20 people is very low, presumably becuase the goal of the study was sleep effects, and they were using sleep labs, so I'd say that budget was a big issue here.

No doubt they wanted to maximise sample size, which is why they used the same 20 individuals for both the test and control groups.