Lowe, Val J.

Abstract Alzheimer's disease (AD) exhibits spatially heterogeneous 3R/4R tau pathology distributions across participants, making it a challenge to quantify extent of tau deposition. Utilizing Tau-PET from three independent cohorts, we trained and validated a machine learning model to identify visually positive Tau-PET scans from regional SUVR values and developed a novel summary measure, THETA, that accounts for heterogeneity in tau deposition. The model for identification of tau positivity achieved a balanced test accuracy of 95% and accuracy of ≥87% on the validation datasets. THETA captured heterogeneity of tau deposition, had better association with clinical measures, and corresponded better with visual assessments in comparison with the temporal meta-region-of-interest Tau-PET quantification methods. Our novel approach aids in identification of positive Tau-PET scans and provides a quantitative summary measure, THETA, that effectively captures the heterogeneous tau deposition seen in AD. The application of THETA for quantifying Tau-PET in AD exhibits great potential.

Abstract There is a longstanding ambiguity regarding the clinical diagnosis of dementia syndromes predominantly targeting executive functions versus behaviour and personality. This is due to an incomplete understanding of the macro-scale anatomy underlying these symptomatologies, a partial overlap in clinical features and the fact that both phenotypes can emerge from the same pathology and vice versa. We collected data from a patient cohort of which 52 had dysexecutive Alzheimer’s disease, 30 had behavioural variant frontotemporal dementia (bvFTD), seven met clinical criteria for bvFTD but had Alzheimer’s disease pathology (behavioural Alzheimer’s disease) and 28 had amnestic Alzheimer’s disease. We first assessed group-wise differences in clinical and cognitive features and patterns of fluorodeoxyglucose (FDG) PET hypometabolism. We then performed a spectral decomposition of covariance between FDG-PET images to yield latent patterns of relative hypometabolism unbiased by diagnostic classification, which are referred to as ‘eigenbrains’. These eigenbrains were subsequently linked to clinical and cognitive data and meta-analytic topics from a large external database of neuroimaging studies reflecting a wide range of mental functions. Finally, we performed a data-driven exploratory linear discriminant analysis to perform eigenbrain-based multiclass diagnostic predictions. Dysexecutive Alzheimer’s disease and bvFTD patients were the youngest at symptom onset, followed by behavioural Alzheimer’s disease, then amnestic Alzheimer’s disease. Dysexecutive Alzheimer’s disease patients had worse cognitive performance on nearly all cognitive domains compared with other groups, except verbal fluency which was equally impaired in dysexecutive Alzheimer’s disease and bvFTD. Hypometabolism was observed in heteromodal cortices in dysexecutive Alzheimer’s disease, temporo-parietal areas in amnestic Alzheimer’s disease and frontotemporal areas in bvFTD and behavioural Alzheimer’s disease. The unbiased spectral decomposition analysis revealed that relative hypometabolism in heteromodal cortices was associated with worse dysexecutive symptomatology and a lower likelihood of presenting with behaviour/personality problems, whereas relative hypometabolism in frontotemporal areas was associated with a higher likelihood of presenting with behaviour/personality problems but did not correlate with most cognitive measures. The linear discriminant analysis yielded an accuracy of 82.1% in predicting diagnostic category and did not misclassify any dysexecutive Alzheimer’s disease patient for behavioural Alzheimer’s disease and vice versa. Our results strongly suggest a double dissociation in that distinct macro-scale underpinnings underlie predominant dysexecutive versus personality/behavioural symptomatology in dementia syndromes. This has important implications for the implementation of criteria to diagnose and distinguish these diseases and supports the use of data-driven techniques to inform the classification of neurodegenerative diseases.

Abstract Given the prevalence of dementia and the development of pathology-specific disease-modifying therapies, high-value biomarker strategies to inform medical decision-making are critical. In vivo tau-PET is an ideal target as a biomarker for Alzheimer’s disease diagnosis and treatment outcome measure. However, tau-PET is not currently widely accessible to patients compared to other neuroimaging methods. In this study, we present a convolutional neural network (CNN) model that imputes tau-PET images from more widely available cross-modality imaging inputs. Participants (n = 1192) with brain T1-weighted MRI (T1w), fluorodeoxyglucose (FDG)-PET, amyloid-PET and tau-PET were included. We found that a CNN model can impute tau-PET images with high accuracy, the highest being for the FDG-based model followed by amyloid-PET and T1w. In testing implications of artificial intelligence-imputed tau-PET, only the FDG-based model showed a significant improvement of performance in classifying tau positivity and diagnostic groups compared to the original input data, suggesting that application of the model could enhance the utility of the metabolic images. The interpretability experiment revealed that the FDG- and T1w-based models utilized the non-local input from physically remote regions of interest to estimate the tau-PET, but this was not the case for the Pittsburgh compound B-based model. This implies that the model can learn the distinct biological relationship between FDG-PET, T1w and tau-PET from the relationship between amyloid-PET and tau-PET. Our study suggests that extending neuroimaging’s use with artificial intelligence to predict protein specific pathologies has great potential to inform emerging care models.

AbstractINTRODUCTIONAtypical variants of Alzheimer's disease (AD) include the visual variant, known as posterior cortical atrophy (PCA), and the language variant, known as logopenic progressive aphasia (LPA). Clinically, rates of disease progression differ between them.METHODSWe evaluated 34 PCA and 29 LPA participants. Structural magnetic resonance imaging and 18F‐flortaucipir positron emission tomography were performed at baseline and at 1‐year follow‐up. Rates of change in tau uptake and grey matter volumes were compared between PCA and LPA with linear mixed‐effects models and voxel‐based analyses.RESULTSPCA had faster rates of occipital atrophy. LPA had faster rates of left temporal atrophy and faster rates of tau accumulation in the parietal, right temporal, and occipital lobes. Age was negatively associated with rates of atrophy and tau accumulation.DISCUSSIONLongitudinal patterns of neuroimaging abnormalities differed between PCA and LPA, although with divergent results for tau accumulation and atrophy.HIGHLIGHTS The language variant of Alzheimer's disease accumulates tau faster than the visual variant. Each variant shows faster rates of atrophy than the other in its signature regions. Age negatively influences rates of atrophy and tau accumulation in both variants.

AbstractBackgroundDementia with Lewy bodies (DLB) is the second most common neurodegenerative dementia, after Alzheimer’s disease (AD). While alpha‐synuclein pathology is the hallmark of DLB, AD and cerebrovascular pathologies are common in DLB patients. However, the contribution of these pathologies towards neurodegeneration in DLB is largely unknown. We investigated associations of amyloid‐beta, tau, and cerebrovascular biomarkers with regional gray matter (GM) volume in DLB patients and cognitively unimpaired (CU) controls.MethodWe included 30 DLB patients (69.3±10.2 years old, 87% men) and 100 CU individuals balanced on age and sex (Table 1). We used 11C‐Pittsburgh Compound‐B (PiB) and 18F‐Flortaucipir positron emission tomography (PET) to assess amyloid‐beta and tau in‐vivo. We used structural MRI to assess white matter hyperintensity (WMH) volume, a marker of cerebrovascular lesion load, and regional GM volume (a marker of neurodegeneration). We used correlations and ANCOVA in the whole dataset and structural equation modelling in the DLB patients to investigate associations of WMH volume and regional amyloid‐beta and tau PET standardized uptake value ratios (SUVRs) with regional GM volume.ResultDLB patients showed lower GM volume across all cortical and subcortical regions except for cuneus, putamen, and pallidum (Fig.1). Greater WMH volume was associated with lower volume in the medial and orbital frontal cortices, insula, fusiform cortex, and thalamus in DLB patients (Fig.2). Greater PiB SUVR was associated with lower volume in the inferior temporal cortex, while Flortaucipir SUVR did not correlate with GM volume (Fig.2). Structural equation modelling showed that higher age and positivity for the APOE e4 genotype were significant predictors of greater WMH, and WMH in turn was a significant predictor of GM volume in medial and orbital frontal cortices, insula, and inferior temporal cortex (Fig.3). In contrast, we observed two distinct paths for the fusiform cortex, with age having an effect through PiB and Flortaucipir SUVRs on the one path, and through WMH on the other path (Fig.3).ConclusionDLB patients have widespread cortical atrophy, most of which likely is influenced by alpha‐synuclein pathology. Although amyloid‐beta, tau, and cerebrovascular pathologies often coexist in DLB, their contributions to neurodegeneration seem to be region specific.

AbstractBackgroundTwo pathways that help explain cognitive health are 1) brain maintenance – referring to the relative absence of brain changes over time; 2) cognitive reserve – a property of the brain that allows for sustained cognitive ability in the presence of brain pathology. Both pathways influence cognitive aging and are best understood using longitudinal data. Our goal was to assess the effect of modifiable vascular/metabolic and lifestyle enrichment factors on these pathways using longitudinal brain changes measured on structural magnetic resonance imaging (MRI) and longitudinal cognitive decline. Lastly, we combined these pathways and their relationships in the same model.MethodWe identified 1250 participants from a well‐characterized sample of older adults from the Mayo Clinic Study of Aging with both longitudinal MRI and cognitive data. Linear mixed models were performed to assess the impact of modifiable factors on brain maintenance and cognitive reserve. Latent growth curve models (SEM) were used to assess the impact of brain structure and cognitive reserve on longitudinal cognition.ResultThe mean age of our sample was 78.8 years, with 45.4% female, and 27.1% APOE4 carriers. Mean follow‐up time was 4.0 years with a mean of 3.2 MRI scans per participant. Linear mixed models showed that poorer vascular/metabolic health (presence of vascular/metabolic conditions) was associated with lower cross‐sectional temporal region cortical thickness (‐0.014 (0.003), p<0.001) and lower global cognition (‐0.042 (0.018), p = 0.024). Higher estimated cognitive reserve at baseline as assessed by education/occupation composite (0.103 (0.01), p<0.001) was associated with higher global cognition. Unexpectedly, late‐life physical activities were related to lower global cognition (‐0.05 (0.007), p = 0.004). Our final latent growth curve SEM fit the data well showing temporal associations between unmodifiable characteristics, cognitive reserve, and brain structural pathways, as well as the influence of modifiable vascular/metabolic factors.ConclusionWe found that brain maintenance and cognitive reserve pathways were related to better cognition over time. We further showed that vascular/metabolic health impacts long‐term cognitive outcomes (i.e. cognitive aging) primarily through the brain maintenance pathway. These data provide a modeling scheme for future studies and can aid in assessing the pathways by which individual factors may influence cognitive aging.

AbstractBackgroundWomen have higher rates of hypertension at the time of menopausal transition than men. Although elevated systolic and diastolic blood pressures (SBP and DBP) are risk factors for white matter (WM) injury both in women and men, WM hyperintensities (WMH), a marker of WM injury, are more common in women than in men after the age of 60. Thus, women may be vulnerable to the effects of elevated BP on WM integrity during menopausal transition. We investigated the association of BP in recently postmenopausal women with good cardiovascular health with WMH volume assessed 13 years later.MethodWomen (n = 212; median age = 67; range 58‐72), who were previously enrolled in a multi‐site randomized menopausal hormone therapy trial, Kronos Early Estrogen Prevention Study (KEEPS), participated in the present observational KEEPS Continuation Study. SBP and DBP were measured at KEEPS baseline (median age = 54 years; range 44‐59). Participants who were on antihypertensive medications at KEEPS baseline were excluded from the analysis (n = 34). Approximately 9 years after the end of KEEPS menopausal hormones versus placebo 4‐year interventions, WM integrity was assessed with automated WMH volume measurements from 3D‐FLAIR MRI. Associations of baseline SBP and DBP with logWMH volume measured 13 years later were tested by linear regression analyses. Model 1 was adjusted for age, study site, and total intracranial volume. Model 2 was adjusted for covariates of model 1, triglycerides, low density lipoproteins, and waist/hip ratio.ResultThe median SBP was 116 (range 83‐158) and DBP was 74 (range 56‐98). Higher SBP was associated with greater WMH volume (model 1:p = 0.005; model 2:p = 0.005). Similarly, higher DBP was associated with greater WMH volume (model 1:p = 0.008; model 2:p = 0.006) (Figure 1). Neither menopausal hormone therapies versus placebo, nor having high/low SBP (≥120/<120) modified these associations. Voxel‐based analyses demonstrate the association of BPs with the spatial distribution of WMHs (Figure 2).ConclusionHigher BPs in recently menopausal women were associated with greater WM injury more than a decade later after adjusting for CVD risk factors. Early postmenopausal stage is a critical time for BP control in women, which may reduce the risk of WM injury later in life.

Abstract Posterior cortical atrophy and logopenic progressive aphasia are atypical clinical presentations of Alzheimer’s disease. Resting-state functional connectivity studies have shown functional network disruptions in both phenotypes, particularly involving the language network in logopenic progressive aphasia and the visual network in posterior cortical atrophy. However, little is known about how connectivity differs both within and between brain networks in these atypical Alzheimer’s disease phenotypes. A cohort of 144 patients was recruited by the Neurodegenerative Research Group at Mayo Clinic, Rochester, MN, USA, and underwent structural and resting-state functional MRI. Spatially preprocessed data were analysed to explore the default mode network and the salience, sensorimotor, language, visual and memory networks. The data were analysed at the voxel and network levels. Bayesian hierarchical linear models adjusted for age and sex were used to analyse within- and between-network connectivity. Reduced within-network connectivity was observed in the language network in both phenotypes, with stronger evidence of reductions in logopenic progressive aphasia compared to controls. Only posterior cortical atrophy showed reduced within-network connectivity in the visual network compared to controls. Both phenotypes showed reduced within-network connectivity in the default mode and sensorimotor networks. No significant change was noted in the memory network, but a slight increase in the salience within-network connectivity was seen in both phenotypes compared to controls. Between-network analysis in posterior cortical atrophy showed evidence of reduced visual-to-language network connectivity, with reduced visual-to-salience network connectivity, compared to controls. An increase in visual-to-default mode network connectivity was noted in posterior cortical atrophy compared to controls. Between-network analysis in logopenic progressive aphasia showed evidence of reduced language-to-visual network connectivity and an increase in language-to-salience network connectivity compared to controls. Findings from the voxel-level and network-level analysis were in line with the Bayesian hierarchical linear model analysis, showing reduced connectivity in the dominant network based on diagnosis and more crosstalk between networks in general compared to controls. The atypical Alzheimer’s disease phenotypes were associated with disruptions in connectivity, both within and between brain networks. Phenotype-specific differences in connectivity patterns were noted in the visual network for posterior cortical atrophy and the language network for logopenic progressive aphasia.

Background: Posterior cortical atrophy (PCA) and logopenic progressive aphasia (LPA) are associated with characteristic patterns of structural network degeneration. Little is known about longitudinal patterns of white matter tract degeneration in these phenotypes. Objective: To assess longitudinal patterns of white matter degeneration and identify phenotype specific cross-sectional and longitudinal diffusion tensor imaging (DTI) biomarkers in PCA and LPA. Methods: Twenty-five PCA, 22 LPA and 25 cognitively unimpaired (CU) individuals were recruited and underwent structural MRI that included a DTI sequence with a follow-up one year later. Cross-sectional and longitudinal mixed effects models were fit to assess the effects of diagnosis on baseline and annualized change in regional DTI metrics. Discriminatory power was investigated using the area under the receiver operating characteristic curves (AUROC). Results: PCA and LPA showed overlapping white matter degeneration profiles predominantly in the left occipital and temporal lobes, the posterior thalamic radiation and sagittal stratum at baseline, as well as the parietal lobe longitudinally. PCA showed degeneration in the occipital and parietal white matter, cross-sectionally and longitudinally, compared to CU, while LPA showed greater degeneration in the temporal and inferior parietal white matter and the inferior fronto-occipital fasciculus cross-sectionally, and in parietal white matter longitudinally compared to CU. Cross-sectionally, integrity of the inferior occipital white matter was best able to differentiate PCA from LPA, with an AUROC of 0.82. Conclusion: These findings contribute to our understanding of white matter degeneration and support usage of DTI as a useful additional diagnostic biomarker for PCA and LPA.