Identifying Executive Function Profiles in Children's Design Fluency: A Clustering Study

Abstract

Executive functions continue to develop throughout childhood and vary considerably across children. Design fluency tasks are commonly used to assess executive functioning through design production, repetitions, and strategy use. When averaged across the full task, these measures summarize performance but do not show how they change as the task progresses. Whether examining these measures minute by minute provides additional information about variability across children remains unclear. We examined this question in 113 neurotypical children aged 7-13 years using a tablet-based design fluency task. To identify performance profiles, we applied K-means clustering separately to design production, repetitions, and strategy use averaged across the task and measured across five one-minute intervals. Overall task measures identified two overlapping profiles. One showed lower design production and lower proportions of repetitions and strategies, whereas the other showed higher production and higher proportions of repetitions and strategies. Minute-by-minute clustering produced nearly identical assignments, but weaker separation, suggesting that temporal information added little to profile identification. However, examining changes over time revealed subtler differences: strategy use followed broadly similar patterns across profiles, but showed a pronounced between-profile difference during the first minute that narrowed thereafter. In contrast, the difference in repetition rates increased as the task progressed, a pattern that may partly reflect differences in overall design production. Profiles did not differ significantly in age, sex, or WISC subtest scores. These findings suggest that overall task measures more clearly distinguish performance profiles, whereas minute-by-minute measures provide complementary information about how these differences evolve over time.