ADHD diagnosis using EEG Signal Patterns Analysis

Abstract

Neurodevelopmental disorders such as Attention Deficit/Hyperactivity Disorder (ADHD) can lead to lifelong negative impacts if not identified and managed early. Affecting approximately 5–7% of children globally, early detection is essential for timely and effective intervention. Traditional diagnostic methods can be subjective and time-consuming; therefore, scientific and data-driven approaches offer more precise alternatives. Among the available techniques, EEG signal analysis stands out for its non-invasive nature, cost-effectiveness, and ability to reflect brain activity patterns associated with ADHD. To address the issue of ADHD recognition, this study used EEG signal analysis in conjunction with machine learning techniques. Due to restrictions on collecting new EEG data, existing data records were used. Raw EEG signals are processed to extract related features that indicate ADHD- related neuronal activity. Several classification algorithms have been explored, including Random Forest, Support Vector Machines (SVMs), and gradient reinforcements. These models were trained on extracted characteristics, and hyperparameters were tuned with cross-validation to optimize performance. The performance of each algorithm was evaluated using standard data science metrics, including classification accuracy, precision, recall, and F1-score. Among the tested models, the Random Forest classifier demonstrated the highest classification accuracy of 79%, outperforming both SVM and Gradient Boosting classifiers. These results underscore the effectiveness of EEG-based machine learning methods in supporting the early and reliable diagnosis of ADHD.