Artificial intelligence and machine learning in pain research: a data scientometric analysis.

Area of Science:

• Pain research
• Health informatics
• Computational science

Background:

• Growing healthcare data necessitates advanced analytical methods for pain therapy and research.
• Classical data analysis approaches face challenges with the increasing volume and complexity of health data.
• Artificial intelligence (AI) and machine learning (ML) offer novel solutions for pain research.

Purpose of the Study:

• To review and evaluate the current literature on AI and ML applications in pain research.
• To identify common ML methods, pain conditions, and publication trends.
• To analyze technical details including sample sizes and data requirements for ML in pain studies.

Main Methods:

• Systematic literature search of AI and ML in pain research.
• Manual curation and evaluation of identified publications.
• Analysis of publication origin, ML algorithms, pain conditions, and sample sizes.

Main Results:

• AI and ML are featured in 475 publications from 18 countries, with 79% published since 2019.
• Commonly studied pain conditions include low back pain, musculoskeletal disorders, osteoarthritis, neuropathic pain, and inflammatory pain.
• Random forests and support vector machines are prevalent ML algorithms; deep learning is used for image-based diagnoses, requiring larger datasets.

Conclusions:

• AI and ML are rapidly advancing pain research by enabling complex data processing.
• While powerful, these methods often require substantial data, particularly deep learning.
• Considerations for big data needs and the interpretability of "black-box" decisions are crucial for future applications.