The use of working memory for task prediction: What benefits accrue from different types of foreknowledge?

Abstract

The assumption that the deployment of executive processes invariably improves task performance is implicit to cognitive theory. In particular, working memory can be used to retain and update historical information about predictable trial sequences (foreknowledge) so that subjects can anticipate and prepare for the upcoming trial more effectively. We review the effects of different types of foreknowledge on response accuracy and latency, particularly in relation to experiments investigating saccadic eye movements in humans. While it is possible to make all aspects of an impending trial predictable, varying the predictability of different components of the trial independently can reveal which cognitive operations are potentially modifiable by foreknowledge. These operations include stimulus processing, retrieval of task-set rules, and response preparation, among others. The available data suggest that, while response preparation can be completed and the response even executed before the stimulus appears (i.e. anticipation) when the subject possesses complete task-foreknowledge (knowing both the stimulus to appear and the response required), foreknowledge of the task-set alone does not permit advance configuration of the task-set rules. A taxonomy for foreknowledge is proposed, including foreknowledge for timing, stimulus, set, response, and task. Work on differentiating these effects in neurophysiology, neuroimaging, and neuropsychology is still in the early stages.