The effect size of the higher level of CO2 on some of the tests is ridiculous. Reminds me of Cochran's speculation on nitrogen content.
Interesting, but has problems with helium supply, even at smaller scales.
Breathing tanks are problematic. If you carry a breathing tank, the simplest approach involves venting a lot of helium. Scrubbers, recycling, and O2 replenishment carries nontrivial risk of death without warning. Capturing the exhaled helium requires heavy, power-hungry compressors.
Building present nontrivial air-quality engineering problems, and need more than airlocks retrofitted on to make them airtight.
Also, it's far from obvious that 20% is the optimum O2 content, though I think it's quite well supported to say that 100% is too much.
If it's worth saying, but not worth its own post, even in Discussion, it goes here.