Is blue light evil? Night lighting, gadget screens, and tenth-assed science

One ill-fated night of lost sleep that I will never forget was caused by a bad news notification on my iPhone. Just as I was supposed to start drifting off, I learned that PainScience.com was offline. This was back in the days when smartphone notifications were still a novelty, and I was still new to running an online business — so I wasn’t yet jaded and unflappable about downtime.

Running a website takes nerves of steel — which I didn’t know when I started, or I might not have started one. I had to learn, slowly and painfully, how to remain calm when the technology was glitching out of my control.

The 2013 “Gunshow” webcomic by artist KC Green that became a meme.

Back then, downtime still freaked me out, and so I was awake for hours waiting for the website to come back to life. If only I hadn’t looked at my phone, I would have slept fine.

But what about the light?

Insomnia is a major topic on PainScience.com because sleep deprivation is one of the most predictable and tractable factors in chronic pain. For instance, obviously we can adjust our gadget usage at night, because they are obviously a source of many potentially engaging or even stressful stimuli, best avoided when you’re trying to wind down.

But should our concern go deeper? Is it about the light? What about the whole blue-light panic? Do our screens blast us with circadian-rhythm wrecking rays?

All modern computers and smartphones now have blue-light filtering features (Night Shift, Night Light, etc). Surely there’s a good reason for that!

Or is it just more over-hyped hand-wringing about the perils of modern living? Which is a weirdly popular hobby in a world where we live three times longer (and better) than our primal ancestors.

I want you to read PainScience.com. But maybe not at night. And maybe with less blue.

Note: This is about blue light at night. People are going to ask, so, yes, I am familiar with Huberman’s podcast and what he’s saying about blue light being important at dawn and dusk for entraining the circadian rhythm, and it’s clearly interesting and relevant. But for this post I chose to focus just on the question of screen light at night. (Also, Huberman is kinda terrible.)

The science of sleep and light at night — half-assed? More like tenth assed

The science is complex, conflicting, and extremely incomplete.

The first ever direct efficacy trial of the iPhone’s Night Shift feature, which makes the screen much more yellow and less blue at night, made headlines in mid-2021 (see Duraccio et al.). It appeared to be a mortal blow — people who had warm, yellow screens did not sleep any better than subjects who did their reading on harsh blue-hued ones.

Every cynic who saw those headlines rolled their eyes, said “I told you so!” … and stopped thinking about it. But genuine critical thinkers rarely close a subject, of course. Skepticism isn’t (just) cynicism.

It was an important study, the right kind of study — testing the effect of the feature on how well people actually sleep. But it also had flaws and limitations, as all studies do, and so it was hardly the last word.

The most important flaw in that study was probably that they used a classic “sample of convenience”: healthy university students, a population that sleeps so much better than people twice their age that they might as well be a different species for this purpose. Many variables and potential confounding factors were not controlled here.

And there’s another flaw that stands out, another “convenience” with a cost, a particularly easy way of measuring sleep: smart watches. This has genuine advantages (chiefly the convenience) and disadvantages, but there’s just no way it can produce definitive data on sleep quality. How much we move in sleep is not a perfect proxy for sleep quality.

I have some of these around the house for (very) warm light at night. I don’t care if it’s evidence-based. I do it because it makes my home look more like a “lair,” & there’s still a bit of “goth” in me, leftover from my ridiculous youth.

And then there’s the other evidence

Just two months after Duraccio et al , a second study of this kind was published, showing …

Basically the opposite (see Schmid et al.). Of course. 🙄 This result did not make headlines.

They were more serious with their measurements — polysomnography for the win! — and they did detect a meaningful effect on “slow-wave-sleep and -activity in the first night quarter.” This study also had flaws and limitations, of course, but it doesn’t even really matter: the point is not what this or that study said, but just that there aren’t enough studies yet.

Just two direct tests of blue filters with conflicting results do not a conclusion make.

My official advice for now

What we have now is a near perfect absence of directly relevant evidence to inform us. We have “circumstantial” evidence and plausibility only, which give us plenty of reasons to be both skeptical and optimistic about blue light filters.

In this situation, apply the cautionary principle. It’s probably worth minimizing blue/intense light at night, because it’s easy, it might help, and it certainly won’t do any harm to live a little more like we did before The Age of Screens.

This post is a condensed excerpt from an update to my full guide to treating insomnia.