Tract-specific statistics based on diffusion-weighted probabilistic tractography

Andrew T Reid,Julia A Camilleri,Simon B Eickhoff,Felix Hoffstaedter

doi:10.1038/s42003-022-03073-w

Abstract

Diffusion-weighted neuroimaging approaches provide rich evidence for estimating the structural integrity of white matter in vivo, but typically do not assess white matter integrity for connections between two specific regions of the brain. Here, we present a method for deriving tract-specific diffusion statistics, based upon predefined regions of interest. Our approach derives a population distribution using probabilistic tractography, based on the Nathan Kline Institute (NKI) Enhanced Rockland sample. We determine the most likely geometry of a path between two regions and express this as a spatial distribution. We then estimate the average orientation of streamlines traversing this path, at discrete distances along its trajectory, and the fraction of diffusion directed along this orientation for each participant. The resulting participant-wise metrics (tract-specific anisotropy; TSA) can then be used for statistical analysis on any comparable population. Based on this method, we report both negative and positive associations between age and TSA for two networks derived from published meta-analytic studies (the “default mode” and “what-where” networks), along with more moderate sex differences and age-by-sex interactions. The proposed method can be applied to any arbitrary set of brain regions, to estimate both the spatial trajectory and DWI-based anisotropy specific to those regions.

Highlights

Diffusion-weighted neuroimaging approaches provide rich evidence for estimating the structural integrity of white matter in vivo, but typically do not assess white matter integrity for connections between two specific regions of the brain
The simplest model of diffusion used for Diffusion-weighted imaging (DWI) analysis is the diffusion tensor, which assumes a Gaussian distribution with one principal and two secondary axes, corresponding to the first three eigenvectors of observed diffusion across gradient orientations[4,5]
For default mode network (DMN) in particular, we found that most positive associations occurred in regions proximal to regions of interest (ROIs), where the potential confound of crossing fibres is likely more pronounced[53]

Summary

Introduction

Diffusion-weighted neuroimaging approaches provide rich evidence for estimating the structural integrity of white matter in vivo, but typically do not assess white matter integrity for connections between two specific regions of the brain. We present a method for deriving tract-specific diffusion statistics, based upon predefined regions of interest. The resulting participant-wise metrics (tract-specific anisotropy; TSA) can be used for statistical analysis on any comparable population. Based on this method, we report both negative and positive associations between age and TSA for two networks derived from published meta-analytic studies (the “default mode” and “what-where” networks), along with more moderate sex differences and age-by-sex interactions. The proposed method can be applied to any arbitrary set of brain regions, to estimate both the spatial trajectory and DWI-based anisotropy specific to those regions. DWI is based on the attenuation of the T2-weighted MRI signal in the presence of a field gradient, which indicates the degree to which diffusion is unrestricted in brain tissue in the direction of that gradient

Objectives

Methods

Results