Progress and trends in neurological disorders research based on deep learning.

Area of Science:

• Clinical Neuroimaging
• Neurological Disorders
• Artificial Intelligence

Background:

• Deep learning (DL) shows significant promise in clinical imaging for diagnosing and treating neurological disorders (NDs).
• Multimodal neuroimaging data analysis is a rapidly advancing area within DL research for NDs.

Purpose of the Study:

• To comprehensively review the role of DL techniques in analyzing neuroimaging data for NDs.
• To categorize and analyze various DL models for their performance in neurology.
• To identify key benchmarks, datasets, challenges, and opportunities in DL for clinical neuroimaging.

Main Methods:

• Systematic literature review of DL applications in neuroimaging for NDs.
• Categorization and critical analysis of DL models (CNNs, LSTM-CNN, GAN, VGG).
• Evaluation of DL model performance across different Neurology Diseases.

Main Results:

• DL models demonstrate potential in analyzing multimodal neuroimaging data for NDs.
• Key benchmarks and datasets for DL model training and testing were identified.
• Effectiveness of DL in real-world clinical scenarios for ND diagnosis and therapy was discussed.

Conclusions:

• DL is transforming clinical neuroimaging, offering enhanced patient care and new discoveries in neurology.
• This review provides insights into current DL applications and guides future development of efficient DL techniques for ND analysis.