A Machine-Learning Model Supporting Return-to-Play Decisions with Children and Adolescents with Suspected Concussions

Validated a machine learning (ML) model for assessing child/adolescent concussions against a pediatric neurologist’s assessment. Using 113 participants, the ML model employed 90 training cases to account for 95.5% variance using predictors from standardized assessments and BKG sensors, perfectly matching the pediatric neurologist’s decisions on the remaining 23 cases.