Anatomically interpretable deep learning of brain age captures domain-specific cognitive impairment

Proceedings of the National Academy of Sciences · 2023 · 116 citations

• Psychology
• Audiology
• Neuroscience

= 359). In individuals with MCI (54% of whom were diagnosed with dementia within 10.9 y from MRI acquisition), BA is significantly better than CA in capturing dementia symptom severity, functional disability, and executive function. Profiles of sex dimorphism and lateralization in brain aging also map onto patterns of neuroanatomic change that reflect cognitive decline. Significant associations between BA and neurocognitive measures suggest that the proposed framework can map, systematically, the relationship between aging-related neuroanatomy changes in CN individuals and in participants with MCI or AD. Early identification of such neuroanatomy changes can help to screen individuals according to their AD risk.

Common, low-frequency, rare, and ultra-rare coding variants contribute to COVID-19 severity

Human Genetics · 2021 · 39 citations

• Computer Science
• Machine Learning
• Biology

The combined impact of common and rare exonic variants in COVID-19 host genetics is currently insufficiently understood. Here, common and rare variants from whole-exome sequencing data of about 4000 SARS-CoV-2-positive individuals were used to define an interpretable machine-learning model for predicting COVID-19 severity. First, variants were converted into separate sets of Boolean features, depending on the absence or the presence of variants in each gene. An ensemble of LASSO logistic regression models was used to identify the most informative Boolean features with respect to the genetic bases of severity. The Boolean features selected by these logistic models were combined into an Integrated PolyGenic Score that offers a synthetic and interpretable index for describing the contribution of host genetics in COVID-19 severity, as demonstrated through testing in several independent cohorts. Selected features belong to ultra-rare, rare, low-frequency, and common variants, including those in linkage disequilibrium with known GWAS loci. Noteworthily, around one quarter of the selected genes are sex-specific. Pathway analysis of the selected genes associated with COVID-19 severity reflected the multi-organ nature of the disease. The proposed model might provide useful information for developing diagnostics and therapeutics, while also being able to guide bedside disease management.

Neurofilament Light Chain as a Biomarker for Cognitive Decline in Parkinson Disease

Movement Disorders · 2021 · 170 citations

• Internal medicine
• Psychology
• Medicine

BACKGROUND: Neurofilament light chain protein (NfL) is a promising biomarker of neurodegeneration. OBJECTIVES: To determine whether plasma and CSF NfL (1) associate with motor or cognitive status in Parkinson's disease (PD) and (2) predict future motor or cognitive decline in PD. METHODS: Six hundred and fifteen participants with neurodegenerative diseases, including 152 PD and 200 healthy control participants, provided a plasma and/or cerebrospinal fluid (CSF) NfL sample. Diagnostic groups were compared using the Kruskal-Wallis rank test. Within PD, cross-sectional associations between NfL and Unified Parkinson's Disease Rating Scale Part III (UPDRS-III) and Mattis Dementia Rating Scale (DRS-2) scores were assessed by linear regression; longitudinal analyses were performed using linear mixed-effects models and Cox regression. RESULTS: Plasma and CSF NfL levels correlated substantially (Spearman r = 0.64, P < 0.001); NfL was highest in neurocognitive disorders. PD participants with high plasma NfL were more likely to develop incident cognitive impairment (HR 5.34, P = 0.005). CONCLUSIONS: Plasma NfL is a useful prognostic biomarker for PD, predicting clinical conversion to mild cognitive impairment or dementia. © 2021 International Parkinson and Movement Disorder Society.

Plasma phosphorylated-tau181 as a predictive biomarker for Alzheimer’s amyloid, tau and FDG PET status

Translational Psychiatry · 2021 · 68 citations

• Medicine
• Pathology
• Oncology

Plasma phosphorylated-tau181 (p-tau181) showed the potential for Alzheimer's diagnosis and prognosis, but its role in detecting cerebral pathologies is unclear. We aimed to evaluate whether it could serve as a marker for Alzheimer's pathology in the brain. A total of 1189 participants with plasma p-tau181 and PET data of amyloid, tau or FDG PET were included from ADNI. Cross-sectional relationships of plasma p-tau181 with PET biomarkers were tested. Longitudinally, we further investigated whether different p-tau181 levels at baseline predicted different progression of Alzheimer's pathological changes in the brain. We found plasma p-tau181 significantly correlated with brain amyloid (Spearman ρ = 0.45, P < 0.0001), tau (0.25, P = 0.0003), and FDG PET uptakes (-0.37, P < 0.0001), and increased along the Alzheimer's continuum. Individually, plasma p-tau181 could detect abnormal amyloid, tau pathologies and hypometabolism in the brain, similar with or even better than clinical indicators. The diagnostic accuracy of plasma p-tau181 elevated significantly when combined with clinical information (AUC = 0.814 for amyloid PET, 0.773 for tau PET, and 0.708 for FDG PET). Relationships of plasma p-tau181 with brain pathologies were partly or entirely mediated by the corresponding CSF biomarkers. Besides, individuals with abnormal plasma p-tau181 level (>18.85 pg/ml) at baseline had a higher risk of pathological progression in brain amyloid (HR: 2.32, 95%CI 1.32-4.08) and FDG PET (3.21, 95%CI 2.06-5.01) status. Plasma p-tau181 may be a sensitive screening test for detecting brain pathologies, and serve as a predictive biomarker for Alzheimer's pathophysiology.

Contribution of Alzheimer's biomarkers and risk factors to cognitive impairment and decline across the Alzheimer's disease continuum

Alzheimer s & Dementia · 2021 · 43 citations

• Psychology
• Internal medicine
• Medicine

INTRODUCTION: Amyloid beta (Aβ), tau, and neurodegeneration jointly with the Alzheimer's disease (AD) risk factors affect the severity of clinical symptoms and disease progression. METHODS: Within 248 Aβ-positive elderly with and without cognitive impairment and dementia, partial least squares structural equation pathway modeling was used to assess the direct and indirect effects of imaging biomarkers (global Aβ-positron emission tomography [PET] uptake, regional tau-PET uptake, and regional magnetic resonance imaging-based atrophy) and risk-factors (age, sex, education, apolipoprotein E [APOE], and white-matter lesions) on cross-sectional cognitive impairment and longitudinal cognitive decline. RESULTS: Sixteen percent of variance in cross-sectional cognitive impairment was accounted for by Aβ, 46% to 47% by tau, and 25% to 29% by atrophy, although 53% to 58% of total variance in cognitive impairment was explained by incorporating mediated and direct effects of AD risk factors. The Aβ-tau-atrophy pathway accounted for 50% to 56% of variance in longitudinal cognitive decline while Aβ, tau, and atrophy independently explained 16%, 46% to 47%, and 25% to 29% of the variance, respectively. DISCUSSION: These findings emphasize that treatments that remove Aβ and completely stop downstream effects on tau and neurodegeneration would only be partially effective in slowing of cognitive decline or reversing cognitive impairment.

Staging tau pathology with tau PET in Alzheimer’s disease: a longitudinal study

Translational Psychiatry · 2021 · 65 citations

• Oncology
• Psychology
• Internal medicine

F-AV-1451 tau PET (baseline age 73.9 ± 7.7 years, 375 female) were stratified into five stages by a topographic PET staging scheme. Cognitive trajectories and clinical progression were compared across stages with or without further dichotomy of amyloid status, using linear mixed-effect models and Cox proportional hazard models. Significant cognitive decline was first observed in stage 1 when tau levels only increased in transentorhinal regions. Rates of cognitive decline and clinical progression accelerated from stage 2 to stage 3 and stage 4. Higher stages were also associated with greater CSF phosphorylated tau and total tau concentrations from stage 1. Abnormal tau accumulation did not appear with normal β-amyloid in neocortical regions but prompt cognitive decline by interacting with β-amyloid in temporal regions. Highly accumulated tau in temporal regions independently led to cognitive deterioration. Topographic PET staging scheme have potentials in early diagnosis, predicting disease progression, and studying disease mechanism. Characteristic tau spreading pattern in Alzheimer's disease could be illustrated with biomarker measurement under NIA-AA framework. Clinical-neuroimaging-neuropathological studies in other cohorts are needed to validate these findings.

Sex and APOE ε4 genotype modify the Alzheimer’s disease serum metabolome

Nature Communications · 2020 · 208 citations

• Medicine
• Biology
• Bioinformatics

Late-onset Alzheimer's disease (AD) can, in part, be considered a metabolic disease. Besides age, female sex and APOE ε4 genotype represent strong risk factors for AD that also give rise to large metabolic differences. We systematically investigated group-specific metabolic alterations by conducting stratified association analyses of 139 serum metabolites in 1,517 individuals from the AD Neuroimaging Initiative with AD biomarkers. We observed substantial sex differences in effects of 15 metabolites with partially overlapping differences for APOE ε4 status groups. Several group-specific metabolic alterations were not observed in unstratified analyses using sex and APOE ε4 as covariates. Combined stratification revealed further subgroup-specific metabolic effects limited to APOE ε4+ females. The observed metabolic alterations suggest that females experience greater impairment of mitochondrial energy production than males. Dissecting metabolic heterogeneity in AD pathogenesis can therefore enable grading the biomedical relevance for specific pathways within specific subgroups, guiding the way to personalized medicine.

Metabolic Network Analysis Reveals Altered Bile Acid Synthesis and Metabolism in Alzheimer’s Disease

Cell Reports Medicine · 2020 · 206 citations

• Biology
• Biochemistry
• Internal medicine

Increasing evidence suggests Alzheimer's disease (AD) pathophysiology is influenced by primary and secondary bile acids, the end product of cholesterol metabolism. We analyze 2,114 post-mortem brain transcriptomes and identify genes in the alternative bile acid synthesis pathway to be expressed in the brain. A targeted metabolomic analysis of primary and secondary bile acids measured from post-mortem brain samples of 111 individuals supports these results. Our metabolic network analysis suggests that taurine transport, bile acid synthesis, and cholesterol metabolism differ in AD and cognitively normal individuals. We also identify putative transcription factors regulating metabolic genes and influencing altered metabolism in AD. Intriguingly, some bile acids measured in brain tissue cannot be explained by the presence of enzymes responsible for their synthesis, suggesting that they may originate from the gut microbiome and are transported to the brain. These findings motivate further research into bile acid metabolism in AD to elucidate their possible connection to cognitive decline.

Utility of the global CDR^® plus NACC FTLD rating and development of scoring rules: Data from the ARTFL/LEFFTDS Consortium

Alzheimer s & Dementia · 2020 · 161 citations

• Psychology
• Medicine
• Clinical psychology

INTRODUCTION: plus NACC FTLD to detect and track early frontotemporal lobar degeneration (FTLD) and to conduct clinical trials in FTLD. METHODS: The CDR plus NACC FTLD rating was applied to 970 sporadic and familial participants from the baseline visit of Advancing Research and Treatment in Frontotemporal Lobar Degeneration (ARTFL)/Longitudinal Evaluation of Familial Frontotemporal Dementia Subjects (LEFFTDS). Each of the eight domains of the CDR plus NACC FTLD was equally weighed in determining the global score. An interrater reliability study was completed for 40 participants. RESULTS: The CDR plus NACC FTLD showed very good interrater reliability. It was especially useful in detecting clinical features of mild non-fluent/agrammatic variant primary progressive aphasia participants. DISCUSSION: The global CDR plus NACC FTLD score could be an attractive outcome measure for clinical trials in symptomatic FTLD, and may be useful in natural history studies and clinical trials in FTLD spectrum disorders.

Challenges and Opportunities with Causal Discovery Algorithms: Application to Alzheimer’s Pathophysiology

Scientific Reports · 2020 · 133 citations

• Computer Science
• Machine Learning
• Artificial Intelligence

Causal Structure Discovery (CSD) is the problem of identifying causal relationships from large quantities of data through computational methods. With the limited ability of traditional association-based computational methods to discover causal relationships, CSD methodologies are gaining popularity. The goal of the study was to systematically examine whether (i) CSD methods can discover the known causal relationships from observational clinical data and (ii) to offer guidance to accurately discover known causal relationships. We used Alzheimer's disease (AD), a complex progressive disease, as a model because the well-established evidence provides a "gold-standard" causal graph for evaluation. We evaluated two CSD methods, Fast Causal Inference (FCI) and Fast Greedy Equivalence Search (FGES) in their ability to discover this structure from data collected by the Alzheimer's Disease Neuroimaging Initiative (ADNI). We used structural equation models (which is not designed for CSD) as control. We applied these methods under three scenarios defined by increasing amounts of background knowledge provided to the methods. The methods were evaluated by comparing the resulting causal relationships with the "gold standard" graph that was constructed from literature. Dedicated CSD methods managed to discover graphs that nearly coincided with the gold standard. For best results, CSD algorithms should be used with longitudinal data providing as much prior knowledge as possible.