Mutual information analysis of EEG of children with attention-deficit/hyperactivity disorder

Attention-deficit/hyperactivity disorder (ADHD) is one of the most prevalent neurobehavioral disorders in children and electroencephalography (EEG) is one of most common used neuroimaging techniques as the most accessible and informative method. EEG can be of great help in ADHD studies. Considering the brain as the most complicated information processing system, information theories can be applied on EEG analysis. In this study, a total of 50 children (9 girls, mean age: 10.44±0.75) with ADHD and 58 age, gender and handedness matched normal children were recruited. This study aimed to (a) characterize the critical differences between ADHD children and a control group in connectivity patterns by applying mutual information (MI) analysis and (b) examine the classification accuracy that can be achieved with EEG MI measures. Statistical and classification results revealed that that in most regions MIs of ADHD children were higher than control group asymmetrically but connections with lower MIs in right frontal and left occipital were also of great importance. Those findings may be related to underlying neuropsychological deviation of ADHD in both coupling and isolation between different brain modules. With 28 selected MI features, the classifier achieved accuracy of 85.7%. Combining with appropriate feature selection algorithm, MI might be an adequate biomarker candidate for assisting diagnoses of children with ADHD.