Abstract
The question of whether visual working memory (VWM) and visual attention share the same representations has been investigated for a long time. Many studies have shown that attention is captured by items that are similar to items currently held in VWM. In such studies, observers are typically first tasked to remember a color. They then perform a visual search task. On some trials, one of the distractors has a unique color. Interestingly, when this distractor is a memory-similar color, search times are slower than when the distractor is a memory-dissimilar color. This capture of visual attention is interpreted as evidence for the view that VWM and visual attention share the same representations. While the bulk of these studies have focused on the effects of VWM representations on visual search performance, the reverse has largely been ignored. Here, we asked whether the reverse is true as well: does the nature of the visual search task affect the quality of the VWM representation? On each trial, observers were tasked to memorize a single color followed by either a retention interval or a search task. The search task could be efficient (easy) or inefficient (hard) and the color of the search items was irrelevant to the search task. There were four possible search conditions: 1) easy search using a memory-similar color or 2) a memory-dissimilar color, 3) hard search using a memory-similar color or 4) a memory-dissimilar color. After each trial, participants reported the color in memory. The results showed that VWM representations were shifted towards the irrelevant color of the search items in the memory-similar condition, and this effect was larger in inefficient search. The results provide further evidence that VWM and visual attention share the same representations and that the interaction between the two systems is bi-directional.