The Effect of Robot-Child Interactions on Solo and Social Multilimb Synchrony in Typically Developing Children and Children with Autism Spectrum Disorders Between 4-8 Years of Age

Background:  

Complex multilimb coordination emerges gradually over development (Getchell and Whitall, 2003). Specifically, children progress from performing consistent dual-limb actions such as clapping to consistent, multilimb actions such as march and clap motions. Children with Autism Spectrum Disorders (ASDs) have significant motor impairments in overall coordination as well as imitation and praxis (Green et al., 2008). These impairments will not only affect an autistic child’s solo multilimb synchrony (movements on their own) but also their social multilimb synchrony (movements with a partner).

Objectives:  

In the present study we aimed to develop a novel intervention tool using robot-child interactions to facilitate solo and social synchrony of typically developing (TD) children and children with ASDs between 4 to 8 years of age.

Methods:  

12 TD children and 4 children with ASDs received 12 training sessions over a period of 6 weeks @ of 2 sessions per week. The training involved interactions of two children with a 24-inch tall humanoid robot called Nao (Aldebran Robotics, Inc.). The 30-45-minute training session comprised of various training conditions: greetings, warm up, rhythmic action, drumming, walking, and farewells. Solo and social coordination were measured using standardized motor measures using the bilateral coordination subtest of the Bruininks-Oseretsky Test of Motor Proficiency (BOTMP). In addition, solo and social synchrony was assessed in a task-specific and generalized synchrony test using kinematic analysis of slow and fast march-clap motions. In the task specific context, children were videotaped while imitating robot actions in session 1, 6 and 12. These were later coded for solo synchrony using measures of arm and leg movement variability. We also assessed social synchrony. In the generalized context, time spent in solo and social synchrony was evaluated using Continuous Relative Phase (CRP) analysis for march and clap actions in solo and social contexts.  CRP values ranged from 0°-180° (Scholz & Kelso, 1989) and were grouped into three bins: 0°-60° (for in-phase coordination), 60°-120° (asynchronous state), and 120°-180° (for anti-phase coordination). In-phase coordination is expected in bilaterally symmetrical or synchronous limbs while anti-phase coordination is expected in bilaterally asymmetrical and alternating limbs.

Results:  

Based on our preliminary analyses, we expect TD children to show improved bilateral coordination scores of the BOTMP during the posttest as compared to the pretest. Following training, we also expect an increase in the total time spent in task-appropriate solo synchrony as well as greater social synchrony during task-specific and generalized synchrony tests.  We expect the children with ASDs to have greater solo and social synchrony impairments as compared to TD children which will improve following training.

Conclusions:

TD children improved their solo and social synchrony following training. Children with ASDs had particular difficulties in social synchrony which is a function of their performance in solo synchrony. However, children with ASDs also showed positive training effects in the form of enhanced social synchrony. Taken together, robot-child interactions may serve as a promising tool to address impairments of solo and social mulitlimb synchrony in children with ASDs.