High resolution MRI of the neonatal brain was acquired on a Philips Achieva 3 T system (Best, The Netherlands) using a dedicated 32-channel neonatal head coil and a neonatal positioning device [42 (link)]. Neonates were scanned during natural sleep and monitored throughout the scan (see Supplementary) whilst a paediatrician with experience in MRI was present.
The full dHCP protocol includes structural, functional, and diffusion imaging, but the current study is focused on the latter. Multi-shell High Angular Resolution Diffusion Imaging was acquired over 20 min using a protocol optimised for the neonatal brain [43 (link), 44 (link)]. For all completed scans, each dataset contained 300 volumes of Echo Planar Imaging slices, sampled with four phase-encode directions on four shells with b-values of 0 (n = 20), 400 s/mm2 (n = 64), 1000 s/mm2 (n = 88) and 2600 s/mm2 (n = 128). Acceleration of multiband 4, SENSE factor 1.2, and partial Fourier 0.86 were used, with acquired resolution 1.5 × 1.5 mm, 3 mm slices with 1.5 mm overlap, and TR/TE of 3800/90 ms. Images were processed with denoising [45 (link)], Gibbs ringing suppression [46 (link)], motion and image distortion correction using slice-to-volume reconstruction to a 1.5 mm isotropic resolution [47 (link)], and inter-slice intensity inconsistency correction [48 ].
The full dHCP protocol includes structural, functional, and diffusion imaging, but the current study is focused on the latter. Multi-shell High Angular Resolution Diffusion Imaging was acquired over 20 min using a protocol optimised for the neonatal brain [43 (link), 44 (link)]. For all completed scans, each dataset contained 300 volumes of Echo Planar Imaging slices, sampled with four phase-encode directions on four shells with b-values of 0 (n = 20), 400 s/mm2 (n = 64), 1000 s/mm2 (n = 88) and 2600 s/mm2 (n = 128). Acceleration of multiband 4, SENSE factor 1.2, and partial Fourier 0.86 were used, with acquired resolution 1.5 × 1.5 mm, 3 mm slices with 1.5 mm overlap, and TR/TE of 3800/90 ms. Images were processed with denoising [45 (link)], Gibbs ringing suppression [46 (link)], motion and image distortion correction using slice-to-volume reconstruction to a 1.5 mm isotropic resolution [47 (link)], and inter-slice intensity inconsistency correction [48 ].