I've been self-experimenting with piracetam the past few months.

I usually study from a site called USMLEWorld with a selection of difficult case-based medical questions. For example, it might give a short story about a man coming into a hospital with a certain set of symptoms, and explain a little about his past medical history, and then ask multiple choice questions about what the most likely diagnosis is, or what medication would be most helpful. These are usually multi-step reasoning questions - for example, they might ask what side effect a certain patient could expect if given the ideal treatment for his disease, and before answering you need to determine what disease he has, what's the ideal treatment, and then what side effects that treatment could cause. My point is they're complicated (test multiple mental skills and not just simple recall) and realistic (similar to the problems a real doctor would encounter on the job).

I've tried comparing my performance on these questions on versus off piracetam. My usual procedure is to do twenty questions, take 2400 mg piracetam + 600 mg lecithin-derived choline, go do something fun and relaxing for an hour (about the time I've been told it takes for piracetam to take effect) then do twenty more questions. It's enough of a pain that I usually don't bother, but in about three months of occasionally doing my study this way I've got a pool of 160 questions on piracetam and 160 same-day control questions. Medicine is a sufficiently large and complicated field that I don't think three months worth of practice effects are a huge deal, and in any case I made sure to do equal piracetam and control questions every day so there wouldn't be a practiced-unpracticed confounder.

I got an average of 65% of questions right in the control condition and 60% of questions right on piracetam, but the difference was not significant.

USMLEWorld also tells you how other people did on each question; I used this information to run a different analysis controlling for the random difficulty variation in the questions. In the control condition I did 2.8% better than average, in the piracetam condition I did 1.3% worse than average; this wasn't a significant difference either.

I do worry that fatigue effects might have played a part; I tried to always rest and relax between conditions, but I was always doing piracetam after control (I wanted to have same-day comparisons to eliminate practice effects, and piracetam lasts too long for me to feel comfortable taking it first and then doing control after it wore off). But I didn't feel fatigued, and I haven't noticed huge fatigue effects when I study a lot without taking piracetam.

In any case, piracetam either has no effect on me in the reasoning domains I'm interested in, or else its effect is so small that it is overwhelmed even by relatively minor fatigue effects.