Justin Grady¹ (), Emma Siritzky¹, Dwight Kravitz^1,2, Stephen Mitroff¹; ¹The George Washington University, ²US National Science Foundation (SBE/BCS)

Research studies commonly average over trials and/or participants to assess general cognitive effects. In visual search, for example, comparing average response times from multiple individuals across two trial types (e.g., low vs. high set size) provides key insights; however, it can also leave meaningful variability untapped and unexplored. An in-depth understanding of visual attention and other related cognitive processes can be gained by assessing factors that may produce trial history effects. Critically, such efforts leverage the fact that performance can be impacted by participants’ experience in previous trials. Previous studies have found that repeating stimulus features can lead to increased accuracy and quicker response times (e.g., Gaspelin et al., 2019), and recent work from our group (Kramer et al., 2022) demonstrated that performance benefits were sensitive to both the amount and the relative proportion of evidence for one stimulus condition over another in previous trials. The current project expanded these efforts by assessing the impact of three different types of evidence accumulation on subsequent performance in a simple visual attention task. In addition to the Binomial Z value used in the previous study (Kramer et al., 2022), the current work also assessed the simultaneous impact on performance of prior evidence for repeated category-level features as well as the impact of prior exposure to visually and semantically related evidence. A big data set (from an object-sorting task within the Airport Scanner mobile game) was used to provide sufficient power to explore the nature of these factors together. Main effects and interactions for accuracy and response times were found for evidence of category membership, visual similarity, and semantic similarity. Taken together, this work highlights how various factors, even within the context of trial history in a simple task, can interact to meaningfully impact subsequent visual attention.

Acknowledgements: Army Research Lab Cooperative Agreement #W911NF-24-2-0188