Fiber clustering versus the parcellation-based connectome

Abstract

We compare two strategies for modeling the connections of the brain’s white matter: fiber clustering and the parcellation-based connectome. Both methods analyze diffusion magnetic resonance imaging fiber tractography to produce a quantitative description of the brain’s connections. Fiber clustering is designed to reconstruct anatomically-defined white matter tracts, while the parcellation-based white matter segmentation enables the study of the brain as a network. From the perspective of white matter segmentation, we compare and contrast the goals and methods of the parcellation-based and clustering approaches, with special focus on reviewing the field of fiber clustering. We also propose a third category of new hybrid methods that combine aspects of parcellation and clustering, for joint analysis of connection structure and anatomy or function. We conclude that these different approaches for segmentation and modeling of the white matter can advance the neuroscientific study of the brain’s connectivity in complementary ways.

Fiberprint: A subject fingerprint based on sparse code pooling for white matter fiber analysis

“…Since tractography may output thousands of fiber trajectories, early work has fo-cused on finding simplified quantitative descriptions of white matter connections by grouping fiber trajectories into anatomically meaningful bundles [43].…”

Do we need to revise the tripartite subdivision hypothesis of the human subthalamic nucleus (STN)? Response to Alkemade and Forstmann

“…Furthermore, the fact that both graduated and abrupt transition zones are present cortically (Schleicher et al, 1999) and subcortically (Haber et al, 2000) has clear methodological implications beyond those raised in the STN (Alkemade, and Forstmann, 2014). The widely used hardclustering methods for brain parcellation (Ruschel et al, 2014;Caspers et al, 2013;O'Donnell et al, 2013;Solano-Castiella et al, 2011) will be unable to accurately model or demonstrate graduated architectural features and instead will artificially provide "anatomically distinct boundaries" as necessary methodological by-products (Gan et al, 2007;Jain, 2010;Accolla et al, 2014). Techniques that are capable of representing greater degrees of subtlety may ultimately provide more anatomically congruent models, but at the cost of significantly increasing the representational complexity.…”

A hybrid approach for optimal automatic segmentation of White Matter tracts in HARDI

“…This method combines for more efficiently clustering based on a similarity measure considering shapes and positions of tracts using a QuickBundles algorithm with a priori anatomical information given by an atlas bundle constructed in this study. Like other authors [8], we believe that hybrid techniques will be instrumental to move the field of automated segmentation techniques forward. Our method is based on two-level of clustering.…”