Ziyi Wang¹ (), Martin Eimer², Anna Grubert1¹; ¹Durham University, ²Birkbeck, University of London

Visual search for known objects is guided by target templates that are held in visual working memory. Such attentional templates are activated in preparation for search. Previous studies have shown that during search for two randomly presented colours, multiple attentional templates can be activated in parallel. But can multiple templates be activated strategically when target colours are repeated versus switched predictably across trials? We presented observers with series of search displays (every 1600ms) which contained two target-colour bars (AB; red and green) and four differently coloured nontargets. Following an AABB target colour sequence, participants reported the orientation of the target-colour bar in each trial. There were clear behavioural switch costs, with delayed target reports on colour-switch trials. Between search displays, we presented a constant stream of probe displays (every 200ms) each of which contained a colour singleton in one of the two target colours (ab; red or green) and five grey items. We measured N2pc components (electrophysiological markers of attentional capture) in response to these colour probes to track the time course of template activation prior to search. In colour-repeat trials, probes that matched the upcoming target colour (aA) produced reliable N2pcs from 800ms prior to search. However, in colour-switch trials, such probes only triggered N2pcs when they immediately preceded a search display. These differences in probe N2pc patterns suggest that the behavioural swich costs are a product of delayed template activation in colour-switch trials. Probes in the search-irrelevant target colour (bA) never triggered N2pcs, suggesting that behavioural switch costs were not caused by competition between simultaneously activated templates. Overall, our results demonstrate perfect colour selectivity and strategic control (i.e., serial template activation) during a two-colour AABB search task, and link colour switch costs to the time course of search template activation.

Acknowledgements: This work was funded by research grants of the Leverhulme Trust (RPG-2020-319) awarded to AG and the Economic and Social Research Council (ES/V002708/1) awarded to ME.