Characterization of predictive abilities in individuals with autism using web-based interception games

  • Awarded: 2018
  • Award Type: Explorer
  • Award #: 602705

Despite progress in elucidating the genetic and neural underpinnings of autism spectrum disorder (ASD), clinical assessments are still the gold standard for diagnosing and evaluating outcomes. Such assessments require significant clinician expertise, resist precise quantification and do not capture children in their familiar environments. To at least partially address this need, Dagmar Sternad’s laboratory at Northeastern University, in collaboration with Pawan Sinha (Massachusetts Institute of Technology) and Robert Joseph (Boston University School of Medicine), has been developing a suite of engaging online tests designed to quantitatively assess an important but understudied aspect of ASD: motor challenges.

The Predictive Impairment in Autism hypothesis posits that the common traits of ASD are manifestations of an individual’s difficulty in making predictions about cause and effect1. Building upon this theoretical framework, Sternad and colleagues hypothesize that the motor challenges in ASD arise from compromised prediction skills. Prediction is required in interacting with dynamic objects, such as catching a ball, and is also fundamental in multi-joint actions, where perturbations due to fast limb movements need to be preempted.

Preliminary work from Sternad’s team has resulted in the development of a suite of child-friendly virtual games to test this hypothesis with quantitative rigor. Pilot laboratory tests in 7- to 12-year-old children revealed selective differences in prediction between ASD and typically developing children. Tests of a larger cohort of typically developing children participating in an outreach activity verified the usefulness of these virtual games in less-controlled environments.

The aim of the current proposal is to assess predictive abilities in larger groups of both ASD and typically developing children using these virtual games in an online, web-based testing platform. Sternad’s team will recruit children from the SPARK cohort and ask them to participate in five games that test their abilities to interact with dynamic objects that require temporal and spatial prediction, interception with occlusion and dynamic interception including prediction and ‘post-diction.’ These games will be deployed online, so children can play them at home, obviating the need for laboratory visits. The fine-grained quantification of motor symptoms will be correlated with genomic data to assess relationships between motor behavior and genetics. Results from this study are expected to promote a hypothesis-driven and unifying understanding of the heterogeneous profile of ASD.


1.Sinha P. et al. Natl. Acad. Sci. USA 111, 15220-15225 (2014) PubMed
Subscribe to our newsletter and receive SFARI funding announcements and news