Abstract
Previous research has demonstrated that monkeys and humans can form an attentional template in visual working memory (VWM) that specifies the features of a target and can be used to guide attention toward matching items. VWM has a capacity of 3–4 items and could presumably store multiple templates, but it is not clear whether multiple VWM representations can simultaneously guide attention. To test this, we recorded eye movements during a visual search task in which observers searched for a Landolt-C target in an array containing 32 items, with 8 in each of four colors: red, yellow, green, and blue. A cue indicated which two colors might be the color of the target; observers were instructed to either search one color and then the other, or to search both colors simultaneously. Some of the cue color pairs were more similar in hue (e.g., red-yellow) and others were dissimilar (e.g., red-green). Searching for two similar colors could potentially be achieved by means of a template with a single intermediate color. When instructed to search one color at a time, observers exhibited no RT difference between similar and dissimilar color pairs, indicating that only one template was active at a time. However, when participants were instructed to search the two cue colors concurrently, RTs were longer when they searched the dissimilar color pairs than the similar color pairs, and they were more likely to direct gaze toward items that were not one of the two possible target colors. Thus, it is difficult (although not impossible) to simultaneously use two templates to guide attention.
NIMH grants R01MH076226 and R01MH065034.