Carbon Dioxide: An Open Door Policy | Slate Star Codex

[Content note: reading this post might cause feelings of suffocation or provoke panic attacks in susceptible individuals. Epistemic status is very speculative.]

Last month I moved into a small cottage behind a big group house. The cottage is lovely. The big group house is also lovely, but the people in it started suffering mysterious minor ailments. Headaches, fatigue, poor sleep – all the things that will make your local family doctor say “Take two placebo and call me in the morning”. Using my years of medical training and expertise, I was able to…remain completely unaware of the problem while my housemates solved it themselves.

There’s been a flare-up of research interest in indoor carbon dioxide levels, precipitated by a Berkeley study (paper, popular article) finding that increasing CO2 concentration from the level of a well-ventilated building to the level of a poorly-ventilated building had profound effects on cognitive ability, cutting various test scores by as much as 50%. This was so dramatic as to be implausible, but seems to match the result of previous Hungarian studies and a later Harvard study on the same subject. The Harvard team later replicated their result with real workers in real offices and found that, controlling for other factors, workers in the best-ventilated offices scored about 25% better on cognitive tests than in the worst-ventilated ones. NASA got really interested in this research because spaceships require a lot of intellectual work and don’t have a lot of open windows. They’re still running tests but they say that “preliminary results suggest differences” between better- and worse- ventilated environments.

On the other hand, a 2017 study failed to find the effect, possibly because their cognitive tests were easier. And bloggers have pointed out that submarines have more CO2 than the worst terrestrial buildings, but don’t have any problems overt enough for the Navy to notice or worry. So it’s a crapshoot of contradictory results and considerations, just like everything else.

Aware of this research, my housemates tested their air quality and got levels between 1000 and 3000 ppm, around the level of the worst high-CO2 conditions in the studies. They started leaving their windows open and buying industrial quantities of succulent plants, and the problems mostly disappeared. Since then they’ve spread the word to other people we know afflicted with mysterious fatigue, some of whom have also noticed positive results.

When I heard about this, my first question was: didn’t any of these people notice they only felt bad at home? Shouldn’t it have been a big red flag when they went to the office, or went for a walk, and all their problems disappeared? This can’t be too big a deal, or else “I feel bad in my house, but fine everywhere else” would be a more common complaint.

My housemate Kelsey referred me to the work on CO2 and sleep. Right now this is just a few papers by a guy named Strøm-Tejsen, but the implications are pretty important. He notes that however bad your carbon dioxide levels are during the day they’re probably much worse at night, when you shut yourself up in a small room, close all the doors and windows, and just breathe for like eight hours straight. Normal outdoor air is about 400 ppm CO2 (more by the time you read this; thanks, fossil fuel industry!) A well-ventilated building during the daytime is about 700 ppm, and a poorly ventilated building during the daytime about 1400 ppm. But the average bedroom at night can be 2000 ppm or more. Friend-Of-The-Blog Gwern got a CO2 monitor to test these findings, and confirmed that while his daytime CO2 was around 500 ppm, nighttime CO2 in his bedroom could get as high as 3000 ppm. MIT’s Joel Jean discussed trying the same in this Medium post, with similar results:


I live in California, and Gwern presumably lives in some kind of formless cybermatrix, so we don’t have to worry about seasons. But Dr. Jean lives in Massachusetts, and he found that during the winter, indoor CO2 went up even further, in some cases exceeding OSHA’s rules for permissible workplace exposure:


If CO2 can affect sleep quality, that would explain how it could produce a whole-day effect. Strøm-Tejsen tests this on sixteen subjects and finds that “objectively measured sleep quality and the perceived freshness of bedroom air improved significantly when the CO2 level was lower, as did next-day reported sleepiness and ability to concentrate and the subjects’ performance of a test of logical thinking.” Good things about this study: subjects were blinded to condition, the paper contains a pilot experiment and a main experiment which mostly replicate each other’s results. Bad things about this study: the experiments were about n = 15 each, the researchers didn’t correct for multiple comparisons, and they admit to manipulating the statistics surrounding their logical reasoning tests to get better results. But if I just look at their tables and try to ignore their manipulation, I’m at least kind of impressed:


And experts seem to take their results seriously – for example, here’s NASA again. And we know from sleep apnea and studies that high physiological levels of carbon dioxide can cause sleep disturbances. I can’t figure out how to convert external ppm to internal likely level of carbon dioxide in the blood, but maybe this could provide a plausible mechanism.

I’m reluctant to be too numerical about all this, because everything about health has massive individual variability. Three people share one of the bedrooms at my group house (look, the Bay Area is really bad). One of them got the typical symptoms of excess CO2 really bad; the other two were fine. Some people are just going to be more sensitive to this kind of thing – the same way three people can drink the same amount of alcohol, two of them will get pleasantly buzzed, and one of them will black out.

I’ve tried sleeping with my door open the past few nights, and I haven’t noticed any difference. Probably I shouldn’t; my house is well-ventilated and I wasn’t feeling too bad beforehand. But I’ve started recommending a few of my patients with mysterious sleep issues try the same thing. It’s too early for results so far, and the science behind it is weak, but it seems like a cheap experiment.

Since the main source of CO2 is human exhalation, I’m most worried about buildings where many people are crammed into small spaces in close proximity (hello, Bay Area readers!). Since the main way CO2 gets cleared is through ventilation, I’m most worried about buildings made to strict environmental standards with great insulation (hello, Bay Area readers again!).

If you’re concerned about this, the best solution is to open a window or an internal door in your bedroom at night. If for some reason this is impossible, the second-best solution is to get certain succulents or other plants that participate in the ominously-named process of “dark fixation” – ie do their plant breathe-in-CO2-and-breathe-out-oxygen thing at night. This is also called “crassulacean acid metabolism” and Googling either term will get you a list of appropriate species. It will probably take like ten succulents to do much to CO2 levels, but a room full of succulents on every flat surface is also kind of #aesthetic.

I’m interested in more data on this, so if you’re planning on experimenting with changes to your nighttime air quality based on this post, please fill out this form to register for an informal quasi-experiment. I’ll follow up with a form for you to give your results in a couple of weeks.

Oh – and sorry for the content warning at the top, but I’ve felt kind of low-grade suffocate-y throughout writing this post, and had to go out and take a couple of breaths of fresh air a few times. Remember – perception is fundamentally Bayesian, and combines external sensation with internal expectations; this is why placebos can have such a profound effect on pain. Perception of air quality vs. suffocation seems to be especially susceptible to this, which is probably one of the major etiological factors behind panic attacks. Just repeat to yourself that it all adds up to normality: the air quality in your room hasn’t changed since when you were feeling just fine before you started reading this article, so you should be okay.