Domain-Derived Biomarkers of Treatment Response in Major Depressive Disorder: An Activation Likelihood Estimation Neuroimaging Meta-Analysis.
TL;DR
Across domains, heightened pretreatment activation in the rAI, a key node in the salience network, emerged as a strong biomarker of poor prognosis in adults with MDD.
Abstract
Background
Neuroimaging research of individuals with major depressive disorder (MDD) has identified varying deficits across affective, cognitive, and resting-state paradigms and distinct patterns of activation associated with these domains. As this heterogeneity may be related to variation in clinical prognosis, we examined convergence across neuroimaging studies of MDD treatment biomarkers within and across domains.
Methods
Using a PubMed search and reference lists from topical reviews, we identified research in which adults with MDD exhibited pretreatment neural responses associated with clinical response to antidepressant medication or cognitive behavioral therapy. We performed activation likelihood estimation meta-analyses of peak voxel coordinates to identify domain-general (i.e., elicited across all tasks) and domain-specific (i.e., affective, cognitive, or resting-state) biomarkers associated with treatment response.
Results
The meta-analyses included 46 experiments with 996 unique participants. Across studies showing pre-treatment activation associated with poor treatment outcome, a cluster in the right anterior insula (rAI) was identified as a domain-general biomarker. For the domain-specific meta-analysis of neuroimaging studies of negatively-valenced emotion processing, significant convergence of baseline activation associated with treatment response emerged in the rAI and in the subgenual anterior cingulate cortex. No significant convergence of biomarkers emerged across resting-state metabolic studies or among studies reporting biomarkers of positive treatment outcomes.
Conclusions
Across domains, heightened pretreatment activation in the rAI, a key node in the salience network, emerged as a strong biomarker of poor prognosis in adults with MDD. Moreover, pretreatment limbic and paralimbic responses elicited by negative emotional stimuli may provide valuable information about subsequent response to treatment.