BasementDwellingGeek wrote:According to Zeo, I had plenty of REM but no "deep sleep" recorded.
The bar graph would suggest that. But the pie chart says 4% and 17 minutes. The bar graph is based on the 5 minute data and the other on 30 second data?
Yes, the bar graph is aggregated data shown in 5-minute windows and the "time in stage" data is based on the 30-second epochs.
For each five minute window in the bar graph, the final assignment of "sleep Stage" is based on a very straightforward algorithm:
- If there is any WAKE scored during the 5-minute window, the bar for that window is labeled as WAKE. In other words, one 30-second epoch of WAKE causes the whole five minute window to show up as WAKE.
- If there is NO WAKE scored during the five minute window, the State assignment is done by "voting" with Winner-take-all. As an example, note that there are ten 30-second epochs that correspond to a particular 5-minute window. Let's suppose that the scoring in one five minute window goes something like this: 3 3 2 2 3 4 3 3 2 4. The five minute period in the bar graph would be labeled as LIGHT (3=Light in the Zeo) because there are five 3's, three 2's and two 4's recorded. But that minute-and-a-half of REM (the 2's) and that minute of DEEP (the 4's) are used in computing the "time in each sleep stage numbers.I'll be honest: I don't know what the zeo does whem there's a tie vote. My guess is that if LIGHT ties ith either DEEP or REM, that LIGHT wins. I have no idea who would be declared the winner of a tie between REM and DEEP.
And also keep in mind that the Zeo is only about 75% accurate according to its makers. (And given the cost of the device and the fact that it only has three rather closely spaced sensors, that's really quite good for accuracy!) And WAKE vs REM is particularly hard to distinguish. But it does mean that you have to use some common sense in interpreting zeo data.
As to how to sync the zeo dat and the SH data: JediMark allows import the thirty-second epoch data directly into SH. You first have to download the Zeo data to your harddrive as a file and then import into SH. Th zeo data uses the Zeo clock's time stamps and the cpap data uses the cpap's time stamp, so syncing the clocks is an issue---particularly for those of us with PR machines. If the clocks are reasonably close, it may just be easier to deal with the horizontal "shift" in the Zeo data the same way S9 users have to deal with shifted data in ResScan (and SH).
