Effects of chronic and acute stimulants on brain functional connectivity hubs

Abstract

The spatial distribution and strength of information processing 'hubs' are essential features of the brain׳s network topology, and may thus be particularly susceptible to neuropsychiatric disease. Despite growing evidence that drug addiction alters functioning and connectivity of discrete brain regions, little is known about whether chronic drug use is associated with abnormalities in this network-level organization, and if such abnormalities could be targeted for intervention. We used functional connectivity density (FCD) mapping to evaluate how chronic and acute stimulants affect brain hubs (i.e., regions with many short-range or long-range functional connections). Nineteen individuals with cocaine use disorders (CUD) and 15 healthy controls completed resting-state fMRI scans following a randomly assigned dose of methylphenidate (MPH; 20mg) or placebo. Short-range and long-range FCD maps were computed for each participant and medication condition. CUD participants had increased short-range and long-range FCD in the ventromedial prefrontal cortex, posterior cingulate/precuneus, and putamen/amygdala, which in areas of the default mode network correlated with years of use. Across participants, MPH decreased short-range FCD in the thalamus/putamen, and decreased long-range FCD in the supplementary motor area and postcentral gyrus. Increased density of short-range and long-range functional connections to default mode hubs in CUD suggests an overrepresentation of these resource-expensive hubs. While the effects of MPH on FCD were only partly overlapping with those of CUD, MPH-induced reduction in the density of short-range connections to the putamen/thalamus, a network of core relevance to habit formation and addiction, suggests that some FCD abnormalities could be targeted for intervention.

Keywords: Addiction; Brain hubs; Cocaine; Dopamine; Functional magnetic resonance imaging; Methylphenidate; Resting-state connectivity.

Figure 1

Methylphenidate plasma (MPH) levels at 45 min and 120 min post drug administration in individuals with cocaine use disorders (CUD) and healthy controls (HC). Values at time 0 (not shown) were null for all participants. Arrow indicates time of the resting-state scan, during peak MPH levels.

Figure 2

Surface rendering showing the distribution of short-range and long-range FCD hubs in the human brain as a function of chronic cocaine use [cocaine use disorders (CUD) and healthy controls (HC)] and acute methylphenidate [MPH and placebo (PL)]. Color maps reflect the average number (k) of functional connections to neighbor (short-range) or remote (long-range) voxels. The images were created using the Computerized Anatomical Reconstruction and Editing Toolkit (CARET) 5.65 (http://brainvis.wustl.edu/wiki/index.php/Caret:About).

Figure 3

Surface rendering showing significant differences in short-range and long-range functional connectivity density (FCD) (a) in individuals with cocaine use disorders (CUD) as compared with healthy controls (HC) and (b) across participants, following a single 20 mg dose of oral methylphenidate (MPH) as compared with placebo (PL). Color maps reflect T-values ranging from 1.5 to 4. Results are significant at P<0.05 cluster corrected for multiple comparisons. The images were created using the Computerized Anatomical Reconstruction and Editing Toolkit (CARET) 5.65 (http://brainvis.wustl.edu/wiki/index.php/Caret:About). Abbreviations: DMN, default mode network; LN, lenticular nucleus; SMA, supplementary motor area; PCC, posterior cingulate; VMPFC, ventromedial prefrontal cortex.

Figure 4

Positive relationship between years of cocaine use and average functional connectivity density (FCD) in regions of the default mode network (DMN). For both the short- and long-range FCD, R>0.46, P<0.05. When age and drug use severity (encompassing frequency of use, perceived control over use, level of craving, and withdrawal symptoms) are included as covariates in partial correlation analyses, the results for short-range FCD remain significant (P=0.024) while the results for long-range FCD are somewhat attenuated (P=0.08).