Regions of the Cerebellum Show Decreased Grey Matter Volume in Women with Autism Spectrum Disorder
Objectives: (1) To examine differences in cerebellar neuroanatomy in females with ASD compared to female neurotypical controls, and (2) to establish the relationship between cerebellar neuroanatomy and the severity of autistic symptoms within the ASD group.
Methods: Structural Magnetic Resonance Imaging (sMRI) data were acquired in 49 females with an Autism Diagnostic Interview Revised (ADI-R)-confirmed diagnosis of ASD, and 47 neurotypical female controls who did not differ significantly in age (28±7 and 27±7 years, respectively) and full-scale IQ (118±11 and 115±7, respectively). These participants were recruited as part of the MRC UK Autism Imaging Multicentre Study (MRC AIMS), and scanned at the Institute of Psychiatry, Psychology and Neuroscience in London, and the Autism Research Centre, University of Cambridge, UK. To examine between-group differences in regional cerebellar neuroanatomy, we utilized the SUIT-toolbox (http://www.diedrichsenlab.org/imaging/suit.htm), which provides a high-resolution template of the human cerebellum and a parcellation for different functionally defined cerebellar subregions. Following cerebellar isolation and normalisation, data were analysed using a voxel-based approach to examine between-group differences at each cerebellar voxel. A cluster-threshold was used to assess cluster-level significance at a P-value that provided the expected number of false positive clusters by chance to be smaller than one (i.e. p=.003). We also examined Pearson correlation coefficients between the individual’s cluster volumes and the different subdomains of the ADI-R, and the Autism Diagnostic Observation Schedule (ADOS).
Results: Individuals with ASD showed significant reductions in regional cerebellar volume in (1) the left lobules Crus II, IX, VIIB, VIIIb, (2) the right lobules IX and VIIIb, and (3) parts of the vermis (VIIb, VIIIa, VIIIb). These clusters functionally sub-serve the default-mode network and the frontoparietal network. Significant negative correlations were found between the cluster volume (1) in the left hemisphere and the ADI-R social (r=-0.313, p=0.44) and communication domain scores (r=-0.396, p=0.009); and (2) between the volume of the right-hemisphere cluster and the ADI-R social domain (r=-0.39, P=0.011), and ADOS social (r=-0,396, P=0.006) and communication scores (r=-0,3, P=0.041). Both clusters also correlated with the total ADI scores (i.e. lower regional volumes were associated with more severe autistic symptoms).
Conclusions: This is the first large-scale multicentre MRI study to investigate differences in cerebellar neuroanatomy between females with ASD and female controls, and to relate neuroanatomical findings to clinical outcomes. We found that females with ASD – like their male counterparts – have regional differences in cerebellar neuroanatomy, and that these are associated with clinical symptom severity. Future research is, however, warranted to compare males and females with ASD directly in order to determine whether these neuroanatomical differences are shared or distinct across sex.