EEG Functional Connectivity Feature-Based Diagnosis of ADHD Using Deep Learning on Raspberry Pi

Attention Deficit Hyperactivity Disorder (ADHD) is marked by hyperactivity, impetuous behavior, and inattention. It poses significant challenges for many children, and emphasizes the crucial need for early identification and treatment to provide timely support and minimize long-term impacts. This study proposes electroencephalography (EEG) as a potential screening tool for automated ADHD diagnosis. The proposed framework utilizing a publicly accessible EEG dataset having 61 ADHD case and 60 control subjects, functional connectivity (FC) features of EEG signals, deep learning (DL) model and Raspberry Pi hardware. The methodology includes preprocessing EEG signals to extract and clean brain signals, generating dynamic functional connectivity correlation features, and creating Pearson’s Correlation (PC) matrix images. These images are analyzed using pre-trained DL models, achieving maximum testing accuracy of 96.12% with GoogLeNet model. The results indicate that the proposed methodology outperforms many state-of-the-art techniques, offering a portable and precise solution for ADHD diagnosis through brain connectivity analysis.