Smart Detection of ADHD in Children Using EEG Signals and AI Models
Paper ID : 1104-ISCH
Authors
Mourad Raafat *1, Mouhamed Moustafa Hassanin2, Anas Taha Zakaria3, Shrouq Metwaly Taha2, Rawan Ahmed Nour eldin2, Aya Osama Mokhtar4, Ahmed Hazem Ahmed3
1Assistant professor, mathematics department, faculty of science, Helwan university
2Mathematics Department, Faculty of Science, Helwan University
3Mathematics Department, Faculty of Science, Helwan university
4Mathematics Department, Faculty of science, Helwan university
Abstract
Attention-Deficit Hyperactivity Disorder (ADHD) is a common neurodevelopmental condition in children that affects attention, impulse control, and behavior. Early and accurate diagnosis is important to ensure proper support and treatment. In this study, we developed an intelligent system that uses brainwave data (EEG) to help identify children with ADHD more reliably.The EEG signals were carefully cleaned to remove noise and unwanted artifacts, then split into short overlapping segments for analysis. To address the imbalance in the dataset, we applied a technique called SMOTE, which helps improve the performance of machine learning models. We focused on signals from key brain regions and extracted meaningful features for classification. We tested eight machine learning (ML) models and one deep learning (DL) model, EEGNet. Among the ML models, Random Forest achieved 91% accuracy. EEGNet performed even better, reaching 97.3% accuracy. These results show that our system, especially EEGNet, can be a valuable tool to assist doctors in diagnosing ADHD more objectively and efficiently.
Keywords
ADHD, Deep Learning, AI, EEG.
Status: Abstract Accepted (Poster Presentation)