Revisiting the electrophysiological correlates of valence and expectancy in reward processing – A Multi-lab replication


Two event-related brain potential (ERP) components elicited during feedback processing are the frontocentral feedback-related negativity (FRN), followed by the posterior P300. According to the Error-Related Reinforcement Learning Theory (Holroyd & Coles, 2002), the FRN amplitude is largest when the outcome is negative and unexpected. Complementing this, studies on the subsequent P300 have often reported larger amplitudes for positive than negative outcomes. In an influential ERP study, Hajcak et al., (2005) manipulated outcome valence and expectancy in a guessing task. However, they found that the FRN component was larger for negative (no-reward) than positive (reward) outcomes, irrespective of expectancy. Conversely, the P300 component was larger for unexpected than expected outcomes, irrespective of valence. These results were at odds with prominent theories and extant literature. Here, we aim to replicate these results within the #EEGManyLabs project (Pavlov et al., 2021). Across thirteen labs we will not only undertake a close replication, but test the robustness of these effects to analytical choices (e.g. quantification of ERPs) and supplement the findings with Bayesian multilevel linear models to test for the reported absence of the effects.