Flexible Resource Allocation in Visual Working Memory: Insights from Pupillary Responses and Behavioral Precision

An ongoing debate exists on how visual working memory (VWM) utilizes its limited resources. This study examined how VWM resource allocation affects memory precision and pupil size. The first aim was to explore if VWM encodes items with a precision relative to their importance according to a flexible resource allocation model or encodes all items equally according to a discrete capacity model. A second aim was to investigate if phasic pupil dilation (PPD) reflected manipulations of item importance. In two experiments, 37 participants briefly memorized four colored items – circles and shapes – and reported on one randomly probed item in each trial. Before presenting items, a cue indicated the probing probability of a circle (50%, 75%, or 100%). The primary outcomes were pupil size during maintenance and response accuracy. As hypothesized, colors of more important items were reported with higher precision, which favors the flexible resource allocation model. Mixture model analysis indicated that differences in guessing rates contributed to the variations, and more important items benefitted from more encoding resources. PPD during memory maintenance revealed nuanced patterns in line with the hypotheses, albeit with non-significant group-level differences across conditions. Interestingly, individual-level analyses exposed mixed results, possibly reflecting differences in PPD sensitivity or task strategies across participants. In sum, this study supports a flexible resource allocation model of VWM and underscores the complexity of measuring resource allocation using PPD