Multi-Modal Fusion and Longitudinal Analysis for Alzheimer's Disease Classification Using Deep Learning.

Area of Science:

• Neuroimaging
• Artificial Intelligence
• Medical Diagnostics

Background:

• Alzheimer's disease (AD) diagnosis presents significant challenges.
• Current diagnostic methods often lack comprehensive data integration.
• There is a need for advanced tools to improve AD classification and monitoring.

Purpose of the Study:

• To introduce FusionNet, a novel framework for enhanced AD classification.
• To integrate multi-modal and longitudinal imaging data for improved diagnostic accuracy.
• To enable early detection and continuous monitoring of Alzheimer's disease.

Main Methods:

• FusionNet synthesizes data from MRI, PET, and CT scans.
• Utilizes machine learning: GANs for data augmentation, lightweight neural networks, and deep metric learning.
• Combines cross-sectional and temporal data with specialized feature extraction and attention mechanisms.

Main Results:

• FusionNet demonstrates superior performance in AD classification.
• Achieved an accuracy of 94%.
• Reported precision of 92% and recall of 93%.

Conclusions:

• FusionNet shows potential as a reliable diagnostic tool for Alzheimer's disease.
• Facilitates early intervention and personalized treatment strategies.
• Offers insights into AD progression, supporting patient care and research.