Network-wise surface-based morphometric insight into the cortical neural circuitry underlying irritability in adolescents

Abstract

Abstract Previous studies examining structural brain correlates of irritability have taken a region-specific approach and have been relatively inconsistent. In a sample of adolescents with and without clinically impairing irritability, the current study examines: (i) cortical volume (CV) in canonical functional networks; (ii) the association between the CV of functional networks and severity of irritability; and (iii) the extent to which IQ mediates the association between structural abnormalities and severity of irritability. Structural MRI and IQ data were collected from 130 adolescents with high irritability (mean age = 15.54±1.83 years, 58 females, self-reported Affective Reactivity Index [ARI] ≥ 4) and 119 adolescents with low irritability (mean age = 15.10±1.93 years, 39 females, self-reported ARI < 4). Subject-specific network-wise CV was estimated after parcellating the whole brain into 17 previously reported functional networks. Our Multivariate Analysis of Covariance (MANCOVA) revealed that adolescents with high irritability had significantly reduced CV of the bilateral control and default-mode networks (p < 0.05) relative to adolescents with low irritability. Multiple regression analyses showed a significant negative association between the control network CV and the severity of irritability. Mediation analysis showed that IQ partially mediated the association between the control network CV and the severity of irritability. Follow-up analysis on subcortical volume (SCV) showed that adolescents with high irritability had reduced bilateral SCV within the amygdala relative to adolescents with low irritability. Reduced CV within bilateral control and default networks and reduced SCV within bilateral amygdala may represent core features of the pathophysiology of irritability. The current data also indicate the potential importance of a patient’s IQ in determining how pathophysiology related to the control network is expressed.