A CNN-LSTM Network for Classification of Attention Deficit Hyperactivity Disorder from EEG Data

Attention deficit hyperactivity disorder (ADHD) is one of the most common neuropsychiatric disorders in children, with core symptoms of inattention, impulsivity, and hyperactivity. As a result, early detection and treatment are critical. At this point, ADHD can be diagnosed using an electroencephalogram (EEG). This study proposes a model for identifying the EEG of children with ADHD. Our proposed model is made up of three parts: temporal convolutional blocks, spatial convolutional blocks, and Long Short Term Memory (LSTM), and because EEG contains more temporal information than other data, we build a model that can extract more temporal information. This study included 24 subjects, including 12 children with ADHD and 12 normal children. The experimental results show that our proposed model outperforms Support Vector Machine (SVM) and EEGNet, with our model achieving correct classification rates of 92.29%, 92.76%, and 90.91% on the three backs, respectively, and 94.25% on the full data, demonstrating the model's feasibility and applicability to the ADHD recognition problem.