For 25 years, one of the most important theories in computational neuroscience has sat in an experimental waiting room. Rao and Ballard's predictive coding framework — which proposes that higher brain ...