Control mechanisms in task switching

Abstract

Daily life often requires to perform multiple tasks either simultaneously or in rapid succession. Usually switching between tasks goes with some temporal costs and errors, known as switch costs. These costs are taken as a reflection of the control mechanisms underling the preparation for and the execution of a task switch. To investigate these control mechanisms, as an experimental tool we propose a new paradigm, the Overlapping Cues Paradigm (OCP). This approach provides the opportunity to look separately (a) at the preparation for a new task while still performing an old task, and (b) at the actual switch from the old task to the new task. In the OCP, two cues are presented within a block of sixteen trials, indicating which of the two tasks has to be performed (in our experiments, either a color or a form match task). Cue1 is presented at the beginning of a block and Cue2 after trial 8. Several trials after Cue2, a star is presented as a warning signal (WS). The two cues are either the same (Cue1 = Cue2, cue non-conflict condition) or different (Cue1 # Cue2, cue conflict condition). Participants are instructed to perform the task indicated by Cue1 until WS and to perform the task indicated by Cue2 after WS. In this way, the WS requires a task switch on cue conflict blocks, but not on cue non-conflict blocks. The main manipulations concern the interval between Cue2 and WS, and the two trials after the WS. The results from the present thesis strongly suggest that two control mechanisms, goal-directed control and stimulus-driven processing guide the preparation for and the execution of a task switch. While goal-directed control is primarily involved in the preparation for a task switch, stimulus-driven processing takes place at the actual execution of a task switch. Furthermore, under certain circumstances these two mechanisms can promote an effortless task switch.