Working memory and collective cognition lab

Models of working memory (WM) suggest that the contents of WM are separated from perceptual input by a gate, that enables shielding information against interference when closed, and allows for rapid updating when open. Recent work in the declarative WM domain provided evidence for this notion, demonstrating the behavioral cost of opening and closing the gate. The goal of the present work was to examine gating in procedural WM, namely in a task-switching experiment. In each trial, participants were presented with a digit and a task cue, indicating whether the required task was a parity or a magnitude decision. Critically, a colored frame around the stimulus indicated whether the task cue was relevant (attend trials), or whether it had to be ignored, and the previous task set should be applied regardless of the present cue (ignore trials). Switching between tasks, and between ignore and attend trials, was manipulated. The results of two experiments demonstrated that the cost of gate opening was eliminated in task switching trials, implying that both processes operate in parallel.