Well, again it depends. I was just reading a few reviews on microglia before replying to your comment, in order to update my knowledge, and I would describe it as very light effort...I'd guess I didn't enter into the lactate territory. It's nothing like coming up with an algorithm or trying to solve a problem.

The only time I've found it has come close is when actually writing a hypothesis or review paper: having to read through the literature in order to come up with a coherent theory while juggling the concepts in my head. It's very similar to solving a programming problem, and does involve significant mental effort.

But it probably also depends on the speed that it's done at. For complex algorithms or prototypes I prefer to just power through and get the whole thing done to the prototype stage in half a day (or whatever it takes, depending on the complexity), then spend the next month (or whatever) fleshing it out. There are probably other ways of doing it that don't require such cognitive effort at the beginning, but I find that for complex problems that's sometimes the best way of solving them, at least in my experience, and it only involves significant mental effort for a very short period of time.