The Quantitative Genetics of Neurodevelopment: A Magnetic Resonance Imaging Study of Childhood and Adolescence

THE QUANTITATIVE GENETICS OF NELRODEVELOPMENT: A MAGNETIC RESONANCE IMAGING STUDY OF CHILDHOOD AND ADOLESCENCE By James E. Schmitt A Dissertation submitted in partial filfillment of the requirements for the degree of Doctor of Philosophy at Virginia Commonwealth University. Virginia Commonwealth University, 2007 Major Directors: Kenneth S. Kendler, M.D. Distinguished Professor, Departments of Psychiatry and Human Genetics and Michael C. Neale, Ph.D. Professor, Departments of Psychiatry and Human Genetics Understanding the causes of individual differences in brain structure may give clues about the etiology of cognition, personality, and psychopathology, and also may identify endophenotypes for molecular genetic studies on brain development. We performed a comprehensive statistical genetic study of anatomic neuroimaging data from a large pediatric sample (N=600+) of twins and family members from the Child Psychiatry Branch at the NINIH. These analyses included variance decomposition of structural volumetric endophenotypes at several levels of resolution, voxel-level analysis of cortical thickness, assessment of gene by age interaction, several multivariate genetic analyses, and a search for genetically-mediated brain-behavioral relationships. These analyses found strong evidence for a genetic role in the generation of individual differences in brain volumes, with the exception of the cerebellum and the lateral ventricles. Subsequent multivariate analyses demonstrated that most of the genetic variance in large volumes shares a common source. More subtle analyses suggest that although this global genetic factor is the principal determinant of neuroanatomic variability, genetic factors also mediate regional variability in cortical thickness and are different for gray and white matter volumes. Models using graph theory show that brain structure follows small-world architectural rules, and that these relationships are genetically-determined. Structural homologues appeared to be strongly related genetically, which was M h e r confirmed using novel methods for semi-multivariate quantitative genetic analysis at the voxel level. Studies on interactions with age were mixed. We found evidence of gene by age interaction on frontal and temporal lobar volumes, indicating that the role of genetic factors on these structures is dynamic during childhood. Analyses on cortical thickness at a finer scale, however, showed that environmental factors are more important in childhood, and environmental changes were responsible for most of the changes in heritability over this age range. When assessing the relationship between brain and behavior, we found weak negative genetic correlations and positive environmental correlations between IQ and cortical thickness, which appear to partially cancel each other out. More complex models allowing for age interactions suggest that high and low