Proactive and reactive cognitive control in the absence of learning and memory confounds: Evidence from a cross-modal trial-unique Stroop task.

Goal-directed behaviour is typically conceptualized as striking a balance between two antagonistic cognitive control states such as proactive and reactive control, as demonstrated by conflict phenomena such as the list-wide proportion congruency and congruency sequence effects. However, control-based explanations for these phenomena have come under criticism due to low-level associative regularities that are frequently confounded with conflict manipulations within these experimental designs. In the present study, a novel Stroop paradigm referred to as the "trial-unique Stroop task" was developed to examine whether these effects could be observed in the absence of low-level associative regularities. On each trial, participants typed a word they heard spoken aloud while ignoring a word visually displayed on the screen. Importantly, each word only appeared in a single trial throughout the experiment, and because stimuli and responses were never repeated, there were no low-level associative regularities across trials. Using this paradigm, we observed both congruency sequence (Experiment 1) and list-wide proportion congruency (Experiment 2) effects, providing the strongest evidence to date for control-based explanations of these phenomena. Split-half analyses revealed much higher reliability than traditional colour-word Stroop tasks for the congruency effect (