A Comprehensive Review on - Machine Learning in Bio-Signal Analysis for Chronic Pain Research

This comprehensive review provides an overview of the application of machine Learning (ML) in bio-signal analysis for chronic pain research. The review addresses various machine learning and deep learning algorithms for bio-signal analysis in the field. This review also discusses the challenges and limitations of using bio-signal-based machine learning analysis in chronic pain research and the potential benefits of this approach. This review focuses on bringing out the relevant publications made in the last decade, which provides greater insights for the readers on this field of research. A comprehensive literature review was conducted focusing on the relevant publications made in the last decade, starting from 2012 to 2023. The articles were identified by scrutinizing the publications through a screening process using the search phrases in publications’ titles and abstracts such as "chronic pain", "electromyography", "machine learning", "supervised learning", "unsupervised learning", and "pain". As conventional methods possess a significantly lower capability for obtaining high-accuracy chronic pain detection, this review encompasses the most recent research that utilizes Machine Learning models. Connected Papers software is used to provide the visualization of the map among the publications in the field, between the years 2015 and 2022. This review discusses the sample size, the type of sensors used, and the type of Machine Learning models used along with results obtained in the most relevant articles. This review also provides future research directions upon which the researchers could focus to obtain a promising methodology for the automatic detection of chronic pain. Machine Learning algorithms pave the way for more advancements in accurate diagnosis for the betterment of chronic pain patients’ quality of life by providing individual treatment plans by their medical professionals.