Validity of Diagnostic Support Model for Attention Deficit Hyperactivity Disorder: A Machine Learning Approach

Area of Science:

• Neuroscience
• Medical Informatics

Background:

• Early diagnosis of attention deficit hyperactivity disorder (ADHD) is crucial for improving patient outcomes and reducing healthcare costs.
• Developing effective screening tools is essential for timely intervention.

Purpose of the Study:

• To develop and evaluate machine learning models for the early diagnosis of ADHD.
• To compare the performance of logistic regression, Classification and Regression Tree (CART), and neural network models in ADHD screening.

Main Methods:

• Three machine learning models (logistic regression, CART, neural network) were developed for ADHD diagnosis.
• Model performance was assessed using receiver operating characteristic (ROC) analysis.
• Sensitivity and specificity were calculated for each model.

Main Results:

• The CART model achieved the highest area under the ROC curve (0.848), outperforming logistic regression (0.826) and the neural network (0.67).
• The CART model demonstrated a sensitivity of 78.8% and a specificity of 50% for ADHD diagnosis.
• Participant enrollment included 74 individuals in the ADHD group and 21 in the control group.

Conclusions:

• The CART model shows significant potential as a diagnostic support tool for ADHD.
• This machine learning approach can be extended to other neurological disorders like autism spectrum disorder, Tourette syndrome, and dementia.
• The developed model offers practical value for future neuroscience research and clinical applications.