Putting Numbers to Sleep
Maria Benet has been tracking her sleep since July, 2014. She wanted to track her sleep so she can learn more about her sleep patterns. A good night’ sleep makes her positive, productive, and active. A night with little sleep leaves her gloomy, sluggish, but it can also inspire her to write lugubrious poetry or to explore the finer points on a pessimistic outlook on life. In this talk, she explores her data to see what affects her sleep quality and whether or not she can trust the data or her own perception.
My name is Maria Benet and I’m here to talk about sleep. Hopefully not in a way that will put you all to sleep. Last July I set out to track my sleep, to learn more about my sleep patterns or even to see if there were such patterns.
On the one hand I was hoping that if there were such patterns I could use them in fitness training. On the other hand I wanted to find out how much my perceived sense of a good or a bad night’s sleep was infact supported by data.
Sleep is a major factor in determining my outlook. A good night’ sleep makes me positive, productive, and active. A night with little sleep leaves me gloomy, sluggish, but it can also inspire me to write lugubrious poetry or to explore the finer points on a pessimistic outlook on life.
I chose the Basis watch as a measuring tool for this project because Basis tracks the amount and the percentages of deep, REM, and light sleep phases of its users. The Basis also has a simple user interface, and then easy on the eyes, graphical analysis on its website and on its mobile app, and also in the weekly email reports it dispatches to its users.
That simplicity however made it difficult for me to do anything more than to simply look at the data each day or wait for that weekly summery, since I don’t have the skills to scrape the data from the site in detail, and at that time Basis didn’t provide export opportunities.
To give my project focus, I took the times I spent in REM and deep sleep states measured by Basis as the quantified indicators of good or bad quality of sleep, and then see how those numbers stacked up against how I felt about that particular stretch of sleep.
To work with the data I used an Access database, and I added some variables to the Basic basics, such as weight, medication I might have taken, alcohol intake. Whether I had whet, since I had been diagnosed for wheat sensitivity, and notes that ranged from recording feelings to recording event logging.
Since at the time I was also training for endurance bike ride events, I was also interested in how sleep affected my resting heartrate. In particular I wanted to see what variations there might be in sleep related data around the time of my big biking events.
Although I had been gathering sleep data for over six months I spent only about three months at that time in my own bed before I had to travel. So the chart I show here only focusses on the data for those three months.
Here you see on this chart my resting heartrate in red, deep sleep in grey, REM sleep in blue, and light sleep in yellow/green from July to early October. The one thing that stood out for me was that I got a lot of light sleep, except when I didn’t those were obviously sleepless nights, and since all the combined phases of sleep take a dive on those days.
Looking a little closer look on the valleys on the deep sleep line, I could see that they were usually followed by small peaks in the next day in the resting heartrate which is nothing unusual.
A couple of those clear drops in all phases of sleep turned out to be the night before my big biking ride events. So it was useful to have a record of all the other variables I tracked in Access. Even if some of them were more subjective and less amenable to quantification. I hope they would still provide some clues for seeing what might have contributed to changes in patterns in sleep.
So here for example in September 12, the day before my big 100Km ride in Sonoma, I went back to my Access database record and saw I woke up with a good resting heartrate, but apparently it was a hot date, I drank too much and I hung out with people too long and I also took an Ativan. So I only got a little over six hours of sleep before the event.
And yet still in spite of all that I managed to best my previous ride time on that route by about 11 minutes.
The next slide here shows another sleepless night a few weeks later before my shorter but more challenging ride on the (Levine Lephanger Grand Fono?) Also in Sonoma. It turns out that that day to it was hot, and my stomach was upset and I ate pasta and had wine, and I ended up with even less sleep. And even though my resting heartrate in the morning the day before was 50, the day of the event which I don’t have the slide because that got recorded the next day. The day of the event my heartrate shot up to 58.
Some of the spike in the morning was probably from the anticipation of riding in such a huge event with so many riders, something I’ve never done before.
So what did I learn about sleep patterns and long bike rides. As much as I like endurance events, it appears the anticipated challenge is stressful enough to take a toll on my sleep time, which might mean that some of my performance at those events is due to more of a rush of adrenaline rather than anything around training or smart resting practice.
By the way, these patterns of sleepless nights also showed up before some of my other longer planned rides, although they didn’t peak as much.
I learned a number of interesting things along the way of tracking my sleep. I assumed that Inderal, which is a beta blocker would lower my heartrate, yet of the 19 times that I took it, my resting heartrate registered below 52 beats only three times. I learned that of course alcohol disrupts my sleep but not if alcohol is in the form of beer!
When I eat wheat and when I eat pizza I seem to be sleeping better strangely enough. Taking Ativan, which is an anti-anxiety medication, tends to disrupt both deep and REM sleep for me. However, in the case of Ativan, my perception of the sleep I have tends to be positive. That is I wake up feeling as if I have had a great sleep even if the numbers don’t support this. Which leads me to the main conclusion from the project so far. Unless I spend a mostly sleepless night, or have fewer than six hours of sleep, how I feel about the quality of that sleep isn’t always correlated with or supported by the data measured by my Basis watch.
It seems that if I have had a good sleep I tend to operate along those lines with more energy to spare. Not to mention feeling well rested in general. But then, when I see the numbers from Basis contradict my positive assessment of my rest, my sense of having rested well is suddenly undermined as if those Basis numbers somehow have the power to make me doubt my own experience.
So putting numbers to sleep, at least for me doesn’t always make for better dreams or brighter day for that matter.