Emma Wiedenmann¹, Myeong-Ho Sohn; ¹The George Washington University, ²The George Washington University

Stimulus frequency has repeatedly been shown to enhance performance across a range of cognitive tasks. However, the mechanisms through which frequency shapes cognitive control remain insufficiently understood. Prior research distinguishes two major learning types: response learning, which is habit-based and reflects direct associations between stimuli and responses, and categorization learning, which is rule-based and requires abstraction over stimulus features (Goodman, 2021). In parallel, the Dimensional Overlap model (Kornblum et al., 1990) identifies two primary forms of conflict: Stimulus–Stimulus conflict, characteristic of the Stroop task, and Stimulus–Response conflict, characteristic of the Simon task. Integrating these frameworks, the present study examines how stimulus frequency influences the engagement of cognitive control across distinct types of conflict and task demands. Across four experiments, participants completed one of the following tasks: (1) a response-learning Stroop task, (2) a response-learning Simon task, (3) a categorization-learning Stroop task, or (4) a categorization-learning Simon task. In each experiment, two colors served as high-frequency stimuli, appearing three times more often than the low-frequency colors, while overall congruency proportions were equated across frequency conditions. This design ensured that any observed effects could be attributed to stimulus frequency rather than differences in conflict frequency. Results showed that under response learning, stimulus frequency modulated cognitive control only in the Simon task, which engages Stimulus–Response conflict, but not in the Stroop task. Under categorization learning, the reverse pattern emerged: frequency modulated performance only in the Stroop task, which engages Stimulus–Stimulus conflict, but not in the Simon task. Taken together, these findings indicate that stimulus frequency attenuates conflict specifically within the domain most relevant to the learning demands of the task.