Atypical Activity in a Left Hemisphere Response Selection System in Autism

Background: While the most commonly used diagnostic tools for assessing autism emphasize social and language abilities, they generally do not explicitly emphasize motor control skills beyond the assessment of orienting, stereotyped and repetitive movements.  This same overall emphasis is reflected in many behavioral studies of autism that report, within a context of developmental milestones in typically developing controls, that differences in basic motor skills are unremarkable compared to atypical linguistic or social skills.  Nevertheless, growing evidence demonstrates that autistics experience a range of difficulties in sensorimotor skill acquisition, manifest as dyspraxia, oromotor problems, atypical or absent speech, and difficulty in action imitation.  While these atypicalities in complex movement are common, their neural mechanisms are poorly understood.

Objectives: Although complex movement differences can arise from many potential sources, our goal was to explore the hypothesis that the neural systems controlling response selection are differentially engaged in the context of tasks that require selection of one from a range of possible actions.  Selection for action has been described as a preferential left hemisphere function involving dorsolateral prefrontal, lateral premotor and posterior parietal cortices.

Methods: We explored the neural systems for action planning in autistics and a matched comparison group using a response selection task, a well-documented method for assessing motor planning operating in the service of voluntary action. A sample of 15 autistic and 17 non-autistic participants, matched on age, sex, IQ and manual preference performed a response selection task that required discrete movements of individual fingers of either the right or left hand to be made in response to one of eight possible visual patterns. Task-related regional patterns of neural activity were estimated using functional MRI.

Results: Although the participants' response time and accuracy did not differ between the autistic and non-autistic groups, greater task-related activity was observed in the non-autistic group in a set of left hemisphere regions associated with visuomotor planning, including posterior parietal cortex, dorsolateral prefrontal cortex, and lateral premotor cortex (p<0.001 FWE corrected). In addition, greater bilateral activity was seen in inferotemporal cortex in the non-autistic group. In contrast, greater task-related activity was observed in the autistic group in regions more closely associated with the execution aspects of movement, including primary motor cortex, the thalamus, putamen and cerebellar cortex.

Conclusions: A group of autistics, not selected on the basis of their motor capabilities, demonstrate clear differences in the cortical mechanisms responsible for visuomotor action planning during a task involving selection of one of multiple possible responses. This differential engagement of a left hemisphere perception-action circuit, observed during performance of a relatively simple visuomotor task, may reflect a more general neural resource limitation encountered by autistics when planning more complex movements. In the context of the higher planning demands associated with complex actions, such as imitation and speech, this resource limitation might lead to easily observable performance differences.