Expectations and experience: Dissociable bases for cognitive control?

Classic theories emphasized the role of expectations in the intentional control of attention and action. However, recent theorizing has implicated experience-dependent, online adjustments as the primary basis for cognitive control--adjustments that appear to be implicit (Blais, Harris, Guerrero, & Bunge, 2012). The purpose of the current study was to evaluate whether explicit expectations play any role in cognitive control above and beyond experience. In a novel precued lists paradigm, participants were administered abbreviated lists of Stroop trials. For half of the lists, precues led participants to validly expect lists of varying proportion congruency (e.g., mostly congruent [MC], mostly incongruent [MI]; Experiments 1 to 4). The Stroop effect was greater in cued MC relative to uncued MC lists. By contrast, the Stroop effect was equivalent in cued MI and uncued MI lists. Only when preparation was encouraged via a speed manipulation (Experiment 3) or incentives (Experiment 4) did we find evidence of heightened control when an MI list was expected, in the form of a short-lived reduction in the Stroop effect on the first (experience-free) trial. These patterns suggest (a) expectations play a role in the relaxation of cognitive control, independent of experience (as also shown in Experiment 5, wherein expectations were varied while holding experience constant across lists), but (b) experience is the dominant basis for the sustained heightening of cognitive control (after the first trial). Theoretical implications of dissociating the contributions of expectations and experience to cognitive control are discussed, including interpretations of the list-wide proportion congruence effect.