Evidence for Selective Damage to Cognitive Cortico-Cerebellar Circuits in Autism Spectrum Disorder

Background:

A growing body of literature suggests that the cerebellum contributes to a wide variety of behaviours, both cognitive and motor. This is largely supported by anatomical evidence from both humans and non-human primates which shows that specific cerebellar lobules are exclusively interconnected with either motor regions (cerebellar lobules HV, HVI, HVIIb, HVIII) or prefrontal and parietal cortices (cerebellar lobules Crus I/II). Cerebellar deficits are routinely found in Autism Spectrum Disorder (ASD), but the functional and behavioural consequences of this are not yet clear. Given that theories of cerebellar function are largely based on its extrinsic connectivity, investigating the differences in functional connectivity within these circuits should help us better understand how cortico-cerebellar deficits contribute to ASD pathology. 

Objectives:  

This study uses resting state fMRI to investigate cortico-cerebellar connectivity changes in ASD.

Methods:

28 right-handed, medication free, male participants with high functioning ASD and 27 right-handed male, age and IQ matched controls participated (Mean age: ASD=17.28 (SD= 3.57); Control=17.15 (SD= 3.64); Mean IQ ASD= 109.25 (SD= 15.04); Control=111.85 (SD= 12.32)). 21 ASD and control participants were retained for the fMRI analyses after excluding subjects for factors such as excessive motion (movements >3mm) and poor data quality. fMRI preprocessing was carried out in SPM8 and functional connectivity analysis was carried out using the CONN toolbox (ROI->Voxel). The cerebellar atlas of Diedrichsen et al (2009) was used to extract timecourses (principle eigenvariate) for each cerebellar lobule. Age and IQ were modeled as regressors of no interest. Results were corrected for multiple comparisons using FDR correction (p<0.05).

Results:

It has been repeatedly shown that cerebellar lobules Crus I and II are interconnected with prefrontal and parietal regions. Our analyses showed that connectivity between Crus II and prefrontal/parietal cortices (specifically bilateral intraparietal sulcus and right middle frontal gyrus) was reduced in ASD compared to controls. We also found connectivity between left Crus I and the vermal lobule IX was reduced in ASD. These two cerebellar regions have been shown to be functionally interconnected using clustering algorithms, and additionally interconnected with prefrontal/parietal cortices (Bernard et al 2012). Finally we found that there was greater connectivity between the right putamen and right lobules I-IV in ASD compared to controls. 

Conclusions:  

The cortico-cerebellar system is known to contribute to both cognitive and motor control, however these results suggest that it is specifically the cognitive cortico-cerebellar circuits that are altered in ASD. It is likely that decreased connectivity of the prefrontal/parietal cortico-cerebellar loop contributes to cognitive deficits in ASD, potentially working memory which has been shown to activate these regions (Stoodley et al 2012). Given that motor cortico-cerebellar circuits do not differ between groups it is unlikely that cortico-cerebellar connectivity is related to motor deficits seen in ASD such as repetitive behaviors (although these may still be caused by other cerebellar abnormalities). Further analyses of neuropsychological data are necessary in order to better relate these differences in cortico-cerebellar connectivity to differences in ASD symptomology.