Assortative mixing in micro-architecturally annotated brain connectomes

Abstract

The wiring of the brain connects micro-architecturally diverse neuronal populations. The conventional graph model encodes macroscale brain connectivity as a network of nodes and edges, but abstracts away the rich biological detail of each regional node. Regions are different in terms of their microscale attributes, many of which are readily available through modern technological advances and data-sharing initiatives. How is macroscale connectivity related to nodal attributes? Here we investigate the systematic arrangement of white-matter connectivity with respect to multiple biological annotations. Namely, we formally study assortative mixing in annotated connectomes by quantifying the tendency for regions to be connected with each other based on the similarity of their micro-architectural attributes. We perform all experiments using four cortico-cortical connectome datasets from three different species (human, macaque and mouse), and consider a range of molecular, cellular and laminar annotations, including gene expression, neurotransmitter receptors, neuron density, laminar thickness and intracortical myelin. Importantly, we disentangle the relationship between neural wiring, regional heterogeneity and spatial embedding using spatial autocorrelation-preserving null models. We show that mixing between micro-architecturally diverse neuronal populations is supported by long-distance connections. Using meta-analytic decoding, we find that the arrangement of connectivity patterns with respect to biological annotations shape patterns of regional functional specialization. Specifically, regions that connect to biologically similar regions are associated with executive function; conversely, regions that connect with biologically dissimilar regions are associated with memory function. By bridging scales of cortical organization, from microscale attributes to macroscale connectivity, this work lays the foundation for next-generation annotated connectomics.

Multimodal, multiscale connectivity blueprints of the cerebral cortex

“…Combining connectivity modes has also been used to better resolve clusters of functional activation in BOLD data [84], and inform the application of deep brain stimulation to psychiatric and neurological diseases [11, 67]. Encouragingly, previous work has found that incorporating multiple perspectives of brain connectivity can result in novel discoveries, including improved generative models of brain connectivity [100], structure-function coupling [56, 105], epicentres of transdiagnostic alterations [57, 60], and the characterization of homophilic wiring principles [16].…”

Spatially heterogeneous structure-function coupling in haemodynamic and electromagnetic brain networks

“…Second, most current studies model a single, globally-uniform structure-function relationship across the brain. Recent reports, however, suggest that structure-function coupling is regionally heterogeneous, with stronger correspondence between structural and functional connectivity in uni-modal cortex, and weaker correspondence in transmodal cortex [12, 108, 151, 157, 163], potentially reflecting underlying molecular and cytoarchitectural gradients [12, 13, 48, 138, 151]. Altogether, a more detailed biological understanding of structure-function relationships – one that takes into account both neurophysiological activity and regional heterogeneity – is necessary [138].…”

The spatial arrangement of laminar thickness profiles in the human cortex scaffolds processing hierarchy

“…Using an alternative approach, another recent study focused on the interrelation between connectivity and the absolute thickness of individual layers in the BigBrain, and showed that regions with thicker layer IV are less likely to connect to regions with higher thickness in layers III, V and VI [78]. Overall, these findings are in line with the wiring principle of "similar prefers similar" [20,26,75,78], which has been observed not only with the similarity of microstructure, but also in association to gene expression patterns [79][80][81][82][83], neurotransmitter receptor profiles [84] and macroscale morphometry [78,85]. It may be argued that this ubiquitous finding simply reflects the fact that nearby cortical regions tend to be similar [49,50], and also are more likely to connect, due to the principle of wiring cost reduction [47,48].…”

“…In our study we extend these findings and show that the probability and strength of connectivity additionally relates to the laminar thickness profiles in the BigBrain. Using an alternative approach, another recent study focused on the interrelation between connectivity and the absolute thickness of individual layers in the BigBrain, and showed that regions with thicker layer IV are less likely to connect to regions with higher thickness in layers III, V and VI [78]. Overall, these findings are in line with the wiring principle of "similar prefers similar" [20,26,75,78], which has been observed not only with the similarity of microstructure, but also in association to gene expression patterns [79][80][81][82][83], neurotransmitter receptor profiles [84] and macroscale morphometry [78,85].…”