Learned target-distractor similarity changes the precision of the target template and visual search behaviors
Poster Presentation: Friday, May 16, 2025, 3:00 – 5:00 pm, Banyan Breezeway
Session: Visual Search: Features, objects
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
Catherine Halpern1 (), Joy Geng1; 1University of California, Davis
Visual search efficiency is generally lower in contexts with high target-distractor similarity. However, expectations about distractor similarity can improve performance. The current study tested the hypothesis that high distractor similarity contexts drive observers to create more precise target templates, which emphasize the parts of the target that are most distinct from distractors. Eighty participants completed a visual search task using a novel circular mushroom space created from 3D objects rendered in Blender. Participants were eye-tracked during search for a repeating target in one of two groups: the similar-distractor group saw 81.25% similar-distractor trials and 12.5% distinct-distractor trials and the distinct-distractor group saw the reversed proportions. All participants completed 6.25% memory probe trials, on which they selected the remembered target from 75 possible shapes on a “mushroom wheel.” Participants in the similar-distractor group were more accurate and faster on similar-distractor trials than those in the distinct-distractor group. They also demonstrated more precise memory representations of the target and had more constrained fixations around parts of the target shape that were maximally differentiated from distractors. However, on distinct-distractor trials they took longer to respond than those in the distinct-distractor group. This difference likely reflected underlying differences in how participants in each group approached the task. Participants in the similar-distractor group serially evaluated each object’s identity, demonstrating longer dwell times on targets and distractors during distinct-distractor trials, but relatively shorter dwell times on similar-distractor trials. In contrast, those in the distinct-distractor group rapidly guided attention to the most likely target – a strategy that succeeded on distinct-distractor trials, but failed on similar-distractor trials. These results emphasize the critical role of context in shaping adaptive search strategies - the predominant distractor-context shaped the precision of the target template and search strategy, which resulted in differing speed-accuracy trade-offs in visual search performance.
Acknowledgements: UCD, JSMF