Functional Brain Mechanisms of Sensorimotor Deficits in Individuals with Autism Spectrum Disorder

Background: Abnormalities in sensorimotor behavior are present in the majority of individuals with ASD and are associated with core symptoms. Cortico-cerebellar networks that control sensorimotor behavior have been implicated in ASD, but little is known about their function during sensorimotor actions.

Objectives: The purpose of this fMRI study was to examine cortical-cerebellar function during feedback-guided motor behavior in ASD.

Methods: Individuals with ASD (11-30 years; N = 18) and age-matched controls (N = 15) completed a visuomotor task of feedback-guided precision gripping during fMRI. Participants pressed with their right thumb and forefinger on a force transducer while viewing a green FORCE bar on a screen that moved upwards with increased force toward a fixed white TARGET bar. Individuals were instructed to maintain the FORCE bar at the level of the TARGET bar for 24 seconds. Target force levels were set at 20% and 60% of each participant’s maximum voluntary contraction (MVC). Force variability was characterized as the coefficient of variation (i.e., standard deviation of the force time series / mean force output; CoV).

Results: Mean force did not differ between groups indicating participants were able to follow task demands. Participants with ASD showed increased force variability (F_(1,30)= 5.214, p= 0.03) at both 20% (d= .45) and 60% (d= .77) MVC compared to controls. Compared to controls, individuals with ASD showed decreased activation in left angular gyrus during the visuomotor task compared to rest (AG; maximum t= 4.31). Individuals with ASD also showed greater visuomotor activation compared to controls in ipsilateral ventral M1, extending anteriorly into posterior ventral pre-motor cortex (PMv; maximum t= -4.06, cluster size = 38 voxels). This difference reflected the finding that control participants showed a selective deactivation of ipsilateral M1/PMv during visuomotor behavior, whereas individuals with ASD did not show this pattern. A significant group x force interaction was observed for contralateral Crus I activation (maximum t= -2.42) that was driven by an increase in activity during 60% compared to 20% MVC in control participants, while individuals with ASD showed no significant change in Crus I activation between force levels.

Conclusions: Increased force variability in individuals with ASD suggests impaired processing of sensory feedback to guide precision motor behaviors. Individuals with ASD did not show deactivation of right motor cortex during visuomotor behavior relative to rest, suggesting reduced ability to selectively modulate motor cortical output. Reduced activation in left AG may reflect an inability to integrate visual, haptic, and proprioceptive inputs to reactively adjust ongoing motor output. Failure to show force-dependent scaling of Crus I in ASD suggests lateral cerebellar circuits do not adapt sensory prediction and error processes to maintain precision motor output during more demanding conditions. Together, our results demonstrate multiple cortical-cerebellar mechanisms associated with sensorimotor imprecision in ASD.