Fig. 2: Organizational differences across PMA and GA at birth.
From: Premature birth changes wiring constraints in neonatal structural brain networks
a Modularity significantly decreased across PMA (p = 4.10 x 10-4) and GA at birth (p = 6.74 x 10-6) in variable density networks, but modularity significantly increased across PMA (p < 2.00 x 10-16) and decreased across GA at birth (p = 2.90 x 10-4) in density-controlled networks. Global efficiency significantly increased across both PMA (p < 2.00 x 10-16) and GA at birth (p = 3.82 x 10-5) in variable density networks. However, with density-controlled networks global efficiency significantly decreased across PMA (p < 2.00 x 10-16) but no significant effect was found across GA at birth (p = 0.462). b The number of rich club connections significantly increased across PMA in both variable density (p < 2.00 x 10-16) and density-controlled networks (p = 2.31 x 10-3). The number of rich club connections also significantly increased across GA at birth in variable density networks (p = 0.022), but no significant effect was found with density-controlled networks (p = 0.059). In addition, the average length of rich club connections (Euclidean distance) significantly decreased across PMA in both variable density (p < 2.00 x 10-16) and density-controlled networks (p < 2.00 x 10-16). The average length of rich club connections significantly increased across GA at birth in both variable density (p = 1.06 x 10-4) and density-controlled networks (p = 0.029). All results presented are generalized additive models. The shaded area indicates 95% confidence intervals. The grey dotted line indicates the cut-off for term birth (37 weeks GA or later is term-born). *** indicates p < 0.001, ** indicates p < 0.01, * indicates p < 0.05.
