Knopman, David S.

AbstractBackgroundLarge neuroimaging studies like ADNI‐4 now include T1‐weighted, T2‐weighted and FLAIR images at isotropic1mm resolution for investigation of both neurodegenerative and cerebrovascular disease. These three input types allow segmentation of GM, WM, DeepGM, WMH, and CSF compartments. Additionally, this supports automated assessment of enlarged perivascular spaces (PVS). Our objective was to propose a unified framework for segmentation of intracranial classes and PVS while reducing human effort.MethodFLAIR and T2‐weighted images are aligned to the T1‐weighted image.. Standard SPM12 unified segmentation using all three input images establishes deformations between subject and template space and bias field corrections. Extended tissue priors including deep GM are deformed into subject space. Adding mean intensities from high probability deep GM voxels as seeds and using the extended prior set, segmentation is re‐started, refining model parameters, deformations and bias‐field estimates. The procedure repeats, adding a WMH channel, resulting in a segmentation including GM, WM, CSF, deep GM, WMH and nuisance classes. Frangi filters are applied to the T1‐ and T2‐ weighted images. Filter response images are smoothed, averaged, masked and inserted in place of the CSF prior. Probability maps of vessel‐like shapes with CSF‐like contrast are estimated to identify PVS. Visual grading was also performed for comparison.ResultThe method was applied to 300 participants. Visual quality control found two instances of poor WMH over‐segmentation. Perivascular space maps were all visually reasonable. Standard segmentation classes were visibly superior to single T1‐based segmentation. Figures 1,2 present examples. Pearson correlations of PVS fractional volume with multiple human raters’ PVS counts in CSO/CR was over 0.7 in a subset of 125 participants (Figure 3).ConclusionModern dementia protocols include high‐resolution T1‐weighted, T2‐weighted, and FLAIR images that can be used together as inputs to extend probabilistic segmentation adding deep grey and WMH classes with automated PVS assessment.

AbstractBackgroundWe have previously shown that 3D T1‐weighted and 3D T2‐FLAIR brain MRI can potentially be re‐identified using face recognition, but this has not been tested for other MRI sequences commonly used in imaging studies of aging and dementia. Here we compare match rates of automated face recognition to re‐identify 3D T1‐weighted, 3D T2‐FLAIR, 3D T2‐weighted, 2D Axial T2‐FLAIR, multi‐echo GRE (T2*), diffusion MRI, functional MRI, ASL, and T2* GRE images.MethodWe used a dataset of 182 Mayo Clinic participants each with MRI and face photographs. MRIs used Siemens and GE vendor‐product sequences similar to ADNI3 and ADNI4 protocols. We created face reconstructions (“basic”, with minimal preprocessing, and “advanced” that automatically filled missing face regions using a template) from each MRI sequence, and we trained a Microsoft Azure PersonGroup face recognition classifier to recognize these MRI‐based faces. We then queried the classifier with five face photographs for each participant, and we recorded how often the classifier chose the correct MRI as the top‐ranked match. For image types with match rates > = 10%, we then repeated the experiment after de‐facing each MRI with mri_reface, our leading software that replaces faces in each image (“re‐faces”) with an average face.ResultResults are presented in Table 1. 3D T2‐FLAIR, 3D T1‐weighted, and 3D T2‐weighted images were matched at 96‐98%, and mri_reface reduced all to 8%. Axial T2‐FLAIR and ME‐GRE were matched at 32‐45%, and mri_reface reduced both to 1%. T2*‐GRE matched at 10%, and mri_reface reduced this to 1%. ASL matched at 8%, while dMRI and fMRI matched at only 0‐2%.ConclusionAmong current‐generation vendor‐product sequences, we recommend that de‐facing be applied before research data sharing of T1‐weighted, T2‐weighted, T2‐FLAIR, T2*, and ME‐GRE images. Minimal match rates for current‐generation product fMRI, ASL, and fMRI images suggest that they present low risk of re‐identification and can be shared without de‐facing, but this conclusion should be re‐evaluated if they are acquired without fat‐suppression, with a larger field of view, or if newer developments reduce the current levels of artifacts and distortion around the face.

AbstractBackgroundCo‐morbid Alzheimer’s disease (AD) pathology in the form of amyloid‐ß plaques and tau neurofibrillary tangles are frequently observed in people with probable dementia with Lewy bodies (DLB). Neurite Orientation Dispersion and Density Imaging (NODDI) is an emerging technique to characterize white matter (WM) tissue microstructure, although NODDI alterations and associations with AD biomarkers remain undefined in DLB. The objectives of this study are to compare NODDI and DTI parameters between DLB and controls and (ii) evaluate the pattern of DTI and NODDI changes associated with amyloid and tau deposition.MethodConsecutive patients within the DLB spectrum (n = 45), including probable DLB (n = 32) and probable Mild Cognitive Impairment with Lewy bodies (MCI‐LB) (n = 13) from the Mayo Clinic ADRC underwent multi‐shell diffusion MRI, [18F]‐Flortaucipir and [11C]‐PiB PET. A comparison group of 45 age‐ and gender‐matched clinically unimpaired controls (CU) was included. Voxelwise non‐parametric permutation models were used to determine pair‐wise group differences and assess the associations of AD biomarkers with fractional anisotropy (FA), mean diffusivity (MD), tissue‐weighted neurite density index (tNDI; NDI x (1 – Free water fraction)) and orientation dispersion index (ODI), controlling for age. Complementary analyses were performed using regions‐of‐interest (ROI) obtained from the JHU atlas.ResultCompared to CU, the DLB spectrum group exhibited widespread decreases in FA and tNDI and increases in MD and ODI, with the most pronounced effects in the cingulum bundles, uncinate, inferior and superior longitudinal fasciculi (Fig. 1). Greater [18F]‐Flortaucipir uptake was also associated with lower FA and tNDI, and higher MD and ODI in the DLB spectrum compared to CU, independently of age and [11C]‐PiB uptake (Fig. 2). In contrast, no associations were found between [11C]‐PiB uptake and any of the DTI or NODDI parameters. Complementary ROI analyses corroborated these findings.ConclusionWidespread loss of WM integrity was observed on DTI and NODDI in probable DLB and prodromal DLB. These findings were associated with greater cortical tau, but not with amyloid‐ß deposition. Although alpha‐synuclein pathology underlies the loss of WM integrity in DLB, amyloid‐independent tau pathology may also be an important influence on white matter integrity within the DLB spectrum.

AbstractBackgroundPredictive models of tau‐PET accumulation have largely relied on the functional connectome to estimate its spatial patterns in Alzheimer’s disease (AD). Patterns of global functional organization represent a promising avenue as they encapsulate many biological properties relevant to neurodevelopment, cognition and degeneration. We compared approaches based on local connectivity (from task‐free fMRI) and global organization (from FDG‐PET) to predict regional tau‐PET across AD phenotypes.MethodWe included 430 cognitively unimpaired (CU) individuals (214 A+T‐, 216 A+T+) and 62 amnestic mild cognitive impairment (aMCI), 64 amnestic AD, 21 dysexecutive AD (dAD), 41 logopenic progressive aphasia (LPA) and 49 posterior cortical atrophy (PCA) patients (all A+T+) from the Alzheimer’s Disease Research Center, Mayo Clinic Study of Aging, and Neurodegenerative Research Groups (referred to as “Mayo cohort”). The connectivity approach consisted of assessing the relationship between a region’s tau level and its functional connectivity to the region of highest tau (epicenter). The global organization approach consisted of 1) performing a principal component analysis on 3000 FDG‐PET images that did not overlap with the Mayo cohort, 2) projecting FDG‐PET images of the Mayo cohort onto the low‐dimensional space generated in Step 1 and extracting corresponding weighted values across the first 100 components, 3) using these weighted values as predictors of regional tau‐PET in a linear model and determining optimal parameters based on predicted R2 values, and 4) using the parameters from the model built in Step 3 to predict tau‐PET in the ADNI cohort (n = 86) (Fig.2).ResultThe connectivity‐based approach yielded adjusted R2 values ranging from 0.19‐0.45 across diagnostic groups (Fig.1). The global organization approach yielded averaged adjusted R2 and predicted R2 values of 0.71 and 0.63, respectively, in the Mayo cohort. Out‐of‐sample median R2 in the ADNI cohort was 0.41 (range 0.10‐0.60) (Fig.3).ConclusionAn unbiased approach based on the global functional organization of the brain and using an imaging modality widely used in clinical settings outperformed a connectome‐based one to predict regional tau‐PET across the AD phenotypic spectrum. Patterns of tau accumulation may better relate to the large‐scale physiology of the brain underlying specific mental functions rather than local connectivity factors.

AbstractBackgroundNeurite orientation dispersion and density imaging (NODDI) measures from diffusion MRI provide information about tissue microstructure. Our goal was to compare the sensitivity of NODDI to structural MRI (sMRI) in capturing neurodegenerative changes in the gray matter due to aging and dementia.MethodWe included 815 participants [644 cognitively unimpaired (CU), 103 MCI, 68 AD] from the Mayo Clinic Study of Aging (MCSA) and Mayo ADRC with sMRI and multi‐shell diffusion MRI. We assessed gray matter volumes scaled by head size and cortical thicknesses from sMRI and neurite density index (NDI) and isotropic volume fraction (ISOVF) from NODDI. Pairwise Spearman correlations were calculated between imaging measures and Tau‐PET SUVRs in Braak I‐II, III‐IV, and V‐VI regions of interest (ROI) in participants = 60 years. Separate multivariable linear regression models were fit for the prediction of brain age using sMRI and NODDI in 515 CU participants. The predicted age gaps were calculated as chronological age minus predicted age in the remaining participants. We compared age gaps computed by sMRI and NODDI using disease surrogates (APOE4 genotype, AD pathologies via PET, WMH via FLAIR‐MRI, vascular risk, cognition via MMSE).ResultBraak ROIs, sMRI, and NODDI measures were correlated. The correlations between ISOVF and Tau‐PET SUVRs were stronger than correlations between NDI, volume, and thickness and Tau‐PET SUVR (Fig.1). MMSE was more associated with ISOVF than Tau‐PET and other imaging measures. The sMRI regression models had a higher mean absolute error (MAE) of 4.13 years than NODDI models with an MAE of 4.02 years. The NODDI model better predicted age gap separation between CU, MCI, and AD compared to the sMRI model (Fig.2). The NODDI predicted age gap showed stronger correlations with global amyloid‐PET SUVR, meta‐ROI Tau‐PET SUVR, CMC, APOE4, and MMSE (Fig. 3).ConclusionNODDI, especially ISOVF, had stronger correlations with MMSE and Tau‐PET SUVRs across all Braak ROIs. The brain age prediction model trained using NODDI was more sensitive in detecting neurodegeneration changes than sMRI based on associations with disease surrogates. Our work suggests that NODDI in gray matter may be a more sensitive surrogate of neurodegeneration compared to sMRI.

AbstractBackgroundAutomated face recognition can potentially re‐identify de‐identified research brain MRI, PET, and CT, and direct differences of brain measurements from original and de‐faced images are statistically significant but very minor, i.e. below scan‐rescan differences. Their effects on analyses correlating brain imaging biomarkers with clinical variables remain unknown.MethodWe sampled 3 age‐and‐sex‐matched groups (CU, MCI, and clinical AD) from Mayo Clinic aging studies, each with 61 participants with same‐visit 3D T1‐weighted and 3D T2‐FLAIR MRI, PiB amyloid PET, and Flortaucipir tau PET (total n = 183). We automatically measured hippocampal volume, PIB global SUVR, and tau temporal meta‐ROI SUVR using a) our in‐house pipeline with SPM12, MCALT, and ANTs, and b) FreeSurfer 7.3.2 (PETSurfer), and we measured voxels associated with clinical group separation using SPM12. We also measured WMH from FLAIR using a) our in‐house pipeline, b) Lesion Segmentation Tool (LST), and c) FreeSurfer’s Samseg. For each measurement, we calculated pair‐wise group separation AUC values and Spearman’s rho with age and CDR‐SOB. We then de‐faced all PET and MRI using mri_reface, our top‐performing automated de‐facing software that replaces imaged faces with an average face (“re‐facing”), and we repeated all measurements and analyses with de‐faced images. We then compared the magnitudes of clinical correlations with original images and those of de‐faced images.ResultAnalyses of original and de‐faced images had very high agreement. Only 2/55 paired comparisons were significant (Tables 1‐2): FreeSurfer‐measured global PIB SUVR from de‐faced images was more correlated with age (rho 0.003 vs ‐0.227, p = 0.046), and Mayo‐measured WMH volume from original images was more correlated with CDR‐SOB (rho 0.321 vs 0.308, p = 0.020). Another comparison‐pair tested p = 0.041 for an AUC change +0.008, which we consider not practically meaningful. All others had p>0.05. In voxel‐based analyses of CU vs AD participants, all comparisons were qualitatively similar between original and de‐faced variants (Figure 1).ConclusionDe‐facing PET and MRI using mri_reface did not significantly affect 53/55 correlations between imaging and clinical variables; of the two affected, one became stronger with re‐faced images. Stronger correlations with clinical variables may be explained by de‐facing’s standardizing face voxels and artifacts that irrelevantly affect brain segmentations.

AbstractBackgroundAlthough amyloid deposition in the brain is often associated with subsequent dementia risk, not everyone with brain amyloid will develop dementia. This discrepancy illustrates the potential importance that risk factors and lived experiences may have in modifying this association. Compared to participants with low social engagement (SE) in mid‐life, we hypothesized that participants with high mid‐life SE will show a weaker association between amyloid burden and incident dementia.MethodWe included data from 310 non‐demented participants of the Atherosclerosis Risk in Communities (ARIC)‐PET study. Social support and isolation were assessed via interviewer‐administered questionnaires (Visit 2; 1990 – 1992). Based upon categorization of both factors, participants were classified as having high, intermediate, or low mid‐life SE (Table 1). Brain amyloid was evaluated with florbetapir PET (Visit 5; 2011‐2014). Elevated amyloid burden was defined as standardized uptake value ratio (SUVR) >1.2 in the global cortex. Incident dementia cases were identified from visit 5 through 2019 through ongoing surveillance utilizing in‐person neurocognitive testing, informant interviews, and hospitalization codes. Relative contributions of mid‐life SE and elevated florbetapir SUVR to incident dementia, independently and with multiplicative interaction terms, were evaluated with Cox regression models, adjusted for demographics, APOEe4 and vascular risk factorsResultAmong 310 participants, 48 developed dementia (median follow‐up: 4.7 years). Mid‐life SE and elevated SUVR each independently predicted dementia risk but did not interact on a multiplicative scale. Participants with high or intermediate mid‐life SE, relative to low mid‐life SE, were less likely to develop dementia (Table 2). Although the interaction between mid‐life SE and elevated SUVR was not significant, stratified models suggested a stronger association between amyloid burden and dementia in participants with high mid‐life SE (Table 3).ConclusionGreater mid‐life SE was associated with lower odds of developing dementia, independent of elevated SUVR. Although protective, there was no strong evidence for effect modification of mid‐life SE on the association between amyloid and dementia risk. Future longitudinal studies evaluating the potential influence of social factors measured throughout the life course are needed to inform our understanding as to what factors may preserve cognition in the presence of elevated brain pathology.