Reward-based prioritization in working memory is distinct from recency and due to a resource trade-off

Prioritized items within working memory often show higher accuracy or faster response times at test. This prioritization benefit is thought to reflect the memory representation being within the focus of attention. The literature on prioritization effects lacks consistency in whether prioritization benefits manifest as a benefit to speed, accuracy, or both, implying that some findings may reflect a speed-accuracy tradeoff rather than improved memory representation. In the present work, we use drift-diffusion modeling to test two theoretical questions regarding prioritization effects without the possibility of a confounding speed-accuracy tradeoff. First, we test whether prioritization effects rely on shifting limited working memory resources away from non-prioritized items. Second, we test whether prioritization and recency effects are two distinct mechanisms or two ways to enter the focus of attention. We find that prioritization effects do reflect shifts in limited working memory resources and that prioritization and recency are two distinct mechanisms. These findings are augmented by model fits suggesting differential effects of prioritization at the perceptual/motor and cognitive levels.