Unpredictable and potentially dangerous aggressive behavior by youth with autism spectrum disorder (ASD) isolates them from foundational educational, social and familial activities, thereby markedly exacerbating morbidity and costs associated with the condition. As many as two-thirds of youth with ASD display aggression, which is one of the primary reasons they use behavioral healthcare services.

Over the past five years, the Autism Inpatient Collection (AIC) has collected phenotypic and biological data from youth with ASD who were hospitalized in six child psychiatry units. The AIC is a unique and expanding collection of an ASD group that has been underrepresented in large extant data collections; the cohort to date includes over 1,400 individuals with confirmed ASD who are minimally verbal (52.4 percent), have intellectual disability (42.6 percent) and/or engage in aggression or self-injurious behavior (more than 90 percent)¹. AIC phenotypic data from 527 individuals is currently available to researchers via SFARI Base. Blood samples (or saliva when blood cannot be obtained) are also being collected from the proband and biological parents. Whole-exome sequencing of DNA extracted from these samples is currently being performed, and deidentified genetic data will be made available to researchers as soon as this is completed.

In the work that Matthew Siegel and his colleagues have performed to date, they have come to appreciate the seminal role that unpredictability plays in magnifying the severe impact aggression has on the lives of individuals with ASD, their families, and the professionals who support them. Several factors make it difficult for youth with ASD to self-report increasing distress prior to aggression; 30–40 percent are minimally verbal, and the fluently verbal often have poor emotional insight and self-awareness. The resultant unpredictability makes aggression in ASD particularly dangerous and is a barrier to individuals accessing the community, therapy services, medical providers and educational placements.

In addition to expanding the cohort to ultimately include data from over 1,600 youth with ASD, the current project plans to utilize Siegel and his colleagues’ unique access and experience with the inpatient ASD population to safely study aggression in situ. Specifically, the team plans to test their abilities to predict the imminent onset of aggressive behavior. This will involve the use of wearable biosensors to detect objective physiological signals and the application of advanced machine-learning algorithms to develop predictive models, as well as the evaluation of whether incorporating information on verbal ability and emotion regulation capacity improves predictions. Finally, the researchers plan to utilize the highest performing classifier to prospectively pilot test the accuracy, reliability and impact of real-time risk prediction alerts for aggression.

In summary, completion of this project will not only provide researchers with an expanded collection of phenotypic and genetic data from an underrepresented segment of the ASD population, but will also lay the groundwork for the future development of interventions that help to prevent aggressive behavior before it develops and reduce its impact when it occurs.