Importance: A large number of structural neuroimaging studies have used voxel-based morphometry (VBM) to identify gray matter abnormalities in youths with conduct problems (CP), but the findings have been disparate and few have been replicated.
Objective: To conduct a meta-analysis of published whole-brain structural neuroimaging studies on youths with CP that used VBM methods to facilitate replication and aid further analyses by researchers.
Data Sources: The PubMed, ScienceDirect, Scopus, Google Scholar, and Web of Science databases were searched for VBM studies published from January 1, 2007, through March 31, 2015. Manual searches were conducted using title and citation information. Authors were contacted for additional data.
Study Selection: A literature search identified 28 studies. Studies were excluded if they (1) failed to use VBM, (2) failed to report a voxelwise comparison between youths with CP and typically developing (TD) youths, (3) used different significance or extent thresholds throughout the brain, (4) included duplicated datasets, and (5) did not provide peak coordinates or parametric maps after contact with the authors. Thirteen studies were deemed eligible for inclusion (394 youths with CP and 350 TD youths).
Data Extraction and Synthesis: Anisotropic effect-size signed differential mapping (SDM) was used for voxel-based meta-analyses. Statistical parametric maps comparing gray matter differences between youths with CP and TD youths were available for 11 of the studies, with peak coordinates available for the remaining studies.
Main Outcomes and Measures: Regional gray matter volume (GMV) differences in youths with CP compared with TD youths.
Results: Youths with CP had decreased GMV in the left amygdala (SDM estimate = -0.218; P < .001) (extending into anterior insula), right insula (SDM estimate = -0.174; P < .001) (extending ventrolaterally into the prefrontal cortex and inferiorly into the superior temporal gyrus), left medial superior frontal gyrus (SDM estimate = -0.163; P = .001) (extending into the right anterior cingulate cortex), and left fusiform gyrus (SDM estimate = -0.146; P = .003). Subgroup meta-analysis assessing age-at-onset effects identified reduced GMV in the left amygdala (SDM estimate = -0.232; P < .001) (extending into anterior insula). Meta-regression analyses revealed that greater scores on measures of callous-unemotional traits were associated with a lower reduction in GMV in the left putamen (SDM estimate = -0.911; P < .001). The proportion of male and female youths in the sample was associated with decreased GMV in the left amygdala (SDM estimate = -0.31; P < .001) and increased GMV in the right inferior temporal cortex (SDM estimate = 0.755; P < .001). While there was no association with comorbid attention-deficit/hyperactivity disorder or IQ, age range was associated with gray matter differences in the left amygdala.
Conclusions and Relevance: We identified gray matter reductions within the insula, amygdala, frontal and temporal regions in youths with CP as well as inconsistencies in sample characteristics across studies that should be addressed in future research.