Disrupted cortical folding and cognitive outcomes in extremely preterm children at mid-childhood

• Extremely preterm birth disrupts cortical development into mid-childhood, with temporal and cingulate regions most affected • Rather than individual metrics, multivariate patterns of cortical morphology across thickness, surface area, curvature, gyrification index, and sulcal depth emerged as robust markers of later cognitive outcomes • Extremely preterm children with cognitive problems show distinct structural connectivity and hub organization compared with peers without problems • Lower gestational age, prolonged ventilator support, and bronchopulmonary dysplasia likely contribute to cognitive heterogeneity among extremely preterm During the third trimester, the cerebral cortex undergoes rapid surface expansion and folding, processes disrupted by preterm birth and associated with later cognitive problems. Although atypical cortical folding has been observed early in extremely preterm children, it remains unclear which alterations persist into mid-childhood and how they relate to cognition. Cognitive outcomes are heterogeneous but it remains unclear whether children with cognitive problems show distinct structural network-level organization. We compared cortical morphometrics at 10 years between extremely-preterm and term-born children, examined associations with cognition at 12 years, and assessed whether structural covariance differed between extremely preterm children with and without cognitive problems. Cortical morphometrics were examined in 54 extremely-preterm (25±1.0 weeks) and 38 term-born (40±1.1 weeks) children using structural MRI processed with FreeSurfer, and cognition was assessed at 12 years using Wechsler Intelligence Scale for Children. PCA was applied to all cortical morphometric measures to examine associations with cognitive outcomes. Extremely preterm children showed widespread cortical thinning, altered surface area, and region-specific sulcal depth differences, shallower in parietal and orbitofrontal, deeper in cingulate and postcentral regions (β= -1.15 to 1.42; all q<0.05). Gyrification was lower in orbitofrontal, temporal, and parietal cortices (β= -0.51 to -1.07; all q<0.05). Four principal components of cortical morphology at 10 years (PC4, PC8, PC11, and PC13) were associated with later cognitive outcomes. Extremely preterm children with cognitive problems exhibited distinct structural covariance and hub organization. These findings suggest lasting cortical reorganization after extremely preterm birth and support multivariate cortical patterns as markers of later cognitive risk.