Structural network topology relates to tissue properties in multiple sclerosis

Svenja Kiljan,Martijn D Steenwijk,Kim A Meijer,Menno M Schoonheim,Linda Douw,Petra J W Pouwels,Geert J Schenk,Jeroen J G Geurts

doi:10.1007/s00415-018-9130-2

Svenja Kiljan, Martijn D Steenwijk + Show 6 more

Open Access

PDF Available

https://doi.org/10.1007/s00415-018-9130-2

Copy DOI

Export

Save

Cite

Abstract
Highlights/Summary
Full-Text PDF
Similar Papers

Abstract

Listen

ObjectiveAbnormalities in segregative and integrative properties of brain networks have been observed in multiple sclerosis (MS) and are related to clinical functioning. This study aims to investigate the micro-scale correlates of macro-scale network measures of segregation and integration in MS.MethodsEight MS patients underwent post-mortem in situ whole-brain diffusion tensor (DT) imaging and subsequent brain dissection. Macro-scale structural network topology was derived from DT data using graph theory. Clustering coefficient and mean white matter (WM) fiber length were measures of nodal segregation and integration. Thirty-three tissue blocks were collected from five cortical brain regions. Using immunohistochemistry micro-scale tissue properties were evaluated, including, neuronal size, neuronal density, axonal density and total cell density. Nodal network properties and tissue properties were correlated.ResultsA negative correlation between clustering coefficient and WM fiber length was found. Higher clustering coefficient was associated with smaller neuronal size and lower axonal density, and vice versa for fiber length. Higher whole-brain WM lesion load was associated with higher whole-brain clustering, shorter whole-brain fiber length, lower neuronal size and axonal density.ConclusionStructural network properties on MRI associate with neuronal size and axonal density, suggesting that macro-scale network measures may grasp cortical neuroaxonal degeneration in MS.

Highlights

Multiple sclerosis (MS) is a disease of the central nervous system characterized by demyelination in white matter (WM) and gray matter (GM), accompanied by neurodegeneration
To assess whether the regions of interest (ROIs) had similar properties on the macro-scale compared to all other nodes in the connectome, we performed the same correlation for this subset of regions
Higher macroscale regional clustering coefficient was correlated with smaller neuronal size (N = 33; rho = − 0.451; p = 0.008; Fig. 3a) and lower micro-scale axonal density (N = 32; rho = − 0.403; p = 0.022; Fig. 3b), while longer average fiber length was correlated with a larger neuronal size (N = 33; rho = 0.458; p = 0.007; Fig. 3c) and higher axonal density (N = 32; rho = 0.409; p = 0.020; Fig. 3d)

Summary

Introduction

Multiple sclerosis (MS) is a disease of the central nervous system characterized by demyelination in white matter (WM) and gray matter (GM), accompanied by neurodegeneration. Disconnection of brain regions due to WM and GM damage disturbs the optimal information flow through the brain and has been proposed to be a substrate of clinical disability in MS [7, 16, 17]. Innovative computational methods originating from graph theory have enabled researchers to discover topological patterns of brain connectivity that contribute to optimal information distribution through the brain in healthy subjects [1, 2, 28, 29]. In disease, this approach has been used to explain clinical symptoms and model disease progression [1]

Methods

Results

Conclusion