This study applied a machine learning (ML) - based model CPT indices for the classification of ADHD.

Using a retrospective factorial fitting, followed by a bootstrap technique, we trained, cross-validated, and tested learning models on complete CPT performance data of 458 children aged 6 - 12 years (213 children with ADHD and 245 typically developed children).

The best performance (87% accuracy, 89% sensitivity and 84% specificity) was achieved by applying the proposed method to predict ADHD using CPT performance measures and relevant control variables (age, gender, time of day, day of the week) as predictors. This performance was 34% higher than the best-achieved benchmark approach.

This study showed that the ML detection model could classify children with ADHD with high accuracy based on CPT performance. ML model of ADHD holds the promise of enhancing, perhaps complementing, behavioral assessment and may assist in early identification, supporting diagnosis, and monitoring treatment outcomes.