EEG Correlates of the Influence of Somatosensory Input, Expectations and Trait-Like Bias on Pain Perception

Background: The weighting of somatosensory input and pain expectation during pain perception is promising for pain phenotyping, with good test–retest reliability. Yet, their concurrent validity with neural and psychological variables requires further investigation. Objectives: In this cross-sectional study, we investigated the concurrent validity of these weights with EEG source correlates of the somatosensory and expectation components during pain processing. Methods: Participants completed a cued pain paradigm, with EEG recorded during pain expectation and perception. We used Bayesian inference to estimate the participant-specific weighting of somatosensory input, expectations and trait-like bias, and identified sources of brain activity at different stages of the cued pain task (early anticipation, late anticipation and post-stimulation). We correlated the estimated weights with EEG source activity across individuals. Results: As hypothesised, the weight placed on somatosensory input correlated with source activity in areas related to attention (middle frontal gyrus) and sensory processing (postcentral gyrus) during late anticipation. The expectation weight positively correlated with activity in areas related to attention (middle frontal gyrus) and semantic processing (medial temporal gyrus). We found no significant correlations between any of the weights and analgesic or hyperalgesic psychological variables (mindfulness, pain catastrophising and attachment). Conclusion: Our findings support the concurrent validity of sensory and expectation weights with related EEG source activity in pain perception, reinforcing their utility in pain phenotyping and paving the way for more personalised pain management. Significance: Our findings support the concurrent validity of sensory and expectation weights extracted through a Bayesian model. This finding supports the use of these weights for pain phenotyping.