A Longitudinal Study of the Amygdala and Hippocampus In Autism

Background: Previous structural MRI studies suggest early life enlargement of the amygdala in autism but it is not clear whether this size increase persists beyond childhood. Interestingly, evidence from cross-sectional imaging investigations examining adolescents and adults with autism suggest that volumetric differences disappear in the older individual. In contrast, structural studies of the hippocampus have been inconsistent, with some studies suggesting decreased volume and others no differences or slight increases.

Objectives: The goal of this investigation was to compare the pattern of amygdala and hippocampal volume changes in older youth with autism and healthy controls using a longitudinal design. Additionally, relationships between structural changes over time and clinical features were also examined.

Methods: Magnetic resonance scans were obtained from male children with autism and age- and gender-matched controls at baseline and 2-year follow-up. Volumes of the left and right amygdala and hippocampus were measured. A comprehensive cognitive and clinical phenotyping was obtained at baseline and included the Autism Diagnostic Observation Schedule (ADOS), Social Responsiveness Scale, and the Wechsler Intelligence Scale for Children. Scores from these measures were used to examine behavioral correlates of brain volumes. Mixed effects regression models were computed with Participant Group (Autism, Control) as a fixed effect factor and Age as a time-varying covariate. Analyses were performed with and without total brain volume and full scale IQ as covariates. Exploratory correlations examined behavioral correlates of volume growth.

Results: Forty-eight male youth (ages 8-13 at baseline) participated in the longitudinal imaging study. Thirty-eight participants had amygdala or hippocampal measurements at baseline and/or 2-year follow-up. Enlargement of the right, but not left, hippocampus was observed in the autism group at baseline, with a trend toward normalization - small increases with age in autism relative to large increases with age in controls. Differences in total amygdala growth were observed (autism 3.9% increase; controls 11.1% increase), but did not reach significance (p<.200).

Despite a very small autism sub-sample, greater social impairment at baseline on the ADOS was associated with less normalization of the volume of the right hippocampus. Interestingly, there were also strong relationships between better eye contact at baseline and larger increases in the right amygdala (r=.86, p=.006) and between better social motivation and higher verbal comprehension at baseline and larger increases in the left amygdala in youth with autism.

Conclusions: Enlargement of the right hippocampus supports previous findings of larger hippocampal volumes in autism and is consistent with the learning and memory deficits often observed in this disorder. Correlations with social dysfunction may suggest a specific role of the right hippocampus in learning social cues or may simply reflect a marker of more severe dysfunction. The amygdala did not show volume differences cross-sectionally or longitudinally which is consistent with evidence of increased size in early childhood and normalization later in life. The relationships of amygdala volume with social and verbal behavior suggest that increases in the size of this structure in the age range studied (ages 8-15) may identify a less impaired autism sub-group.