A new computerized assessment measuring pediatric patients' motor imitation abilities in just 1 minute may help distinguish autism spectrum disorder from attention-deficit/hyperactivity disorder, according to research conducted at the Kennedy Krieger Institute.
In the preprint study, researchers at the Center for Neurodevelopmental and Imaging Research found that the Computerized Assessment of Motor Imitation (CAMI) could identify autism spectrum disorder (ASD)-specific motor imitation difficulties with 80% accuracy when comparing pediatric patients who had ASD with neurotypical pediatric patients, and 70% accuracy when distinguishing ASD from attention-deficit/hyperactivity disorder (ADHD).
The cross-sectional study included 183 patients aged 7 to 13 years—35 of whom had ADHD (without ASD), 63 of whom had ASD and co-occurring ADHD, 21 of whom had ASD only, and 65 of whom were neurotypical patients. All of the participants had IQ scores of 70 or higher.
"Regardless of co-occurring ADHD, [patients] with an ASD diagnosis had significantly poorer CAMI performance as compared to neurotypical [patients]," the study authors reported.
The researchers found that the performance of patients with ADHD was "indistinguishable" compared to the performance of neurotypical children, suggesting the tool's specificity to ASD.
The CAMI assessment involved patients copying dance-like, whole-body movements of a video avatar while standing. Two Xbox Kinect cameras recorded the movements, and computer vision methods analyzed the data "virtually without any human input for data processing," eliminating the need for time-consuming human observation coding.
Among the patients with ASD, poorer CAMI performance was specifically associated with core ASD traits measured by the Autism Diagnostic Observation Schedule (ADOS-2), particularly in social affect and restricted and repetitive behaviors. Notably, CAMI performance was not associated with ADHD traits or general motor ability in the ASD group.
The researchers emphasized CAMI's potential clinical utility, writing that "establishing motor imitation as a biomarker for [ASD] requires robust methods with high sensitivity and specificity." However, they noted several limitations, including the relatively small sample sizes for the ADHD (n = 35) and ASD-only (n = 21) groups.
The researchers suggested that future research should adapt CAMI in younger patients, given that earlier ASD diagnosis may enable more effective interventions. Because the current study only included patients with IQ scores above 70, they also recommended exploring the tool's utility in those with intellectual disabilities.
Distinguishing between ASD and ADHD remained challenging: 50 to 70% of ASD cases have co-occurring ADHD. This difficulty in differential diagnosis can lead to delayed ASD identification and reduced access to appropriate interventions.
The researchers concluded that CAMI shows promise as "a reliable and highly scalable method for detecting ASD based on motor imitation differences." They emphasized that the tool's brevity, automated analysis, and high accuracy make it potentially valuable for clinical practice; however, additional research is needed prior to clinical implementation.
This research was funded by grants from the National Science Foundation, National Institutes of Health, and Simons Foundation. The study authors declared no conflicts of interest.