Paul M. Thompson - US grants

Brain Modeling Challenges. Investigations into brain structure and function require a diverse array of tools to create. analyze, visualize, and interact with models of the brain. There is a rapidly growing need for brain models comprehensive enough to represent brain structure and function as they change across time,. as they change vary across large populations, in different disease states, across imaging modalities, across age and gender. and even across species. This mandates the development of 3D and 4D brain models that are adaptable to a wide variety of applications. including the detection and analysis of structural change and abnormality, in all their spatial and temporal complexity. The range and sophistication of these strategies will match the broad scope of studies that will focus on mapping and modeling the dynamically changing brain. Develop dynamic brain modeling tools, based on the concept of parameterization in 2, 3, and 4 dimensions, to represent and analyze brain structure as it changes in different disease states. across large populations, across age and gender, across imaging modalities, and across species. Computational brain modeling tools will be created to track and analyze complex patterns of three-and four-dimensional structural change in the brain during a variety of neurodevelopmental and degenerative disease processes. Tools will correlate structural indices with databases of clinical, behavioral and neuropsychiatric test data, to expand investigations of brain structure-function relationships to four dimensions. Specific Aim 2. Create mathematical strategies to analyze the structure of the cerebral cortex, as it changes across time in single subjects and groups. Novel computerized approaches will be developed to map temporal patterns of cortical development and degeneration, to encode patterns of cortical variation in human populations, and to detect abnormal gyral and sulcal patterns in individual patients and groups. Statistical anatomic models will map cortical change in four dimensions and detect anomalies in patient populations with Alzheimer's Disease, schizophrenia, and neurodevelopmental disorders. Specific Aim 3. Develop and extend software for automated extraction and analysis of brain structure models. Parameterization of anatomic models is critical for comparative neuroanatomy, as it makes models comparable at different time-points. Robust tools to extract surface models for a comprehensive range of neuroanatomic structures will accelerate and expand the scope of brain modeling projects and diagnostic applications in which brain structure models must be created rapidly and robustly. Specific Aim 4. Develop brain mapping tools and mathematical image analysis algorithms that adapt and learn from archives of three and four-dimensional neuroanatomic models. Immense archives of computational models resulting from in-house and collaborative modeling projects will be structured so that they can guide and inform mathematical algorithms which analyze future neuroanatomic data of the same type. Model-driven tools will include software for structure extraction. delineation of structural change and abnormality, analysis of functional brain image data, and nonlinear registration tools that integrate brain data from subjects and groups with brain structure differences.

[unreadable] DESCRIPTION (provided by applicant): [unreadable] The overall goal of this project is to develop a powerful computational framework to map disease and genetic effects on the brain. Blending neuroimaging and genetics techniques, we will go beyond the mapping and visualization of disease-specific patterns in the brain to develop a framework to map how genes affect brain structure in health and disease. In pilot projects, we will develop the mathematics and software necessary to link large-scale brain imaging and genetic studies of the brain. These algorithms will detect, map, and help understand patterns of abnormality in subjects at genetic risk for disease. They will also empower the identification and investigation of quantitative trait loci (QTLs) that confer vulnerability for disease. First, we will encode how brain structure varies in large populations. Novel algorithms will chart how the brain changes dynamically with age, gender, in childhood, and in health and disease. Specialized methods will track average, group-specific anatomical patterns in the cerebral cortex. These patterns will be stored in a computational/statistical brain atlas, and linked with cognitive, clinical and therapeutic parameters. To detect and map how genes affect brain development and disease, we will test new tools in genetically informed designs. These include large neuroimaging data components from (1) the Finnish twin registry, (2) normally developing young twins (ages 6-20) scanned longitudinally, (3) identical and fraternal twins discordant for schizophrenia, and (4) subjects with known risk genotypes (including a schizophrenia risk allele on chromosome 1q). Following up on our recent findings, our algorithms will be tested for uncovering deficits and gene effects in schizophrenia, but they will be designed to be applicable to any brain disease (Alzheimer's, autism, bipolar disorder, drug addiction alcoholism, and other neurodevelopmental/psychiatric disorders). We will mathematically combine algorithms from computational anatomy, partial differential equations, pattern theory, random field theory, and harmonic maps, to detect gene effects on the brain with maximal power. We will also map the heritability of brain structure. Novel tools for automated segmentation and labeling of brain structures will also accelerate large scale studies with these techniques. Validated on unique datasets, these tools will greatly empower biomedical studies that bridge imaging and genetics. They will help investigate the genetic transmission, triggers, and dynamics of disease in whole human populations. [unreadable] [unreadable]

DESCRIPTION (provided by applicant): We will develop a powerful computational framework to map disease effects on the brain. Many degenerative and developmental diseases affect the cerebral cortex, but its 3-dimensional geometry is so complex and variable across subjects that disease effects are hard to detect or compare across subjects and groups. Creating a new direction in the field of computational anatomy, we will build on revolutionary advances in the field of partial differential equations (PDEs) that allow geometric and statistical manipulation of surfaces. Extending these tools, we will build a general mathematical framework to analyze, compare, and process neuroimaging data represented on the cortical surface. New algorithms will detect and map disease effects on human brain structure, visualizing systematic deficits in gyral/sulcal patterns, cortical shape and tissue distribution, and gray matter thickness, all of which are sensitive indicators of disease. Using novel mathematics of implicit functions and mappings between manifolds, we will build tools that elastically transform cortical surfaces from multiple subjects to a common anatomic template, encoding complex differences in neuroanatomy across subjects (Aim 1). Adapting these PDEs to map the profile of gray matter thickness at the cortical surface (Aim 2), we will encode how brain structure varies in large populations, and chart how the brain changes dynamically with age, in health and disease. New surface-based signal processing methods based on Laplace-Beltrami filters and adaptive grids will optimize detection of subtle or diffuse effects on cortical structure and function (Aim 3). Specialized methods will track average, group-specific anatomic patterns in the cerebral cortex. Building on our recent findings, our algorithms will be tested for uncovering deficits in Alzheimer's disease, including brain changes over time (Aim 4). These tools will be designed to be applicable to any brain disease (schizophrenia, autism, bipolar disorder, drug addiction/alcoholism, and other developmental/psychiatric disorders). We will mathematically combine algorithms from computational anatomy, PDEs, pattern theory, random field theory, and harmonic maps, to detect disease effects on the brain with maximal power. Validated on unique MRI datasets, these tools will be made publicly available and will find immediate application in a range of neuroimaging collaborations: they will chart the dynamics and spatial profiles of disease and medication effects in whole human populations.

Neuroimaging offers a powerful strategy to chart the dynamic path of aging and Alzheimer's disease (AD) in the brain. This project will apply strategies with unprecedented sensitivity to detect, map, and analyze patterns of brain change (structural and functional) in AD and those at risk. Our novel brain maps will quantify disease progression to advance identification of early brain changes, and monitor treatment and gene effects. Using new mathematics and supercomputing technology, we will build on our work to create detailed 4D (dynamic) maps of brain changes based on serial magnetic resonance imaging (MRI). To increase sensitivity to detect and analyze drug and gene effects on brain changes, we will use (1) large cohorts scanned repeatedly over long intervals, and (2) improved image analyses using nonlinear deformation techniques and surface-based statistics. Algorithms will map the profiles, and rates, of cortical thinning, hippocampal change, gray matter loss, and volumetric atrophy. Collecting these maps from multiple populations, we will compare rates of brain change in patients with AD and FTD (Specific Aims 1 and 2), and MCI, CIND and healthy individuals with known risk genes (Aim 3). Probabilistic maps and statistics of these changes will be stored in a normative database, or brain atlas, and analyzed for group differences. We will stratify this population atlas by cohort, medication, symptom profiles, and risk genotype to help understand the dynamics of disease onset and progression, and how treatment affects these changes. We will determine how structural brain changes correlate with cognitive/metabolic decline (PET data) measured in the same subjects (Aim 4). In collaboration with an ongoing randomized, double-blind clinical trial (Aim 5), we will test whether drug treatment with donepezil and/or vitamin E decelerates atrophic rates in MCI subjects. The products of these efforts (engineering and neuroscience) will shed light on how dementia emerges in the brain and will improve our ability to track it. These tools will be publicly available, and will greatly expand our ability to investigate the dynamics of AD, detect its onset, compare patterns of therapeutic response, and understand its spatial/temporal selectivity.

[unreadable] DESCRIPTION (provided by applicant): This is a competitive renewal of a 10-year project that has provided enormous insight into Alzheimer's Disease (AD), a disease that costs $113 billion/yr in the U.S. alone, with no known cure. The project develops a multidimensional, computational atlas of AD. Our 5-year plan of work provides the most powerful computational tools to track AD emerging and spreading in the living brain - years before symptoms begin. We will correlate 3 perspectives on AD in an atlas coordinate framework - serial MRI, novel PET tracers, and 3D pathology. This will provide scientists will the most sensitive approach ever created to gauge which factors affect disease progression (drug treatment, genetic risk, etc.), and how effectively treatments slow the transition to AD. First, we chart the anatomic trajectory of AD with novel analyses of serial MRI (Aims 1,2). Our tools to detect brain changes (cortical thickness mapping, tensor-based morphometry) provided the first time-lapse maps of the disease spreading in the living brain. Here we apply them to MCI subjects (who are at five-fold higher risk of converting to AD in any given year) to identify the best predictors of imminent disease onset, and to predict changes in specific cognitive domains. This will greatly advance drug trials by better identifying candidates for early treatment, who can benefit most before irreversible damage sets in. Next, we will use novel PET tracer molecules to reconstruct the dynamic sequence of AD pathology as it builds up and spreads in the living brain (Aim 3). Hailed as a breakthrough in the AD community, our newly-developed PET (positron emission tomography) tracer compound, [18F]-FDDNP, visualizes amyloid plaques and neurofibrillary tangles (NFTs) - hallmarks of AD previously only detectable at autopsy. Our sensitive surface-based 3D analytic techniques will map the spatio-temporal trajectory of plaque and tangle build-up in aging, mild cognitive impairment, and AD, comparing groups to identify brain changes that predict imminent cognitive deterioration or transition to AD; we will compare and correlate these signals with MRI measures of atrophic rates and cortical degeneration to create joint MR-PET measures of disease burden. We will pioneer 3D cryosection imaging (in 3 subjects/year) to establish how 3D reconstructed maps of tangle density and betaamyloid distribution throughout the brain correlate with imaging measures from living patients. This groundtruth data will be a resource to the AD community, revealing the cellular correlates of imaging signals whose physiological meaning is poorly understood. We will map individuals and populations, revealing group patterns of cortical thinning and plaque and tangle pathology that predict outcomes. Identifying predictors of imminent decline and disease onset, our atlas will identify candidates with emerging pathology for early drug treatment, and will store statistical data to quantify how well treatments resist AD in those at risk. We will share all images, protocols, and algorithms with our 100+ collaborating laboratories. [unreadable] [unreadable] [unreadable]

[unreadable] DESCRIPTION (provided by applicant): This project significantly extends the power of MRI and diffusion tensor imaging (DTI) at ultra-high magnetic field strengths (7T) to resolve previously unseen features of brain structure and fiber properties, providing unique power to investigate disease. We will exploit the extreme spatial resolution and signal contrast at 7T, to compute sensitive biomarkers of aging, mild cognitive impairment (MCI), and Alzheimer's disease (AD). Combining data from 80 human subjects scanned at both 7T and 3T, we will compute the profile of cortical gray matter thickness (Aim 1), a measure sensitive to subtle changes in aging, development and a key target of drug trials in dementia and neuropsychiatric research. Extending DTI to 7T, we will assess white matter microstructure and fiber integrity with unprecedented power (Aim 2), and how it changes with aging and dementia. We will quantify how much 7T (versus 3T) boosts the detection sensitivity, signal to noise, stability and effect size of these key biomarkers of brain disease. Uniting the UCLA/University of Minnesota imaging centers, we will develop novel signal processing computations and mathematics to extract maximal information from the 7T images, advancing their current resolution and detection limits. New denoising, registration, and statistical techniques will detect individual and group differences in cortical thickness and fiber integrity. Advances in the mathematics of partial differential equations (PDEs), harmonic maps, and Riemannian manifolds, will provide unique power to denoise tensor- and vector-valued imaging signals (DTI). New statistics will detect disease-sensitive changes in the brain's fiber pathways. We will quantify how much effect sizes increase at 7T versus 3T, and what benefits our novel algorithms yield. In the first high- field study of AD and MCI, we will exploit the higher contrast and resolution at 7T to map key brain changes, undetected at 3T. We will unravel the geometry of cortical surfaces in the brain, and map how cortical thickness changes in aging (Aim 3), AD and MCI (Aim 4). We will map individuals and populations, encoding group patterns of cortical thinning and fiber architecture to detect local or diffuse brain changes. These gains will immediately advance clinical trials of anti-dementia and anti-psychotic drugs that depend on MRI for their power. Validated on unique data, our tools will help monitor disease progression, and map how brain diseases begin and spread in human populations. We will share all images, protocols, and algorithms with our network of 100+ collaborating laboratories. [unreadable] [unreadable] [unreadable]

Affinity; Assay; Binding; Binding (Molecular Function); Bioassay; Biocompatible; Biologic Assays; Biological Assay; Biosensor; Blood Serum; Body Tissues; Boronic Acids; CRISP; Cancers; Carbohydrates; Cell surface; Colon Cancer; Colon Carcinoma; Colonic Carcinoma; Complex; Computer Retrieval of Information on Scientific Projects Database; Detection; Development; Diagnostic; Diagnostic tests; Disease; Disorder; Funding; Glycans; Grant; Institution; Investigators; Lead; Lectin; Malignant Neoplasms; Malignant Tumor; Methods; Methods and Techniques; Methods, Other; Molecular Interaction; NIH; National Institutes of Health; National Institutes of Health (U.S.); Nature; Numbers; Optics; Pb element; Peptide Library; Peptide/Protein Chemistry; Peptides; Polysaccharides; Process; Protein Chemistry; Randomized; Reagent; Research; Research Personnel; Research Resources; Researchers; Resources; Series; Serum; Solid; Source; Spinal Column; Spine; Techniques; Therapeutic; Tissues; United States National Institutes of Health; Vertebral column; backbone; base; boronic acid; combinatorial; design; designing; disease/disorder; heavy metal Pb; heavy metal lead; malignancy; neoplasm/cancer; new diagnostics; next generation diagnostics; novel; novel diagnostics; randomisation; randomization; randomly assigned; response; sensor; sensor (biological); small molecule; sugar; tool

DESCRIPTION (provided by applicant): This project advances the state-of-the-art in High-Angular Resolution Diffusion Imaging (HARDI), a powerful new imaging approach that can resolve fiber pathways in the brain with spectacular precision. Uniting expertise at an NIH-funded National Neuroimaging Resource (at UCLA), the University of Minnesota Center for Magnetic Resonance Research, and Siemens Corporate Research, we aim to demonstrate that HARDI provides new and vital information in assessing clinically important brain degeneration in HIV/AIDS and Alzheimer's Disease (AD), extending our initial findings that revealed how these diseases spread dynamically in the living brain. HARDI applies magnetic field gradients to the brain, in up to 256 different directions, to precisely detail the directions, pathways, and integrity of fibers in the brain. HARDI datasets cannot yet be compared across subjects without new mathematics that treats these signals as lying in Riemannian manifolds. This project provides those tools. Our research will (1) advance the mathematics - based in part on geometry, statistics, and Riemannian manifolds - to extract information from HARDI, and (2) quantify how much HARDI can improve our understanding of brain degeneration, and what factors affect it. Using the extra detail in HARDI images, we will develop a method to enable large-scale multi-subject comparison of HARDI images, by fluidly aligning 3D images across subjects (Aim 1;Multi-subject Alignment). This is the first step towards population studies of disease, e.g., comparing fiber integrity across patient populations to examine gene or treatment effects, or comparing a patient with a normative database. Validation on phantoms and synthetic data is a key part of all Aims. In Aim 2 (Segmentation and Connectivity Mapping), we will develop algorithms to map white matter connectivity, and identify clinically important fiber pathways in the brain, based on the full angular information of HARDI. In Aim 3 (HARDI Statistics), optimized voxel-based statistics will compare HARDI data, point-by-point, across populations, to identify systematic fiber deficits, comparing fiber integrity and connectivity with a normal reference population. In Aim 4 (HARDI Maps of Brain Degeneration), we will evaluate HARDI for revealing new descriptors of AD and HIV-related brain degeneration: two illnesses on which we have published prolifically, where measures of white matter degeneration are sorely lacking. The societal burden of AD and HIV is growing;HIV affects 40 million people worldwide, and AD affects 4.5 million individuals in the U.S. alone;everyone is at risk. Our powerful markers of brain white matter degeneration will help us determine how much benefit HARDI's added resolution provides. This new analytic approach will greatly advance our ability to understand pathological brain degeneration, providing sensitive new measures to track it. This has immediate value for drug trials and patient monitoring. As always, we will share all algorithms, protocols, and images, with 50+ collaborating laboratories. PUBLIC HEALTH RELEVANCE: This project develops tools that unleash the full power of HARDI (high-angular resolution diffusion imaging) to advance clinical studies of the brain. HARDI applies magnetic field gradients to the brain in up to 256 different directions to precisely detail the directions, pathways, and integrity of fibers and their connections. We will evaluate HARDI for understanding and revealing new descriptors of Alzheimer's Disease and HIV-related brain white matter degeneration - with immediate value for drug trials and patient monitoring in HIV, which affects 40 million people worldwide, and in AD, which affects 4.5 million individuals in the U.S. alone.

DESCRIPTION (provided by applicant): The recent discovery of new Alzheimer's disease (AD) risk genes has re-ignited efforts to understand how genes interact to impact brain vulnerability to AD. Deeper genetic analysis of AD risk using sophisticated imaging biomarkers will enable us to (1) predict risk for AD in younger adults to initiate prevention strategies in those at risk and (2) boost power in drug trials by selecting those at greatest risk of decline. Our goal is to assess genetic control over two measures that show deficits in AD patients: (1) functional "connectivity" (synchronicity) between the posterior cingulate cortex (PCC) and medial frontal cortex (MFC) at rest and (2) white matter integrity in the associated cingulum tract and splenium of the corpus callosum. Our project advances the study of AD genetic risk by investigating how multiple risk genes interact to disrupt brain connectivity. In this first-ever genomic analysis of structural and functional connectivity, we use both genome-wide association scanning (GWAS) and a candidate gene analyses of two large healthy cohorts: (1) 1150 young adult Queensland twins (QTwin;age: 20-29), and (2) 273 older healthy and cognitively impaired adults in the Alzheimer's Disease Neuroimaging Initiative cohort (ADNI;age: 55-90). Carriers of a common AD risk gene (CLU) have deficits in white matter integrity even as young adults and AD pathology may later exploit this vulnerability. We now expand our investigation of DTI genetics to assess structural connectivity. We also study how risk genes may impair functional synchronicity between PCC and MFC using resting state fMRI (rs-fMRI). We will: 3/4 Use a candidate gene approach to reveal how (1) known risk genes impair connectivity and fiber integrity in the splenium and cingulum in the young and elderly (QTwin and ADNI) and how (2) PCC-MFC synchronicity is influenced in both cohorts by top identified AD risk SNPs and SNPs that affect white matter integrity in the splenium and cingulum. 3/4 Use GWAS to identify new risk genes, whose carriers have impaired brain connectivity. We previously used GWAS to identify genes related to deficits in gray and white matter. We extend these findings to determine SNPs associated with (1) reduced FA and structural connectivity in the splenium and cingulum and (2) reduced synchronicity of PCC-MFC fMRI signal in the QTwin and ADNI samples. Promising hits will be proposed for verification and meta-analysis in Enigma (http://enigma.loni.ucla.edu) (N=10,000). 3/4 Use new multi-locus genetic models (ridge regression, PC regression, vGeneWAS) to evaluate how multiple common risk SNPs and multiple genes interact to impair brain connectivity. The product of these efforts will be (1) new assessments of genetic risk for brain dysconnectivity in healthy adults, (2) a means to boost clinical trial power, based on imaging and multi-SNP modeling of liability. PUBLIC HEALTH RELEVANCE: The strength of brain connections is reduced in people with Alzheimer's disease (AD), a disease whose onset and progression are influenced by many known risk genes as well as lifestyle and environmental factors. We will use MRI measurements of this connectivity as an AD risk proxy to better characterize how known AD risk genes affect the brain, helping researchers to improve treatment focus. We will also use this proxy to identify new possible AD risk genes, allowing researchers to assess more homogeneous samples of people. Covarying for individual genetic risk will empower the evaluation of treatments and preventative measures, and identify subjects early in life who are vulnerable to impaired brain connectivity.

DESCRIPTION (provided by applicant): This project, Empowering Personalized Medicine: Integrating Imaging, Genetics and Biomarkers, responds to RFA-MH-12-020, entitled Integrating Multi-Dimensional Data to Explore Mechanisms Underlying Mental Disorders. By bringing together experts in neuroimaging, genetics, and mathematics, we plan to create an advanced, portable framework to combine diverse biomedical data from 3D neuroimaging (MRI, amyloid/FDG-PET), gene expression networks, genome-wide association studies (GWAS), and other multidimensional data (e.g., physiological biomarkers, epigenetic data, etc.). Our overall goal is to improve diagnosis and prognosis of disease by combining multiple levels of biological information (personalized medicine). In doing so, novel mathematical tools will automatically discover which biomarkers are most helpful in different contexts. To discover and test relationships between very high-dimensional measures (such as images and genomes), we use novel concepts for data reduction such as penalized regression (elastic nets), adaptive hierarchical clustering, Bayesian networks, and support vector machines. Avoiding the limitations of current work that tests individual gene effects independently, we extend the analysis of gene expression networks to images, to relate signs of disease to their genetic underpinnings and to all available biomarkers. Aim 1 empowers discovery genetic variants (identified in GWAS, whole-exome and whole-genome sequencing) that modulate measures of disease. We will use compressive coding models to discover and verify which sets of genetic variants affect multidimensional images (e.g., co-registered MRI & PET, DTI). We will verify our predictions using k-fold cross-validation and independent replications in new samples and controllable test data. Aim 2 extends our work using weighted gene co-expression network analysis (WGCNA) from single traits to entire databases of 3D images (MRI/PET). Our framework will merge GWAS, eQTL analysis, and expression-phenotype analysis but will be broadly applicable to any future high-throughput biological information (e.g. methylation profiles, DTI, fMRI). In Aim 3, we will quantify the added predictive value derivable from genotyping, gene expression profiling, and multimodal neuroimaging for personalized prognosis and diagnosis. For example, which biomarkers (gene expression, CSF, MRI) are most useful in which cases? To maximize impact of this effort, we and our collaborators will test our tools on existing and new datasets from a range of neuropsychiatric disorders including frontotemporal dementia, Alzheimer's disease, schizophrenia, bipolar disorder, and autism (see Support Letters). All tools will be disseminated and linked to web-accessible databases that store and ease access to high-throughput genetic, genomic, and imaging datasets.

DESCRIPTION (provided by applicant): In affluent countries where combination anti-retroviral therapy (CART) is widely used, life expectancy with HIV has increased well beyond 10 years, but this has greatly increased the number of people living with HIV/AIDS. 40 million people are now at risk for progressive HIV-related damage to the brain (Navia 2011), and 40% show some neurological or cognitive impairment. Building on our recent discoveries and innovations in MRI and diffusion tensor imaging, our project charts the dynamics of HIV disease in the brain, revealing factors that predict clinical decline and brain decline. By analyzing data from the HIVNC (HIV Neuroimaging Consortium) cohort (218 HIV+ subjects scanned longitudinally with MRI every 26-32 weeks for 3 years; 900 scans), we will use our sensitive image analysis method, tensor-based morphometry to (1) map rates of brain tissue loss over time, determining which brain systems lose tissue fastest; (2) relate these loss patterns to neurocognitive impairment (NCI); and (3) determine which host and disease-related factors, at baseline, predict higher atrophic rates over the 3-year follow-up interval. To deeply probe the causes of dysfunction, we will use whole-brain tractography based on diffusion tensor imaging in 267 subjects to determine where HIV+ patients have reduced fiber integrity. We will use each patient's DTI scan to compute a whole-brain connection matrix, based on a state-of-the-art whole-brain tractography method we developed. We will combine the best neuroimaging measures from Aims 1 and 2 into our support vector machine method to predict (1) future rates of atrophy, and (2) cognitive decline over the 3-year follow-up. With our best predictors of decline from Aims 1 and 2, we will predict which HIV+ patients will show imminent decline. We will estimate the sample sizes needed for a neuroimaging-based drug trial to detect a 5%, 10%, or 25% slowing in the rate of atrophy, and the same percents of slowing in the rate of cognitive decline. We will test whether our predictors generalize to the large, independent Charter and Miriam datasets beyond HIVNC (see Support Letters from Drs. Igor Grant and Ron Cohen), and when used by our many HIV research collaborators now using our methods; see Pilot Data). As shown in new pilot data, we assess how imaging protocol differences affect the measures; our innovations to reduce scan protocol confounds (e.g., adjusted FA) will guide selection of the most robust predictors. We will evaluate how useful these new measures are for predicting cognitive decline, and boosting power in an antiretroviral drug trial. In other words, to what extent can a DTI scan, and an MRI-based map of atrophy, help to make clinical predictions of imminent cognitive impairment? Can they help select a sample for a drug trial? Our activities will make clinical trial design more efficient by selecting subjects with greater potential to respond t future therapies. As always, we will disseminate our methods to both experts and trainees in medicine, neuroscience, engineering, and to our network of over 100 collaborating labs (including two national HIV consortia: HIVNC, CHARTER, and the Miriam HIV cohort), to accelerate their work.

DESCRIPTION (provided by applicant): The ENIGMA Center for Worldwide Medicine, Imaging and Genomics is an unprecedented global effort bringing together 287 scientists and all their vast biomedical datasets, to work on 9 major human brain diseases: schizophrenia, bipolar disorder, major depression, ADHD, OCD, autism, 22q deletion syndrome, HIV/AIDS and addictions. ENIGMA integrates images, genomes, connectomes and biomarkers on an unprecedented scale, with new kinds of computation for integration, clustering, and learning from complex biodata types. ENIGMA, founded in 2009, performed the largest brain imaging studies in history (N>26,000 subjects; Stein +207 authors, Nature Genetics, 2012) screening genomes and images at 125 institutions in 20 countries. Responding to the BD2K RFA, ENIGMA'S Working Groups target key programmatic goals of BD2K funders across the NIH, including NIMH, NIBIB, NICHD, NIA, NINDS, NIDA, NIAAA, NHGRI and FIC. ENIGMA creates novel computational algorithms and a new model for Consortium Science to revolutionize the way Big Data is handled, shared and optimized. We unleash the power of sparse machine learning, and high dimensional combinatorics, to cluster and inter-relate genomes, connectomes, and multimodal brain images to discover diagnostic and prognostic markers. The sheer computational power and unprecedented collaboration advances distributed computation on Big Data leveraging US and non-US infrastructure, talents and data. Our projects will better identify factors that resist and promote brain disease, that help diagnosis and prognosis, and identify new mechanisms and drug targets. Our Data Science Research Cores create new algorithms to handle Big Data from (1) Imaging Genomics, (2) Connectomics, and (3) Machine Learning & Clinical Prediction. Led by world leaders in the field who developed major software packages (e.g., Jieping Ye/SLEP), we prioritize trillions of computations for gene-image clustering, distributed multi-task machine learning, and new approaches to screen brain connections based on the Partition Problem in mathematics. Our ENIGMA Training Program offers a world class Summer School coordinated with other BD2K Centers, worldwide scientific exchanges. Challenge-based Workshops and hackathons to stimulate innovation, and Web Portals to disseminate tools and engage scientists in Big Data science.

The ENIGMA Center for Worldwide Medicine, Imaging and Genomics galvanized the brain imaging and genomics communities worldwide to pool all their data, talents, and infrastructure to work on previously intractable computational and biomedical goals. ENIGMA'S Administrative Core offers support, leadership, and democratic policies to create the largest brain imaging genomics studies in history. Our Aims are: Aim 1. Coordinate the largest worldwide genomic analyses of images. The ENIGMA Center and its Support Groups will coordinate work by 287 scientists at 125 institutions. Our expertise, administrative support, and analytical resources will accelerate worldwide studies of the brain across the U.S., Europe, Australia, Asia and Africa. Our Support Groups distribute computations on worldwide genomic, imaging and clinical data, in ever-increasing power and depth. Aim 2. Coordinate Worldwide ENIGMA Working Groups on Disease. ENIGMA coordinates 9 mutually supportive Working Groups on 9 major worldwide brain diseases: schizophrenia, bipolar disorder, major depression, ADHD, OCD, autism, 22q deletion syndrome, HIV/AIDS and addictions. These multinational activities use large-scale distributed computation to draw on massive infrastructure and expertise from 287 scientists from 20 countries. Guiding principles are: clear and unified medical goals, protocol harmonization, consortium science, and meta-analysis to improve disease diagnosis and prognosis worldwide. Aim 3. Implement ethical worldwide collaboration. We will assure ethical handling of biomedical Big Data, authorship, credit, and democracy. We encourage secondary proposals to work with ENIGMA data. Projects use Memoranda of Understanding (MoUs), with clear policies for timelines, embargo handling and conflict resolution. Aim 4. Sustain ENIGMA'S growth. We introduce metrics to evaluate how our ENIGMA Center impacts the scientific community. We synergize with other Big Data efforts, helping with training, and scientific exchange, saving costs. Sustained funding will involve philanthropy and will leverage multi-continent support and our partners' non-US infrastructure.

Our ENIGMA Consortium is a truly international consortium involving researchers from 20 countries and 125 institutions worldwide. The consortium is organized into (1) the largest imaging genomics effort in history, analyzing brain MRI scans from 26,000 individuals with genome-wide data, (2) technical experts in the mathematics of Big Data, and (3) 9 Working Groups on schizophrenia, bipolar disorder, depression, ADHD, autism, OCD, 22q, HIV, and addictions. The opportunities for training and dissemination - and the global reach of the ENIGMA project - lie at the heart of ENIGMA'S ultra-collaborative training efforts. ENIGMA spans imaging, genetics, medicine, and mathematics, with diverse opportunities for cross-disciplinary training for researchers at all levels, to advance Big Data science. Aim 1. ENIGMA Scientist Exchange Program. This innovative program will train researchers to analyze Big Data, focusing on imaging and genomics. Aim 2. International Summer School. Building on our experience hosting summer schools on mathematics in medical imaging for hundreds of students, our annual summer school will host around 100 students for 2 weeks. These lectures will be archived on a YouTube Channel to facilitate access and allow us to update content. Meeting venues will rotate annually for maximum reach. Aim 3. Workshops for Novices in ENIGMA Themes, Big Data Science. These 1-day hands-on workshops offer training in tool use for people analyzing Big Data. Aim 4. ENIGMA Challenges. Expert users will compete to solve technically tough problems (ENIGMA Challenges) using the Hackathon/MICCAI Challenge model - yielding a published proceedings of short papers. Aim 5. ENIGMA Website and Portal to ENIGMA Tools (ENIGMA-Vis, SLEP) and Training Materials. We will offer a single portal (http://enigma.ini.usc.edu) to access (1) ENIGMA'S tools, protocols, and software, (2) web-based queries of gene effects on brain measures, via the ENIGMA-Vis tool, and (3) multimedia updates on discoveries across ENIGMA, papers, video tutorials, radio interviews, news, and publicity materials to interest trainees, visitors, and new participants.

Via our BD2K Consortium Activities Plan (with 4 Specific Aims), we will support the NIH BD2K Consortium and Program in its quest to propel Big Data Science across disciplines to create a movement of scientific discovery that no one BD2K Center could achieve on its own. We are extremely proactive in Consortium Building and Consortium Support: The ENIGMA Center for Worldwide Medicine, Imaging and Genomics is already a global Consortium uniting 287 scientists from 20 countries, endeavoring to examine 9 major human brain diseases by integrating images, genomes, connectomes, and biomarkers using new kinds of computation. ENIGMA performed the largest brain imaging studies in history (N>26,C)00 subjects; Stein -1-207 authors. Nature Genetics, 2012) and it continues to evolve, establishing links to other powerhouses of scientific discovery. The BD2K Consortium will revolutionize how Big Data is handled, shared and optimized - with common themes likely in all proposals. Our 4 Specific Aims will help maximize synergies across funded Centers, targeting programmatic goals of BD2K funders across NIH. Our 4 Aims are: Aim 1. Establish a Trans-BD2K Scientific Exchange Program. Having supported Scientific Exchange Programs for 20 years, we propose a Trans-BD2K Scientific Exchange Program where trainees learn Big Data Science at multiple BD2K Centers, offering a new education in Big Data Science. Trainees will develop project proposals, projects, and initiatives that foster and exploit trans-BD2K connections. Aim 2. Workshops and Partnerships with NIH BD2K Funders. In regular NIH meetings to evaluate, direct, and synergize Center activities, we will work with NIH ICs program directors whose missions our ENIGMA Center targets (NIMH, NIBIB, NICHD, NIA, NINDS, NIDA, NIAAA, NHGRI) to set priorities for future work, and adapt discoveries in Big Data Science at other BD2K Centers. Such a dialog avoids the pitfalls of initiating a program without continual feedback from the BD2K funders, designers, and stake-holders. Aim 3. Connect with Other Big Data Consortia To Bring their Expertise and Data to BD2K. With our BD2K Center partners, we will proactively seek collaborations, joint workshops, and joint traineeships with other emerging Big Data consortia to connect other funded efforts with the NIH BD2K Program mission; these include vast genomics efforts covered in our ENIGMA-PGC Working Group (with the Psychiatric Genomics Consortium), our ENIGMA-ILAE Working Group (International League Against Epilepsy) and others as they emerge. Aim 4. Collaborative Big Data Algorithm Repository. As our tested and validated protocols, algorithms, and software will all be made publicly available, we will contribute to a central Open Source Tool Repository with other centers with similar computational needs and goals.

Our comprehensive Data Science Research program is organized into 4 complementary Algorithm Research Cores: (1) Imaging Genomics, (2) Connectomics, (3) Machine Learning & Clinical Prediction, and (4) ENIGMA Disease Working Groups. World leaders in each field lead each Core, tackling computational questions on a scale not previously imagined or attempted. BigData tools will fuel ENIGMA'S worldwide scientific discoveries. We combine innovations in mathematics, machine learning, genomics, consortium science, and expertise from >20 countries and >125 institutions. ENIGMA is not a project, it is a scientific movement of rapidly and constantly Interacting collaborations that support each other. ENIGMA cohorts boost each other's power with gigantic datasets, tools and expertise to maximally exploit each other's data, performing some of the world's largest disease studies, beyond what any one site could perform alone. ENIGMA is distributed computation at its best, drawing on gigantic datasets and expertise. We create massive economic savings - drawing on worldwide computational and infrastructural resources vastly beyond what any one site in any one country would apply to a targeted biomedical goal. We bring BigData Science and the ENIGMA Consortium together to advance Worldwide Medicine. In Cores 1 and 2 we mine images and connectomes for genetic markers that re-wire the brain or boost brain tissue loss, using new mathematics to prioritize and organize trillions of computations, jointly searching images and genomes. In Core 3, we unleash multi-task sparse learning to predict diagnosis and prognosis from vast high-dimensional biomarker data in the largest neuroimaging genetics datasets ever. In Core 4 we use these tools in a massive distributed computation: a vast, mutually interacting set of 9 Worldwide Working Groups, led by experts in 9 major diseases of the brain - schizophrenia, bipolar, major depression, ADHD, autism, OCD, 22q deletion syndrome, HIV/AIDS, and addictions. We will discover what genes, medications, and lifestyle factors promote or resist brain disease worldwide. Our ENIGMA Center is a worldwide movement in mutually supportive discovery in medicine - spurred on by tools to perform gigantic computations never before Imagined.

DESCRIPTION (provided by applicant): The aging brain suffers from progressive brain tissue loss, putting hundreds of millions of us at risk for memory loss and dementia with no known cure. We recently discovered, in studies related to physical activity, growth factors, and homocysteine, several new leads about what may promote brain integrity. These variables interact with each other, with inflammation, and with brain structure and disease risk in complex and striking ways. To resist the looming epidemic of degenerative disease, we must determine how such variables deter brain decline. We will map, with unprecedented sensitivity, how these vital processes, which are targetable with interventions, relate to brain structure. We also combine brain measures to predict, with machine learning who will imminently decline. We relate brain and cognitive decline in 3 populations to (1) an inflammatory marker and related neurodegenerative risk genes, (2) growth factor and homocysteine levels, and (3) physical exercise. We will discover how these variables interact to promote or deter brain disease, in our regions of interest (ROIs) - the hippocampus, prefrontal cortex (PFC), and cingulate gyrus. This work is crucial to combat diseases such as Alzheimer's disease (AD) and schizophrenia (SZ), among others that are both related to these variables and marked by deficits in these ROIs. We will also survey the whole brain with novel brain mapping methods. Physical activity promotes human brain regeneration in these regions, in part through its effects on growth factors. It decreases inflammation and homocysteine levels, both of which relate to hippocampal integrity. Our project will boost power in clinical trials by identifying targetable disease-driving mechanisms. Building on our recent discoveries, we use (1) genotyped variants in inflammation-related genes, (2) a controllable environmental variable (exercise), and (3) peripheral measures related to inflammation, neurogenesis, and homocysteine to assess integrity of the hippocampus, PFC, and cingulate gyrus, and their progressive decline. We examine three populations already scanned with MRI: the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort (566 AD, MCI, and normal elderly subjects), the Cardiovascular Health Study (CHS; 517 AD, MCI, and control subjects; of whom 85 have our targeted serum measures), and the Dominantly Inherited Alzheimer Network, composed of younger adults from families with autosomal dominant AD mutations (DIAN; ~400: ~20% with AD; ~50% mutation carriers). We will assess: 3/4 how candidate gene variants associated with AD and inflammation (see Approach) relate to a serum marker of inflammation (tumor necrosis factor alpha;TNF?), and to hippocampal volume. 3/4 how serum levels of growth factors and homocysteine relate to hippocampal, PFC, and cingulate volume. 3/4 how exercise levels affect hippocampal volume; how growth factors and TNF? modulate this relationship. 3/4 how baseline serum levels of growth factors, TNF?, and homocysteine, polymorphisms in related genes, and exercise together predict brain and cognitive changes over a 2-year follow-up period.

? DESCRIPTION (provided by applicant): Patients with Mild Cognitive Impairment (MCI) are at high risk of progression to dementia. MCI offers an opportunity to target the disease process early. Clinicians and researchers are intensifying their efforts to detect MCI pre-symptomatically in order to develop preventive treatments. These efforts generate a large amount of data - brain images of multiple modalities, and proteomics, genetic, and neurocognitive data that provide unprecedented opportunities to investigate MCI-related questions with greater precision and predictive power. Understanding its importance, NIH in 2003 funded the Alzheimer's Disease Neuroimaging Initiative (ADNI) to facilitate scientific evaluation of various biomarkers for the onset and progression of MCI and AD. To realize such an ambitious vision, there is an urgent need for multi-source fusion and disease biomarker discovery frameworks. While promising, large volumes of incomplete data from multiple heterogeneous data sources pose huge challenges to scientists and engineers. For instance, the ADNI-1 data (like many other large datasets) exhibit a block-wise missing pattern: most subjects have MRI, genetic information; about half of the subjects have CSF measures; a different half of the subjects have FDG-PET; and some subjects have proteomics data. Although many bioinformatics tools are available, no existing tools offer an effective way to fuse multi-source incomplete data for disease biomarker discovery. Here we aim to develop a novel computational framework to integrate and analyze multiple, heterogeneous, large volume, incomplete biomedical data for early detection of MCI. Our 4 primary aims are: (1) Develop novel structured sparse learning formulations for multi-source fusion. The computational methods will identify biomarkers to correlate multi-source data with MCI. Novel sparse screening methods will be developed to scale the proposed formulations to very high-dimensional data. (2) Develop computational methods to integrate network data. We will develop novel methods for incorporating existing biological knowledge such as pathways represented as networks into the prediction model. The network structure will be used as prior knowledge to constrain model parameters, to further improve predictive power. (3) Develop computational methods to integrate multiple incomplete data sources. The proposed computational framework will integrate multiple heterogeneous data with a block-wise missing pattern. The proposed framework formulates the multiple incomplete data source fusion problem as a multi-task learning problem by first decomposing the prediction problem into a set of tasks, then building the models for all tasks simultaneously. (4) Develop and disseminate software tools for multi-source fusion and biomarker identification. The software tools will be used for early detection of MCI and will be validated by several clinical research projects. Our open source software will be made freely available to the research communities, including our large community of existing users. One of our current packages, SLEP, has ~4,500 active users from ~25 countries. Our software tools will be easily adaptable for analyzing multi-source data from other neurological and psychiatric disorders.

ABSTRACT Exposure to trauma and abuse during childhood, a critical neurodevelopmental period, is a major risk factor for adult psychopathology. However, not all children exposed to childhood trauma will develop adult psychopathology. Variability in the risk for trauma-related pathology is expected to arise in part from genetic susceptibility. Several genes have recently been identified that interact with childhood trauma to increase rates of anxiety and mood disorders in adulthood. This risk can be more easily detected by examining endophenotypes such as brain measures obtained from MRI because of a simpler underlying genetic architecture with fewer individual genes or pathways than the multiple factors driving overall risk for psychopathology. Understanding the molecular-genetic contributors to brain structure that conspire with early-life environment (psychological trauma) and lead to adult psychopathology, will require large-scale collaborative efforts which harness big-data methodologies. Our goal is to conduct a GWAS of relevant structural brain measures in individuals exposed to childhood trauma, with the long-term goal of identifying genetic modulators of brain structure that are informative for early prediction and treatment for a range of psychiatric disorders where childhood trauma is a major risk factor. We hypothesize that (1) childhood trauma will interact with specific genetic markers to produce structural brain alterations and adult psychopathology, (2) that unique genetic variants, in the context of genetic vulnerability to childhood trauma, will influence the onset of specific disorders (e.g. depression vs PTSD), as well as (3) the presentation of specific symptom constructs (e.g. sustained threat) across disorders. Finding disease-associated genetic variation that point to molecular mechanisms of pathogenesis has proven challenging due to the polygenicity of clinical phenotypes. Leveraging neuroimaging phenotypes may offer a more direct path than clinical phenotypes in identifying these elusive genetic markers and relevant neurobiological pathways. Ultimately, the promise of finding genetic contributors of any psychiatric disorder is in identifying the presence of new biologic pathways for which targeted interventions may be devised and deployed.

Alzheimer?s disease (AD) is the most common form of dementia, with the number of affected Americans expected to reach 13.4 million by the year 2050. While it is well known that AD leads to progressive neuronal death, the exact mechanism of AD remains elusive. Currently, a definitive diagnosis can only be reached by autopsy or brain biopsy, and the neurodegenerative processes in two AD patients can follow very different courses. Further, treatment options for AD remain limited, let alone cure. For this reason, non-invasive neuroimaging has been extensively investigated in the hope that it may provide more sensitive markers for screening and early detection of AD. Yet, despite the amount of resources devoted to AD imaging research, CSF Tau and A?42 continue to outperform any non-invasive imaging markers. Multimodal connectomics, including functional and structural connectome (derived from fMRI and diffusion MRI respectively), has the potential to gain system-level structure- function insights into the mechanisms of AD and thus offers a novel platform for developing new diagnostic strategies. Despite a number of interesting connectome findings in recent years, few of these connectome results have been replicated independently or proven clinically relevant, which can be partially explained by the sensitivity to parameter settings during preprocessing and connectome construction, such as the choice of brain parcellation and the type of fMRI time series correlations (full versus partial) or tractography (deterministic or probabilistic). Moreover, conventional connectome approaches usually focus on scalar summary statistics (e.g., nodal or edge-wise measures) using linear statistical techniques, which fit at each node (or edge) independent of other nodes (or edges) and thus discard important informative graph structure. Instead, this proposal will develop a multi-view connectome framework that homogenizes multiple instances of stable and reproducible high-level connectome properties across modalities and across spatiotemporal scales. This framework will be applied and cross-validated using two independent AD cohorts (Alzheimer's Disease Neuroimaging Initiative or ADNI and Wisconsin Alzheimer?s Disease Research Center cohort or Wisconsin ADRC). The identified connectome features can serve as the potential non-invasive markers for guiding the AD diagnosis.

PROJECT SUMMARY - OVERALL The LONIR is focused on developing innovative solutions for the investigation of imaging, genetics, behavioral and clinical data. The LONIR structure is designed to facilitate studies of dynamically changing anatomic frameworks, e.g., developmental, neurodegenerative, traumatic, and metastatic, by providing methods for the comprehensive understanding of the nature and extent of these processes. Specifically, TR&D1 (Data Science) focuses on methodological developments for the management and informatics of brain and related data. This project will develop and issue new methods for robust scientific data management to create an environment where scientific analyses can be reproduced and/or enhanced, data can be easily discovered and reused, and analysis results can be visualized and made publicly searchable. TR&D2 (Diffusion MRI and Connectomics) seeks to advance the study of brain connectivity using diffusion imaging and its powerful extensions. This project will go beyond traditional tensor models of diffusion for assessing tissue and fiber microstructure and connectivity, develop tract-based statistical analysis tools using Deep Learning, introduce novel adaptive connectivity mapping approaches, using L1 fusion of multiple tractography methods, and provide mechanisms to study connectivity and diffusion imaging over 10,000 subjects. (This technology and these methods will be managed and executed by the TR&D1 framework to distributed datasets totaling over 10,000 subjects). Lastly, our TR&D3 (Intrinsic Surface Mapping) develops a general framework for surface mapping in the high dimensional Laplace-Beltrami embedding space via the mathematical optimization of their Riemannian metric. Our approach here overcomes fundamental limitations in existing methods based on spherical registration by eliminating the metric distortion during the parameterization step, thus achieving much improved accuracy in mapping brain anatomy. Coupled with a mature and efficient administrative structure and comprehensive training and dissemination, this program serves a wide and important need in the scientific community.

PROJECT SUMMARY- TR&D2: DIFFUSION MRI AND CONNECTOMICS In our new TR&D2 project, Diffusion Imaging & Connectomics, we extend our decades of work in the mathematics of diffusion MRI to create new and useful tools. Our work draws on our last 4 years of productive collaborations, which led to over 100 papers. We focus on 3 tasks: (1) modeling white matter microstructure, (2) automatically extracting maps of fiber bundles, and (3) modeling brain connectivity in more powerful and adaptive ways. Each method seeks to overcome a major problem that affects the validity of diffusion MRI today. In Aim 1, we develop mathematical metrics of white matter microstructure that can better use the available information in low-quality dMRI, but drastically boost the analytical power of higher-quality dMRI. With our diverse range of Collaborative Projects - including one that collects ultra-high resolution diffusion spectrum imaging at high-field (7T DSI - we test extensions of our tensor distribution function model of dMRI, which outperforms the standard tensor model and the most popular metric, DTI-FA. We extend the TDF to bi-exponential and multi- compartmental models, using TV-L1 fusion to merge all metrics across the image to better detect disease. In Aim 2, we extend our tools for automatic tract labeling. Our new approach, FiberNET, uses deep learning to avoid the manual intervention required by our initial tool, autoMATE. We will compare it head to head for accuracy, power, and utility on all our Collaborative Project datasets. These include diffusion MRI data from HIV+ children, from elderly people with various types of dementia, and adolescents at risk for psychiatric disorders. In Aim 3, we use two approaches - L1 fusion (dFUSE) and connectivity based on continuous functions (ConCon) - to overcome 2 very serious problems with the analysis of brain networks - the problem of large number of false positive fibers in tractography, and the arbitrariness of picking a cortical parcellation. These problems affect all downstream network metrics, to the point where the networks are very hard to compare across studies, with current methods. ConCon models the connectivity directly as a density, making it amenable to registration and statistics without defining specific borders in the cortex; dFUSE combines the reliable fibers across many tractography methods, boosting the contribution of methods that are more reliable on a given dataset. All our tools will be widely disseminated. With our broad collaborative network, we hope to advance the neuroscience of brain connectivity by using novel mathematics, tested head to head against existing methods.

ABSTRACT This application is the first, large scale, concerted effort to study sex as a biological variable in human neuroscience through a worldwide brain initiative. We created it to respond to the NIH mandate to address major gaps in knowledge on why some brain disorders are more prevalent in women than men, and vice versa. Women and men differ in aging trends and in the prevalence of major depression, anxiety disorders, substance abuse disorders (SUDs), alcoholism, psychosis, post-traumatic stress disorder (PTSD), and neurodegenerative diseases such as Alzheimer's disease. There are significant sex differences in the brain disorders' mean age of onset, treatment response, symptom severity and disease trajectory. Yet science knows alarmingly little about why. To respond to this, we launch a new worldwide brain initiative: ?ENIGMA-SD: Understanding Sex Differences in Global Brain Health through ENIGMA?. ENIGMA's Sex Differences Initiative pools data worldwide to address the ?crisis of reproducibility? from underpowered studies (Ioannidis 2014). We bring two sources of innovation: (1) the vast, unprecedented power of a global study across 35 countries, yielding exceptional sensitivity to sex differences and (2) a lifespan approach, to chart disease emergence throughout life, to pick up sex-specific `shifts' in the age of onset, severity, and trajectory of brain disease. We will consistently analyze imaging, genomic, and clinical data from hundreds of institutes - to understand (1) sex differences in brain structure and function across life; and (2) how brain abnormalities differ by sex in five major psychiatric illnesses: major depression (MDD), bipolar disorder (BD), schizophrenia (SCZ), PTSD, and SUDs. We will leverage the success of our ENIGMA consortium - a global study of 18 major brain diseases. Our Aims are: Aim 1. Chart Sex Differences in Brain Aging throughout Life. Building on our largest-ever normative study of the brain in 10,144 people (ENIGMA-Lifespan; Dima 2017), we will create sex-specific lifespan ?charts? showing how the brain ages, on average, in women and in men. Using MRI, DTI and resting state fMRI on a worldwide scale, we measure cortical/subcortical brain structures, white matter tracts, and age- sensitive metrics of functional brain synchrony. We develop statistical norms based on age and sex; we test interactions between biological sex and brain aging metrics. Aim 2. Determine Sex Differences in the Trajectory of Brain Deficits in MDD, BD, SCZ, PTSD, and Addictions. Building on our global neuroimaging studies of 5 prevalent brain disorders - we will ask: (1) Are the brain deficits the same in women and men with each disease? (2) Are brain differences greater in women at the same level of clinical severity? Aim 3. Sex and Genetic Risk. In a unique collaboration across our 5 global consortia ? ENIGMA-MDD, -BD, -SCZ, -PTSD, and -SUDs ? we will compute polygenic risk scores for each disease, and ask: What is the effect on the brain of carrying a higher genetic risk for one of these disorders? Does this effect differ by sex? We will do the groundwork to submit an ENIGMA-wide U01 on Sex Differences in Brain Disease, in 2022.

Abstract Core B The Imaging Core B (Paul Thompson, Director) provides brain image analysis for Projects 1-4, which analyze neurodegenerative responses to air pollution. Sub-core B1 (M. Braskie) supports Projects 1 and 2 on human brain MR images. Sub-core B2 (B. Zlokovic) supports Projects 3 and 4 on mouse brain. Sub-core B1 Human Neuroimaging Aim 1: Perform standardized structural image analyses in the ADNI, WHIMS and VETSA cohorts to assess voxelwise cortical thickness, volumes of selected Alzheimer's disease (AD)-relevant subcortical structures (hippocampus and amygdala), and regional white matter using FreeSurfer software. We will also assess white matter hyperintensity volumes. In ADNI and VETSA we will analyze diffusion tensor imaging scans. Aim 2: Use machine learning to generate structural MRI cortical signatures and associated predicted risk scores related to risk for preclinical AD, mild cognitive impairment, and AD. Aim 3: Support Projects 1 & 2 in their proposed mediation analyses using both targeted approach in structural equation models (SEM) and agnostic approach with high-dimensional analyses. Sub-Core B2, Mouse brain imaging. Aim 1. Perform in vivo mouse brain imaging to assess cerebral blood flow, blood brain barrier (BBB) permeability, vascular angiography, and MRI tractography. Aim 2. Postmortem studies of the neurovascular unit & BBB by confocal microscopy for immunofluorescence. Aim 3. Ultra-highfield MRI to detect mouse connectivity changes through diffusivity and tractography maps & quantification of BBB permeability by dynamic contrast enhanced (DCE)-MRI. Aim 4 Analyze mouse data from Projects 3 and 4.

This revised application, submitted in response to NIH PA-18-741 ?Secondary Analyses in Obesity, Diabetes, and Digestive and Kidney Diseases,? relates to the public health problem of childhood obesity, with a specific focus on the characterization, determinants and role of energy balance homeostasis-related brain circuitry in the human newborn. The proposal describes a comprehensive plan designed to advance neonatal MRI analytical methods, developmental systems neuroscience, and fetal programming of health and disease risk. The critical importance of the hypothalamic-limbic-cortical brain circuitry that regulates energy homeostasis is well established, and MRI-based measures of energy homeostasis-related brain circuits have been associated with obesity outcomes. However, it is unclear whether the observed differences in these brain regions and circuitry in obese relative to normal-weight individuals are a cause, consequence, or both, of the obese state. Moreover, relatively little is known about the developmental ontogeny of these brain regions and circuitry, particularly during the fetal period, and their prospective role in shaping propensity for childhood obesity. This project addresses this fundamental knowledge gap. We will develop novel measures and conduct analyses using newborn brain imaging and other data elements from 4 inter-linked NIH-funded projects on prenatal stress and fetal programming of brain development and infant body composition (R01 HD-060628; R01 MH-091351; R01 HD-065825; UG3 OD-023349). The importance of selecting the newborn brain as the starting time point derives from the logic that brain circuitry at this time could not yet have been influenced by postnatal factors such as diet/feeding, thereby enabling the study to disentangle the temporality of effects. In a recent position paper on the pathogenesis of obesity, the U.S. Endocrine Society emphasized the need to conceptualize obesity as a disorder of the energy homeostasis system and elucidate its underlying mechanisms and developmental influences. Towards this objective, and using a population of ~100 mother-infant dyads followed from early gestation through birth till 5 yrs age, we will integrate research aims that leverage the resources of these projects to advance our understanding of the origins of childhood obesity. Aim 1. Develop measures of energy homeostasis brain circuitry using anatomical, diffusion and functional MRI. Because such measures have not yet been established in newborn homeostasis circuitry, this aim will fulfill an important need in terms of not only scientific knowledge but also technical capability. Aim 2. Address the physiological relevance and clinical significance of these novel MRI-based newborn brain measures by testing the hypothesis that measures of the human newborn?s energy homeostasis brain circuitry are prospectively associated with infant adiposity and subsequent childhood obesity risk. Aim 3. Identify the prenatal (gestational biology) determinants of variation in the measures of newborn brain energy homeostasis circuitry that are associated with infant adiposity. Significance. By identifying the role and determinants of energy homeostasis-related brain circuitry in the human newborn, these findings will ultimately provide the basis for the subsequent development of strategies aimed at the primary prevention of childhood obesity.

ABSTRACT Our revised proposal launches the India ENIGMA Initiative for Global Aging & Mental Health - a globally coordinated study of brain aging and Alzheimer's disease (AD), created response to the NIH FOA: Global Brain and Nervous System Disorders Research Across the Lifespan (R01; PAR-18-834; https://grants.nih.gov/grants/guide/pa-files/PAR-18-835.html). Our overall goal is to identify predictive markers in the blood, genome, and epigenome that influence brain aging in India, to better understand prognosis, and to support personalized risk evaluations on each continent. We plan to identify etiological pathways to resilience using the rich biobanking strategy developed by our partners at NIMHANS in India. To do this, we will leverage our global consortium, ENIGMA (http://enigma.ini.usc.edu), to partner with dementia research pioneers in India, creating new links between international biobanks, and building research capacity. By 2020, 70% of the world's population over age 60 will live in developing countries, with 14% in India (Mathuranath 2012). Recently, attention has been drawn to a ?diversity? crisis in brain research, as most brain research is conducted in Caucasian populations from relatively wealthy backgrounds (LeWinn 2017). This lack of ethnic diversity means that: (1) we do not know if predictors of health (and disease) generalize to other ethnic groups, and (2) we fail to collect vital data that could teach us how AD progresses in populations with different genetic and environmental backgrounds. Our coordinated analyses in US/EU and Indian biobanks will help identify brain aging predictors specific to India and those that are universal. Specifically, we will: Aim 1. Create Lifespan Charts of brain aging Trajectories in India using MRI, DWI and Resting State Functional MRI. Aim 2. Identify Blood and Epigenetic Markers that Predict Brain Aging and AD in India. Aim 3. Using a combination of multimodal imaging, blood markers, and clinical data to predict clinical decline in India. We test structural equation models that hypothesize how brain aging depends on lifestyle and psychosocial factors (diet, family support, drug abuse, literacy, sleep, and depression), as well as sex, education, and AD genetic risk. With novel machine learning methods, will analyze blood markers and plasma proteomic analytes, to define processes that are harmful to brain aging. In Capacity Building Aims, we will leverage ENIGMA's successful strategies to train emerging and established scientists in India to analyze their data with high quality control and precision, with targeted biostatistical and imaging workshops to bolster capacity. This collaborative India-US initiative will enable future science initiatives, and equip the NIMHANS team with the necessary tools to train new scientists and independently conduct high impact research bridging efforts into numerous international partnerships.

ABSTRACT One in three seniors dies with Alzheimer's disease (AD) or another dementia - diseases that cost the nation $259 billion, to rise to $1.1 trillion by 2050 (Alzheimer's Association, 2017). Despite the vast personal and economic cost of these diseases, two major barriers stall efforts to discover key biological mechanisms that influence brain aging. First, the sheer cost of data collection means that most national initiatives have limited power to detect factors that affect brain aging. Even in datasets of N=1,000+ people (e.g., ADNI) ? the power to discover modulators of brain aging is limited and may not generalize worldwide. Second, with the crisis of reproducibility, we do not always know if a finding will replicate; and if not, if this is due to true population heterogeneity or problems with methods. ENIGMA offers a coordinated global approach to solve these problems. ENIGMA's World Aging Center is a global brain aging study that builds on our vast and highly productive ENIGMA consortium - a global network of 340 institutions in 45 countries. ENIGMA published the largest-ever genetic studies of the brain (Nature 2017; Science 2020), and the largest neuroimaging studies of 5 major psychiatric disorders. ENIGMA's World Aging Center is a concerted global effort to pool all available data, methods, expertise and capital infrastructure to discover factors that affect brain aging. Our long-term goal is to identify personalized biological predictors of brain structural and functional decline and assess how they generalize globally. We have 4 aims: Aim 1: ENIGMA-Lifespan. Develop Lifespan Charts for Brain and Neural Tract Aging in 20,000 people. We will create charts showing how MRI brain measures change throughout life in 20,000 people, aged 1-92. We will compute a composite brain aging score, `Brain Age', from available MRI, DTI, rsFMRI data, that measures how much the brain deviates from expected values, for a person's age and sex. Aim 2: ENIGMA-Epigenetics. Relate genome-wide methylation levels to brain metrics in 10,000+ people, to discover epigenetic markers of accelerated brain aging. We discovered 2 epigenetic loci promoting brain aging in pilot studies. We will compute a ?epigenetic clock? and test if it predicts brain metrics better than simple biological age. Aim 3: ENIGMA-Plasticity. Discover genomic loci that promote or mitigate brain tissue loss, in > 37 worldwide cohorts with longitudinal MRI. Aim 4: ENIGMA-Alzheimer's Disease (New Aim). Meta-analyze the role of APOE, AD polygenic risk, and a new risk score for accelerated atrophy on neuroimaging biomarkers in aging and AD, including amyloid and FDG PET. These aims seek to analyze worldwide imaging, epigenetic, and clinical data with harmonized methods. We aim to create new aging ?clocks? and reveal targetable risk factors and modifiers of brain aging in the genome and epigenome, test how and when they shift AD biomarkers, and test their generalizability worldwide.