Semantically related objects act as spatial predictors during visual search

Poster Presentation: Sunday, May 19, 2024, 8:30 am – 12:30 pm, Pavilion
Session: Attention: Spatial selection 2

Makayla Souza-Wiggins1, (), Joy Geng2; 1University of California, Davis

In our daily lives, we navigate complex environments in search of target objects with remarkable efficiency. Successful detection relies on strategic attention allocation informed by semantic and structural information, but how this occurs remains unclear. In two studies, we tested the hypothesis that target search is facilitated by associated objects acting as semantic primes and spatial predictors. In Study 1 (N=43), we used an online RSVP task. Each trial featured five displays with two lateralized objects. The target (e.g., “toothbrush”) appeared, preceded by a local or thematic “prime.” The local prime was unrelated but consistently appeared before the target (e.g., “refrigerator”). The thematic prime was a semantically related object (e.g., "sink"). Results revealed quicker target detection with the thematic prime. Study 2 (N=114) replicated the previous study, placing the prime in a naturalistic spatial position relative to the target. We aimed to investigate if the thematic prime improved performance solely as a semantic cue or also functioned as a spatial predictor. Object pairs were extracted from Unity scenes for naturalistic object placement. The target appeared in four conditions: (a) thematic prime in a congruent spatial location (e.g., toothbrush on the sink), (b) local prime in a congruent spatial location (e.g., toothbrush next to the refrigerator), (c) thematic prime in an incongruent position (e.g., toothbrush on the floor by the sink), or (d) local prime in an incongruent position (e.g., toothbrush on top of the refrigerator). Participants exhibited a larger congruency effect for targets with a thematically related prime compared to a local prime, indicating semantic primes not only captured attention but also created a spatial prediction. Overall, these studies deepen our understanding of efficient attention allocation in a semantically rich world.