Decoding the perception of pain from fMRI using multivariate pattern analysis

► Subjects received a series of near-threshold pain stimuli while undergoing fMRI. ► We studied multivariate patterns of brain activity underlying pain perception. ► PAG, VLPFC, DLPFC, anterior insula, and OFC were highly predictive of pain. ► Yet the highest accuracies were afforded by combinations of these regions. ► This suggests that pain is encoded in a distributed network.