A deep learning approach with graph neural network for unveiing ad pathology through MRI bio markers

Alzheimer's disease is a progressive neurodegenerative disorder that significantly impacts cognitive function, making early and accurate diagnosis crucial. In this study, we propose an Alzheimer's Disease Graph Neural Network (AD-GNN) to classify MRI brain images into five categories: Alzheimer's Disease (AD), Cognitive Normal (CN), Early Mild Cognitive Impairment (EMCI), Late Mild Cognitive Impairment (LMCI), and Mild Cognitive Impairment (MCI). The model transforms MRI data into a graph structure, where each pixel represents a node, and edges between neighboring nodes capture local anatomical connectivity. A key advantage of graph-based learning is its ability to capture the underlying structural and relational information within brain regions. By treating the brain as a graph, the model can analyze complex spatial relationships, making it particularly effective in recognizing subtle changes in brain connectivity that are critical for diagnosing different stages of Alzheimer's disease, enhancing the prediction accuracy. Experimental results demonstrate that AD-GNN effectively distinguishes between different stages of Alzheimer's and cognitive states. Overall, this research offers a promising solution for earlier and more accurate AD detection, with potential benefits for patient care and research in neurodegenerative diseases. Keywords: Alzheimer's Disease, GNN, CN, EMCI, LMCI, MCI