Computerized assessment of motor imitation (CAMI): Advancing the validity and scalability of a promising phenotypic biomarker for autism
- Awarded: 2020
- Award Type: Pilot
- Award #: 724867
In the current project, Stewart Mostofsky and colleagues plan to use computer vision technology to examine motor imitation ability, which is known to be crucial for social-communicative development in neurotypical populations and to be affected in autism spectrum disorders (ASDs). The team’s approach is brief (one minute), engaging (videogame format) and scalable, making it highly tractable for assessing a large number of children across a wide range of clinical and home settings.
Mostofsky and his team have performed a pilot study in which they performed initial steps towards developing and validating their method, ‘computerized assessment of motor imitation’ (CAMI), to assess imitation performance in children with ASD and typically developing children1. As a first step, they applied 3-D motion Kinect Xbox cameras to assess how well a child imitated an avatar’s on-screen movements. The CAMI scores alone successfully identified ASD diagnosis with over 87 percent accuracy—a 10-point increase from traditional human observation methods.
In the current project, they propose to advance the CAMI method to increase its scalability and establish it as a biomarker of ASD. Specifically, they plan to: (i) establish the test-retest reliability of CAMI and (ii) optimize the efficiency and scalability of CAMI by using low-cost 2-D cameras. Individuals will be asked to imitate a video avatar as the avatar performs dance-like movements in six brief (15-second) blocks. They aim to recruit 50 children with ASD (ages 6 to 12 years) as well as expanding the study to adults with ASD. Recruitment will be done through the SPARK clinical site at Kennedy Krieger Institute.
The researchers anticipate that their improved CAMI method and the portable, simple and fun dance task will enable widespread assessment of motor imitation without requiring lengthy and costly training for clinicians. Objective quantification of imitation impairments will provide a foundation for developing targeted behavioral interventions, as enhanced imitation can help improve the acquisition and development of social-communicative skills.