An Objective Assessment of Adult Attention Deficit/Hyperactivity Disorder Based on Surface Electromyography

Objective: : This exploratory study aimed to develop an objective and efficient diagnostic assessment driven by surface electromyography (sEMG) for adult attention-deficit/hyperactivity disorder (ADHD) and evaluate its reliability and validity compared to the Conners Adult ADHD Diagnostic Interview (CAADI). Methods: : A case-control study was conducted with 49 adults with ADHD and 54 healthy controls. Participants performed eight motor tasks, including fine motor skills and body coordination assessments, while sEMG signals were recorded from their forearms. Machine learning models (K-Nearest Neighbor, Support Vector Machine, Decision Tree, and Ensemble Learning-AdaBoost) were trained to classify ADHD and control participants, and diagnostic performance was compared to CAADI using sensitivity, specificity, receiver operating characteristic (ROC) curves, and area under the curve (AUC). Results: : The Ensemble Learning-AdaBoost model achieved 94.17% accuracy, with 93.88% sensitivity and 96.30% specificity. No statistically significant difference was observed between the sEMG-driven assessment and CAADI (sensitivity: χ2 = 0.211, p > 0.05; specificity: χ2 = 0.706, p > 0.05). The sEMG-driven assessment showed a slightly higher AUC (0.951) compared to CAADI (0.932), though not statistically significant (p > 0.05). Conclusion: : The sEMG-driven diagnostic assessment demonstrated comparable performance to CAADI in identifying adult ADHD. This objective, efficient method may serve as a useful auxiliary tool for ADHD diagnosis in clinical settings.