Introduction: Fix Your Sleep With Low Blue Light!
I struggled for decades with an erratic sleeping schedule. I thought that my irregular ‘body clock’ was dooming me to a life of falling to sleep at 2-3am, waking up 5-6 hours later, feeling numb and defeated before the day even begun. Even on weekends, sleeping in until 1-2pm, I rarely ever felt like I was getting ‘sound sleep’.
About a year ago, my friend introduced me to the concept of ‘low blue light’, which is based on the connection between blue light and the sleep-regulating hormone melatonin. The core idea is that exposure to sky blue part of the light spectrum suppresses the production of melatonin, which then interrupts the body’s cycle of tired->sleep->alert. Constant, night-time exposure to the sky blue part of the light spectrum (eg. from a computer screen or from most ordinary, white lights) tells the body it is constantly daytime outside.
So, I experimented with my friends advice and took a few simple steps to limit my nighttime exposure to blue light, and the results have made me feel like a ‘sane’ sleep again. After a while of using the bully pulpit to preach the low blue light gospel, I wanted to have numerical data to show skeptical, analytical people. Without data I sounded like the ICP talking about magnets.
Like any decent experiment write-up, I will first detail the 3 small modifications in my life I made to reduce blue light exposure at night, and then present the results!
More info: https://www.lowbluelights.com/index.asp
Teachers! Did you use this instructable in your classroom?
Add a Teacher Note to share how you incorporated it into your lesson.
Step 1: Materials
For the sake of completeness, here are the 'materials' I used for this experiment.
- -- Orange Glasses, to block blue light
- -- f.lux computer program, to make computer screens less sleep-destructive.
- -- Orange Lightbulbs, to not produce blue light in the first place!
- -- Gmail, for measuring first sent email time.
- -- Plotly, to generate the graphs.
Step 2: Install F.lux or Other Program to Simulate Sunset/sunrise
f.lux is a program that simulates sunset/sunrise by changing how colors display on your monitor. You tell f.lux your time-zone, and the program takes care of the rest. As the f.lux website says: “During the day, computer screens look good—they're designed to look like the sun. But, at 9PM, 10PM, or 3AM, you probably shouldn't be looking at the sun.”
That seems right to me. My first ‘ah-ha’ moment with experimenting with low blue light came the first night after I installed f.lux. I switched it to nighttime mode (which is orange-ish like a fire or the setting sun), and then, an hour later, switched it briefly back to ‘daytime’. It was agony.
More info: https://justgetflux.com/
Step 3: Wear Orange Glasses at Night!
I still needed some way to filter out the blue spectrum from the ordinary white lights in my house (and everywhere else). As a solution, I bought a pair of these orange goggle-like things to wear at night before I sleep. I bought this type because they fit over my prescription glasses, but there are many other styles.
You may be asking yourself, “But, Nathan, do you actually wear those hideous goggle-things if you go out to get a drink? Or for a late-night dinner?”. The answer is No... but I wish I had the stones too. Still, for the times when I do spend the majority of the night at home, they work very well.
Glasses those pictured above can be found, cheaply, on Amazon, or expensively at a number of other sites.
Step 4: Install Some Orange Lightbulbs
While not entirely necessary, I wanted to change some of my ordinary white lightbulbs to hipster-orange low-blue-lightbulbs. Wearing the orange glasses for hours some nights is annoying. So I swapped the lightbulbs in my bedroom (my girlfriend vetoed any further low blue light room annexations).
Step 5: Data Collection Methods
There you have it: 3 (somewhat) simple steps to sleeping sanity. But what about the data?!
First, let me explain. Because I have not recorded my exact time of getting out-of-bed each and every day (my OCD finds its bliss in other ways), I have no way to directly measure the dependent variable of this experiment.
To explain my workaround, requires a bit of backstory. Over the past 6 years, I have run my own successful business in web analytics/SEO. I spend most of the work day emailing clients…accumulating tons of email. Therefore, I decided that the time of my first email sent each day would be a suitable proxy for how earlier I had gotten out of bed that particular day. (Technical note: ‘6am’ was the cut-off time I used as the beginning of each day, to prevent data contamination with late-night computer bingeing).
With that, in mind, I went to work gathering the data from my gmail archive.
Step 6: Results
After almost a year had passed, I graphed the data to see if there had been significant effect on my sleep.
The first shows a scatterplot of the time of the first email I sent each day. I thought it would be easier to see the trend by smoothing the data, so I put in a 30-day moving average line as well. Notice how the 30-day MA is consistently below 10AM after going on the LBL regimen.
The second graph shows the data as a histogram of first email times, before and after low blue light. You can see that both have a roughly l normal distribution -- but with means almost 2 hours apart!
If you want to interact with the graphs on plotly itself, here are the links:
Step 7: Errata
This instructable is a replication of study my friend Brian Toomey did, along with Seth Roberts. He is currently in the process of writing up his own report on his findings, which I will link to here when finished.
If you are interested in a great read (that I found funny in the same way conspiracy videos are funny), I recommend the following book:
I would love to have more people replicate this experiment in a way that would work for them. Obviously, there are some pretty serious flaws that make this experiment less that rigorous for publication, but that wasn't my goal. My goal was to show how a little creativity can lead to interesting personal science results.
First Prize in the
Data Visualization Contest