Experiment 2 - Automatic value learning results in counterproductive human behavior

To adequately learn from reinforcement, individuals need to refrain from assigning credit to features of their actions that they deem to be irrelevant for outcome obtainment. Outcome-irrelevant learning is defined as a tendency to assign credit to features of the action that are known with high certainty as having no causal relation to the outcome. Previously, we found individuals to assign credit for outcome-irrelevant action features despite explicit instructions. Here, we hypothesized this may result from an ongoing feature learning process making it necessary for participants to experience “at first hand” that locations are indeed random and irrelevant for outcome prediction. Thus, we test whether following extensive experience of three sessions participants learn to disregard locations. Furthermore, we ask participants for their motivation when doing the task in each session to evaluate whether this correlated with outcome-irrelevant learning. We predicted that despite extensive training, and high motivation to complete the experiment, participants will still show outcome-irrelevant learning.