Shaoying Wang¹, Srimant Tripathy², Haluk Öǧmen¹; ¹University of Denver, ²University of Bradford

Working memory (WM) has limited capacity, making efficient updating crucial for everyday tasks. Prior studies using categorical stimuli (consonants) showed that updating WM involves an active removal process that is dependent on the spatial layout of stimuli (Ecker et al., 2014; Wang et al., 2024). We investigated whether these findings generalize beyond categorical stimuli that are easy to rehearse and maintain in WM. In Experiment 1, we used an articulatory suppression (verbally repeating “the”) approach to minimize the effect of rehearsal. Participants viewed three random consonants in black-outlined separate boxes. Before new consonants appeared, some box outlines turned red (removal cue) to indicate which consonants participants (N=13) needed to update in WM. The interval between the onset of the removal-cue and the appearance of new letters varied from 50 to 1500 ms. Participants pressed a key when the removal and update were complete and their Reaction Times (RTs) were recorded. In Experiment 2, instead of categorical stimuli, we used a continuous orientation-judgment task. Procedures were similar to Experiment 1, but instead of consonants, we used randomly oriented Gabor-patches as stimuli and did not include articulatory suppression. Participants viewed these patches and reported their orientations via a method of adjustment. Results were similar between the two experiments: accuracies were high; RTs decreased with increasing removal-cue durations; the presence of a spatial gap between items to be updated increased RTs; under similar gap conditions, updating the leftmost item took more time compared to the rightmost item. For continuous stimuli, global updates required more time than partial updates (contrary to previous findings for categorical stimuli), likely because continuous stimuli require shorter removal time but longer encoding time. The results largely replicate those found using categorical-stimuli and suggest that similar spatio-temporal properties of WM removal- and update-processes apply to both categorical and continuous stimuli.