A phenotypic structure and neural correlates of compulsive behaviors in adolescents

Montigny, Chantale and Castellanos-Ryan, Natalie and Whelan, Robert and Banaschewski, Tobias and Barker, Gareth J and Büchel, Christian and Gallinat, Jürgen and Flor, Herta and Mann, Karl and Paillère-Martinot, Marie-Laure and others
PLoS One, 8(11): e80151, 2013


A compulsivity spectrum has been hypothesized to exist across Obsessive-Compulsive disorder (OCD), Eating Disorders (ED), substance abuse (SA) and binge-drinking (BD). The objective was to examine the validity of this compulsivity spectrum, and differentiate it from an externalizing behaviors dimension, but also to look at hypothesized personality and neural correlates.


A community-sample of adolescents (N=1938; mean age 14.5 years), and their parents were recruited via high-schools in 8 European study sites. Data on adolescents’ psychiatric symptoms, DSM diagnoses (DAWBA) and substance use behaviors (AUDIT and ESPAD) were collected through adolescent- and parent-reported questionnaires and interviews. The phenotypic structure of compulsive behaviors was then tested using structural equation modeling. The model was validated using personality variables (NEO-FFI and TCI), and Voxel-Based Morphometry (VBM) analysis.


Compulsivity symptoms best fit a higher-order two factor model, with ED and OCD loading onto a compulsivity factor, and BD and SA loading onto an externalizing factor, composed also of ADHD and conduct disorder symptoms. The compulsivity construct correlated with neuroticism (r=0.638; p≤0.001), conscientiousness (r=0.171; p≤0.001), and brain gray matter volume in left and right orbitofrontal cortex, right ventral striatum and right dorsolateral prefrontal cortex. The externalizing factor correlated with extraversion (r=0.201; p≤0.001), novelty-seeking (r=0.451; p≤0.001), and negatively with gray matter volume in the left inferior and middle frontal gyri.


Results suggest that a compulsivity spectrum exists in an adolescent, preclinical sample and accounts for variance in both OCD and ED, but not substance-related behaviors, and can be differentiated from an externalizing spectrum.