Relationship between quantitative digital behavioral features and clinical profiles in young autistic children

Abstract

This study explored whether behaviors measured by a computer are related to clinician-administered and caregiver-reported ratings of behavior. An iOS application (app) called SenseToKnow was created to elicit a wide range of autismrelated behaviors using strategically designed movies and games. The app recorded videos of the child while they watched the movies and their touch data and computer vision analysis measured behaviors including the child's attention to social stimuli, such as people and speech, and fine motor control. We found that several behaviors digitally measured via the app are related to commonly used and standardized caregiver-report and clinician-administered measures of social, language, and adaptive behaviors. These results suggest that the digital app may be of clinical value in identifying specific clinical profiles associated with autism. Future research is needed to determine the usefulness of the app as an autism screening tool and for monitoring changes in behavior over time.

Early detection of autism using digital behavioral phenotyping

“…Such information could be used prescriptively to identify areas in which behavioral intervention should be targeted. This approach is supported by a recent study that included some participants in the present sample that examined the concurrent validity of the individual digital phenotypes generated by the app and reported significant correlations between specific digital phenotypes and several independent, standardized measures of autism-related behaviors, as well as social, language, cognitive and motor abilities 29 . Notably, the app quantifies autism signs related to social attention, facial expressions, response to language cues and motor skills, but does not capture behaviors in the restricted and repetitive behavior domain.…”

“…Inclusion criteria were age 16–38 months, not ill and caregiver’s language was English or Spanish. Exclusion criteria were sensory or motor impairment that precluded sitting or viewing the app, unavailable clinical data and child too upset at their well-child visit 29 . Table 1 describes sample demographic and clinical characteristics.…”

“…Children were administered the M-CHAT-R/F 2 , a parent survey querying different autism signs. Children with a final M-CHAT-R/F score of >2 or whose parents and/or provider expressed any developmental concern were provided a gold standard autism diagnostic evaluation based on the Autism Diagnostic Observation Schedule-Second Edition (ADOS-2) 31 , a checklist of ASD diagnostic criteria based on the American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), and Mullen Scales of Early Learning 32 , which was conducted by a licensed, research-reliable psychologist who was naïve with respect to app results 29 . Mean length of time between app screening and evaluation was 3.5 months, which is a similar or shorter duration compared to real-world settings.…”

“…App completion took approximately 10 min. English and Spanish versions were shown 29 . The stimuli (brief movies) and game used in the app are illustrated in Fig.…”

“…CVA was used for the identification and recognition of the child’s face and the estimation of the frame-wise facial landmarks, head pose and gaze 19 . Several CVA-based and touch-based behavioral variables were computed, described next 29 .…”