The Effects of Robot-Child-Child Interactions on Joint Attention and Verbalizations Patterns of Typically Developing Children and Children with ASDs Between 4 to 8 Years of Age

Background:  

Verbal and nonverbal communication delays are primary impairments of Autism Spectrum Disorders (ASDs) and are often addressed by traditional autism interventions. Currently, we are evaluating how an embodied social intervention involving interactions between a humanoid robot and two children affects the social communication skills of typically developing (TD) children and children with ASDs. Specifically, we would like to see whether such a motivating context enhances nonverbal and verbal communication between the two children interacting with a sophisticated 24-inch tall, humanoid robot called, Nao (Aldebaran Robotics, Inc.).

Objectives:  

To examine the effects of robot-child-child interactions on the rates of Joint Attention (JA) bids and the percent duration of verbalizations in typically developing children and children with ASDs between 4 to 8 years of age.

Methods:  

12 typically developing children and four children with ASDs were examined. Each child received 12, 45-minute training sessions involving robot-child-child interactions across six weeks. The 30-45-minute training session comprised of various training conditions: greetings, warm up, rhythmic action, drumming, walking, and farewells. Each movement context was also divided into “copy robot” and “move together” trials. We examined the rates of JA bids to the other child, percent duration of attention to the robot and the other child, as well as the percent duration of verbalization directed to the other child during the first, mid, and last sessions. We divided each code into spontaneous and responsive forms of communication. 

Results:  

Based on preliminary analysis, we expect the percent of spontaneous verbalizations to the other child to increase during the last session as compared to the mid and first training sessions. In terms of non-verbal communication, we expect the children to have greater rates of JA bids during the “move together” trial as compared to the “copy robot” trial. Overall, we expect the rates of JA bids to increase within the “move together” context across the several weeks of training.

Conclusions:  

Our preliminary data suggest that TD children increase their social communication skills following training as seen by increased spontaneous verbalizations to the other child. Relatively smaller improvements are expected in children with ASDs who may need a more extended and intense training protocol. Our next study will conduct a larger randomized controlled trial using such an intense protocol to examine the effects of robot-child interactions on the verbal and non-verbal communication skills of children with ASDs.