Shaoying Wang¹ (), Srimant Tripathy², Haluk Ogmen¹; ¹University of Denver, ²University of Bradford

Working memory (WM) has limited capacity, which necessitates an efficient update process. Previous studies showed that updating WM involves an active removal process whereby information that is no longer relevant is discarded to make room for new information (Ecker et al, 2014a, 2014b). We focused on determining the dynamics of this active removal and updating process, particularly in cases of partial modifications within WM. We used a modified version of Ecker’s (2014b) experimental design. In Experiment 1, the stimuli consisted of three random letters placed in boxes with black boundaries. Before new letters appeared, some of these box boundaries turned red to indicate which letters were to be updated. The duration between the onset of this “removal cue” and the appearance of the new letter(s) varied from 50 to 1500 ms. After this delay, only the new, updated letters were displayed to ensure a partial update. Subjects pressed the space-bar to indicate when they finished their update. Results showed that, as the duration of the removal cue increased, the reaction time for updating decreased. This suggests that subjects were actively removing information from their WM. The time required to remove one letter was similar to that required for two letters. To investigate the relationship between WM capacity and the removal/update dynamics, Experiment 2 increased set-size to four, while keeping all other parameters unchanged. Our findings revealed that set-size and the spatial layout of the stimuli influenced the removal/update dynamics. We found that the time required for WM content removal and update is influenced by two factors: 1) the presence of a spatial gap, i.e. an unchanged letter, which may prevent the grouping/chunking of to be removed/updated items leads to longer removal/update times; and 2) under similar gap conditions, updating the leftmost letter takes more time compared to the rightmost letter.