Note: Most Internet Explorer 8 users encounter issues playing the presentation videos. Please update your browser or use a different one if available.

Playing with a Robot: Enhancing Social Communication and Interaction

Friday, 3 May 2013: 09:00-13:00
Banquet Hall (Kursaal Centre)
A. S. Roberts1 and S. M. Shore2, (1)Clinical, Boston Higashi School, Randolph, MA, (2)Special Education, Adelphi University, Garden City, NY

Intelligent robots were first developed in the 1950’s. Initially the focus was on cognition and problem-solving abilities. The use of robots with social-emotional intelligence, either socially evocative or socially competent, only developed more recently. These potential uses are just beginning to be explored. Some initial studies using such robots to teach play skills to autistic preschoolers found they successfully mediated turn-taking, joint attention, imitation, and proactive behavior (Dautenhahn, 2007). Recent reports of “humanoid” robots (KASPAR and Bandit) suggest they are successful in generating social attention and social smiles from more isolated severely autistic children (Dautenhahn, Nehaniv, Walters, Robins, Kose-Bagci, Mirza and Blow, 2009; Woolston, 10-17-11). Furthermore, initial results using the “humanoid” robot NAO indicate future potential for using this technology during traditional therapy sessions (Shamsuddin, 2012). PARO, a baby harp seal robot, has been used in nursing homes and hospitals to promote positive social interaction from withdrawn and socially isolated individuals (Wada and Shibata, 2007).


To determine if interaction with PARO will be effective for children with autism in…

• stimulating individual/social play, language and emotional expression, attention and joint attention, and appropriate sensory play while decreasing stereotypical behavior.

• aiding students demonstrating an initial fear of PARO (with a history of fear of dogs/small animals), in decreasing that fear through repeated exposure as measured by the behavioral variables of the study.


18 students at the Boston Higashi School (15 boys and 3 girls) with diagnoses on the autism spectrum, aged 8 to 14, participated. 8 of these students (44%) were day students while 10 (56%) were residential students, participating 24/7 based on severity of need. The students, divided into groups of 4 or 5, met once a week to “play with PARO”. Over 9 sessions, each group began with 3 Free Interaction for the initial “A”, then 3 each for the Facilitated Interaction and Representational Play for the “B” or intervention, and then 1 Free Interaction session for the second “A” generalization session.  A repeated measures analysis of variance, other relevant statistical tools, and qualitative data was analyzed to determine if there was a significant difference in how students interacted with PARO before and after the intervention.


Preliminary findings suggest...

  • Providing structure (“B”) aides students in developing emotional interactions with PARO.
  • Social interaction and communications increased during the intervention phase and sensory seeking decreased as hypothesized.
  • Most variables returned to pre intervention levels when structure was removed except for seeking peers out which remained improved.
  • Self stimulatory behavior increased during the intervention stages, contrary to expectations, but fell below baseline levels during the generalization session. 
  • Fearful students decreased their levels of fear and anxiety over the repeated trials, measured by their proximity to PARO, willingness to engage in contact and observable facial expressions.


It appears there are significant differences among various variables including social interaction and self-stimulatory behaviors during the intervention stage.  Further research is indicated to increase the generalization effect of the intervention for students at the severe end of the spectrum.

| More