A web-based therapy can benefit children with autism by plugging into their natural affinity for technology and robots.
The Robot-Inspired Computer-Assisted Adaptive Autism Therapy (RoboCA3T) uses robot avatars and integrates them with computer-assisted therapies.
It reinforces the role of technology in improving engagement and developing skills in children with autism spectrum disorder (ASD).
The novel therapy, outlined in the Journal of Computer Assisted Learning, also demonstrates the importance of automated algorithms in minimizing assessment bias and human error.
“The research contributes significantly to the ongoing effort to develop cost-effective, time-efficient, evidence-based treatments for children with autism spectrum disorder,” said corresponding author Sara Ali, PhD, from the National University of Sciences and Technology, in Islamabad, Pakistan.
“RoboCA3T prioritizes personalized content delivery along with integration of AI-based automatic gaze and pose detection algorithms.”
ASD presents challenges in social skills and behavior, the authors note. These include joint attention—which is the ability to coordinate attention and share a point of reference with another person—and imitation skills, which involve observing and copying other people’s behavior.
While existing therapies such as applied behavior analysis and social skills training have shown benefits for ASD, there are limitations such as high cost and limited access.
RoboCA3T builds on the principles of applied behaviour analysis, aiming to maximize the engagement of children with ASD while also improving their behavioral and social skills.
It was tested on 11 children with ASD, who participated in 30 sessions per module, divided into two halves, over the course of eight months.
In total, there were 660 experimental trials, 110 familiarizations, and 110 follow-up sessions.
Joint attention was evaluated in a module that tracked the children’s eye movements to follow their gaze pattern in response to four least-to-most robot-generated cues.
The imitation module assessed the children’s imitated actions using computer vision techniques for pose estimation.
Comparing scores on the Childhood Autism Rating Scale before and after the intervention revealed significant enhancements in joint attention and imitation skills.
Automatic gaze and pose detection algorithms were integrated to ensure accurate results and address the challenges posed by human error and observation bias in assessing a child’s progress.
The authors report that their research responds to a need for more effective, technology-driven therapies for autism, filling gaps in existing methods.
“Our study introduces an innovative therapy method, emphasising technology’s role and advocating for inclusive, predictable, and personalized therapy approach for better engagement and skill development for ASD children,” they maintained.
