Complementary MR measures of white matter and their relation to cardiovascular health and cognition

Abstract

The microstructural and macrostructural integrity of white matter (WM) underpins efficient brain function, and is known to decline with age and vascular burden. Key aspects of WM health include axonal fibre density, myelination, free-water content, and the presence of tissue damage or lesions. Magnetic Resonance Imaging (MRI) offers multiple complementary sequences to non-invasively estimate these properties in vivo. For example, diffusion-weighted imaging (DWI) provides sensitive measures of microstructure, while T1-weighted and T2-weighted MRI can estimate total WM volume and hyper-intensities, and magnetisation transfer imaging (MT) and T1:T2 ratios can indicate myelin content. In this study, we leveraged all of these MRI-derived measures in a large population-based cohort (Cam-CAN) to identify latent WM factors and test how these factors relate to cardiovascular health and cognitive performance. Among 11 commonly-used WM metrics [Fractional Anisotropy (FA); Mean Signal Diffusion (MSD); Mean Signal Kurtosis (MSK); Neurite Density Index (NDI); fibre Orientation Dispersion Index (ODI); Free water volume faction (F_iso); spread of Mean Signal Diffusivity values (MSDvar); Magnetisation Transfer Ratio (MTR); T1:T2 ratio; volume of White Matter Hyper-Intensities (WMHI); White Matter Volume (WMV)], latent factor analysis showed that four factors were needed to explain 89% of the variance, which we interpreted in terms of (1) fibre density/myelination, (2) free-water / tissue damage, (3) fibre-crossing complexity and (4) microstructural complexity. These factors showed distinct effects of age and sex. To test the validity of these factors, we related them to measures of cardiovascular health and cognitive performance. Specifically, we ran path analyses linking (1) cardiovascular factors to the WM factors, and (2) the WM factors to cognitive measures. Even after adjusting for age and sex, we found that a vascular factor related to pulse pressure predicted the WM factor capturing free-water/tissue damage, and that several WM factors made unique predictions for fluid intelligence and processing speed. Our results show that there is both complementary and redundant information across common MR measures of WM, and their underlying latent factors may be useful for pinpointing the differential causes and contributions of white matter health in aging.

Fig. 1

Example Brain maps of the 8 WM measures from a single subject aged 22 included in the analyses. We show a young participant’s data from 6 diffusion metrics (FA, MSD, MSK, NDI, ODI, F_iso), plus the additional measures of MTR and T1/T2 ratio. Not shown are the raw T1w + T2w images from which total WM volume (TWM) was calculated, and the other two global measures of White Matter Hyperintensities (WMHI), which were computed across WM ROIs using the T1w and T2w images and the SAMSEG algorithm, and PSMD, which was computed across the WM ROIs using the DWI FA and MSD measures.

Fig. 2

Values for all 11 WM metrics by sex and across age after removing outliers. (A) Fractional Anisotropy (FA); (B) Mean Signal Diffusion (MSD); (C) Mean Signal Kurtosis (MSK); (D) Neurite Density Index (NDI) from NODDI; (E) Orientation Dispersion Index (ODI) from NODDI; (F) Volume Fraction of Free isotropic water (F_iso) from NODDI; (G) MSDvar - range of MSD values across 27 WM tracks; (H) MTR values across age; (I) T1w vs. T2w ratio; (J) Volume of White Matter Hyper-Intensities (WMHI) normalised by head size. (K) total White Matter Volume (WMV) volume corrected for head size. The proportion of variance (R² explained by the Linear (A) and Quadratic (Q) effects of age (after accounting for main effect of sex and sex-by-age interactions) is shown at the top of each panel.

Fig. 3

Correlation matrix for all 11 WM measures across participants before (left) and after (right) correcting for sex, linear and quadratic age effects, and the sex-by-age interactions. Metrics are ordered according to the dendrogram of agglomerative similarity between measures.

Fig. 4

PCA cross-validation. The figure shows the ekf cross-validated error as a function of number of principal components retained, along with the residual variance (inverse of variance explained) in orange. Based on the ekf, we selected 4 PCs and performed factor analysis with 4 factors.

Fig. 5

Loadings and Factor scores across age. We show factor loadings across all 11 WM measures and how the factor scores varied with age. We interpret Factor 1 (WMF1) as a microstructural properties/myelination factor. Factor 2 (WMF2) as free-water/tissue damage factor. Factor 3 (WMF3) as fibre-crossing complexity factor. Factor 4 (WMF4) as microstructural complexity.

Fig. 6

Path models relating cardiovascular factors to WM factors (left panels) and WM factors to Cognitive factors (right panels), without (top panels) and with (bottom panels) adjustment of outcome variables for Age and Sex effects. Solid coloured lines indicate significant paths. R-square values are shown as percentages below each dependent variable. Age and Sex were included as confounds of no interest, represented by dashed lines. LVF = latent vascular factor; WMF = white matter factor; IQ = fluid intelligence; PS = processing speed; Mem = episodic memory.

Fig. 7

Path Model for effects of WM on Cognition, now additionally controlling for polygenic scores (PGS) for cognitive ability.