Deep learning techniques for Alzheimer's disease detection in 3D imaging: A systematic review

Area of Science:

• Neurology
• Artificial Intelligence
• Medical Imaging

Background:

• Alzheimer's disease (AD) is a progressive neurological disorder impacting cognition and memory.
• Early and accurate AD diagnosis is crucial for effective treatment and patient management.
• Deep learning (DL) offers advanced computational approaches for analyzing complex medical data.

Purpose of the Study:

• To systematically review the application of deep learning (DL) algorithms in diagnosing Alzheimer's disease (AD).
• To evaluate the current state, efficiency, and potential enhancements of DL methods using 3D imaging.
• To provide insights into improving the rapid and precise diagnosis of AD through DL.

Main Methods:

• Systematic literature search of online repositories (IEEE Xplore, Elsevier, PubMed, etc.) for studies published between 2020-2024.
• Adherence to PRISMA guidelines for systematic reviews to ensure data organization and comprehensibility.
• Analysis of 87 selected articles focusing on DL techniques for AD detection using 3D imaging data.

Main Results:

• Convolutional Neural Networks (CNNs), including 3D CNNs, are widely used for analyzing spatio-temporal features in volumetric data.
• Techniques like transfer learning, multimodal data integration, and attention mechanisms enhance DL model precision.
• The review analyzed 87 articles, with 31 focusing on theoretical models and 56 on practical implementation challenges.

Conclusions:

• DL, particularly CNNs, demonstrates significant potential for accurate Alzheimer's disease detection via 3D imaging.
• Further research and development in DL methods can enhance diagnostic speed and precision.
• This review offers a critical evaluation of DL-based AD assessment, outlining future research directions.