Trevor Bissert¹ (), Candice Koolhaas², Zsuzsa Kaldy³, Erik Blaser⁴; ¹University of Massachusetts Boston

Naturalistic visuomotor tasks, like assembling a new bookshelf from a diagram, allow individuals to modulate their use of external resources (e.g., looking at the diagram) versus internal resources (memorizing steps); a sampling-remembering trade-off (Ballard et al., 1995; van der Stigchel, 2020; Liang et al., 2025). There are large individual differences in the strategies participants use in the sampling-remembering trade-off (Koolhaas et al., 2025) - here we tested whether these individual differences are stable across tasks. We used two games that were superficially different but identical in underlying task structure. In our tablet-based ‘Shopping Game’ and ‘Zoo Game’, participants (N=30, ages 18-45, 83% female) were given a list of 10 food items (or animals, in Zoo Game) and asked to locate them in a store (or field) of 20 items (10 target items plus 10 distractors). The list and the store (or field) were not visible simultaneously, but participants could toggle freely between them. We measured participants' study time of the list per visit as a measure of their cognitive strategy: the more they choose to encode in their working memory (internal resource), the longer they will spend studying the list. Study times per block ranged from 1.05s to 11.1s (M=4.1s, SD=1.8s) and, despite these large individual differences, our results show that participants were highly stable in their study time both across tasks (Kendall’s tau = 0.57, p < .001) and over time in the same task (tau = 0.52, p < .001): people who prefer to remember a lot and sample less tend to stay ‘heavy-loaders’ while those that prefer a lighter load stay light-loaders. We are also using task-evoked pupil responses (using a Pupil Labs Neon eyetracker) and a set of metacognitive debriefing questions to determine the implicit and explicit factors affecting these strategies.

Acknowledgements: The project was supported by NIH R15HD115244.