Reactivation of learned reward association reduces retroactive interference from new reward learning.

Learning to associate specific objects with value contributes to the human's adaptive behavior. However, the intrinsic nature of associative memory posits a challenge that newly learned associations may interfere with the old ones if they share common features (e.g., a reward). In the present study, we conducted a set of behavioral experiments and demonstrated that retroactive interference in reward learning can be reduced by reactivating originally learned reward associations before the new learning. We used the well-known effect, attentional capture driven by reward-associated feature, as the index of reward learning (Experiments 1A and 1B) and showed that learning a new reward-color association impaired the old learning as indicated by the reduced capture effect of the old reward-color associations (Experiment 2). Interestingly, the retroactive interference was significantly reduced if a brief reactivation of the old reward-color associations was introduced immediately before the new reward learning (Experiment 3). However, the retroactive interference reemerged if the new learning was conducted outside a reconsolidation window, indicating the critical period during which reactivation protects learned reward salience from the interference of new reward learning (Experiment 4). These findings suggest that reactivation could serve as an effective procedure to reduce mutual interference between multiple learnings. (PsycInfo Database Record (c) 2021 APA, all rights reserved).