Mapping population-based structural connectomes

Abstract

Advances in understanding the structural connectomes of human brain require improved approaches for the construction, comparison and integration of high-dimensional whole-brain tractography data from a large number of individuals. This article develops a population-based structural connectome (PSC) mapping framework to address these challenges. PSC simultaneously characterizes a large number of white matter bundles within and across different subjects by registering different subjects' brains based on coarse cortical parcellations, compressing the bundles of each connection, and extracting novel connection weights. A robust tractography algorithm and streamline post-processing techniques, including dilation of gray matter regions, streamline cutting, and outlier streamline removal are applied to improve the robustness of the extracted structural connectomes. The developed PSC framework can be used to reproducibly extract binary networks, weighted networks and streamline-based brain connectomes. We apply the PSC to Human Connectome Project data to illustrate its application in characterizing normal variations and heritability of structural connectomes in healthy subjects.

Test–retest reproducibility of white matter parcellation using diffusion MRI tractography fiber clustering

“…Lopes, Leclerc, Derambure, & Tyvaert, 2014;Bonilha et al, 2015;Buchanan, Pernet, Gorgolewski, Storkey, & Bastin, 2014;Dennis et al, 2012;Duda, Cook, & Gee, 2014;Schumacher et al, 2018;Smith et al, 2015;Vaessen et al, 2010;Zhao et al, 2015;Zhang, Descoteaux, et al, 2018) and on anatomical fiber tracts (Besseling et al, 2012;Cousineau et al, 2017;Heiervang, Behrens studies of fiber clustering, to our knowledge. Studies have suggested that fiber clustering approaches have advantages in parcellating the white matter in a highly consistent way, aiming to reconstruct fiber parcels/tracts corresponding to the white matter anatomy (Ge et al, 2012;Sydnor et al, 2018;Zhang, Wu, Norton, et al, 2018;Ziyan, Sabuncu, Eric, Grimson, & Westin, 2009), while corticalparcellation-based methods could be less consistent considering factors such as the variability of intersubject cortical anatomy, the dependence on registration between dMRI and structural images, and the presence of false positive/negative connections (Amunts et al, 1999;Fischl, Sereno, Tootell, & Dale, 1999;Maier-Hein et al, 2017;Sinke et al, 2018;Zhang, Descoteaux, et al, 2018). However, to our knowledge, test-retest reproducibility of the cortical-parcellation-based and fiber clustering white matter parcellation strategies has not yet been quantitatively compared.…”

“…Fifth, in the present study, we compared two fiber clustering and cortical-parcellation-based strategies that are widely used but relatively traditional approaches. In the past few years, there have been methods designed to improve test-retest reproducibility for constructing cortical-parcellation-based connectomes by including additional pre-and/or postprocessing steps, for example, dilating gray matter regions (Zhang, Descoteaux, Zhang, et al, 2018), constructing continuous connectome matrices (Moyer, Gutman, Faskowitz, Jahanshad, & Thompson, 2017), and filtering out implausible fiber streamlines (Smith et al, 2015). A further study could include comparison of test-retest reproducibility between the fiber clustering and cortical-parcellation-based strategies with advanced processing.…”

“…Other groups have used diffusion measures such as fractional anisotropy (FA) and mean diffusivity (MD) computed from the voxels through which the parcellated fibers pass (Besseling et al, ; Ciccarelli et al, ; Duan, Zhao, He, & Shu, ; Kristo et al, ; Papinutto, Maule, & Jovicich, ; Pfefferbaum, Adalsteinsson, & Sullivan, ; Vollmar et al, ; Yendiki, Reuter, Paul, Diana Rosas, & Fischl, ). While multiple studies have assessed test–retest reproducibility of white matter parcellations using cortical‐parcellation‐based strategies, for example, on the brain connectome network (Besson, Lopes, Leclerc, Derambure, & Tyvaert, ; Bonilha et al, ; Buchanan, Pernet, Gorgolewski, Storkey, & Bastin, ; Dennis et al, ; Duda, Cook, & Gee, ; Schumacher et al, ; Smith et al, ; Vaessen et al, ; Zhao et al, ; Zhang, Descoteaux, et al, ) and on anatomical fiber tracts (Besseling et al, ; Cousineau et al, ; Heiervang, Behrens, Mackay, Robson, & Johansen‐Berg, ; Kristo et al, ; Lin et al, ; Papinutto et al, ; Tensaouti, Lahlou, Clarisse, Lotterie, & Berry, ; Wang et al, ; Yendiki et al, ), there are no existing studies of fiber clustering, to our knowledge. Studies have suggested that fiber clustering approaches have advantages in parcellating the white matter in a highly consistent way, aiming to reconstruct fiber parcels/tracts corresponding to the white matter anatomy (Ge et al, ; Sydnor et al, ; Zhang et al, ; Zhang, Wu, Norton, et al, ; Ziyan, Sabuncu, Eric, Grimson, & Westin, ), while cortical‐parcellation‐based methods could be less consistent considering factors such as the variability of intersubject cortical anatomy, the dependence on registration between dMRI and structural images, and the presence of false positive/negative connections (Amunts et al, ; Fischl, Sereno, Tootell, & Dale, ; Maier‐Hein et al, ; Sinke et al, ; Zhang, Descoteaux, et al, ).…”

Heritability of the Mouse Brain Connectome

“…Other recent studies have detected significant global heritability of FA, AD and RD in human populations [6,30] . And studies have demonstrated heritability of specific regions of connectivity, in human [31,32]. However, the work described here is the first to bring together a robust study of heritability of all these metrics (volume, FA, AD, RD, connection strength) in the mouse.…”

Symmetric Bilinear Regression for Signal Subgraph Estimation

“…We applied our method to the Human Connectome Project (HCP) dataset [26], exploring the association between the brain connectome and two cognitive abilities, auditory language comprehension ability and oral reading ability. The dataset contains sMRI and dMRI data for 1065 subjects and for each subject, a weighted brain network of fiber counts among 68 regions was constructed by a state-of-the-art dMRI processing pipeline [27].…”