Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 Mar 1;28(3):924-935.
doi: 10.1093/cercor/bhw416.

The Rich-Club Organization in Rat Functional Brain Network to Balance Between Communication Cost and Efficiency

Affiliations
Free PMC article

The Rich-Club Organization in Rat Functional Brain Network to Balance Between Communication Cost and Efficiency

Xia Liang et al. Cereb Cortex. .
Free PMC article

Abstract

Network analyses of structural connectivity in the brain have highlighted a set of highly connected hubs that are densely interconnected, forming a "rich-club" substrate in diverse species. Here, we demonstrate the existence of rich-club organization in functional brain networks of rats. Densely interconnected rich-club regions are found to be distributed in multiple brain modules, with the majority located within the putative default mode network. Rich-club members exhibit high wiring cost (as measured by connection distance) and high metabolic running cost (as surrogated by cerebral blood flow), which may have evolved to achieve high network communications to support efficient brain functions. Furthermore, by adopting a forepaw electrical stimulation paradigm, we find that the rich-club organization of the rat functional network remains almost the same as in the resting state, whereas path motif analysis reveals significant differences, suggesting the rat brain reorganizes its topological routes by increasing locally oriented shortcuts but reducing rich-club member-involved paths to conserve metabolic running cost during unimodal stimulation. Together, our results suggest that the neuronal system is organized and dynamically operated in an economic way to balance between cost minimization and topological/functional efficiency.

Keywords: functional connectome; module; motif; rat brain; rich club.

Figures

Figure 1.
Figure 1.
(A) The weighted rich-club coefficient values for a range of strength s are shown for the rat functional connectome (black line) and random networks (gray line, averaged across 100 comparable random networks). The normalized rich-club coefficients ϕnorm(s) are also shown (red line). The rich-club curve suggests a significant rich-club organization of the rat functional connectome for a range of s from 2.11 to 4.83 (*P < 0.05, Bonferroni corrected). (B) Map of rich-club (RC) regions in resting-state functional brain networks of rats. (C) Module structure of rat functional connectome with brain regions color-coded according to their modular affiliation. (D) RC regions were found to be distributed across 4 brain modules as depicted in the pie graph, with a large percentage of hubs (40%) belonging to the DMN. SM, somatomotor areas; THA, thalamus; TP, temporal/parietal areas; ST, striatum areas; OLF, olfactory cortex.
Figure 2.
Figure 2.
(A) Topological performance of rich-club and nonrich-club regions. Bar graphs show that the nodal strength, averaged functional connectivity, betweenness centrality and PC of rich-club nodes were significantly higher than those of nonrich-club ones. (B) Wiring and running cost of rich-club and nonrich-club regions. Bar graphs show the averaged physical distance (left panel) and the rCBF of rich-club and nonrich-club nodes, with rich-club nodes displaying a significantly longer wiring distance as well as higher rCBF supply as compared with nonrich-club nodes. ***P < 0.001; **P < 0.01; *P < 0.05.
Figure 3.
Figure 3.
(A) Schematic illustration of path motif and local (L), feeder (F), and rich-club (R) connections. (B) Proportions of local, feeder, and rich-club connections contributing to whole network, intermodule, and intramodule communications. (C) The percentage occurrence of path motifs of all shortest paths connecting regions across the whole network, between modules and within single modules, respectively. (D) The significance profiles are plotted as normalized significance level (Z score) of path motifs of all shortest paths connecting regions across the whole network, between modules and within single modules, respectively (*P < 0.05, Bonferroni corrected). Solid lines are significance profiles for group-averaged brain network, whereas dashed lines represent significance profiles for each individual rat brain network.
Figure 4.
Figure 4.
Forepaw stimulation effects on rich-club organization of rat functional brain network. (A) Fingerprint of path motif occurrence of whole-brain network and intermodule communications during both resting and forepaw stimulation states. Path motif occurrences are shown as their log values for easier visualization. *P < 0.05, Bonferroni corrected. (B) The occurrences of the 2 existing kinds of path motifs, “L” and “LF,” of the shortest paths connecting the right S1FL are plotted for both resting and forepaw stimulation states. The inset plot shows the normalized significance level (Z score) of the 2 path motifs comparing with random networks, and only the “L” path motif is found to occur significantly more than that in random networks during both resting and stimulation states (P < 0.01, Bonferroni corrected). In comparison to at resting state, the “L” motif occurs significantly more frequently during the forepaw stimulation (P = 0.001, permutation test). (C) The increased occurrence of the “L” motif correlated significantly with the stimulation-induced activation in right S1FL.

Similar articles

See all similar articles

Cited by 9 articles

See all "Cited by" articles

Publication types

Feedback