To be or not to be relevant: Comparing short- and long-term consequences across working memory prioritization procedures

Priority-based allocation of attentional resources has shown robust effects in working memory (WM) with both cue-based and reward-based prioritization. However, direct comparisons between these effects in WM are needed. Additionally, the consequences of WM prioritization for remembering in the long term remain unclear for both prioritization procedures. Here, we tested and compared the immediate and long-term memory (LTM) effects of cue-based versus reward-based retrospective prioritization of WM content. Participants encoded four memory items and were then indicated to prioritize one of the items through the presentation of either a retro-cue or a reward pattern. We then tested their immediate and delayed memory. The results of the first experiment showed better memory for prioritized than for unprioritized information in WM and LTM, but the WM effect was driven solely by the retro-cue, making it difficult to interpret any reward-based effects in LTM. In the second experiment, using a more explicit and meaningful reward-based manipulation, the results showed a prioritization benefit in WM for both prioritization procedures. In LTM, however, the prioritization effect was predominantly driven by the retro-cue manipulation. Taken together, we found that (1) the way in which attention is directed in WM impacts the size of the prioritization benefit in WM, (2) WM prioritization generally results in a prioritization effect in LTM, and (3) that the effect in LTM is more robust for cue-based prioritization. Exploratory analyses indicated that the LTM effect of cue-based prioritization reflected a cost in performance for noncued items rather than a benefit for cued items.