Fig. 1: Rich-club hub nodes in the neonatal brain.

A The regions of the µBrain atlas (coloured regions; top) are subdivided into parcels of ~90 vertices (bottom). B The group consensus connectivity matrix thresholded to retain the top 15% of edges. Nodes are ordered according to the µBrain region of which they are a subset. C The degree distribution (top) of the thresholded, group consensus network. The rich-club coefficient (middle) is calculated over all degree thresholds in the empirical data (black line) and compared to the rich-club coefficient of degree sequence preserving null networks (grey line). Normalised rich-club coefficient (bottom; red line) values > 1 indicate greater rich-club organisation than expected by chance. The dashed vertical line indicates the 90th percentile for node degree (k = 161), above which nodes were considered network hubs. D Edge length distributions for rich (connections between hub nodes), feeder (connections between hubs and non-hubs), and local (connections between non-hubs). E The location of hub nodes (red) in the neonatal brain. F The hubness for µBrain regions, calculated as the proportion of a region’s subparcels that were identified as a hub (hubness thresholded at 0.1 for visualisation).