Conditional automaticity of memory-driven attention: Dissociable effects on reaction time and accuracy.

It is well known that attention is biased toward stimuli matching the contents of working memory (WM). However, it remains unclear at which processing stage this memory-driven attentional capture (MDAC) exerts its influence. Whereas some studies have taken slowed reaction time (RT) as evidence that WM-matching distractors disrupt perceptual processing, RT alone cannot distinguish perceptual interference from later decisional or response-related costs. The present study tested whether the influence of WM-matching distractors extends to target perception by measuring both RT and accuracy in two visual search tasks that differed in filtering demands and task structure. Accuracy was emphasized as the primary index of target-report fidelity. Across both experiments, WM-matching distractors reliably slowed responses, indicating a robust attentional-priority signal from WM. Accuracy also showed sensitivity to distractor presence; in the T/L search, distractor presence produced a reliable overall accuracy cost across set sizes, while in the line-search task the effect of distractor presence on accuracy depended on set size. Critically, however, WM-matching distractors did not produce an additional accuracy decrement relative to nonmatching distractors. Thus, although WM-matching stimuli reliably biased attention, their effect was expressed in RT rather than in target-report accuracy. More broadly, these findings suggest that, under the present paradigm, the consequence of MDAC was expressed at a later stage, or downstream of initial target encoding, rather than as a uniquely perceptual impairment, thereby constraining strong claims that WM-matching distractors uniformly penetrate perceptual processing.