Congruency sequence effect is modulated by the similarity of conflicts

Congruence effect can be modulated by adjacent conflict conditions, producing congruency sequence effect (CSE). However, there are boundary conditions preventing the transfer of cross-conflict CSE. Intensive discussions have been made on whether CSE reflects top-down control or bottom-up associative learning, but neither could perfectly interpret the various boundary conditions. Their imperfection recently inspired an integrative associative learning account of cognitive control, which predicted that conflict similarity affects the magnitude of the cross-conflict CSE. We examined this hypothesis by introducing a compound condition containing both spatial Stroop and Simon components in Experiment 1. The conflict similarity was defined by the amount of component overlap, as manipulated by the polar angle of the target arrow in Experiments 2a and 2b and by the Euclidean distance of the target arrow in Experiments 3a and 3b, respectively. Mixed-effect modeling analyses indicated that in all experiments, the cross-conflict CSEs were positively correlated to the similarity of the two conflict conditions. Specifically, the compound condition with the same amount of Stroop and Simon components generated comparable CSE with both Stroop and Simon conditions (Experiment 1). When the compound condition was more similar to the Stroop than the Simon condition, a trend of larger CSE was observed between the compound conflict and the Stroop condition than between the compound conflict and the Simon condition, and vice versa (Experiments 2 and 3). Our study revealed that the continuum of the cross-conflict CSE was modulated by conflict similarity, hence supporting the associative learning account of cognitive control.