SHAKES

Identification of Social Problem

SHAKES (Screening for High functioning Autism at KindErgarten age) is a digital tool development project that aims to create a serious game system. Its objective is to create an interactive digital screening system (a prototype) to more effectively identify cases of autism with relatively mild symptoms and good language and cognitive abilities in preschool age, and (2) demonstrate the feasibility of the concept/prototype through empirical research. As a significant development for the project implementation, during the indicated period, we conducted and successfully completed exploratory statistical analyses on our emotional facial expression database. In this process, we performed comparative statistical analyses on the emotional facial expression data collected from children diagnosed with autism and typically developing children, for each scene of the diagnostic serious game, to be validated.

Result of the Development/Research

A screening algorithm has been developed, which measures three data points of the preschool child during the targeted usage scenario: their emotional facial expressions through a webcam, their open-ended responses, and their eye movements and visual exploration behavior. The system feeds these collected data into a data processing pipeline, where an intelligent decision-making component performs risk assessment and categorizes the child into one of the three categories of autism severity (high, low, not present).

Presentation of the Method

These analyses covered 56 scenes and focused on the intensity of 8 different emotional facial expressions. We compared the two diagnostic groups and conducted a total of 448 intergroup tests. From the 56 scenes, we found statistically significant differences in at least one emotional dimension between the group of children with autism and the typically developing group in 13 cases. Additionally, we found trend-level differences along an emotional dimension for 24 scenes between the two groups.

Contact