An age-specific atlas for delineation of white matter pathways in children aged 6-8 years

Introduction: Diffusion magnetic resonance imaging (MRI) allows noninvasive assessment of white matter connectivity in typical development and of changes due to brain injury or pathology. Probabilistic white matter atlases allow diffusion metrics to be measured in specific white matter pathways, and are a critical component in spatial normalization for group analysis. However, given the known developmental changes in white matter it may be suboptimal to use an adult template when assessing data acquired from children. Methods: By averaging subject-specific fiber bundles from 28 children aged from 6 to 8 years, we created an age-specific probabilistic white matter atlas for 12 major white matter tracts. Using both the newly developed and Johns Hopkins adult atlases, we compared the atlas with subject-specific fiber bundles in two independent validation cohorts, assessing accuracy in terms of volumetric overlap and measured diffusion metrics. Results: Our age-specific atlas gave better overall performance than the adult atlas, achieving higher volumetric overlap with subject-specific fiber tracking and higher correlation of fractional anisotropy (FA) measurements with those measured from subject-specific fiber bundles. Specifically, estimates of FA values for corticospinal tract, uncinate fasciculus, forceps minor, cingulate gyrus part of the cingulum, and anterior thalamic radiation were all significantly more accurate when estimated with an age-specific atlas. Discussion: The age-specific atlas allows delineation of white matter tracts in children aged 6-8 years, without the need for tractography, more accurately than when normalizing to an adult atlas. To our knowledge, this is the first publicly available probabilistic atlas of white matter tracts for this age group.