Reducing failures of visual working memory with tailored feedback

Poster Presentation 43.324: Monday, May 20, 2024, 8:30 am – 12:30 pm, Banyan Breezeway
Session: Visual Memory: Working memory and development, individual differences, capacity, resolution

Olga Kozlova1 (), Kirsten Adam1; 1Rice University

Although individuals can hold up to 3-4 items in visual working memory, they frequently experience severe lapses of performance (e.g., 0-1 items correct). An initial study found that performance feedback can improve working memory performance if the feedback encourages participants to perform consistently and avoid lapses (i.e., get 3+ items correct or lose points; Adam & Vogel, 2016). However, past work used a single performance goal that matched group-averaged capacity (3 items). Here, we hypothesized that performance feedback would be most effective when the performance goal is aligned with individuals’ capacity. In Experiment 1 (N = 28), participants performed a whole-report memory task with performance feedback and either a well-aligned goal (get 3 items correct) or a poorly aligned goal (get 1 or 6 items correct). During each trial, participants first saw a reminder of the performance goal. Then, a memory array composed of six different colored squares was presented, and participants were instructed to remember the items over a blank delay. Finally, participants recalled each item’s color and either gained or lost points depending on if their performance exceeded the performance goal (e.g., at least 3 correct). In Experiment 2 (N = 25), participants completed an identical task, but with a more fine-grained goal range (varying from 1–5). Across both experiments, we found that a performance goal tailored to working memory capacity (3 items) is most effective. Compared to sub-optimal goals, an optimal goal improved performance by 0.5 items. Moreover, in Experiment 2, tailoring the performance goal to an individual’s specific capacity improved performance by 1 item relative to a sub-optimal goal. Performance was particularly poor when the goal was too high, with lapse rates increasing from 3% to 22%. Together, our results suggest that individualized goals can encourage optimal performance by targeting the frequency of lapses.