Our new paper asks whether autism is linked to the way people learn from rewards. We’ve previously shown that people not only learn to value the features that predict reward, but also assign credit to features of their actions that they know are irrelevant (in this case, the card's location).
