Alzheimer's Disease Diagnosis using a Multimodal Approach with 3D MRI and PET

2026-06-18

AI Summary

This study addresses the challenge of insufficient model robustness in early Alzheimer’s disease diagnosis at the mild cognitive impairment (MCI) stage, primarily caused by heterogeneity across patients and imaging sites. To this end, the authors propose a multimodal 3D deep learning framework that integrates MRI and PET scans. The approach innovatively combines gated multimodal units, gated self-attention mechanisms, and a sparse gated mixture-of-experts (MoE) classifier to enable sample-adaptive feature fusion and expert routing, thereby enhancing both diagnostic accuracy and model interpretability. The framework achieves classification accuracies of 80.46%, 82.08%, and 95.47% on NC vs. MCI, MCI vs. AD, and NC vs. AD tasks, respectively. Ablation studies further confirm the critical contribution of the MoE component to overall performance improvement.

Abstract

Alzheimer's disease (AD) is an irreversible neurodegenerative disorder and a leading cause of death worldwide. Early diagnosis plays an important part especially at the Mild Cognitive Impairment stage, where timely intervention can help slow its progression before it advances to AD. Neuroimaging data, like Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) scans, can help detect brain changes early by providing structural and functional brain changes related to the disease. Yet, many multimodal models still fuse MRI and PET with static concatenation and apply identical computation to all subjects, which limits robustness to patient/site heterogeneity and can waste computation. To address these limitations, we present the first study of combining 3D convolutional feature extractors with three fusion strategies - concatenation, Gated Multimodal Unit (GMU), and gated self-attention - and a sparsely gated Mixture-of-Experts (MoE) classifier that performs input-adaptive routing, activating only the most informative experts per case. Finally, we utilize Grad-CAM to visualize disease-related regions, ensuring model interpretability. Experiments are performed across three binary classification tasks (NC vs. MCI, MCI vs. AD, and NC vs. AD). Results show that GMU achieves accuracies of 80.46 % (NC vs. MCI) and 95.47 % (NC vs. AD), while gated self-attention attains 82.08 % on MCI vs. AD. Ablations show that removing the MoE consistently degrades accuracy across all tasks. These findings underscore the value of input-adaptive, multimodal modeling for AD diagnosis by leveraging the complementary nature of MRI and PET.

Problem

• Alzheimer's Disease
• Early Diagnosis
• Multimodal Fusion
• MRI
• PET

Innovation

• Multimodal Fusion
• Mixture-of-Experts
• Input-Adaptive Routing
• 3D CNN
• Gated Self-Attention