ZusammenfassungPopulation-based neuroimaging has gained considerable attention over the past years, following a paradigm change from small single-center to large-scale studies for studying brain structure and function and its variability in relation to environmental and genetic influencing factors. Understanding the normal variability in the general population is essential for delimiting disease-related changes and for identifying (pre-)clinical imaging biomarkers. Using standardized workflows for coherent data extraction is essential for performing future analyses with phenotypic or other imaging-derived NAKO data from other organ systems, including data provenance tracking. Since structural and functional neuroimage data processing requires days per subject to calculate all relevant parameters, modern high-performance computing approaches are necessary for fast and optimized data processing. This allows for preparing the NAKO neuroimaging data for all typical analysis settings: (i) correlations between global or regional imaging-derived phenotypes of the brain’s structure and function; (ii) correlative voxel-wise analyses using respective software packages; and (iii) machine learning-based predictions of age, gender, behavioral phenotypes, traits. These approaches are the basis for estimating the amount of variability of brain structure and function in different age strata, between the sexes and in subjects with different performance or personality traits and explaining them by environmental factors such as lifestyle, anamnestic information or ambient air pollution. The present proposal aims at implementing such standardized workflows, extracting all relevant parameters of brain structure and function based on state-of-the-art neuroimaging tools, optimizing machine learning based prediction algorithms in the large population-based setting and demonstrating usability for variability estimation in exemplary constellations.

Schlüsselwörter-

EinrichtungenForschungszentrum Jülich, Helmholtz Zentrum München - Deutsches Forschungszentrum für Gesundheit und Umwelt