Control of cognitive processes in task-switching

Electronic versions

Documents

  • James A. Grange

Abstract

The explicitly-cued task-switching paradigm is a popular tool for measuring the
processes that enable flexible cognitive control. In this paradigm, participants are typically presented with a stimulus that affords multiple tasks ( e.g. numbers). Participants know which task to perform on a given trial by way of a task cue. In the present thesis, I investigate what cognitive control processes allow the correct task to be performed in the face of conflicting irrelevant tasks. Participants were presented with a cued attention-switching paradigm, wherein most aspects of the task-set were kept constant (i.e. stimulus display, response processes etc.), and participants merely switched attention between cue-target pairings.
Across five experimental chapters, I provide evidence that task performance
involves activation of relevant working memory (WM) representations that guide
behaviour, together with inhibition of irrelevant representations. In the first half
of the thesis, I investigate inhibitory mechanisms during task-switching. Although typically believed to suppress response-related processes of trial performance, I provide converging evidence that inhibition can also target earlier, cue-based preparatory stages of task performance. Specifically, I suggest that inhibition targets WM representations that are formed after cue-presentation, and that inhibition does not target the cue itself. In the second half of the thesis I investigate in more detail the activation processes that serve to form these WM representations by separating cue-related processes from task-switching processes using two cues per task. By varying the degree to which a cue provided the necessary WM representation (cue-transparency), I was able to show that cue-switch costs arise from cognitive control processes. Additionally, I provide evidence that cue-transparency has no carry-over effect on target processing by separating cue- and target-related processes using an extended runs design.
The findings suggest control of cognitive processes requires a balance between
inhibitory and activation processes that serve to manage access to a WM
representation of what to do.

Details

Original languageEnglish
Awarding Institution
  • Bangor University
Supervisors/Advisors
Award dateMar 2010