Can Robotic Interaction Improve Nonverbal Communication and Social Anxiety of Children with Autism Spectrum Disorders?

Friday, May 15, 2015: 5:30 PM-7:00 PM
Imperial Ballroom (Grand America Hotel)
H. Kumazaki1,2, Y. Yoshikawa3, Y. Matsumoto4, S. Mizushima5, T. X. Fujisawa1, H. Kosaka1, A. Tomoda1, S. Nemoto6, M. Nakano7, M. Miyao7, T. Maeda2, H. Ishiguro3, T. Muramatsu2 and M. Mimura2, (1)Research Center for Child Mental Development, University of Fukui, Yoshida-gun, Fukui Prefecture, Japan, (2)Department of Neuropsychiatry, School of Medicine, Keio University, Tokyo, Japan, (3)Graduate School of Engineering Science, Osaka University, Toyonaka, Japan, (4)The National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan, (5)Research Center for Child Mental Development of University of Fukui, University of Osaka, Yoshida-gun, Fukui Prefecture, Japan, (6)Donguri clinic for developmental disorders, Tokyo, Japan, (7)National Center for Child Health and Development, Tokyo, Japan
Background:  Robot-mediated interventions for children with autism spectrum disorder (ASD) have recently received more attention. The use of robots has several advantages; robots are engaging and motivate children with ASD to interact. In addition, their behavior is consistent and highly predictable. Robot-mediated interventions may encourage children with ASD to exhibit social behaviors, such as imitation, eye contact, and joint attention. An android is a robot that looks human, but remain predictable and repeatable. The benefit of using androids in ASD therapy is that they may have the greatest potential for generalization.

Objectives:  This study aimed to investigate the changes in non-verbal communication and social anxiety seen in ASD with a structured communication protocol using an android. We hypothesized that children with ASD would show improved nonverbal communication and decreased social anxiety after they communicated with an android.

Methods:  The inclusion criteria were a chronological age of 6 to 16 years and previous diagnosis of high-functioning ASD (IQ ≥ 70). Twenty participants were randomly assigned to two experimental groups. Participants in one group received therapy with an android in the first session and with a human in the second. The other group received therapy with a human in both sessions. The investigator monitored the sessions from a neighboring room using a laptop. Remote controlled cameras were set on standby mode, ready to record and children were brought to the room by a clinical psychologist. We used the android Geminoid F, which is a female type, tele-operated android that looks like a person. Its artificial body has the same proportions, facial features, hair color, and hairstyle as its model such that, at the first instance and from a distance, it is difficult to distinguish the android from the model. The android is capable of a range of movements (moving limbs up & down and turning the head to the sides). We used hidden, Wizard of Oz-style, real-time, human remote control of the robot, a popular design paradigm in human-robot interaction research, in order to elicit each participant’s belief that the robot was behaving and responding autonomously.

Results:  Based on video recordings of the interactions, a quantitative and qualitative analysis was conducted. Some elementary behavioral criteria, such as eye contact, touching, pointing, smiling, attending to sounds, and following instructions were evaluated throughout the trials. In one group, eight out of 10 children showed improvements in their social interactions after they received therapy with the android. In the other group, only two out of 10 showed improvements in their social interactions (after therapy with a human.).

Conclusions: These results suggested that androids might help improve the nonverbal communication and social anxiety of children with ASD. Therapeutic approaches using androids are promising and warrant further study.