Abstract: PURPOSE: Rapidly locating and identifying targets in complex environments is a fundamental ability for human survival. The attention templates are stored in visual working memory (VWM) to help guide attention in performing search tasks. In daily life, individuals often need to search for multiple targets simultaneously. Classic laboratory tasks have asked observers to search for a simple visual target (such as colored letters or shapes) in an array of distractors scattered randomly on a display to explore the mechanism of templates. This study used real-world scene images as stimuli to get closer to those in real-world searches. The purpose of the current research is to address two questions: firstly, whether two templates could be maintained simultaneously in the working memory when two objects were searched at the same time, and secondly, to further clarify whether the dual-target template guide search in a simultaneous or switched pattern.
METHODS: In Experiment 1, we used a dot-probe paradigm to investigate whether both target templates were maintained in VWM simultaneously during dual-target search tasks. We observed the consistency effect (i.e., faster responses to a dot presented on the side of the cued object compared with the non-cued object) under single- and dual- conditions separately, indicating attentional capture by template-matching stimuli. In Experiment 2, we utilized the ERP technique to explore further the fundamental neural mechanism of how representations are maintained in working memory. We aimed to assess whether the load-dependent component of working memory storage, as measured by the contralateral delay activity (CDA), is influenced by the varying number of activated target template representations during the search preparation phase of the visual search.
RESULTS: The results of Experiment 1 showed that both single-target and dual-target template conditions exhibited a consistency effect, and the consistency effect of the dual-target template was not significantly lower than that of the single-target template. This indicates that both attention templates were maintained in VWM to guide attention and performance of the search task during a dual-target search in real-world scenes. Experiment 2 showed no significant difference in the CDA component between the single- and dual-target template conditions. This reflected the idea that dual-template guide the search using a switching method, and only one representation would be activated for each search task.
CONCLUSIONS: It can be concluded that in dual-target search tasks in real-world scenes, both attention templates can guide attention simultaneously. Still, only one template is activated at a time, and the two templates alternate to guide attention in a switching pattern.

Key words: visual search, attentional template, visual working memory, CDA