Neural Correlates of Hand Gesture Imitation in Children with Autism Spectrum Disorder

Background: Motor skill and motor planning impairments are frequently observed in children with ASD, which is associated with impaired acquisition of social communication skills linked to poor imitation learning. Learning an action requires integration of sensorimotor information and encoding models to replicate the action and understand those actions when performed by others. Learning skilled gestures is particularly reliant on integration of visual and proprioceptive input, and may provide critical developmental foundations for social communication. Exploring the diagnostic differences in imitation and motor planning may provide insight into mechanisms of social communication deficits.

Objectives: We used a gesture imitation task to examine the neural correlates of motor planning and to investigate whether differential neural patterns of motor planning during imitative learning exist between typically developing (TD) and high functioning children with autism (HFA).

Methods: fMRI data were collected on 22 subjects, HFA (N=10) (9M, 10.89+/-1.20 years) and TD children (N=12)(9M, 10.29+/-0.98 years). Participants were trained to imitate a series of 14 videos of novel meaningless-gestures performed by an actor using her left hand, establishing a similar level of performance crucial for investigating diagnostic differences. Participants were instructed to observe the gesture, plan how to copy, copy the gesture with their right hand (mirror image), and rest, according to trained colour cues.

Results: Behavioural analysis showed comparable gesture accuracy performance between groups (HFA: 77.8%±8.00; TD: 85.7%±7.14, p=.972). Preliminary analyses reflect significant effect of diagnosis at the p<0.001 level using a 10 voxel cluster-level threshold. In the observe condition greater activation in bilateral sensorimotor areas was observed in the HFA group than the TD, although the difference was more apparent in the right hemisphere. In the plan condition, it was the TD group that shows greater activation in the postcentral gyrus. In copy condition, there was greater activation for the HFA group localized to the left posterior cingulate gyrus, the right parahippocampus and anterior insula, whereas TD showed greater activation in V3.

Conclusions: While behavioural analyses demonstrated comparable gesture imitation, imaging analyses revealed significant group differences in the neural activation during observation, motor planning and execution. During observe, the HFA children recruited areas in the right somatosensory region relating to the left hand not seen in the TD group. The TD children showed recruitment of similar sensorimotor areas when required to plan the gesture, but in the left hemisphere, recruiting areas related to the right hand, used in copying the gesture. This differential recruitment in timing and hemispheres suggests that HFA children may have a delayed visuomotor transformation, initially encoding the anatomical hand (left) before transferring to the mirror hand (right). In copy, the posterior cingulate gyrus, involved in motor imagery tasks, is recruited in the HFA group, indicating mental imagery processing that is not apparent in TD children. These results suggest differential motor planning processing in imitative learning in children with autism, and may help to explain behavioural observations of autism-associated impairments in imitative skills. Further study needs to be done to explore possible differences in imitative encoding timecourses in autism.