ADHD Classification Through EEG-derived Centrality Measures and Machine Learning: A Neurobiological Perspective

Attention Deficit Hyperactivity Disorder (ADHD) is a common neurodevelopmental disorder characterized by symptoms of inattention, hyperactivity, and impulsivity. Traditional diagnostic methods rely heavily on subjective behavioral assessments. This research explores an objective, data-driven approach to ADHD classification using centrality measures from Electroencephalography (EEG) signals. The objective of this study is to develop an objective method for ADHD classification by analyzing EEG signals using graph-theoretical centrality measures. The aim is to enhance the accuracy and objectivity of ADHD diagnosis. The dataset consists of EEG recordings from 61 children with ADHD and 60 healthy controls, ages 7–12, during visual attention tasks. EEG signals were used to create a graph based on the correlation between electrode signals, with edges established for correlations surpassing a set threshold. Key centrality measures-degree, betweenness, and closeness centrality-were computed and aggregated into centrality profiles. These profiles served as feature vectors for training three classifiers: Support Vector Machine (SVM), Random Forest Classifier, and K-Means Clustering. Preliminary findings indicate that the SVM model effectively distinguishes between ADHD and non-ADHD cases based on the derived centrality profiles. Comparisons with the performances of Random Forest Classifier and K-Means Clustering suggest that this graph-theoretical approach to analyzing EEG data can enhance diagnostic accuracy. This study demonstrates that using centrality measures from EEG signals can provide a promising, objective tool for ADHD diagnosis. The graph-theoretical approach has the potential to improve the accuracy and objectivity of ADHD classification, offering valuable insights for clinical application.