Dezso Nemeth¹; ¹INSERM

Human predictive processing and learning rely on multiple cognitive systems supported by partially distinct brain networks. These systems do not always cooperate; instead, they may compete as the brain seeks to optimize performance. Prior research shows that reducing the engagement of prefrontal, attention-driven control processes can enhance visual statistical learning and prediction. Competition theory provides a powerful framework for interpreting developmental changes in implicit statistical learning and the formation of predictive representations. In our studies, we examined the developmental trajectory of statistical learning using both cross-sectional and longitudinal designs. Across both approaches, we consistently observed an age-related decline in statistical learning performance. Crucially, executive functions can predict the developmental trajectory of this fundamental learning mechanism. Stronger executive functions—typically improving with age—were associated with the decreasing trajectory of statistical learning. Together, our developmental and cognitive neuroscientific results reveal a systematic antagonistic relationship between prefrontal executive control and visual statistical learning. These findings support a developmental “less-is-more” account in statistical learning and predictive processing, emphasizing that changes in the balance between cognitive control and statistical learning shape how prediction emerges across childhood, adolescence, and adulthood.