When I moved back to Berkeley from Beijing last spring, I noticed that my sleep was worse in Berkeley, months after arrival. I woke up less rested than in Beijing. There was no obvious explanation. My life was similar in the two places, especially on dimensions that influence sleep. I had expected my health to be better in Berkeley than Beijing because of Beijing pollution.
Wondering why my sleep was different, I realized my Beijing bedroom was probably darker than my Berkeley bedroom. In Beijing I live in an apartment complex and cover most of my bedroom windows to block outside light and for privacy. In Berkeley, I live in a house. My bedroom window looks out over an enclosed backyard. That my Berkeley bedroom might not be dark enough had never occurred to me. It was fairly dark — no street light, no alley light, no light from neighbors.
Did the (likely) difference in darkness contribute to the difference in sleep? I tested this possibility by making my Berkeley bedroom much darker. Later I made it lighter, then darker again (an ABAB design). I measured sleep quality by rating how rested I felt when I woke up on a 0-100 percentage scale where I estimated how rested I felt compared to completely rested (= 100%). I have used this scale for many years. Here are the results:
To my surprise, when I made my bedroom darker my sleep improved. It got worse when I returned my bedroom to its original darkness. It improved again when I again made it darker. Until I graphed the data, I hadn’t realized that my baseline ratings probably shifted shortly before I made my bedroom darker. (I kept a paper record of my sleep, which made it hard to graph the data. Failure to notice this baseline shift was the last straw….I have gone paperless.) In spite of the baseline problem, the data are convincing that even at low intensities, light intensity mattered.
Depth of sleep (controlled by the amplitude of a circadian rhythm) is surely controlled by the amplitude of the light/dark rhythm. Below a certain threshold of light intensity, however, reducing light at night won’t make a difference. These observations implied that the threshold was lower than I’d thought. Support for the idea that the threshold is low — lower than other people realize, too — comes from a study published last summer after my experiment. Researchers reanalyzed old data to see if there was a correlation between lunar phase and sleep quality. Their subjects had slept in a windowless laboratory room. Nevertheless, sleep was worse during a full moon. One researcher was baffled. “What I can’t get my head around is, what would that cue be?” he said. In other words, how could the phase of the moon influence sleep? I’m not puzzled. The subjects spent only a few nights in the sleep lab. I believe there was carryover from when subjects slept at home, in rooms open to moonlight. Light from a full moon reduced the amplitude of sleep. This affected sleep later in the lab for the same reason jet lag lasts several days.
Is your bedroom dark enough? The light at night in Person X’s bedroom will differ in many ways from the light at night in someone else’s bedroom so a one-size-fits-all rule (your bedroom should be darker than . . . ) makes little sense. What does make sense is personal science: measure your sleep and test different levels of darkness.