Impaired Voluntary Imitation of Biological Motion in Autism Spectrum Conditions

Background: Imitation is intricately linked to developing social relationships and motor learning. These interpersonal contexts require the appropriate imitation of biological motion, which is suggested to be impaired in autism spectrum conditions (autism) due to processes underpinning social modulation and motor control. These suggestions were based on examining biological motion during automatic imitation, with human models, and manipulating movement speed. Here, we examined voluntary imitation using a novel methodology that displayed three models with the same movement time, but different biological motion kinematics. To control for modulatory effects of social context, we used a non-human agent model (single white dot).

Objectives: The objectives were to: (1) examine imitation of biological motion kinematics; (2) examine underlying motor control processes learned from acquiring a 3-segment motor sequence timing task (motor task).

Methods: Twelve adults with autism, diagnosed by a clinical assessment and ADOS, plus twelve adults (control) participated in the study, which was conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee. In Phase 1, each participant imitated, using a stylus on a graphics tablet, different models displaying distinct velocity profiles: unnatural velocity (biological); natural velocity (biological); constant velocity (non-biological). The movement time of each model was 1700 ms. In Phase 2, each participant physically practised, with feedback, a motor task that had a movement time goal of 1700 ms.

Results: Compared to the control group, data from Phase 1 indicated that participants with autism imitated movement time over trials to become significantly closer to the model time (p<.05). The kinematic data showed no difference between the participants with autism and control participants when imitating natural and constant velocities (ps>.05). However, the control participants imitated unnatural velocity more accurately than the participants with autism (ps<.05). In Phase 2, the performance data indicated the autism and control groups became significantly more accurate and consistent over trials (ps<.05). The kinematic data indicated that over trials the participants with autism did not increase peak limb acceleration in the ballistic phase of the movement as per the control group (p<.05). A correlational analysis indicated a relationship between imitation accuracy of unnatural velocity and peak acceleration for the control group (p=.058), but not the autism group (p>.05).

Conclusions: Using non-human agent models, we observed that imitation of biological motion kinematics is impaired in autism, but that imitation of the temporal properties of the model became more accurate over trials. These findings suggest the deficit in imitating biological motion might be related to persons with autism orientating attentional focus to the temporal goal at the expense of kinematics. Although the general sensorimotor processes associated with learning a motor task are functional in autism, a difference was observed in planning and executing the ballistic phase of the movement. When correlated with imitation accuracy, no relationship was observed between motor control and imitation of biological motion in participants with autism. These findings suggest impairments in imitating biological motion in autism spectrum conditions is most likely related to focusing attention to the temporal goal.