Predicting Problematic Behavior in Autism Spectrum Disorder Using Medical History and Environmental Data

Area of Science:

• Neuroscience and Artificial Intelligence
• Developmental Psychology and Behavioral Science

Background:

• Autism spectrum disorder (ASD) presents with social, communication, and behavioral challenges, affecting approximately 1 in 36 children.
• Co-occurring conditions such as sleep, immune, and gastrointestinal (GI) disorders, alongside sensory sensitivities, are common in individuals with ASD.
• Certain individuals with ASD may experience challenging behaviors, including aggression and self-injurious behavior (SIB), posing risks to themselves and others.

Purpose of the Study:

• To explore the efficacy of artificial intelligence (AI) models in predicting behavioral episodes in individuals with ASD.
• To identify key co-occurring conditions and environmental factors that may serve as predictors for these behaviors.

Main Methods:

• Utilized AI models to analyze historical data from 80 individuals in a residential setting.
• Data included co-occurring conditions, environmental factors, and recorded behavioral episodes.
• Models were trained to predict occurrences of both behavioral and non-behavioral states.

Main Results:

• AI models achieved prediction accuracies as high as 90% for certain individuals.
• Environmental factors and gastrointestinal issues emerged as significant predictors across the study population.
• Demonstrated the potential of AI in forecasting behavioral episodes based on associated health and environmental data.

Conclusions:

• AI models show promise in accurately predicting behavioral episodes in individuals with ASD.
• Environmental and GI factors are important considerations in understanding and predicting these behaviors.
• Accurate behavioral prediction holds potential for enhancing the quality of life for individuals with ASD.