similar independent testing situation.

From what you wrote, I guess you can trust it somewhat. If it's useful enough to compare with Rescan data and you can draw useful conclusions, then I guess it is worth its sale price.

[Lizistired]

Randy, once I get to sleep the Zeo seems pretty consistant. The REM is spot on to the breathing patterns I had already identified, and the deep sleep patterns make sense. There are posting of results in Jay's $1000 thread where we overlayed Zeo, oximeter and Rescan data. There are times when I wake up that it doesn't record as awake, but it has only happened once that it was of any duration. I go back to sleep pretty fast when I turn over or whatever.

[Jay Aitchsee]

Dave, last night the zeo.dat file uploaded was 507KB. Now, I think that should be the entire .dat file (78 days) on the card. Yes, I just looked at the card properties and it shows 544KB used.

Randy, I understand your question. I haven't seen any independent tests - other than those backed by Zeo. I know it isn't 100% accurate, but it does correlate very nicely with my other data sources - ResScan and Video. Is every instance reported by Zeo correct? No, but I do think it gets it right most of the time. It is consistent, and my sleep patterns as reported make sense. But it's just a tool. For years, I've kept a sleep log, entered data on a spreadsheet and charted it. Now, the Zeo has become my log and chart. I'm interested in things like: I've exercised every day for 3 weeks-how has that effected my sleep? Do I feel better? Has my deep sleep improved? What is the trend? Or, I'm taking a new drug-what is it doing to my sleep? Do I have more or fewer awakenings? How do I feel? The zeo helps me objectively answer some of those questions and provides a place to log variables, and results with graphic representations. It does not tell me why I don't get any deep sleep. It only confirms that I don't (as likewise confirmed by PSG). The biggest shortcoming, in my opinion, is its inability to measure micro aroursals. Can't tell for sure if that apnea or leg jerk actually disturbes sleep.

My sleep Dr. said she was familiar with Zeo, didn't say she studied it or offer an opion on its accuracy, but we use the results from Zeo and ResScan to determine therapies.

So for me, I think it was worth the price and I'd buy it again. For others, I'm sure that would not be so. You can review these threads for some of the published chart work. viewtopic.php?f=1&t=68747&st=0&sk=t&sd=a#p638020, viewtopic.php?f=1&t=68424&st=0&sk=t&sd=a#p635357

Jay

P.S. if i find that independent study, I'll be sure to post it

[Jay Aitchsee]

Here's what I call my summary sheet from last night. Notice the awake time around 3 and the rescan pressure and leaks at that time, that;s when I got up. The multi color bar above the sleep stages is from my video - pink is up, green is left side, blue is back, and red would be right side.

[davelikesbeer]

Thanks for the data Jay!

Last night I thought of a couple of more questions. The first is how does their mobile app manage power. There are quite a few apps on my Nexus One that heat the phone up quite a bit (GPS, listening to music). I honestly don't know if there is a possibility of damage to the device with repeated uses, but it is a something for me to consider. This is a huge unknown given the mobile version hasn't shipped and there are no reviews.

Another question for the bedside version (and mobile for that matter) is the how bright is the display at night?? I'm very sensitive to even the stupid "humidifier is connected led" on the PRS1. Are there options to dim or black out the display?

Finally, their site suggests replacing the headband every 90 days. Is this realistic or can you get alot of more time out of them than three months?

And for Jay, I know you use a camera to record yourself sleeping. How do you take that data and turn it into sleep position?

Thanks,
Dave.

[Jay Aitchsee]

Dave,

Don't know about phone pwr/usage

The bedside unit is dimmable in steps to off.

Sleep position: I video record with a time-lapse set at 30 sec capture rate with time stamp. Playback is about 25 frames per sec. so that yields about a 30-40 sec video. I've got a spread sheet I use as a template with each cell the equivalent of 5min (corresponds to Zeo). Hours are indicated by the yellow background columns. I just watch the video, note the time of each change on the spreadsheet, fill the background accordingly, snip it and add it to the drawing making up my "summary sheet" . This snip below shows the last two nights. Using the 5 min time blocks allows me to easily align ResScan, Zeo, and position.
When I start my my summary sheet, I use a template that already has the dashed line hour markers that you see in the one that's finished. Then all i have to do is snip and size the object to fit appropriately and everything is synched. Of course, occasionally I have to reset the times of the Autoset and Zeo so they don't drift too far away from the computer. (see here to build infrared camera/light source: viewtopic.php?f=1&t=65101&st=0&sk=t&sd= ... ed#p607701)

Jay

[M.D.Hosehead]

RandyJ, I suspect the detection of REM isn't very difficult. The Zeo electrodes are right above the eyes. Eye movements create very strong signals, measured in millivolts (EEG is in microvolts) so REM should be easy to detect. The large, slow delta waves of deep sleep are fairly unique, too. I suspect telling the shifts between NREM stages and between awake and light NREM is more error-prone.

Just speaking theoretically. I don't know anything about the ZEO.

[davelikesbeer]

Dave,

...
Of course, occasionally I have to reset the times of the Autoset and Zeo so they don't drift too far away from the computer.
...

Bummer. It's been years since I've set a clock of consequence. I guess the mobile version would be better off in this sense.

My PRS1 gives no notion whatsoever of the time.

What about the headbands? How often do you replace them?