Diffusion MRI Indices and Their Relation to Cognitive Impairment in Brain Aging: The Updated Multi-protocol Approach in ADNI3

Brain imaging with diffusion-weighted MRI (dMRI) is sensitive to microstructural white matter (WM) changes associated with brain aging and neurodegeneration. In its third phase, the Alzheimer’s Disease Neuroimaging Initiative (ADNI3) is collecting data across multiple sites and scanners using different dMRI acquisition protocols, to better understand disease effects. It is vital to understand when data can be pooled across scanners, and how the choice of dMRI protocol affects the sensitivity of extracted measures to differences in clinical impairment. Here, we analyzed ADNI3 data from 317 participants (mean age: 75.4 ± 7.9 years; 143 men/174 women), who were each scanned at one of 47 sites with one of six dMRI protocols using scanners from three different manufacturers. We computed four standard diffusion tensor imaging (DTI) indices including fractional anisotropy (FADTI) and mean, radial, and axial diffusivity, and one FA index based on the tensor distribution function (FATDF), in 24 bilaterally averaged WM regions of interest. We found that protocol differences significantly affected dMRI indices, in particular FADTI. We ranked the diffusion indices for their strength of association with four clinical assessments. In addition to diagnosis, we evaluated cognitive impairment as indexed by three commonly used screening tools for detecting dementia and AD: the AD Assessment Scale (ADAS-cog), the Mini-Mental State Examination (MMSE), and the Clinical Dementia Rating scale sum-of-boxes (CDR-sob). Using a nested random-effects regression model to account for protocol and site, we found that across all dMRI indices and clinical measures, the hippocampal-cingulum and fornix (crus)/stria terminalis regions most consistently showed strong associations with clinical impairment. Overall, the greatest effect sizes were detected in the hippocampal-cingulum (CGH) and uncinate fasciculus (UNC) for associations between axial or mean diffusivity and CDR-sob. FATDF detected robust widespread associations with clinical measures, while FADTI was the weakest of the five indices for detecting associations. Ultimately, we were able to successfully pool dMRI data from multiple acquisition protocols from ADNI3 and detect consistent and robust associations with clinical impairment and age.


Supplementary Figure 1. Effect sizes (d-values) from 12 CN sex-and age-matched controls from each ADNI protocol (Supplementary
show the direction of dMRI associations with age in the full WM were consistent across protocols. While direct comparisons of effect sizes were underpowered, findings suggested larger effect sizes for protocol S127, the protocol with greatest total number of diffusion-weighted (b = 1000 s/mm 2 ) and non-diffusion sensitized (b 0 ) gradients, followed by S55, the protocol with the second greatest number of diffusion-weighted and b 0 gradients.

Supplementary Figure 2. (A) d-values
from the ANCOVAs assessing differences in dMRI indices between protocols, for each of the 24 ROIs, after ComBat harmonization; across dMRI indices, the CST was the only region where significant protocol differences remained. (B) We report the number of times each protocol and each dMRI index showed significant differences in pairwise tests between protocols after ComBat harmonization (out of 504 tests per index and 720 tests per protocol). After ComBat, the number of pairwise tests for which each protocol showed significant differences in dMRI indices decreased by 93.8%.

Associations with Age After ComBat
Supplementary Regions that were significant after FDR (q = 0.05) or Bonferroni (α = 0.05) multiple comparisons correction are delineated by a dotted or solid line respectively.

dMRI Effect Sizes Before and After ComBat
Supplementary Figure 3. Effect sizes (beta-values with error bars that represent the standard error) are plotted for the association between each diffusion index and age in CN controls from ADNI2 and ADNI3 protocols pooled together before and after ComBat harmonization. Compared to pre-ComBat analyses, effect sizes were marginally different across indices, but still within the standard error bounds. All associations were significant (FDR q = 0.05) except for FA DTI in the CST and PCR.

Full WM Associations with Age
Supplementary Figure 4. For each protocol, the beta-values (error bars represent the standard error) are plotted for the association between each diffusion index in the full WM and age in CN controls, before and after ComBat harmonization. Compared to pre-ComBat analyses, effect sizes were marginally different across indices, but still within the standard error bounds.

Fx/ST dMRI Associations with Age
Supplementary Figure 5. For each protocol, the beta-values (error bars represent the standard error) are plotted for the association between each diffusion index in the fornix (crus) / stria terminalis (Fx/ST) and age in CN controls, before and after ComBat harmonization. Compared to pre-ComBat analyses, effect sizes were marginally different across indices, but still within the standard error bounds.

GCC dMRI Associations with Age
Supplementary Figure 6. For each protocol, the beta-values (error bars represent the standard error) are plotted for the association between each diffusion index in the corpus callosum genu (GCC) and age in CN controls, before and after ComBat harmonization. Compared to pre-ComBat analyses, effect sizes were only marginally different across indices and protocols, but still within the standard error bounds.

CDR-sob Associations
Supplementary

CGH Clinical Associations
Supplementary Figure 8. For each of six ADNI3 protocols individually and pooled, effect sizes (d-values) are shown for associations between cognitive scores or diagnosis and dMRI indices in the CGH -one of two ROIs that consistently showed significant associations across all four clinical tests and dMRI indices. We note that due to differences in sample size between protocols, effect sizes should not be directly compared. However, the direction of associations was consistent across protocols.

Fx/ST Clinical Associations
Supplementary Figure 9. For each of six ADNI3 protocols individually and pooled, effect sizes (d-values) are shown for associations between cognitive scores or diagnosis and dMRI indices in the Fx/ST -one of two ROIs that consistently showed significant associations across all four clinical tests and dMRI indices. We note that due to differences in sample size between protocols, effect sizes should not be directly compared. However, the direction of associations was consistent across protocols.   Figure 14. Effect size (absolute d-value) maps of WM regions that showed significant differences between CN participants and those with MCI (FDR q = 0.05). Positive associations were detected between MCI diagnosis and (A) AxD DTI , (B) MD DTI , and (C) RD DTI , where higher diffusivity was associated with greater cognitive impairment. Lower (D) FA DTI and (E) FA TDF were associated with greater impairment, but FA DTI associations were detected in fewer regions with weaker effect sizes compared to FA TDF . Light green regions show the largest effect sizes.