EEG Signatures in Alzheimer and FTD's Diseases Using GNN Deep Learning Technique

Electroencephalography (EEG) is a noninvasive neuroimaging technique that records the electrical activity generated by the brain's neuronal networks. The main objective of this study aims to evaluate the efficiency of EEG signals as a diagnostic tool for Alzheimer's disease (AD) and Frontotemporal Dementia (FTD). Specifically, we seek to determine whether EEG signatures can be effectively utilized to differentiate AD patients from healthy controls and FTD patients combined. To achieve this, we will employ Graph Neural Networks (GNNs) within a Deep Learning (DL) framework and compare their performance to other machine learning (ML) techniques. We aim to develop a robust machine learning model capable of accurately differentiating between AD patients and healthy controls (HC) and FTD patients utilizing a dataset comprising more than 80 subjects. Our results demonstrate that the network-based approach, specifically the GNN model, outperformed traditional machine learning algorithms in classifying EEG data from patients with Frontotemporal Dementia (FTD) and healthy control (HC).