Heritability of individualized cortical network topography

Abstract

Human cortex is patterned by a complex and interdigitated web of large-scale functional networks. Recent methodological breakthroughs reveal variation in the size, shape, and spatial topography of cortical networks across individuals. While spatial network organization emerges across development, is stable over time, and is predictive of behavior, it is not yet clear to what extent genetic factors underlie interindividual differences in network topography. Here, leveraging a nonlinear multidimensional estimation of heritability, we provide evidence that individual variability in the size and topographic organization of cortical networks are under genetic control. Using twin and family data from the Human Connectome Project (n = 1,023), we find increased variability and reduced heritability in the size of heteromodal association networks (h² : M = 0.34, SD = 0.070), relative to unimodal sensory/motor cortex (h² : M = 0.40, SD = 0.097). We then demonstrate that the spatial layout of cortical networks is influenced by genetics, using our multidimensional estimation of heritability (h²-multi; M = 0.14, SD = 0.015). However, topographic heritability did not differ between heteromodal and unimodal networks. Genetic factors had a regionally variable influence on brain organization, such that the heritability of network topography was greatest in prefrontal, precuneus, and posterior parietal cortex. Taken together, these data are consistent with relaxed genetic control of association cortices relative to primary sensory/motor regions and have implications for understanding population-level variability in brain functioning, guiding both individualized prediction and the interpretation of analyses that integrate genetics and neuroimaging.

Keywords: function brain networks; functional connectome; heritability; individualized parcellation; resting-state.

Fig. 1.

Individualized network size is more variable in heteromodal relative to unimodal cortex. (A) Individualized parcellations are composed of 17 canonical functional networks present in all HCP individuals, as defined by Kong and colleagues (13). (B) The ridge plot shows distributions of network size across all individuals, measured in mm² and separated by hemisphere (top ridge = right hemisphere, bottom ridge = left hemisphere). (C) Variability of individualized network size across all participants, measured with coefficient of variation, which corrects for differences in the total average size of each network. (D) Network sizes are significantly more variable within heteromodal (M = 24.6, SD = 5.3) relative to unimodal cortices (M = 20.5, SD = 2.4; F(1,32) = 2.51, P = 0.017). Hetero, heteromodal cortex; Uni, unimodal cortex.

Fig. 2.

Heritability of individualized network size is greater in unimodal relative to heteromodal networks. (A) Heritability of individual network size (controlling for total surface area) was estimated across 17 canonical functional networks using SOLAR (49), separately for each hemisphere. Error bars reflect 95% CIs. (B) The amount of variance explained by genetics (h²) for each network was consistent across hemispheres, as revealed by a correlation of left- and right-hemisphere h² values (r = 0.62, P = 0.0086). Each dot in the correlation plot is a functional network. (C) Heritability of normalized individual network size was higher among unimodal/sensory networks relative to heteromodal association networks (P = 0.025). Each dot represents one of 17 cortical networks, split by hemisphere (n = 34). Hetero, heteromodal cortex; Uni, unimodal cortex.

Fig. 3.

Individualized network topography is heritable across all networks. (A) The ridge plot displays distributions of interindividual Dice coefficients across each network. Higher Dice coefficients reflect higher spatial overlap of a network for a given pair of individuals. Topography of unimodal networks were overall more similar across individuals, relative to heteromodal cortex. (B) Significant heritability was observed across all examined 17 cortical networks (q < 0.01; range = 0.12 to 0.19, mean = 0.14), which was symmetric across hemispheres (r_s = 0.68, P = 0.0032). (C) Boxplots show higher Dice similarity of overall network organization between MZ pairs, relative to DZ, sibling, and unrelated participant pairings. (D) Individual examples illustrate HCP participants with a high and low dice overlap for Default B (high = 0.78; low = 0.29) and Visual C (high = 0.93; low = 0.59) networks. SIB, sibling; UNR, unrelated.

Fig. 4.

Local heritability of individualized network topography. (A) Multidimensional heritability of network topography estimated for every vertex, using an ROI at each point on the cortical sheet (radius = 10 vertices). Individualized network labels in each ROI were used to compute a participant-to-participant Dice similarity matrix, reflecting the similarity of network assignments within a given cortical area. Warmer colors indicate higher heritability of network assignments, for instance reflecting greater similarity among twins and siblings than unrelated individuals. (B) Example participant pairs with high and low Dice overlap of network labels for an ROI in the precuneus. Dice similarity was higher between MZ twins, relative to DZ, sibling, and unrelated participant pairs. SIB, sibling; UNR, unrelated; LH, left hemisphere; RH, right hemisphere.