Genetic influences on hub connectivity of the human connectome

Brain network hubs are both highly connected and highly inter-connected, forming a critical communication backbone for coherent neural dynamics. The mechanisms driving this organization are poorly understood. Using diffusion-weighted magnetic resonance imaging in twins, we identify a major role for genes, showing that they preferentially influence connectivity strength between network hubs of the human connectome. Using transcriptomic atlas data, we show that connected hubs demonstrate tight coupling of transcriptional activity related to metabolic and cytoarchitectonic similarity. Finally, comparing over thirteen generative models of network growth, we show that purely stochastic processes cannot explain the precise wiring patterns of hubs, and that model performance can be improved by incorporating genetic constraints. Our findings indicate that genes play a strong and preferential role in shaping the functionally valuable, metabolically costly connections between connectome hubs.


Reporting for specific materials, systems and methods
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response. Generic_human_ncbiIds_noParents.an.txt.gz on April 29, 2021. Gene Ontology terms and definitions were automatically downloaded by ErmineJ on April 29, 2021 as go.obo (data version 2021-02-01) and can be downloaded from http://release.geneontology.org/2021-02-01/ontology/index.html. Custom MATLAB code, and other data necessary to generate the figures in this work are available at https://github.com/BMHLab/GeneticBrainHubs and associated repository https://doi.org/10.5281/zenodo.4724186 respectively.
All available data surpassing the quality control criteria were used in each analysis. Group-representative connectome was generated based on the 972 subjects from the Human Connectome project; Heritability analysis included 117 pairs of genetically confirmed monozygotic (MZ) twin pairs and 69 of their non-twin siblings, as well as 60 dizygotic (DZ) same-sex twin pairs and 48 of their non-twin siblings. Grouprepresentative connectome for Monash sample was generated based on the data from 424 subjects.
Subjects from both DWI datasets were excluded based insufficient connectome density (more than 3SD below the mean of the sample). 5 subjects from Monash sample were excluded due to issues with cortical surface segmentation.
We replicated the main heritability and transcriptional coupling results using different data processing choices (brain parcellations and connectome densities) by applying the same analysis to data derived from 2 brain parcellations (180 and 250 regions per hemisphere respectively) and across three connectome density thresholds. Across all iterations connections between hub regions (rich links) were more heritable compared to other link types (results presented in Figure S4). For transcriptional coupling analysis, connected pairs of hubs demonstrated increased transcriptional coupling (results presented in Figure S7). The results demonstrating the increase in microstructural profile covariance were not replicated using random parcellation containing 500 regions. This discrepancy likely reflects the fact that the HCPMMP1 parcellation more closely approximates boundaries between functional zones of the cortex, as it is based on a fusion of multimodal imaging data. The random parcellation makes no attempt to capture such boundaries and may blur different cytoarchitectonic regions within the same network node, thus resulting in noisier microstructural profile covariance (MPC) estimates. In this way, the MPC results appear to depend on accurate approximation of cytoarchitectonic boundaries in cortex.
No experimental conditions requiring randomization were applied in the study.
No procedures requiring blinding were applied in the study. Participants for the Monash Sample were recruited through online advertisements. The sample is likely to include a relatively high number of university students which is representative of healthy young population that was required for this study and should not impact the results of the study. Details regarding the recruitment process of the HCP dataset is provided in the following paper: https://doi.org/10.1016/j.neuroimage.2013.05.041/. In short, the sample includes healthy adults with no neuropsychiatric or neurological disorders including siblings and twin pairs. Twins born prior to 34 weeks gestation and nontwins born prior to 37 weeks gestation are excluded in order to avoid potential confounding factors. Subject selection procedure should not influence the results of this study.
The experimental protocol was approved by Monash University's Human Research Ethics Committee and was carried out in accordance with the approved guidelines. Informed consent was obtained from all participants before testing. For HCP dataset subject recruitment procedures and informed consent forms, including consent to share de-identified data, were approved by the Washington University institutional review board. Connectomes reconstructed in subject-specific spaces.
HCP: correction for EPI susceptibility and signal outliers, eddy-current-induced distortions, slice dropouts, gradient nonlinearities and subject motion. Monash Sample: Distortions in the Monash DWI data were corrected with TOPUP in FSL, using the forward and reverse phase-encoded b = 0 images to estimate the susceptibility-induced off-resonance field. We corrected for eddy-current distortions, volume-to-volume head motion, within-volume head motion, and signal outliers using eddy tool in FSL [version 5.0.11]. This implementation of EDDY significantly mitigates motion-related contamination of DWI connectivity estimates. DWI data were subsequently corrected for B1 field inhomogeneities using FAST in FSL.