Jan Tünnermann¹ (), Anna Schubö¹; ¹Philipps University Marburg

Visual foraging, where participants gather items through clicks or touches, gained popularity as a naturalistic paradigm in visual search and selective attention research. Unlike single-target search tasks with discrete trials and button responses, foraging involves multiple targets, continuous interactions with the search environment, and naturalistic goal-directed actions. Notably, the latter offers an opportunity to investigate the interplay between selective attention and actions, a characteristic of foraging not leveraged until now. In the present tablet-based study, we asked how selective attention influences the collecting actions in visual foraging. Earlier foraging experiments with feature conjunctions revealed same-type selection runs. For instance, with red circle and green square targets among green circle and red square distractors ("conjunction foraging"), foragers adhered to one target type (e.g., green squares) for extended runs to avoid costly switches. For the present experiment, we introduced a novel stimulus: "conjunction objects". Targets were composed, for instance, of red rectangular parts joined with green half-circles, and distractors had opposite arrangements. Anticipating the robust same-type run effect seen in conjunction foraging, we expected foragers to repeatedly attend to the same object parts, a bias that should be reflected in the selection actions. Indeed, we found such a bias, with an over-proportional number of stylus clicks landing on the same object parts. However, we also observed a bias toward the center of mass of the overall objects. Hence, the attended location might be fed into motor control, where it combines with motoric influences such as stable contact point selection. We argue that visual foraging is a fruitful tool for examining the relationship between attention and action. In high-paced foraging selection sequences, occasional errors bear minimal consequences, allowing organisms to program motor responses based on attentional selection without involving decision-level processes that might otherwise diminish the direct impact of attention on actions.