A replicable and generalizable neuroimaging-based indicator of pain sensitivity across individuals

Abstract

Developing neural indicators of pain sensitivity is crucial for revealing the neural basis of individual differences in pain and advancing individualized pain treatment. To identify reliable neural indicators of pain sensitivity, we leveraged six large and diverse functional magnetic resonance imaging (fMRI) datasets (total N=1046). We found replicable and generalizable correlations between nociceptive-evoked fMRI responses and pain sensitivity for laser heat, contact heat, and mechanical pains. These fMRI responses correlated more strongly with pain sensitivity than with tactile, auditory, and visual sensitivity. Moreover, we developed a machine learning model that accurately predicted not only pain sensitivity but also pain reduction from different interventions in healthy individuals. Notably, these findings were influenced considerably by sample sizes, requiring >200 for univariate correlation analysis and >150 for multivariate machine learning modelling. Altogether, we demonstrate the validity of decoding pain sensitivity from fMRI responses, thus facilitating interpretations of subjective pain reports and promoting more mechanistically informed investigation of pain physiology.