Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2013 Nov;37(11):1891-900.
doi: 10.1111/acer.12181. Epub 2013 Jul 26.

The Effects of Alcohol on the Nonhuman Primate Brain: A Network Science Approach to Neuroimaging

Free PMC article

The Effects of Alcohol on the Nonhuman Primate Brain: A Network Science Approach to Neuroimaging

Qawi K Telesford et al. Alcohol Clin Exp Res. .
Free PMC article


Background: Animal studies have long been an important tool for basic research as they offer a degree of control often lacking in clinical studies. Of particular value is the use of nonhuman primates (NHPs) for neuroimaging studies. Currently, studies have been published using functional magnetic resonance imaging (fMRI) to understand the default-mode network in the NHP brain. Network science provides an alternative approach to neuroimaging allowing for evaluation of whole-brain connectivity. In this study, we used network science to build NHP brain networks from fMRI data to understand the basic functional organization of the NHP brain. We also explored how the brain network is affected following an acute ethanol (EtOH) pharmacological challenge.

Methods: Baseline resting-state fMRI was acquired in an adult male rhesus macaque (n = 1) and a cohort of vervet monkeys (n = 10). A follow-up scan was conducted in the rhesus macaque to assess network variability and to assess the effects of an acute EtOH challenge on the brain network.

Results: The most connected regions in the resting-state networks were similar across species and matched regions identified as the default-mode network in previous NHP fMRI studies. Under an acute EtOH challenge, the functional organization of the brain was significantly impacted.

Conclusions: Network science offers a great opportunity to understand the brain as a complex system and how pharmacological conditions can affect the system globally. These models are sensitive to changes in the brain and may prove to be a valuable tool in long-term studies on alcohol exposure.

Keywords: Brain Networks; Functional Magnetic Resonance Imaging; Graph Theory; Nonhuman Primates.


Figure 1
Figure 1. Schematic of network and reproducibility analysis
Voxel time courses are extracted from the fMRI time series and computing the Pearson correlation between all voxel pairs, a correlation matrix is produced. A threshold is applied to the correlation matrix to generate a binary adjacency matrix. From the adjacency matrix, graph metrics are calculated at every node in the network and averaged to produce mean graph metrics.
Figure 2
Figure 2. Graph metric reproducibility in rhesus macaque
Average graph metrics in rhesus macaque show no significant differences across fMRI runs for clustering coeffiecnt (C, p=0.92), path length (L, p=0.70), local efficiency (Eloc, p=0.94), global efficiency (Eglob, p=0.39). A plot of the degree distribution shows both runs are similar and follow a truncated power law.
Figure 3
Figure 3. Hub maps of top 20% of nodes of highest degree in rhesus macaque
(a-b) Hubs appear in the medial prefrontal cortex, cingulate cortex, temporal lobe and visual cortex. (c) Comparing the two runs, nodes from Run 1 (red) were overlaid onto nodes in Run 2 (blue). Regions appearing in purple highlight the overlap of high degree nodes across both runs.
Figure 4
Figure 4. Binary hub maps for two fMRI runs across thresholds
(a) Binary images of the network hubs were generated for each run with an increasing threshold from the top 25% to 10% of the most connected nodes. Nodes only in the first run appear in blue, nodes only in the second run appear in green, while nodes in both runs appear in red. (b) The percentage overlap shows that as much as 54% down to 41% of the nodes for either run is shared. Across both maps, high degree hubs appear to be stable across run with no major differences at increasing threshold levels.
Figure 5
Figure 5. Module organization of default mode communities in rhesus macaque
Two modules were found to be associated with the default mode network. However, in this study, regions in the premedial frontal cortex were not connected to the precuneus/posterior cingulate cortex.
Figure 6
Figure 6. Overlap map across 10 vervet monkeys
Hub maps of top 20% high degree nodes were overlaid to determine areas with a high consistency of high degree nodes. Hubs were found in the medial prefrontal cortex, anterior cingulate, cingulate cortex, visual cortex, parietal lobe, temporal lobe, and superior temporal gyrus.
Figure 7
Figure 7. Average graph metrics across vervet monkey cohort
Across subjects, average metrics were found to be highly reproducible across animals. The degree distribution for each animal also appear to be similar and follow a truncated power law.
Figure 8
Figure 8. Average graph metrics pre- and post- ethanol bolus
Average metrics did not significantly differ across metrics for clustering coefficient (C, p=0.70), path length (L, p=0.71), local efficiency (Eloc, p=0.83), global efficiency (Eglob, p=0.70). The degree distribution does not appear to significantly change across runs.
Figure 9
Figure 9. Network changes in rhesus macaque brain following ethanol challenge
A rhesus macaque was given 1 g/kg of ethanol after an initial scan. (a) As seen in the hub map, the network is altered with much of the hub structure disappearing (b) While the community organization shows regions associated with the default mode network before the ethanol bolus, no discernible pattern.

Comment in

  • Graphs of brain networks.
    Zahr NM. Zahr NM. Alcohol Clin Exp Res. 2013 Nov;37(11):1813-5. doi: 10.1111/acer.12293. Alcohol Clin Exp Res. 2013. PMID: 24164166

Similar articles

See all similar articles

Cited by 4 articles

Publication types