Brain Volumes Associated with High Levels of Aggression in ASD

Background:   Some level of aggression is present in around half of children and adolescents diagnosed with autism spectrum disorders (ASD). Comorbid aggression in autism causes additional family stress including increased isolation, parental exhaustion, safety concerns, and financial strain. Previous neuroimaging research has suggested structural abnormalities in ASD in prefrontal and medial temporal regions of the brain. Similar structures have also been implicated in aggressive and hyperactive behavior in non-autism samples. The definition of narrower behavioral phenotypes will be useful in elucidating brain networks that are more specific to autism subtypes. However, we know of no studies investigating the potential association of aggression with atypical brain development in ASD. 

Objectives:  We analyzed associations between MRI-based volumetric measurements and behavioral reports of aggression in ASD. We hypothesized that aggression would predict abnormal growth, relative to matched controls, in prefrontal and medial temporal regions of the brain.

Methods:   We used data from the Utah Autism Research Project, described in Allen-Brady et al. (2010). Fifty-eight ASD participants (ages 3-36) were compared with 33 neurotypical controls matched for IQ and age. We conducted whole-brain analyses of structural MRI data. The ASD aggression group was defined using a score of >17 on the Irritability subscale of the Aberrant Behavior Checklist (ABC). Statistical analyses consisted of general linear models with brain measures as dependent variables predicted by group (low aggression controls, low aggression ASD, and high aggression ASD) while controlling for age and intracranial volume.

Results:   Various frontal regions emerged as significant areas associated with increased aggression in ASD. The high aggression ASD group had decreased white matter volume in total frontal lobe bilaterally. Additionally they had increased gray matter volumes in left frontal pole and right mOFC. In addition to frontal areas, the high aggression ASD group had decreased white matter in the right temporal pole.  An unexpected decrease in brainstem volume was also observed for the high-aggression group. 

Conclusions:  Consistent with morphological studies of ASD as well as studies of aggression, these data suggest abnormal neural organization in the frontal and temporal regions for ASD individuals with comorbid aggression. The decrease in key white matter volumes suggest a possible disruption in the neural network of the brain leading to increased aggression. Decreased brainstem volume may indicate the contribution of a low-level regulatory system contribution to impulse control difficulties associated with aggression.