Convergent network-level brain alterations across drugs of abuse: A meta-analysis of structural MRI studies

Hill-Bowen LD, College on Problems of Drug Dependence (2020).

Lauren gave this 10 minute pre-recorded talk for the virtual CPDD conference in June, 2020.


Aim: Neuroimaging studies have provided insight into the structural and functional brain alterations linked with substance abuse; yet, these studies often assess such alterations within the context of individual drugs (e.g., nicotine). Rather, neurobiological theories of addiction emphasize common network-level brain alterations across drug classes. Using emergent meta-analytic neuroimaging techniques, we sought to identify common structural brain alterations across drugs and to characterize the functionally-connected networks with which such structurally-altered regions interact.

Methods: We performed a literature search and identified 87 studies characterizing gray matter volume (GMV) differences between substance users versus non-users. Using the Activation Likelihood Estimation (ALE) algorithm we identified convergent GMV reductions across all studies. Next, we characterized an “extended addiction network” (eAN) by identifying the brain areas that were functionally-coupled with these structurally-altered regions. To do so, we performed resting-state functional connectivity (rsFC) and meta-analytic connectivity (maC) analyses using each structurally-altered region as a seed and computed whole-brain functional connectivity profiles as the union of both rsFC and maC maps. We considered regions to constitute nodes in the eAN where at least 3 of the functional connectivity profiles overlapped. Finally, we performed hierarchical clustering with all of the eAN node’s functional connectivity profiles to identify sub-networks linked with substance abuse.

Results: We identified the left insula, cingulate, and medial/superior frontal gyri as regions showing convergent GMV reductions among users. The overlap of these region’s functional connectivity profiles identified the cingulate, bilateral insulae, putamen, and superior temporal gyrus as regions of the eAN. Hierarchical clustering identified 3 sub-networks closely corresponding to the default-mode (posterior-cingulate, angular gyrus), salience (dorsal anterior cingulate, caudate), and executive control networks (lateral prefrontal and parietal cortex).

Conclusions: These outcomes indicate that structurally-altered brain regions linked with substance abuse are functionally connected to canonical intrinsic connectivity networks (default-mode, salience, and executive control).