Adults with Autism Spectrum Disorders Imitate Means to an End: The Effects of Sensorimotor Integration and Interference

Thursday, May 12, 2016: 5:30 PM-7:00 PM
Hall A (Baltimore Convention Center)
M. Andrew1, S. J. Bennett2, D. Elliott2 and S. J. Hayes2, (1)Liverpool John Moores University, Rossendale, England, United Kingdom, (2)Research Institute for Sport and Exercise Sciences, Liverpool John Moores University, Liverpool, United Kingdom

When imitating a novel action, the sensorimotor system is configured by representing and executing observed biological motion kinematics. Although this process is functional in early development, individuals with autism spectrum disorders (autism) show significant difficulties imitating biological kinematics. Rather than this deficit being underpinned by a core dysfunction in the sensorimotor system, it could be a function of how the lower-level sensorimotor processes are engaged to integrate information during imitation. Here we present three experiments that used a novel ‘true’ imitation protocol designed to facilitate (Exp. 1), and disrupt (Exp. 2 and 3) sensorimotor integration during the imitation of biological motion kinematics in individuals with autism.


Examine whether the imitation of biological kinematics is influenced by sensorimotor integration and interference.


Fifteen (neurotypical) adults, and fifteen adults with autism, diagnosed by a clinical assessment and ADOS, participated in each experiment of this study. The study was approved by an ethics committee. Participants imitated atypical and typical biological motion. A control stimulus displayed typical biological motion where peak velocity occurred at 44% of the movement. This is typical of most “bell-shaped” velocity profiles exhibited during goal-directed aiming. An experimental model displayed atypical biological motion where peak velocity occurred at 18% of the movement. Although this peak occurred earlier, it was still achievable, and thus allowed the direct examination of imitation of biological kinematics. In Experiment 1, the stimuli were presented in a fixed predictable trial order (e.g., 30 trials of atypical motion, followed by 30 trials of typical motion). This order facilitates integration across trials because afferent and efferent sensorimotor information from trial n can be processed and integrated during the inter-trial delay, and used to plan trial n+1. In Experiments 2 and 3, we disrupted sensorimotor integration using a secondary interference task that was placed in the inter-trial delay (Exp. 2), or by presenting the stimuli in a random unpredictable (Exp. 3) trial order (i.e., 60 trials of atypical and typical motion randomised).


Individuals with autism imitated atypical biological motion to a similar level of accuracy as matched neurotypical controls (p > 0.05), and significantly (p < 0.01) different to the typical control stimulus. Following the secondary task (Exp. 2), and random trial order (Exp. 3), the autism groups became significantly less accurate than the neurotypical control groups at imitating atypical biological motion (ps < 0.05). The interference effects resulted in the autism groups exhibiting comparable timing of peak velocity in the atypical and typical models (p > 0.05).


Poor voluntary imitation in autism, and its underlying sensorimotor aetiology, has received a great deal of attention and is still a matter of controversy. Experiment 1 established robust imitation effects when adults had opportunity to integrate, consolidate and represent atypical biological kinematics in a predictable environment. Experiments 2 and 3 disrupted sensorimotor integration and consequently attenuated imitation fidelity. Rather than there being a core deficit in imitation, the findings indicate intact sensorimotor processing of biological motion in autism when the system has opportunity to repeatedly integrate information.