Heritability of Behavioral and Brain Measures in a Large Cohort of Healthy Twin and non-Twin Subjects

Abstract

This research investigated comprehensively the effects of genetics on behavioral traits, brain structure and function, and their associations in a large cohort of monozygotic (MZ) twins, dizygotic (DZ) twins, non-twin siblings (SIB) and non related (NR) individuals (N = 1206, total) provided by the Human Connectome Project (HCP). All primary measures available are of the highest quality and quantitatively assessed. They include the following for each individual: (a) Measures of behavioral traits in 5 domains (motor, sensory, cognitive, emotion, and personality); (b) volumes of 70 cortical brain areas extracted from high-resolution (0.7 mm isotropic) structural magnetic resonance imaging (sMRI) data; (c) resting-state blood oxygenation level dependent (BOLD) activity of the same areas extracted from long-duration (1200 volumes), fast-acquisition (every 0.72 s), high-resolution (2 mm isotropic) functional MRI (fMRI) data; and (d) white matter integrity measures (fractional anisotropy [FA] and mean diffusivity [MD] for 7 brain regions regions) derived from high angular resolution diffusion imaging (HARDI) MRI (dMRI) data at 1.25 mm spatial resolution and very strong magnetic field gradients at (100 mT/m). Data extraction and preprocessing was performed using a dedicated 704-processor high-performance computer cluster at the Brain Sciences Center using Matlab. Univariate and multivariate statistical analyses were carried out in personal computers using Matlab and IBM-SPSS (version 24). These analyses include the following. (a) Computation of common univariate statistics (mean, variance, etc.); (b) computation of intra class correlation (ICC) for each of the 4 genetic groups (MZ, DZ, SIB, NR) and its z-transform [zICC = atanh(ICC)] for each primary measure above; (c) analysis of variance (ANOVA) of zICC across genetic groups for each measure; (d) computation of heritability using Falconer’s formula; (e) multidimensional scaling (MDS) and hierarchical tree clustering (HTC) of this heritability for the different data sets (behavioral, sMRI, fMRI, dMRI). These analyses yielded substantial new information on the effects of genetics on brain and behavior, and partially elucidated underlying associations among the various diverse measures above. To our knowledge, this is the first such comprehensive study carried out.