Improved Communication Outcomes Using a Socially Assistive Robot

Background:  Recent research has employed socially assistive robots (SAR) as a tool for promoting communication and socialization skills in young children with ASD (Greczek et al, 2013, Bekele et al, 2013, Welch et al, 2010).  A number of studies describe observed therapeutic outcomes such as increased speech, social interaction, joint and directed attention (Robins et al, 2009, Feil-Seifer et al, 2008), but few quantify communication increases with assessment instruments accepted by autism and speech therapy interventionists.  Even fewer provide a between-group statistical analysis of outcome measures achieved with an added robot intervention compared to a control group not receiving it.

Objectives:  To evaluate the efficacy of a robot-assisted intervention by quantitatively (1) assessing the effectiveness of a robust, interactive robot for increasing spontaneous speech, communication and social skills in children with ASD and (2) comparing communication and social skills increases obtained through therapies augmented with a robot-assisted intervention to speech therapy as usual.

Methods:  Study participants (N=8; 8 males; age M = 4.83 years, SD =0.83 years) diagnosed with ASD and a speech deficiency as confirmed by the ADOS and a speech pathologist. Each child participated in two, 30-minute robot-assisted interventions per week for six weeks. A control group (N=3; 2 males, 1 female; M = 4.39 years; SD=0.38) with an ASD diagnosis and a documented speech delay were also recruited to compare outcomes obtained through speech therapy alone.  A low-cost robot prototype (CHARLIE) with safety features such as a snap-off head, two snap-off arms and a camera for face and hand detection was used for the intervention (Boccanfuso et al, 2011).  The study group received pre- and post-intervention measures including the Vineland Adaptive Behavioral Scale II (VABS-II), the Mean Length Spontaneous Utterance Determination (MLSUD), Motor Imitation Scale (Stone et al., 1997), Unstructured Imitation Assessment (UIA) (Ingersoll et al., 2010) and the Expressive Vocabulary Test 2 (EVT2) (Williams, 2007).

Results:  Paired-samples t-test statistic were performed to determine the therapeutic effect of CHARLIE of on adaptive functioning, imitation, and basic language skills.  The results indicate the CHARLIE intervention improved adaptive functioning in the VABS-II Socialization (p<0.0184) and Communication (p<0. 0348) Domains. The results also showed improvement in MLSUD (p<0.0092), UIA Social Interaction (p<0.0193), UIA Requesting (p<0.0068), and UIA Joint Attention (p<0.0145).  An independent samples t-test, performed to compare treatment and control groups, showed significance in VABS-II subdomains Play and Leisure (p<0.0469), and Receptive Language (p<0.0421). There were trends toward significance on the VABS-II subdomains Coping Skills (p<0.0667), and Interpersonal Scale (p>0.0618).

Conclusions:  While this is a pilot with a relatively small N, the results suggest one can achieve measured improvements in communication and social interaction using an augmented intervention with a low-cost robot. This study provides critical insight as to the viability and efficacy of using a simple robot as a tool for various autism interventions.  Our technique employs a robot that is widely-accessible, robust, can withstand some physical manipulation by children and is easily operable by therapists, teachers and parents.