The consequences of strategic prioritization in working memory

Within working memory, a subset of information can be strategically prioritized when considered highly relevant. This relevance may stem from objective task demands or more subjective factors, both of which can improve memory performance. We systematically compared three commonly used approaches to induce strategic prioritization in visual working memory (test-relevance cueing, rewarding, and refreshing) within a single paradigm to test whether the memory boosts they produce are similar and can be understood as reflecting a common underlying mechanism, as often assumed. We conducted three experiments and, overall, we found that these approaches, as implemented in prior studies, do not produce the same memory boosts, nor do they consistently reflect the same combination of benefits to prioritized information and costs to unprioritized information. This indicates that the different approaches are not interchangeable. Our findings suggest that strategic prioritization operates in at least two modes, depending on the predictive value of the priority signal and task context. These modes could reflect either two distinct mechanisms, or a single mechanism operating at different intensities. One is a focused mode, likely involving strong attentional resource allocation, consistently engaged by test-relevance cueing. The other is a more diffuse mode, enhancing prioritized information through subtle resource shifts or alternative processes, engaged variably by rewarding and refreshing. These findings highlight the complexity of strategic prioritization in working memory and suggest that not all approaches leverage the focus of attention in the same way or to the same extent to prioritize information in working memory.