The Effects of Robot-Child Interactions on the Solo and Social Synchrony of Children with Autism Spectrum Disorders

Background: Children with Autism Spectrum Disorders (ASDs) have significant motor impairments in dual and multilimb coordination as well as imitation and praxis (Green et al., 2008; Fournier et al., 2010). Motor incoordination in this population not only impacts their performance of daily activities but also affects their social participation skills (Bhat et al., 2011).

Objectives: In the present study, we evaluated the efficacy of a novel movement based intervention using robots on the solo and social coordination skills of children with ASDs between 5 and 12 years of age.

Methods: 24 children with ASDs between 5 and 12 years of age were matched on age and level of functioning and randomly assigned to either a robot group or an academic group. Children in both groups received 16 training sessions over 8 weeks (2 sessions per week). Pretest and posttest sessions were conducted in the first and last weeks of the study. The robot group engaged in imitation games involving whole body movements and drumming with a humanoid robot, Nao (Aldebran Robotics, Inc.) and an adult model. The academic group engaged in tabletop academic and fine motor activities such as reading, building, and arts & crafts with an expert trainer and an adult model. Generalized motor performance was examined using the standardized Bruininks-Oseretsky Test of Motor Proficiency (BOTMP) and the Sensory Integration and Praxis Tests (SIPT) in the pretest and posttest. Specifically, we used the bilateral coordination, fine motor precision, and fine motor integration subtests of the BOTMP and the bilateral motor coordination subtest of the SIPT. Training-specific changes in social synchrony between the child with ASD and the adult model were assessed during an early, mid, and late training session using standard rhythmic actions. Specifically, we assessed the percent of time children spent in-phase, anti-phase, and off-synchrony with the model.

Results: Preliminary data from one child with ASD who received robotic training suggest the presence of generalized and training-specific improvements in solo and social synchrony. Specifically, the child improved motor performance on the bilateral coordination subtest of the BOTMP by 18 points and reduced the total number of errors in the SIPT by 16 points in the posttest compared to the pretest. In terms of social synchrony, the child with ASD increased the percent of time spent “in-phase” with the adult model from 14.61% in the early session to 32.27% in the late session. Given the nature of the training activities in the academic group, we hypothesize that children in this group will show greater improvements in the fine motor subtests of the BOTMP post-training compared to the robot group. We are continuing to analyze all the collected data.

Conclusions: Our final results may provide supporting evidence for the use of robot-child interactions within group-based motor activities to improve the solo and social synchrony abilities of children with ASDs and could serve as a promising tool for addressing the motor and social impairments of this population.