Integrating Neuroimaging and Multi-omics to Advance Studies of Individual Variation in Human Brain Organization

Abstract

Distributed networks of structure and function in the human brain support complex perception, cognition, and behavior that ultimately make us human. This motivates neuroscientific studies of the biological basis of individual human brain organization. Consequently, variations in this organization could reflect individuality of health and serve as promising biomarkers of neurological and psychiatric disorders. This work seeks to advance studies of individual brain organization through multi-dimensional modeling of neuroimaging and clinical traits, as well as genomics and transcriptomics (“multi-omics”). Existing studies on gene-trait associations in the brain largely focus on one or two of these dimensions, mainly due to the limited availability of these data for large human cohorts. In particular, there are few gene expression resources available for whole-brain analyses, despite their major role in genetically-driven brain individuality. By contrast, my work developed an integrated approach that 1) leverages gene expression prediction models to identify gene associations with structural brain traits, 2) benchmarks these associations against genome-wide association studies, 3) compares gene-structure patterns against brain-related clinical measures, and 4) reveals polygenic links with complex functional brain traits. Additionally, the work developed a python-based software that provides a user-friendly interface for performing neuroimaging and multi-omics association tests in a streamlined manner. Collectively, this work enables a new direction for studying the genetic underpinnings of individual variation in human brain organization.