EEG-Enabled Neurodegenerative Disease Prediction Using Attention-Enhanced Neural Architectures

Abstract

Neurodegenerative diseases such as Alzheimer’s disease, Parkinson’s disease, and dementia are progressive disorders that severely impact cognitive and motor functions. Early diagnosis is critical for improving patient outcomes and enabling timely therapeutic interventions. Electroencephalography (EEG) provides a non-invasive and cost-effective method for monitoring brain activity, but EEG signals are highly nonlinear, noisy, and complex, making automated disease prediction challenging. This study proposes an EEG-Enabled Neurodegenerative Disease Prediction framework using Attention-Enhanced Neural Architectures (AENA). The proposed model integrates deep neural networks with attention mechanisms to capture both spatial and temporal dependencies in EEG signals while enhancing feature interpretability. The attention module dynamically focuses on the most relevant neural patterns associated with disease progression, improving classification accuracy and robustness. The model is evaluated using standard EEG datasets, and performance is measured using accuracy, precision, recall, F1-score, and ROC-AUC. Experimental results demonstrate that the proposed attention-enhanced architecture significantly outperforms traditional machine learning and baseline deep learning models. The framework is suitable for real-time neurodiagnostic systems and clinical decision-support applications.