Martin Constant¹ (), Dirk Kerzel¹; ¹University of Geneva

It has recently been shown that salience can have a massive impact on visual working memory (Constant & Liesefeld, 2021; https://doi.org/gjk9jh). That is, in an array of vertical bars, the color of strongly tilted bars was recalled better than the color of less tilted bars. This effect is also hard to erase or overrule with manipulations of relevance, even at long encoding times (Constant & Liesefeld, 2023; https://doi.org/gr6xzr). Here, across three pre-registered experiments, we tried to determine whether feature pre-cues, spatial pre-cues and retro-cues can counteract the effect of salience. Pre-cues were presented about 500ms before onset of the memory display and retro-cues 750ms after its offset. In all experiments, the memory array consisted of three tilted bars (12°, 28° and 45°) among 33 vertical bars and was presented for 350ms. Participants recalled the color of one of the tilted bars using a colorwheel. In the experiment with feature pre-cues, the target orientation was pre-cued by a black tilted bar in the center. While the cue improved overall memory performance when comparing valid (66%) to invalid (17%) trials, the effect of salience remained in both conditions. In the experiment with spatial pre-cues, a 66%-valid black circle was briefly presented at the future location of one of the targets. Again, overall memory performance improved but the effect of salience remained. In the experiment with retro-cues, the cue was either neutral (vertical bar presented centrally) or 100% valid (tilted bar presented at the location of the target). Once again, overall memory performance improved, but the effect of salience remained virtually unaffected. These results suggest that the effects of salience on visual working memory may not be purely attentional. That is, lower-salience targets were probably found and attended when they were cued, yet performance remained worse for these targets.