Abstract
When an item is held active in visual working memory (VWM), visually similar information in the environment will capture attention, even if it is task-irrelevant. This capture is commonly measured through the modulation of visual-search distractor costs. Memory-matching distractors slow response times more than unrelated distractors, suggesting that placing an item in VWM is equivalent to establishing an attentional control set (ACS) for that item's features. Here we examined this possibility by testing a different prediction of ACSs: Beyond facilitating capture by matching stimuli, ACSs should eliminate capture by non-matching stimuli, at least when measured using spatial cueing effects (i.e., response time differences when a task-irrelevant cue is presented at the target location versus a non-target location). In contrast to this control over cueing effects, establishing an ACS does not eliminate the cost of non-matching search distractors, ostensibly because distractor costs measure an additional "non-spatial" component of capture. Across two experiments, participants completed a cued visual-search task while maintaining a color in VWM. The task-irrelevant cue stimulus could either match the color of the VWM item or not, and could appear at the target location (cued trial), a non-target location (un-cued trial), or not at all (no-cue trial), allowing us to measure both spatial cueing effects (cued vs. un-cued) and search distractor costs (un-cued vs. no cue). Consistent with previous search tasks, all cues produced distractor costs, and these costs were larger for memory-matching stimuli. Cueing effects by non-matching stimuli were not eliminated, however. In fact, both matching and non-matching cues produced robust cueing effects, with little evidence for VWM-based modulation. By showing that VWM-based guidance does not prevent non-matching stimuli from capturing spatial attention, these findings extend recent proposals that adopting an ACS requires more than the storage of information in VWM.
Meeting abstract presented at VSS 2016