A large sample size of people isn't necessary to verify the ability.

Case in point, the woman who can smell Parkinson's disease, to some ridunculous p-value.

http://www.scientificamerican.com/article/one-woman-s-abilit...