A computational method for longitudinal mapping of orientation-specific expansion of cortical surface in infants

Abstract

The cortical surface of the human brain expands dynamically and regionally heterogeneously during the first postnatal year. As all primary and secondary cortical folds as well as many tertiary cortical folds are well established at term birth, the cortical surface area expansion during this stage is largely driven by the increase of surface area in two orthogonal orientations in the tangent plane: 1) the expansion parallel to the folding orientation (i.e., increasing the lengths of folds) and 2) the expansion perpendicular to the folding orientation (i.e., increasing the depths of folds). This information would help us better understand the mechanisms of cortical development and provide important insights into neurodevelopmental disorders, but still remains largely unknown due to lack of dedicated computational methods. To address this issue, we propose a novel method for longitudinal mapping of orientation-specific expansion of cortical surface area in these two orthogonal orientations during early infancy. First, to derive the two orientation fields perpendicular and parallel to cortical folds, we propose to adaptively and smoothly fuse the gradient field of sulcal depth and also the maximum principal direction field, by leveraging their region-specific reliability. Specifically, we formulate this task as a discrete labeling problem, in which each vertex is assigned to an orientation label, and solve it by graph cuts. Then, based on the computed longitudinal deformation of the cortical surface, we estimate the Jacobian matrix at each vertex by solving a least-squares problem and derive its corresponding stretch tensor. Finally, to obtain the orientation-specific cortical surface expansion, we project the stretch tensor into the two orthogonal orientations separately. We have applied the proposed method to 30 healthy infants, and for the first time we revealed the orientation-specific longitudinal cortical surface expansion maps during the first postnatal year.

Mapping hemispheric asymmetries of the macaque cerebral cortex during early brain development

“…One reason for these similar hemispheric asymmetric clusters of the sulcal depth and surface area is that the uneven deepening of sulci is usually accompanied by uneven expansion of surface area. However, as uneven widening or lengthening of sulci or gyri would also result in the expansion of surface area (Xia, Wang, et al, ), there are still many different asymmetric clusters between these two features. The hemispheric asymmetries of vertex positions can provide important information on cortical shape asymmetries between right and left hemispheres.…”

Fetal cortical surface atlas parcellation based on growth patterns

“…The STS appears between 25 and 27 GW and develops between 28 and 29 GW (Clouchoux, Du Plessis, et al, ). Our parcellation maps are based on the growth patterns of local surface area, which can be caused by multiple complicated changing patterns of surface area, for example, the emerging of new folds, growth along the length of existing folds, and growth along the depth of existing folds (Xia et al, ). Hence, the growth patterns of local surface area are related to the gross shapes of cortical anatomical structures.…”

Robust 3D brain segmentation in magnetic resonance image with weighted feature fusion

“…According to previous research on supervoxels applied on brain image segmentation [14], the number of supervoxels is empirically set to 5000. Moreover, according to our other research on diffusion method [30], the smooth parameter α is set as 0.5 empirically. The experiments were performed on a macbook pro with Intel Core 2.5 GHz and a 16 GB RAM, with NVIDIA GeForce GT 750M 2 GB.…”