Advanced EEG-Based Analysis for ADHD Identification Utilizing Convmixer and Continuous Wavelet Transform

Abstract

Children with ADHD may experience challenges such as attention deficits, behavioral problems, educational problems, and low self-confidence. This study summarizes research aiming to evaluate the diagnosis of attention deficit hyperactivity disorder (ADHD) with electroencephalography (EEG) signals. The research used EEG data from 30 children diagnosed with ADHD and 30 healthy control groups. EEG data was first processed for noise reduction purposes and then classified using deep learning models such as ConvMixer, ResNet50, and ResNet18. The findings show that ConvMixer demonstrates high accuracy in classification. while requiring low computational resources. Additionally, the effects of different channels on the usability of EEG signals in the diagnosis of ADHD were examined, and the T8 channel was found to be particularly effective. In conclusion, the study emphasizes the effectiveness of lightweight models and underscores the significance of specific EEG channels in diagnosing ADHD using EEG signals.