Early childhood development of cortical structural covariance networks

Abstract

BACKGROUND: Structural covariance networks of cortical thickness are composed of regions with highly correlated variation. These networks change with age and are altered in neuropsychiatric disease. Very little is known about the development of structural covariance networks in early childhood, a period of rapid cortical thickness growth. METHODS: Structural and maturational covariance of the default, dorsal attention, visual and sensorimotor networks was studies in 116 children who had longitudinal 3T MRI scans after birth, and at 1 and 2 years. Vertex-based cortical thickness was estimated using longitudinal tissue segmentation and surface reconstruction methods. Condition-by-covariate analyses using seed regions in each network determined significant correlations in the whole-brain patterns of structural covariance and maturational coupling for each network. RESULTS: Each structural covariance network had isolated correlation patterns, suggesting that in early childhood, the cortex has structural covariance within local but not distributed regions. The thickness growth of the posterior cingulate cortex PCC is significantly correlated with the growth of lateral parietal, temporal and frontal regions. The maturational covariance of intra-parietal sulcus (IPS) seed has significant correlations in temporal and frontal regions. The primary visual seed was mostly correlated with local regions and the primary motor seed was mostly correlated with the pre- and post-motor cortices, and the lateral occipital regions. CONCLUSIONS: Structural covariance of default and dorsal attention networks are not present in the first years of life and evolve from local to distributed connections during development. In contrast, the structural covariance of primary visual and sensorimotor networks in early childhood exhibits adult-like patterns.