The interaction effect of MAOA and Maltreatment on Aggression subtypes, and their Neural Correlates

No Thumbnail Available
Journal Title
Journal ISSN
Volume Title
Aggression poses a major problem for society, through both financial costs and emotional distress. Although the genetics and etiology of human aggression are poorly understood, a polymorphism in the gene encoding monoamine oxidase A, MAOA, has been implicated through its effect on MAOA-expression. In an influential example of a gene-environment interaction, males with low MAOA-expression show a larger increase in aggressive behavior in response to childhood maltreatment than those with high MAOA-expression. It has not yet been studied how, and through which biological mechanisms, this interaction affects the reactive and proactive subtypes of aggression described in literature. Several subcortical brain regions have been implicated in aggression, most notably the amygdala, hippocampus, caudate nucleus and nucleus accumbens. We hypothesized that MAOA and maltreatment exert their effect on behavior through influencing volume of these brain structures. We aimed to test the interaction effect of MAOA and maltreatment on subtypes of aggression, as well as the volumes of subcortical regions of interest. We also investigated whether the volumes of these regions were associated with aggression measures in our sample. Additionally, we tested whether functional connectivity in aggression-related brain circuits was associated with aggression subtypes. The 30bp MAOA uVNTR was genotyped in a sample of healthy adults with available structural and resting-state MRI data. For partially overlapping subsets of the sample, information was available on maltreatment, assessed by the List of Threatening Events and on aggression, measured by the Reactive Proactive Questionnaire, the short form Buss-Perry Aggression Questionnaire and the Inventory of Callous-Unemotional traits. We performed confirmatory factor analysis in order to test the two-factor and three-factor models of aggression proposed in literature (N = 661). General linear model and logistic regression were used to assess the effect of MAOA-genotype and maltreatment on aggression measures (N = 82) and the subcortical volumes of interest (N = 258). General linear model was also used to examine the association between aggression measures and the volume of these subcortical regions (N = 574). Permutation testing was used to test the association between resting-state connectivity and aggression subtypes (N = 124). Confirmatory factor analysis showed a satisfactory fit for both the two-factor and the three-factor model, with a superior fit for the latter. No interaction between MAOA-genotype and maltreatment was found for any of the aggression measures. However, this interaction did affect the left nucleus accumbens, where maltreatment was associated with a volume increase in low- but not in high-MAOA-expressing subjects. Aggression measures were not significantly associated with the volume of subcortical regions of interest or with resting-state functional connectivity. The current findings show that the reactive-proactive aggression distinction is valid, but that aggression is better described by a three-factor model in the current sample. A novel finding is that maltreated subjects with low MAOA-expression show a volume increase in the left nucleus accumbens, which may, in light of earlier findings, mediate increased aggression. No subcortical regions of interest were associated with aggression subtypes in the current sample. Resting-state functional connectivity was also not significantly associated with reactive or proactive aggression. Future research is needed to identify how, and through which biological mechanism the interaction of MAOA-genotype and maltreatment affects aggression subtypes. Additionally, more research is needed to characterize the neural correlates of aggression subtypes.
Faculteit der Sociale Wetenschappen