Date of Award
Fall 2023
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Public Health
First Advisor
Zhang, Heping
Abstract
The exploration of the associations between genetic variants or single nucleotide polymor- phisms (SNPs), brain connectivity, and mental disorders is crucial for comprehending the biological origins of psychiatric and behavioral conditions. Biomedical technologies, such as magnetic resonance imaging (MRI) and next-generation sequencing (NGS), can help elu- cidate the underlying biological processes. Resting state functional MRI (rsfMRI) studies have demonstrated altered functional connectivity in numerous psychiatric and behavioral disorders, which is moderately heritable and can be linked to genetic variants. Diffusion tensor imaging (DTI) can be employed to investigate the structural foundation of func- tional networks. However, challenges remain in analyzing brain imaging and genetic data together, including weak genetic signals in complex high-dimensional data, intricate connec- tivity structures, complex relationships, and high computational complexity. To enhance our understanding of the neural and genetic factors involved in these disorders and develop more effective treatments, innovative statistical methods are needed to overcome these hurdles.First, Chapter 2 introduces the Ball Covariance Ranking and Aggregation (BCRA), a SNP-set hypothesis test, which incorporates functional connectivity matrix structure to detect significant SNPs while controlling the false discovery rate. A faster version is also proposed to reduce the computational burden of high-dimensional data. Second, to integrate the cluster/network structure of connectivity, Chapter 3 presents the network-based mediation model (NMM) to estimate the effect of genetics on behavioral out- comes or diseases mediated by functional connectivity. Chapter 4 presents a semi-constrained network-based statistic (scNBS) to address the issue of intricate connectivity structures, a method that employs a data-driven selection process to associate functional connectivity with clinical outcomes, achieving increased power and validity through benchmarking studies. Lastly, we present a chapter with a continuous coupling measure to identify genetic loci associated with structural-functional connectivity coupling (SC-FC coupling), with potential implications for neurological and psychiatric disorders. All four projects emphasize the links between genetics, brain connectivity, and behav- ioral and psychiatric disorders. The first three projects (Chapters 2-4) focus specifically on functional connectivity, tackling major statistical challenges in analyzing this data. Specif- ically, Chapter 2 uses the BCRA approach to incorporate the complex matrix structure of functional connectivity, improving SNP identification. Chapters 3 and 4 employ NMM and scNBS approaches to address the intricate, multi-scale network-level structure of functional connectivity. In contrast, the final project (Chapter 5) expands the concept of brain connec- tivity from functional regions to structural landscapes, further investigating the structural basis supports functional organization and how genetics govern this interrelationship.
Recommended Citation
Dai, Wei, "Integrating Genetics and Brain Connectivity: Overcoming Statistical Challenges to Uncover the Neural Basis of Psychiatric and Behavioral Disorders" (2023). Yale Graduate School of Arts and Sciences Dissertations. 1181.
https://elischolar.library.yale.edu/gsas_dissertations/1181
