Abstract
Visual search for a target progresses from a "coarse" to "fine" scale strategy (Over et al., 2007; Godwin et al., 2014). That is, people scan a global layout first, then focus on local areas to find the target. In the current study, participants searched for an oddly oriented bar among other bars (distractors) and reported target presence. Participants had to process overall distractors, because the odd target orientation was defined by the distractor orientations. In Experiment 1, the target orientation was either 45° clockwise or counter-clockwise to the vertical meridian, and it was randomly selected from any orientation in Experiment 2. Thus, participants could not form top-down knowledge about the target orientations in Experiment 2. The difference between the mean distractor orientation and the target orientation was kept constant in every trial. Distractor orientations were either smoothly varied around their mean (smooth distribution) or equally divided into two clusters having extreme orientations (distinct distribution). It has been reported that processing overall distractors is easier when the distractors form a smooth distribution (Utochkin & Yurevich, 2016). In both Experiments, people responded faster in the smooth distribution condition than in the distinct distribution condition. Critically, reaction times were shorter when the current trial was preceded by a correct target-absent trial than by a correct target-present trial. We did not find a significant interaction between the previous trial's target presence and the distribution type. This suggests that participants were using a "coarse" strategy when a search was terminated with a "target absent" response, which may help the global processing of distractors in the following trial. On the other hand, when participants finished a search with a "fine" scale strategy after a successful target detection, it may have cost them to switch to a "coarse" strategy in the following trials.
Meeting abstract presented at VSS 2018