Skip to main content
. 2013 Jul 3;7:302. doi: 10.3389/fnhum.2013.00302

Figure 1.

Figure 1

The illustration of analysis steps to construct networks from MRI. The cortical thicknesses are estimated in native spaces (A, section 2.2.1), normalized into the template space (B, section 2.2.2), and smoothed by a heat kernel based on Laplace-Beltrami eigenfunctions (C, section 2.2.3). Partial correlations are computed over 148 ROIs (D, section 2.2.4) and binarized to obtain adjacency matrices at a certain rewiring cost (E, section 2.2.5).