To Sleep-Perchance To Remember
For about the past 10 years, Ariel was consistently and chronically fatigued and tired and I really didn’t know why. Ariel went to many doctors and got many diagnoses and a lot of medications and nothing really seemed to be helping. Ariel then went into a sleep study, a polysomnography, and it came out that she had sleep apnea with zero percent deep sleep and would waken up about 10 times an hour. According to the deep sleep, her 30 years old is actually a 60 years old.
CPAP machine | zeo
So then I went in for a sleep study, a polysomnography, and it came out that I had sleep apnea with zero percent deep sleep and would waken up about 10 times an hour. So what is sleep apnea? Sleep apnea is the cessation of breathing due to airway blockage or constriction while sleeping and it’s very disruptive to sleep quality.
So the treatment is continuous positive airway pressure otherwise known as the CPAP machine and at least for me it’s very uncomfortable, and even although it solved some issues of the sleep apnea it just creates its own disruption to my sleep anyway, so it didn’t really help that much.
Anyway, so I went for another device, a dental appliance, which the same basic principle. It tries to hold ones airway open during sleep when you lose muscle tone. That also didn’t seem to work very well.
So I’ve been tracking my sleep with a Zeo and I’m sure everyone’s familiar with this. So one of the first things I found was my sleep is very fragmented. The Zeo confirmed it and I already guessed that. So I wanted to look at my sleep cycles on a per night basis. And for most individuals there’s well defined peaks and valleys to their sleep cycle when different stages of sleep are more likely to occur. My own sleep cycle is significantly flattened, so for instance I’ll have any stage of sleep at almost any time.
So that was interesting to but nothing much I could do with it per se, but it’s a nice bit of data. Then using that same thing I looked at the length of my sleep cycle, which for most people is about 90 minutes. For me it turns out to be 100 minutes, which is within the normal range and nothing really special about that except for the fact that it is very flattened in general.
So monitoring my deep sleep over the course of treatment there has been very little change. According to Zeo I get about 43 minutes a night and of course on a nightly basis there’s a huge amount of variation. But overall the average has stayed remarkably constant even with the CPAP and the dental appliance treatments. So it has gone from 7.9 to 8.1% over about a year, and normal is about 17 for my age.
The same thing with REM, so my REM I realized was much higher according to Zeo again much higher than normal, but also has not changed in the past 450 days.
Then looking over a period now of weeks, days, and months, I see that my deep sleep kind of cycles every 70 days, the amount of deep sleep, and the amount of REM every 35, and awakenings about every like basically every two weeks. So I’m not really sure what that means, or what the significance is but I’m going to look more into that, these sorts of you know weekly and monthly long cycles.
Comparing Zeo versus CPAP, so my CPAP also collects data, and on this particular night my CPAP told me that I had about 22 awakenings. Whereas the Zeo only picked up eight of those, so there’s a little bit of a discrepancy there. Now the AHI, which is like the little orange bar is the Apnea Hypopnea Index and that’s basically how many times you stop breathing a night.
So then seeing if I can use my Zeo data to predict my CPAP data and the correlation isn’t really that strong. So predicting like the AHI, the Apnea Hypopnea Index, based on the Zeo awakenings per hour only gives you R squared of 0.04, so there’s really not that strong of a correlation between the Zeo data and the CPAP data.
Also, the thing that’s kind of most important to me is my brain functioning, so I’m using a brain workshop which is an N-N back to test my brain functioning in some way , and here’s the difference correlations between various data with the brain workshop.
So the best predictor of my brain workshop score is actually the Zeo deep sleep time with R square of 0.2. And what I thought what would be a great predictor is the AHI, but it turned out to have very low correlation with my actual brain workshop score. So the Zeo deep sleep was in particular by far the best predictor of that.
I’ve also tried different pre-sleep interventions. Mostly interested in how I can increase my amount of deep sleep and it’s kind of hard to see. So I found that the results mirrored what’s found in medical literature for certain medications. And one thing that I did find interesting is that melatonin is actually associated with a lower amount of deep sleep. I mean it helps me get to sleep, but once I am asleep the quality of sleep might be reduced.
This night I removed my CPAP machine you know at the beginning of the night just to see how many awakenings the Zeo would pick up under those circumstances and it only go like maybe 10 or so. Whereas according to my sleep study it should have been closer to 10 per hour.
Okay, so these are the normal values. I’m a 30 year old so I should fit in that category and get about 70 minutes of sleep, but according to my deep sleep I’m actually a 60 year old. And my ZQ of 93 is very high as far as these things go, and I don’t think it reflects my personal experience of sleep quality accurately at all.
So the ZQ in general, it doesn’t seem to correlate that well with what my CPAP tells me and what I personally experience objectively, but it does seem to predict the brain workshop score so it’s good at that. And then okay, this is how the ZQ score is correlated with different components of sleep. So total sleep seems to be the largest contributor to ZQ score at least for me, and in future I want to try other biometrics, oxygenation, actigraphy, breathing, and maybe do some more with the raw EG analysis because perhaps the ZQ score is optimized for typical individuals and I’m not a typical individual and maybe I can glean something else if I have access to the raw data.
And I want to thank people who gave me insights after my previous QS talk, and knowledge by itself is not power but it is a good first step and I’d just like to thank everybody.