Assessing Social Communication and Collaboration in Autism Spectrum Disorder Using Intelligent Collaborative Virtual...

Area of Science:

• Technology and Medicine
• Autism Spectrum Disorder Research
• Human-Computer Interaction

Background:

• Current autism spectrum disorder (ASD) intervention research relies on subjective human ratings for social communication outcomes, limiting real-world applicability.
• Virtual reality (VR) and collaborative virtual environments (CVE) offer potential for objective, scalable, and cost-effective measurement platforms.
• Intelligent agent technology integrated with peer interactions can enhance the utility of CVEs for research.

Purpose of the Study:

• To develop and pilot a novel collaborative virtual environment and intelligent agent (CRETA) for assessing social communication and collaboration in individuals with ASD.
• To evaluate the accuracy of CRETA in classifying user statements and the agreement of observed skills in human-agent versus human-human interactions.
• To explore the potential of CRETA as an autonomous measurement system for future ASD research.

Main Methods:

• Development and piloting of the CRETA system, a collaborative virtual environment with an integrated intelligent agent.
• Utilizing CRETA to assess social communication and collaboration skills through system and peer interactions.
• Classification of user statements and comparison of communication/collaboration skills between human-human and human-agent interactions within the CVE.

Main Results:

• The CRETA system demonstrated moderate to high accuracy in classifying user statements.
• A moderate to high agreement was observed in displayed communication and collaboration skills between human-human and human-agent interactions.
• Pilot data suggests CRETA's feasibility for assessing social communication in ASD.

Conclusions:

• CRETA presents a novel and promising approach for the autonomous assessment of social communication and collaboration in autism spectrum disorder research.
• The system's ability to achieve agreement in skill assessment between human-agent and human-human interactions supports its potential for objective measurement.
• Further development of CRETA could lead to more replicable and generalizable findings in ASD intervention studies.