Abstract
Recently, several magnetic resonance imaging contrast mechanisms have been shown to distinguish cortical substructure corresponding to selected cortical layers. Here, we investigate cortical layer and area differentiation by automatized unsupervised clustering of high-resolution diffusion MRI data. Several groups of adjacent layers could be distinguished in human primary motor and premotor cortex. We then used the signature of diffusion MRI signals along cortical depth as a criterion to detect area boundaries and find borders at which the signature changes abruptly. We validate our clustering results by histological analysis of the same tissue. These results confirm earlier studies which show that diffusion MRI can probe layer-specific intracortical fiber organization and, moreover, suggests that it contains enough information to automatically classify architecturally distinct cortical areas. We discuss the strengths and weaknesses of the automatic clustering approach and its appeal for MR-based cortical histology.
Highlights
There has been a century-old dominance of the cytoarchitectonic classification of human cortex, mostly based on Brodmann’s cortical parcellation scheme (Brodmann, 1909), there is currently a renewed appreciation for the richness of intra-cortical detail visible in its myeoloarchitecture (Nieuwenhuys, 2013)
When applied to post mortem samples, in particular, high spatial resolution Diffusion magnetic resonance imaging (dMRI) has been used to highlight different layers or layer-complexes in human cortical gray matter (OrosPeusquens et al, 2012; Roebroeck et al, 2012; Bastiani et al, 2013; Kleinnijenhuis et al, 2013; Leuze et al, 2014; Aggarwal et al, 2015). This results from the sensitivity of dMRI to characteristic radial and tangential orientations of myelinated and unmyelinated axons and dendrites. These studies have shown that at high enough spatial resolution cortical layers can be manually delineated based on the average organization of local neurite orientation and this lamination can be distinguished between preselected cortical areas
In this work we investigate whether high-resolution dMRI data acquired post mortem can support automatic segmentation of human cortical layer-complexes and area boundaries in human cortex by unsupervised clustering of their diffusion characteristics
Summary
There has been a century-old dominance of the cytoarchitectonic classification of human cortex, mostly based on Brodmann’s cortical parcellation scheme (Brodmann, 1909), there is currently a renewed appreciation for the richness of intra-cortical detail visible in its myeoloarchitecture (Nieuwenhuys, 2013). When applied to post mortem samples, in particular, high spatial resolution dMRI has been used to highlight different layers or layer-complexes in human cortical gray matter (OrosPeusquens et al, 2012; Roebroeck et al, 2012; Bastiani et al, 2013; Kleinnijenhuis et al, 2013; Leuze et al, 2014; Aggarwal et al, 2015) This results from the sensitivity of dMRI to characteristic radial and tangential orientations of myelinated and unmyelinated axons and dendrites (i.e., neurites). These studies have shown that at high enough spatial resolution cortical layers can be manually delineated based on the average organization of local neurite orientation and this lamination can be distinguished between preselected cortical areas. Beyond manual delineation of cortical layer and areas boundaries, dMRI’s well-structured 3D data seems to lend itself to automatic clustering of architectural properties and localization of boundaries (Nagy et al, 2013)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
