Transcriptomic divergence of network hubs in the prenatal human brain.

Connections in the human brain are not uniformly distributed; instead, a dense network of long-range projections converge on highly connected hub regions located in paralimbic and association cortices. Hub connectivity is strongly influenced by genetic factors but the molecular cues guiding the foundation of these structures remain poorly understood. Here, we combined high-resolution diffusion MRI data acquired from 208 term-born neonates with spatially resolved prenatal gene expression data to investigate the molecular correlates of network hub formation at mid-gestation. We identified robust hub architecture in the neonatal connectome and mapped these structural hubs to corresponding cortical regions in the µBrain prenatal digital brain atlas. Transcriptomic analysis revealed differential gene expression in network hubs at mid-gestation, with genes positively associated with hub status supporting the establishment of early neuronal circuitry and predominantly expressed in the transient subplate and intermediate zones. Hub genes were expressed by excitatory neurons, including subplate neurons and intratelencephalic projection neurons in deep cortical layers, overlapped with markers of cortical expansion and interhemispheric connectivity in adulthood and were associated with common neurodevelopmental disorders. Our study identifies prenatal transcriptomic signatures of network hubs in the neonatal human brain, demonstrating how early gene expression programs can shape brain network connectivity from mid-gestation.