Genetic Contributions to Regional Variability in Human Brain Structure: Methods and Preliminary Results

• Published in: NeuroImage (Orlando, Fla.) Vol. 17; no. 1; pp. 256 - 271
• Main Authors: Wright, I.C., Sham, P., Murray, R.M., Weinberger, D.R., Bullmore, E.T.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.09.2002; Elsevier Limited
• Subjects: Algorithms; Analysis of Variance; asymmetry; Brain - anatomy & histology; Brain - growth & development; brain anatomy; Cerebral Ventricles - anatomy & histology; Female; Functional Laterality - physiology; genetic correlation matrix; Genetics; Gestational Age; heritability; Humans; Image Interpretation, Computer-Assisted; Magnetic Resonance Imaging; MRI; path analysis; Phenotype; Pregnancy; Principal Component Analysis; structural equation modeling; twins; Twins, Dizygotic; Twins, Monozygotic; path analysis; twins; MRI; structural equation modeling; brain anatomy; asymmetry; genetic correlation matrix; heritability
• DOI: 10.1006/nimg.2002.1163
• ISSN: 1053-8119

Twin studies provide one approach for investigating and partitioning genetic and environmental contributions to phenotypic variability in human brain structure. Previous twin studies have found that cerebral volume, hemispheric volume, ventricular volume, and cortical gyral pattern variability were heritable. We investigated the contributions of genetic and environmental factors to both global (brain volume and lateral ventricular volume) and regional (parcellated gray matter) variability in brain structure. We examined MR images from 10 pairs of healthy monozygotic and 10 pairs of same-sex dizygotic twins. Regional gray matter volume was estimated by automated image segmentation, transformation to standard space, and parcellation using a digital atlas. Heritability was estimated by path analysis. Estimated heritability for brain volume variability was high (0.66; 95% confidence interval 0.17, 1.0) but the major effects on lateral ventricular volume variability were common and unique environmental factors. We constructed a map of regional brain heritability and found large genetic effects shared in common between several bilateral brain regions, particularly paralimbic structures and temporal–parietal neocortex. We tested three specific hypotheses with regard to the genetic control of brain variability: (i) that the strength of the genetic effect is related to gyral ontogenesis, (ii) that there is greater genetic control of left than of right hemisphere variability, and (iii) that random or fluctuating asymmetry in bilateral structures is not heritable. We found no evidence in support of the first two hypotheses, but our results were consistent with the third hypothesis. Finally, we used principal component (PC) analysis of the genetic correlation matrix, to identify systems of anatomically distributed gray matter regions which shared major genetic effects in common. Frontal and parietal neocortical areas loaded positively on the first PC; some paralimbic and limbic areas loaded negatively. Bilateral insula, some frontal regions, and temporal neocortical regions functionally specialized for audition and language loaded strongly on the second PC. We conclude that large samples are required for powerful investigation of genetic effects in imaging data from twins. However, these preliminary results suggest that genetic effects on structure of the human brain are regionally variable and predominantly symmetric in paralimbic structures and lateral temporal cortex.