Comparing scoring algorithms for NIH Toolbox executive function tasks in children with and without ADHD

Executive function (EF) supports goal-directed behavior, yet little is known about how different scoring algorithms influence the measurement of executive function performance in pediatric Attention-Deficit/Hyperactivity Disorder (ADHD) remains unclear. We compared four algorithms: Accuracy, Speed, Rate Corrected Score (RCS), and Linear Integrated Speed-Accuracy Score (LISAS) in 51 children with ADHD (23.5% female; M_age = 12.9, SD_age = 2.8) and 102 matched controls (23.5% female; M_age = 12.8, SD_age = 3.1) using the NIH Toolbox Dimensional Change Card Sort (DCCS) and Flanker Inhibitory Control and Attention (Flanker) tests. Results showed that while RCS and Speed demonstrated high internal consistency across tasks, Accuracy and LISAS indices were less reliable, particularly during cognitive flexibility trials in the ADHD group. Children with ADHD exhibited lower Accuracy and higher switching costs on the DCCS, whereas no group differences were observed on the Flanker task. Group differences in cognitive flexibility were detected by accuracy-based metrics but remained obscured by scores based on speed alone. Findings suggest that integrated methods, particularly accuracy-dominant metrics like RCS, may offer reliable, clinically sensitive alternatives to single-metric scores. Their utility depends on task demands, emphasizing that relying solely on response speed may underestimate cognitive flexibility deficits in ADHD.