Multiscale functional connectivity patterns of the aging brain learned from harmonized rsfMRI data of the multi-cohort iSTAGING study.

To learn multiscale functional connectivity patterns of the aging brain, we built a brain age prediction model of functional connectivity measures at seven scales on a large fMRI dataset, consisting of resting-state fMRI scans of 4186 individuals with a wide age range (22 to 97 years, with an average of 63) from five cohorts. We computed multiscale functional connectivity measures of individual subjects using a personalized functional network computational method, harmonized the functional connectivity measures of subjects from multiple datasets in order to build a functional brain age model, and finally evaluated how functional brain age gap correlated with cognitive measures of individual subjects. Our study has revealed that functional connectivity measures at multiple scales were more informative than those at any single scale for the brain age prediction, the data harmonization significantly improved the brain age prediction performance, and the data harmonization in the functional connectivity measures' tangent space worked better than in their original space. Moreover, brain age gap scores of individual subjects derived from the brain age prediction model were significantly correlated with clinical and cognitive measures. Overall, these results demonstrated that multiscale functional connectivity patterns learned from a large-scale multi-site rsfMRI dataset were informative for characterizing the aging brain and the derived brain age gap was associated with cognitive and clinical measures.

Five dominant dimensions of brain aging are identified via deep learning: associations with clinical, lifestyle, and genetic measures.

Brain aging is a complex process influenced by various lifestyle, environmental, and genetic factors, as well as by age-related and often co-existing pathologies. MRI and, more recently, AI methods have been instrumental in understanding the neuroanatomical changes that occur during aging in large and diverse populations. However, the multiplicity and mutual overlap of both pathologic processes and affected brain regions make it difficult to precisely characterize the underlying neurodegenerative profile of an individual from an MRI scan. Herein, we leverage a state-of-the art deep representation learning method, Surreal-GAN, and present both methodological advances and extensive experimental results that allow us to elucidate the heterogeneity of brain aging in a large and diverse cohort of 49,482 individuals from 11 studies. Five dominant patterns of neurodegeneration were identified and quantified for each individual by their respective (herein referred to as) R-indices. Significant associations between R-indices and distinct biomedical, lifestyle, and genetic factors provide insights into the etiology of observed variances. Furthermore, baseline R-indices showed predictive value for disease progression and mortality. These five R-indices contribute to MRI-based precision diagnostics, prognostication, and may inform stratification into clinical trials.

Association of Cerebral Blood Flow With Longitudinal Changes in Cerebral Microstructural Integrity in the Coronary Artery Risk Development in Young Adults (CARDIA) Study.

Importance: Decreased cerebral tissue integrity and cerebral blood flow (CBF) are features of neurodegenerative diseases. Brain tissue maintenance is an energy-demanding process, making it particularly sensitive to hypoperfusion. However, little is known about the association between blood flow and brain microstructural integrity, including in normative aging. Objective: To assess associations between CBF and changes in cerebral tissue integrity in white matter and gray matter brain regions. Design, Setting, and Participants: In this longitudinal cohort study, magnetic resonance imaging was performed on 732 healthy adults from the Coronary Artery Risk Development in Young Adults (CARDIA) study, a prospective longitudinal study (baseline age of 18-30 years) that examined participants up to 8 times during 30 years (1985-1986 to 2015-2016). Cerebral blood flow was measured at baseline (year 25 of the CARDIA study), and changes in diffusion tensor indices of fractional anisotropy (FA) and mean diffusivity (MD), measures of microstructural tissue integrity, were measured at both baseline and after approximately 5 years of follow-up (year 30). Analyses were conducted from November 5, 2020, to January 29, 2022. Main Outcomes and Measures: Automated algorithms and linear mixed-effects statistical models were used to evaluate the associations between CBF at baseline and changes in FA or MD. Results: After exclusion of participants with missing or low-quality data, 654 at baseline (342 women; mean [SD] age, 50.3 [3.5] years) and 433 at follow-up (230 women; mean [SD] age, 55.1 [3.5] years) were scanned for CBF or FA and MD imaging. In the baseline cohort, 247 participants were Black (37.8%) and 394 were White (60.2%); in the follow-up cohort, 156 were Black (36.0%) and 277 were White (64.0%). Cross-sectionally, FA and MD were associated with CBF in most regions evaluated, with lower CBF values associated with lower FA or higher MD values, including the frontal white matter lobes (for CBF and MD: mean [SE] ß = -1.4 [0.5] × 10-6; for CBF and FA: mean [SE] ß = 2.9 [1.0] × 10-4) and the parietal white matter lobes (for CBF and MD: mean [SE] ß = -2.4 [0.6] × 10-6; for CBF and FA: mean [SE] ß = 4.4 [1.1] × 10-4). Lower CBF values at baseline were also significantly associated with steeper regional decreases in FA or increases in MD in most brain regions investigated, including the frontal (for CBF and MD: mean [SE] ß = -1.1 [0.6] × 10-6; for CBF and FA: mean [SE] ß = 2.9 [1.0] × 10-4) and parietal lobes (for CBF and MD: mean [SE] ß = -1.5 [0.7] × 10-6; for CBF and FA: mean [SE] ß = 4.4 [1.1] × 10-4). Conclusions and Relevance: Results of this longitudinal cohort study of the association between CBF and diffusion tensor imaging metrics suggest that blood flow may be significantly associated with brain tissue microstructure. This work may lay the foundation for investigations to clarify the nature of early brain damage in neurodegeneration. Such studies may lead to new neuroimaging biomarkers of brain microstructure and function for disease progression.

Disentangling Alzheimer's disease neurodegeneration from typical brain ageing using machine learning.

Neuroimaging biomarkers that distinguish between changes due to typical brain ageing and Alzheimer's disease are valuable for determining how much each contributes to cognitive decline. Supervised machine learning models can derive multivariate patterns of brain change related to the two processes, including the Spatial Patterns of Atrophy for Recognition of Alzheimer's Disease (SPARE-AD) and of Brain Aging (SPARE-BA) scores investigated herein. However, the substantial overlap between brain regions affected in the two processes confounds measuring them independently. We present a methodology, and associated results, towards disentangling the two. T1-weighted MRI scans of 4054 participants (48-95 years) with Alzheimer's disease, mild cognitive impairment (MCI), or cognitively normal (CN) diagnoses from the Imaging-based coordinate SysTem for AGIng and NeurodeGenerative diseases (iSTAGING) consortium were analysed. Multiple sets of SPARE scores were investigated, in order to probe imaging signatures of certain clinically or molecularly defined sub-cohorts. First, a subset of clinical Alzheimer's disease patients (n = 718) and age- and sex-matched CN adults (n = 718) were selected based purely on clinical diagnoses to train SPARE-BA1 (regression of age using CN individuals) and SPARE-AD1 (classification of CN versus Alzheimer's disease) models. Second, analogous groups were selected based on clinical and molecular markers to train SPARE-BA2 and SPARE-AD2 models: amyloid-positive Alzheimer's disease continuum group (n = 718; consisting of amyloid-positive Alzheimer's disease, amyloid-positive MCI, amyloid- and tau-positive CN individuals) and amyloid-negative CN group (n = 718). Finally, the combined group of the Alzheimer's disease continuum and amyloid-negative CN individuals was used to train SPARE-BA3 model, with the intention to estimate brain age regardless of Alzheimer's disease-related brain changes. The disentangled SPARE models, SPARE-AD2 and SPARE-BA3, derived brain patterns that were more specific to the two types of brain changes. The correlation between the SPARE-BA Gap (SPARE-BA minus chronological age) and SPARE-AD was significantly reduced after the decoupling (r = 0.56-0.06). The correlation of disentangled SPARE-AD was non-inferior to amyloid- and tau-related measurements and to the number of APOE ε4 alleles but was lower to Alzheimer's disease-related psychometric test scores, suggesting the contribution of advanced brain ageing to the latter. The disentangled SPARE-BA was consistently less correlated with Alzheimer's disease-related clinical, molecular and genetic variables. By employing conservative molecular diagnoses and introducing Alzheimer's disease continuum cases to the SPARE-BA model training, we achieved more dissociable neuroanatomical biomarkers of typical brain ageing and Alzheimer's disease.

Deep Generative Medical Image Harmonization for Improving Cross-Site Generalization in Deep Learning Predictors.

BACKGROUND: In the medical imaging domain, deep learning-based methods have yet to see widespread clinical adoption, in part due to limited generalization performance across different imaging devices and acquisition protocols. The deviation between estimated brain age and biological age is an established biomarker of brain health and such models may benefit from increased cross-site generalizability. PURPOSE: To develop and evaluate a deep learning-based image harmonization method to improve cross-site generalizability of deep learning age prediction. STUDY TYPE: Retrospective. POPULATION: Eight thousand eight hundred and seventy-six subjects from six sites. Harmonization models were trained using all subjects. Age prediction models were trained using 2739 subjects from a single site and tested using the remaining 6137 subjects from various other sites. FIELD STRENGTH/SEQUENCE: Brain imaging with magnetization prepared rapid acquisition with gradient echo or spoiled gradient echo sequences at 1.5 T and 3 T. ASSESSMENT: StarGAN v2, was used to perform a canonical mapping from diverse datasets to a reference domain to reduce site-based variation while preserving semantic information. Generalization performance of deep learning age prediction was evaluated using harmonized, histogram matched, and unharmonized data. STATISTICAL TESTS: Mean absolute error (MAE) and Pearson correlation between estimated age and biological age quantified the performance of the age prediction model. RESULTS: Our results indicated a substantial improvement in age prediction in out-of-sample data, with the overall MAE improving from 15.81 (±0.21) years to 11.86 (±0.11) with histogram matching to 7.21 (±0.22) years with generative adversarial network (GAN)-based harmonization. In the multisite case, across the 5 out-of-sample sites, MAE improved from 9.78 (±6.69) years to 7.74 (±3.03) years with histogram normalization to 5.32 (±4.07) years with GAN-based harmonization. DATA CONCLUSION: While further research is needed, GAN-based medical image harmonization appears to be a promising tool for improving cross-site deep learning generalization. LEVEL OF EVIDENCE: 4 TECHNICAL EFFICACY: Stage 1.

Blood pressure, executive function, and network connectivity in middle-aged adults at risk of dementia in late life.

Midlife blood pressure is associated with structural brain changes, cognitive decline, and dementia in late life. However, the relationship between early adulthood blood pressure exposure, brain structure and function, and cognitive performance in midlife is not known. A better understanding of these relationships in the preclinical stage may advance our mechanistic understanding of vascular contributions to late-life cognitive decline and dementia and may provide early therapeutic targets. To identify resting-state functional connectivity of executive control networks (ECNs), a group independent components analysis was performed of functional MRI scans of 600 individuals from the Coronary Artery Risk Development in Young Adults longitudinal cohort study, with cumulative systolic blood pressure (cSBP) measured at nine visits over the preceding 30 y. Dual regression analysis investigated performance-related connectivity of ECNs in 578 individuals (mean age 55.5 ± 3.6 y, 323 female, 243 Black) with data from the Stroop color-word task of executive function. Greater connectivity of a left ECN to the bilateral anterior gyrus rectus, right posterior orbitofrontal cortex, and nucleus accumbens was associated with better executive control performance on the Stroop. Mediation analyses showed that while the relationship between cSBP and Stroop performance was mediated by white matter hyperintensities (WMH), resting-state connectivity of the ECN mediated the relationship between WMH and executive function. Increased connectivity of the left ECN to regions involved in reward processing appears to compensate for the deleterious effects of WMH on executive function in individuals across the burden of cumulative systolic blood pressure exposure in midlife.

White matter lesions: Spatial heterogeneity, links to risk factors, cognition, genetics, and atrophy.

OBJECTIVES: To investigate spatial heterogeneity of white matter lesions or hyperintensities (WMH). METHODS: MRI scans of 1,836 participants (median age 52.2 ± 13.16 years) encompassing a wide age range (22-84 years) from the cross-sectional Study of Health in Pomerania (Germany) were included as discovery set identifying spatially distinct components of WMH using a structural covariance approach. Scans of 307 participants (median age 73.8 ± 10.2 years, with 747 observations) from the Baltimore Longitudinal Study of Aging (United States) were included to examine differences in longitudinal progression of these components. The associations of these components with vascular risk factors, cortical atrophy, Alzheimer disease (AD) genetics, and cognition were then investigated using linear regression. RESULTS: WMH were found to occur nonuniformly, with higher frequency within spatially heterogeneous patterns encoded by 4 components, which were consistent with common categorizations of deep and periventricular WMH, while further dividing the latter into posterior, frontal, and dorsal components. Temporal trends of the components differed both cross-sectionally and longitudinally. Frontal periventricular WMH were most distinctive as they appeared in the fifth decade of life, whereas the other components appeared later in life during the sixth decade. Furthermore, frontal WMH were associated with systolic blood pressure and with pronounced atrophy including AD-related regions. AD polygenic risk score was associated with the dorsal periventricular component in the elderly. Cognitive decline was associated with the dorsal component. CONCLUSIONS: These results support the hypothesis that the appearance of WMH follows age and disease-dependent regional distribution patterns, potentially influenced by differential underlying pathophysiologic mechanisms, and possibly with a differential link to vascular and neurodegenerative changes.

Sex hormones and brain volumes in a longitudinal study of middle-aged men in the CARDIA study.

INTRODUCTION: Several findings suggest that testosterone (T) is neuroprotective and that declining T levels during aging are associated with cognitive and brain pathologies; however, little is known on T and brain health in middle-age. We examined the relationships of total T, bioavailable T, and sex hormone binding globulin (SHBG) levels with total and regional gray matter (GM) and white matter (WM) volumes in middle-aged men. We also evaluated the association of sex hormone levels with cognitive function. METHODS: Analysis included 267 community-dwelling men participating in the Coronary Artery Risk Development in Young Adults (CARDIA) brain magnetic resonance imaging (MRI) substudy. Total T, bioavailable T, and SHBG levels were measured at three times from the 2nd to 4th decade of life; brain volumes were measured at the ages of 42-56. Associations were estimated using linear regression models, adjusted for several potential confounders. RESULTS: Higher SHBG levels were associated with greater total WM volume (+3.15 cm3 [95% confidence interval [CI] = 0.01, 6.28] per one standard deviation higher SHBG). Higher SHBG levels were associated with lower total and regional GM volumes overall and significantly with smaller parietal GM volume (-0.96 cm3 [95%CI = -1.71, -0.21]). T levels were not related to brain volumes. Neither T nor SHBG levels were associated with cognitive function. CONCLUSION: Results suggest a role for SHBG in structural brain outcomes in men and emphasize the value of investigating SHBG levels as modulators of sex hormone and metabolic pathways regulating brain and behavioral characteristics in men.

Relative differences in resting-state brain connectivity associated with long term intensive lifestyle intervention.

A number of studies have reported that type 2 diabetes mellitus (T2DM) is associated with alterations in resting-state activity and connectivity in the brain. There is also evidence that interventions involving physical activity and weight loss may affect brain functional connectivity. In this study, we examined the effects of nearly 10 years of an intensive lifestyle intervention (ILI), designed to induce and sustain weight loss through lower caloric intake and increased physical activity, on resting-state networks in adults with T2DM. We performed a cross-sectional comparison of global and local characteristics from functional brain networks between individuals who had been randomly assigned to ILI or a control condition of health education and support. Upon examining brain networks from 312 participants (average age: 68.8 for ILI and 67.9 for controls), we found that ILI participants (N=160) had attenuated local efficiency at the network-level compared with controls (N=152). Although there was no group difference in the network-level global efficiency, we found that, among ILI participants, nodal global efficiency was elevated in left fusiform gyrus, right middle frontal gyrus, and pars opercularis of right inferior frontal gyrus. These effects were age-dependent, with more pronounced effects for older participants. Overall these results indicate that the individuals assigned to the ILI had brain networks with less regional and more global connectivity, particularly involving frontal lobes. Such patterns would support greater distributed information processing. Future studies are needed to determine if these differences are associated with age-related compensatory function in the ILI group or worse pathology in the control group.

Educational attainment, MRI changes, and cognitive function in older postmenopausal women from the Women's Health Initiative Memory Study.

The relationship between neuropathology and clinically manifested functional and cognitive deficits is complex. Clinical observations of individuals with greater neuropathology who function better than some individuals with less neuropathology are common and puzzling. Educational attainment, a proxy for "cognitive reserve," may help to explain this apparent contradiction. The objective of this study is to determine if educational attainment is correlated with cognitive decline, brain lesion volume, and total brain atrophy. One thousand three hundred ninety of the 7,479 community-dwelling women 65 years of age and older enrolled in the Women's Health Initiative Memory Study, two parallel randomized, placebo-controlled clinical trials comparing unopposed and opposed postmenopausal hormone therapy with placebo, were studied. Study participants received annual assessments of global cognitive function with the Modified Mini Mental State exam. One thousand sixty-three participants also received supplemental neurocognitive battery and neuroimaging studies. Magnetic resonance imaging was used to calculate total ischemic lesion and brain volumes. Incident cases of probable dementia and mild cognitive impairment were centrally adjudicated. After adjustment for total lesion and total brain volumes (atrophy), higher educational attainment predicted better cognitive performance (p < 0.001). Following conversion to dementia/MCI, higher education predicted steeper declines in cognitive function (p < 0.001). Thus, higher educational attainment was associated with a delay in diagnosis of dementia/MCI in the face of a growing neuropathological load.

Relationship of hypertension, blood pressure, and blood pressure control with white matter abnormalities in the Women's Health Initiative Memory Study (WHIMS)-MRI trial.

This paper evaluates the relationship of blood pressure (BP) levels at Women's Health Initiative (WHI) baseline, treatment of hypertension, and white matter abnormalities among women in conjugated equine estrogen (CEE) and medroxyprogesterone acetate and CEE-alone arms. The WHI Memory Study-Magnetic Resonance Imaging (WHIMS-MRI) trial scanned 1424 participants. BP levels at baseline were significantly positively related to abnormal white matter lesion (WML) volumes. Participants treated for hypertension but who had BP > or = 140/90 mm Hg had the greatest amount of WML volumes. Women with untreated BP > or = 140/90 mm Hg had intermediate WML volumes. Abnormal WML volumes were related to hypertension in most areas of the brain and were greater in the frontal lobe than in the occipital, parietal, or temporal lobes. Level of BP at baseline was strongly related to amount of WML volumes. The results of the study reinforce the relationship of hypertension and BP control and white matter abnormalities in the brain. The evidence to date supports tight control of BP levels, especially beginning at younger and middle age as a possible and perhaps only way to prevent dementia.