Prenatal neural network organization and later executive function development.

Executive function (EF), which includes inhibitory control, working memory, and cognitive flexibility, supports adaptive behavior and predicts long-term academic, social, and mental health outcomes. While EF-related neural networks mature throughout childhood, their earliest developmental origins remain unclear. Large-scale brain systems begin organizing in utero, but whether fetal connectivity prospectively relates to EF has not been tested. We collected resting-state fMRI from 113 fetuses (26-39 weeks gestation). Executive function at age 5 was assessed using the Behavior Rating Inventory of Executive Function-Preschool Version (BRIEF-P), resulting in 52 cases with both usable imaging and outcome data after applying motion and birth-outcome criteria. Whole-brain functional connectivity matrices (197 ROIs) were parcellated into 15 networks, and enrichment analysis identified network pairs with a higher-than-expected density of brain-behavior associations. Six enriched network pairs were associated with global EF, including Frontal, Default Mode Network (DMN), Cerebellar, Salience, and Visual networks. Stronger positive connectivity in Frontal-Salience, DMN-Visual, and Cerebellar-DMN predicted poorer EF, whereas greater negative connectivity in Cerebellar-Visual and Salience-Salience predicted better EF. Subscale analyses revealed both shared and distinct associations across EF domains, with Cerebellar and Salience networks consistently implicated. These findings provide the first evidence linking fetal functional architecture to later EF, highlighting distributed, beyond-frontal systems as early scaffolds of executive control.