Detecting ADHD Children using Attention Continuity as Nonlinear Feature of EEG

Attention Deficit Hyperactivity Disorder (ADHD) is the current description of the most prevalent psychiatric disorder of childhood. The essential feature is the developmentally inappropriate degree of inattention, impulsiveness and hyperactivity. Manifestations of ADHD usually appear in most situations, including home, school, work, sporting and social settings.
Since the essential feature of ADHD is inattention manner also nonlinear features of EEG may be equivalent to the attention we investigate nonlinear features of the EEG. We evaluated 29 children with AD/HD who were diagnosed by DSM-IV criteria and 20 age-sex matched controls. During recording EEG we showed images to children and asked them to concentrate on those images and number them. Using this method we stimulated visual attention of children.
In this study, we use an MLP neural network as a classifier. By investigating these nonlinear features, we obtained a classification with 96.7% accuracy, using frontal lobe electrodes as the best result.
Results showed a significant difference between the accuracy of the frontal region, and other regions. This result can confirm the defect in the anterior segment of the brain of ADHD children.