Predicting Traumatic Brain Injury Through Behavioral Pattern Analysis in Youth with Autism

Abstract

This practicum, conducted at HANDS in Autism® in collaboration with the Indiana NeuroDiagnostic Institute (NDI), explored the relationship between patient behavioral profiles and the presence of Traumatic Brain Injury (TBI) in individuals with Autism Spectrum Disorder (ASD). Using a combination of pre-admission data from Cerner and post-discharge data from REDCap, a comprehensive dataset of 58 patients was coded and analyzed to identify behavioral patterns. Prediction models, including Logistic Regression and Random Forest, were developed in Python to assess the likelihood of TBI based on specific behavior indicators. Results revealed statistically significant correlations between certain behavioral patterns and the presence of TBI. This work supports the potential for predictive modeling to improve early identification and intervention strategies for patients with ASD and co-occurring neurological conditions.