Probing the prioritization of multiple spatial locations held in working memory.

Poster Presentation 43.444: Monday, May 22, 2023, 8:30 am – 12:30 pm, Pavilion
Session: Visual Working Memory: Space, features, objects

Jordan Garrett1,2, Daniel Thayer1,2, Thomas Sprague1,2, Barry Giesbrecht1,2; 1University of California, Santa Barbara, 2Institute for Collaborative Biotechnologies

Working memory (WM) is a capacity limited system that utilizes attention-based mechanisms to prioritize information critical for accomplishing behavioral goals. Internal representations held within an active state are insulated from distractions, have a higher probability of successful retrieval, and can bias behavioral responses. There is evidence that multiple prioritized representations can simultaneously facilitate target detection in visual search tasks compared to non-prioritized counterparts, but it is unclear if all items in WM contribute equally. Here, we investigated whether the degree of prioritization differs across multiple goal-relevant representations maintained in WM by assessing the ability of each item to incidentally direct attention during a secondary visual search task. Participants (n=27) performed a spatial WM task that required them to memorize the sequential order of four stimuli (500ms each) within an array of possible locations. After a brief delay interval (1000ms), participants then completed a 2AFC visual search task. Subsequent to the search task, participants were required to recall the memorized sequence. Based on the serial position effect, locations presented at the start and end of the memory sequence were expected to have the highest degree of prioritization, as evidenced by search task performance. Recall differed as a function of serial position, such that accuracy was highest for the first item in the sequence versus subsequent items (Bayes Factor, BF=5.41). Reaction time (RT) was faster on the search task when the target appeared in a location maintained in WM (M=1.28, SEM=0.09s) compared to those outside of memory (M=2.09, SEM=0.19s) (BF=3.91e4), despite the remembered locations being irrelevant to the search task. Notably, when the target was presented in a memorized location, RT did not differ as a function of serial position (BF=0.02). These results suggest that the degree of prioritization across multiple goal-relevant WM representations may be equivalent.

Acknowledgements: Research was sponsored by the U.S. Army Research Office and accomplished under contract W911NF-19-D-0001 for the Institute for Collaborative Biotechnologies, a Sloan Research Fellowship, and a UCSB Senate Research Grant.