Before conducting a costly experiment, neuroscientists need to identify potentially useful hypotheses. Boyu Xu shows how ...