Functional Connectivity Hubs and Networks in the Awake Marmoset Brain.

Annabelle M Belcher,Yihong Yang,Nora D Volkow,Thomas J Ross,Dardo Tomasi,Afonso C Silva,Lucia Notardonato,Cecil Chern-Chyi Yen,Elliot A Stein

doi:10.3389/fnint.2016.00009

Abstract

In combination with advances in analytical methods, resting-state fMRI is allowing unprecedented access to a better understanding of the network organization of the brain. Increasing evidence suggests that this architecture may incorporate highly functionally connected nodes, or “hubs”, and we have recently proposed local functional connectivity density (lFCD) mapping to identify highly-connected nodes in the human brain. Here, we imaged awake nonhuman primates to test whether, like the human brain, the marmoset brain contains FC hubs. Ten adult common marmosets (Callithrix jacchus) were acclimated to mild, comfortable restraint using individualized helmets. Following restraint training, resting BOLD data were acquired during eight consecutive 10 min scans for each subject. lFCD revealed prominent cortical and subcortical hubs of connectivity across the marmoset brain; specifically, in primary and secondary visual cortices (V1/V2), higher-order visual association areas (A19M/V6[DM]), posterior parietal and posterior cingulate areas (PGM and A23b/A31), thalamus, dorsal and ventral striatal areas (caudate, putamen, lateral septal nucleus, and anterior cingulate cortex (A24a). lFCD hubs were highly connected to widespread areas of the brain, and further revealed significant network-network interactions. These data provide a baseline platform for future investigations in a nonhuman primate model of the brain’s network topology.

Highlights

The last decade has witnessed a precipitous increase in the number of fMRI reports exploring patterns of correlated signals that occur spontaneously in the human brain
The probability of the local functional connectivity density (lFCD), P(k), decreased exponentially with k, the strength of the lFCD, such that there were few highly connected hubs and numerous weakly connected nodes, a pattern that was highly significant across the eight repetitions and ten animals (p < 0.001, T-score > 3)
The t-values of the lFCD were larger for cortical regions than for subcortical regions, as within-subject variability in the strength of lFCD was higher for the subcortical hubs than for the cortical hubs (Figure 2)

Summary

Introduction

The last decade has witnessed a precipitous increase in the number of fMRI reports exploring patterns of correlated signals that occur spontaneously in the human brain. In a thoughtful and critical review of the utility of RSfMRI, Buckner et al (2013) argue that intrinsic functional connectivity uniquely provides information about relations between networks at a whole-brain level: a feature which, at minimum, merits its use as a tool for generating testable hypotheses that can only be explored with external methods of probing brain function. For ethical reasons, these methods preclude studies with humans, demanding more research with preclinical models. High-order cognitive processes, such as those that are under putative study with the use of RSfMRI, are best modeled in animals whose behaviors and neuroanatomical complexity are most proximate to humans: nonhuman primates

Objectives

Methods

Results

Conclusion