Genetic influences on individual differences in longitudinal changes in global and subcortical brain volumes: Results of the ENIGMA plasticity working group

Abstract

Structural brain changes that occur during development and ageing are related to mental health and general cognitive functioning. Individuals differ in the extent to which their brain volumes change over time, but whether these differences can be attributed to differences in their genotypes has not been widely studied. Here we estimate heritability (h² ) of changes in global and subcortical brain volumes in five longitudinal twin cohorts from across the world and in different stages of the lifespan (N = 861). Heritability estimates of brain changes were significant and ranged from 16% (caudate) to 42% (cerebellar gray matter) for all global and most subcortical volumes (with the exception of thalamus and pallidum). Heritability estimates of change rates were generally higher in adults than in children suggesting an increasing influence of genetic factors explaining individual differences in brain structural changes with age. In children, environmental influences in part explained individual differences in developmental changes in brain structure. Multivariate genetic modeling showed that genetic influences of change rates and baseline volume significantly overlapped for many structures. The genetic influences explaining individual differences in the change rate for cerebellum, cerebellar gray matter and lateral ventricles were independent of the genetic influences explaining differences in their baseline volumes. These results imply the existence of genetic variants that are specific for brain plasticity, rather than brain volume itself. Identifying these genes may increase our understanding of brain development and ageing and possibly have implications for diseases that are characterized by deviant developmental trajectories of brain structure. Hum Brain Mapp 38:4444-4458, 2017. © 2017 Wiley Periodicals, Inc.

Keywords: ENIGMA plasticity working group; heritability; individual brain plasticity; longitudinal magnetic resonance imaging; twins.

Figure 1

(A) Univariate twin model. The path loadings a₁, c₁, and e₁ represent the genetic, common environmental influences and unique environmental influences on change rates, respectively. (B) Latent change model. The factors A_L, C_L, and E_L and corresponding factor loadings a_L, c_L, and e_L represent the influences of genetic, common environmental, and unique environmental influences on “level,” the nonchanging component of the volumes. Likewise, A_C, C_C, and E_C and corresponding factor loadings a_C, c_C, and e_C reflect the influences that are unique to change. The contributions to level and change are allowed to be different for left and right, and baseline and follow‐up, represented by the factor loadings f₁ and f₂. The overlap between the factors for baseline and change are modelled by the paths a_LC, c_LC, and e_LC. Heritability of change in this model is computed as $\left({\mathrm{a}}_{\text{LC}}{}^2+{\hspace{0.17em}\mathrm{a}}_{\mathrm{C}}{}^2\right)/\left({\mathrm{a}}_{\text{LC}}{}^2+{\hspace{0.17em}\mathrm{a}}_{\mathrm{C}}{}^2+{\hspace{0.17em}\mathrm{c}}_{\text{LC}}{}^2+{\hspace{0.17em}\mathrm{c}}_{\mathrm{C}}{}^2+{\hspace{0.17em}\mathrm{e}}_{\text{LC}}{}^2+{\hspace{0.17em}\mathrm{e}}_{\mathrm{C}}{}^2\right)$. (C) Bivariate twin model including baseline volume and change rate. The path loadings a₁₁, a₁₂, and a₂₂ represent the influences of genetic factors, the path loadings c₁₁, c₁₂, and c₂₂ represent common environmental influences and e₁₁, e₁₂, and e₂₂ represent unique environmental influences. The factors A₁, C₁, and E₁ are shared between baseline volume and change rate, and A₂, C₂, and E₂ represent influences on change rate that are independent from the factors influencing baseline volume. (D) Bivariate twin model including baseline volume and volume at follow‐up. The path loadings a₁₁, a₁₂, and a₂₂ represent the influences of genetic factors, c₁₁, c₁₂, and c₂₂ represent common environmental influences and e₁₁, e₁₂, and e₂₂ represent unique environmental influences. The factors A₁, C₁, and E₁ are shared between baseline volume and followup, and A₂, C₂, and E₂ represent influences on follow‐up volume that are independent from the factors influencing baseline volume. Heritability of change rate is derived from this model by computing $\left({\mathrm{a}}_{11}{}^2+{\hspace{0.17em}\mathrm{a}}_{12}{}^2+{\hspace{0.17em}\mathrm{a}}_{22}{}^2–{\hspace{0.17em}2\mathrm{a}}_{11}{\mathrm{a}}_{12}\right)/\left({\left({\mathrm{a}}_{11}{}^2+{\hspace{0.17em}\mathrm{c}}_{11}{}^2+{\hspace{0.17em}\mathrm{e}}_{11}{}^2\right)}^{}+\hspace{0.17em}\left({\mathrm{a}}_{12}{}^2+{\hspace{0.17em}\mathrm{a}}_{22}{}^2+{\hspace{0.17em}\mathrm{c}}_{12}{}^2+{\hspace{0.17em}\mathrm{c}}_{22}{}^2+{\hspace{0.17em}\mathrm{e}}_{12}{}^2+{\hspace{0.17em}\mathrm{e}}_{22}{}^2\right)\right)\hspace{0.17em}–\hspace{0.17em}2\left[\left({\mathrm{a}}_{11}{\mathrm{a}}_{12}+{\hspace{0.17em}\mathrm{c}}_{11}{\mathrm{c}}_{12}+{\hspace{0.17em}\mathrm{e}}_{11}{\mathrm{e}}_{12}\right)\right])$.

Figure 2

Variance components of global change rate per year for four cohorts. The upper panel displays variance components for the meta‐analyses in the full group, and child/adult cohorts separately. The lower panel shows results for the individual cohorts. The colours of the bars represent the different structures. Heritability estimates (bottom) are displayed in the darkest, solid colour. Common environmental influences (middle) are displayed with diagonal shading. Unique environmental influences are displayed in the lightest colour. Significant h ² and c ² estimates in the meta‐analyses are marked with a star.

Figure 3

Heritability of subcortical change rates per year for the five cohorts. The upper panel displays variance components for the meta‐analyses in the full group, and child/adult cohorts separately. The lower panel shows results for the individual cohorts. The colours of the bars represent the different structures. Heritability estimates (bottom) are displayed in the darkest, solid colour. Common environmental influences (middle) are displayed with diagonal shading. Unique environmental influences are displayed in the lightest colour. Significant h ² and c ² estimates in the meta‐analyses are marked with a star.