Chronic pain has a high incidence, high disability, causing severe burden for people throughout the world. There are 300 million participants suffering from chronic pain in China. The number is increasing at a rate of 10-20 million per year. Pain has been the third healthy problem followed by cardiovascular disease and cancer. However, the causes of chronic pain and its prognosis are largely unknown. Treatment of chronic pain is often unsatisfactory.
Accumulating evidence has suggested that chronic pain is a complex condition associated with a combination of biological, psychological and social factors. Chronic pain is often accompanied by cognitive decline and elevated dementia risk, as well as abnormal brain morphology and function. Neuroimaging is increasingly used in the pain study to unravel the brain-pain interaction mechanisms.
However, most neuroimaging studies for chronic pain are carried out with small sample sizes, making them difficult to reproduce in large and more diverse groups. In addition, these previous prospective studies have rarely been validated in out of-sample patients and the generalizability of the findings to new patients remain largely unknown. The UK Biobank provided a large-scale longitudinal dataset with multimodal magnetic resonance imaging data. This project will explore the brain abnormalities in chronic pain with the large-scale dataset and the dynamic alterations using the longitudinal data.
Machine learning and deep learning approaches have shown quite excellent performance in risk prediction, early diagnosis and prognosis evaluation of many diseases. Building on these ideas, we aim to use machine learning or deep learning to explore the risk factors contribution to the development of chronic pain and the associated cognitive dysfunction. This project will integrate comprehensive data, such as clinical psychological measures, imaging data, behavioral factors, environmental data.
We aimed to conduct this study in 3 years by using data from the UK Biobank. The result will provide models to find the risk factors of chronic pain and the concomitant cognitive decline. We also want to use this large dataset to explore the brain abnormalities in chronic pain and their dynamic alterations with longitudinal data. We will compare the results with previous small-sample and cross-sectional data. Ultimately, the results will help to clarify the brain modulation mechanism of chronic pain in large-scale datasets. Our results will add important knowledge to the management of chronic pain by clinicians.