Abstract
Decoding subjective pain perception from functional magnetic resonance imaging (fMRI) data using machine learning technique is gaining a growing interest. Despite the well-documented individual differences in pain experience and brain responses, it still remains unclear how and to what extent these individual differences affect the performance of between-individual fMRI-based pain prediction. The present study is aimed to examine the relationship between individual differences in pain prediction models and between-individual prediction error, and, further, to identify brain regions that contribute to between-individual prediction error. To this end, we collected and analyzed fMRI data and pain ratings in a laser-evoked pain experiment. By correlating different types of individual difference metrics with between-individual prediction error, we are able to quantify the influence of these individual differences on prediction performance and reveal a set of brain regions whose activities are related to prediction error. Interestingly, we found that the precuneus, which does not have predictive capability to pain, could also affect the prediction error. This study elucidates the influence of interindividual variability in pain on the between-individual prediction performance, and the results will be useful for the design of more accurate and robust fMRI-based pain prediction models.
Highlights
IntroductionSelf-report is the golden standard in clinical applications
Pain is a subjective unpleasant experience (Loeser and Treede, 2008)
It is interesting to note that, the precuneus was not predictive of subjective pain perception. This implies that the precuneus may modulate pain perception in an indirect manner, which will be discussed. It is well-known that the accuracy of between-individual prediction is usually lower because of significant individual differences in the subjective pain perception and neural responses, but it remains unclear how and to what extent these individual differences determine the between-individual prediction error
Summary
Self-report is the golden standard in clinical applications. Studies have found that physiological signatures of pain could be used to develop new pain assessment tools that complement self-report (Wager et al, 2013; Woo et al, 2017; Reddan and Wager, 2018). Identifying objective physiological signatures of pain is highly desired in clinical practice and basic research. Decoding an individual’s subjective pain perception from BOLD responses in the “pain matrix” is considered to be a potential and promising pain assessment technique (Marquand et al, 2010; Brown et al, 2011; Brodersen et al, 2012; Schulz et al, 2012)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
