abstract Will Alexander

Will Alexander (Department of Experimental Psychology, Ghent University)

Error representation in dorsolateral prefrontal cortex.

Coactivation of anterior cingulate cortex (ACC) and dorsolateral prefrontal cortex (dlPFC) is frequently observed in the context of cognitive control and decision making. While a sizable body of work has investigated the role of these regions in supporting behavioral control, a detailed computational account of how these regions interact is still lacking. ACC has been variously implicated in performance monitoring, error detection, resolution of conflict, and a host of other processes, while dlPFC has been associated with working memory, implementing top-down control, and the representation of rules and task-sets that govern behavior. In this talk, I present a new computational model that proposes how ACC and dlPFC may interact to learn complex cognitive tasks. The Hierarchical Error Representation (HER) model suggests that error signals generated by ACC train error representations in dlPFC, which are then associated with task-related stimuli. In turn, error representations maintained in dlPFC modulate predictions of likely outcomes maintained by ACC. In a series of simulations, I demonstrate the ability of the HER model to account for a range of data from behavioral, fMRI, EEG, single-unit, and lesion studies investigating the neural bases of control.