Corticocerebellar Resting State Correlations in Autism

Background: Altered cerebellar function has been suggested as the source of speech and movement atypicalities in autism. The cerebellum is strongly connected with motor control systems in the cerebral cortex. Evidence from both neurophysiologic and neuroimaging studies have identified four spatially distinct body maps in the cerebellum, with two representations in each hemisphere, one in the anterior cerebellum (Lobule IV/V) and the other in the posterior cerebellum (Lobule VIII). These body maps are connected to cortical motor areas via the thalamus and receive reciprocal connections via corticopontine and pontocerebellar projections. Recent work has demonstrated that functional MRI can be used to record spontaneous BOLD-contrast fluctuations at rest. Modeling these data using correlation techniques allows estimation of low frequency interregional influences in corticocerebellar systems responsible for supporting a range of sensorimotor and cognitive activities.

Objectives: Our goal was to examine the resting state correlations in four different corticocerebellar systems in autistics. The strength of these correlations may provide information concerning the operation of the structural and functional infrastructure supporting movement in autistics.

Methods: Our sample included 19 autistic and 21 non-autistic participants, matched for age, sex, manual preference and IQ. Using a 3T MRI system, we examined interregional BOLD-contrast bivariate correlations in time series collected over a 10 min period while participants rested with eyes closed. Noise sources were removed using linear regression, including estimated head motion, and average ventricular and white matter signals. Seeds were placed in various parts of the visuomotor control system, including primary motor cortex, supplementary motor area, thalamus, posterior parietal cortex, and cerebellar lobules IV/V and VII. We calculated Pearson’s correlation coefficient between the band-pass filtered time series of the seed regions and all brain voxels. The resulting r-maps were then converted to z-scores using the Fisher transform for between-group comparisons.

Results: The between-group analysis was restricted to regions which exhibited positive correlations with the seed regions. Compared to the non-autistics, the autistics had atypical interregional correlations in both hemispheres in frontocerebellar loops, with the lobule IV/V -> thalamus correlations increased, the thalamus -> primary motor cortex correlations increased and the primary motor cortex -> lobule IV/V correlations decreased. The parietocerebellar loops also showed differential influences, with the lobule VIII -> thalamus correlations unchanged, the thalamus -> superior parietal lobule correlations increased and the superior parietal lobule -> lobule VIII correlations increased. In addition, the autistics exhibited reduced correlations between the left and right thalami, suggesting a relative decoupling of their respective hemispheric systems.

Conclusions:

The observed pattern of atypical interregional resting state correlations encompassed anterior and posterior corticocerebellar systems in both hemispheres. Examination of interregional neural influences at rest provides useful information about the functional organization of the distributed neural subsystems responsible for movement execution. In autistics, the corticocerebellar systems involving both frontal and parietal cortex are characterized by generally higher synchrony, perhaps reflecting developmental differences in local synaptic organization. These differences in interregional corticocerebellar activity may be the neural substrate for the delays and atypicalities in speech and movement commonly seen in autism.