AlzGraph-CL: Multi-modal Graph Neural Network with Contrastive Learning for Alzheimer’s Disease Classification

Read full text Summarise with VIRAT
Authors: Shruti VijayKumar Hegdekar, Prathibhavani P Maruthi, Venugopal Kuppanna Rajuk
Cited by: 0
References: 27

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

In recent years, multi-modal approaches have become important for accurate alzheimer's disease classification, as they integrate complementary structural and clinical information.Traditional models fail to capture subtle interactions between brain anatomy and cognitive measures, which limits their predictive performance.In this research, a multi-modal Graph Neural Network (GNN) with Contrastive Learning (CL), namely AlzGraph-CL is proposed for alzheimer's disease classification that integrates structural Magnetic Resonance Imaging (MRI) and clinical/demographic features.Initially, the data are collected from a ADNI dataset, including structural MRI scans and clinical as well as demographic attributes such as age and sex.MRI scans are encoded using a 3D Convolutional Neural Network (3D-CNN), while clinical features are processed by Multi-Layer Perceptron (MLP) to obtain modality-specific embeddings.Then, CL aligns these embeddings into a shared latent space by capturing subjectspecific pathology.Finally, graphs are constructed to model inter-subject similarities, and a GNN propagates information to enhance classification.Finally, a Softmax layer predicts diagnostic categories, like Alzheimer's disease (AD), Mild Cognitive Impairment (MCI), or Normal Control (NC), jointly optimized with cross-entropy and contrastive losses for accurate, robust and interpretable classification.For ADNI dataset, the proposed AlzGraph-CL achieved better results in terms of accuracy 97.98%, precision 98.04%, specificity 98.19%, and F1-score 97.02% than existing approaches.

Fields of Study & Subjects

Dementia and Cognitive Impairment Research Functional Brain Connectivity Studies Alzheimer's disease research and treatments

Publication details

Journal / source
International journal of intelligent engineering and systems
Volume
19
Issue
4
Pages
424–441
Publication date
23 March 2026
Language
EN