Isolating brain mechanisms of expectancy effects on pain: Cue-based stimulus expectancies versus placebo-based treatment expectancies

Clinical trials, laboratory experiments, and neuroimaging studies provide converging evidence that pain is highly sensitive to expectations, whether based on the psychosocial context surrounding treatment (e.g., placebo analgesia) or transient cues that provide information about painful events (e.g., pain-predictive cues). We asked whether placebo analgesia and pain-predictive cues modulate pain through the same mechanisms or dissociable brain pathways. Forty healthy volunteers rated pain in response to noxious heat during functional magnetic resonance imaging (fMRI) scanning. We crossed pain-predictive cues, which induce expectations for high or low pain on a trial-by-trial basis, with administration of an inert placebo cream or a control cream. Behavioral analyses revealed a significant interaction, such that predictive cues had weaker effects on pain during placebo blocks than control blocks. This interaction was accompanied by interactions in the insula, pons, and other brain regions. We also observed distinct neural substrates when we compared pure cue effects with pure placebo effects, and only predictive cues modulated responses in nociceptive regions and the Neurologic Pain Signature (NPS; (Wager et al., 2013). The only regions that were influenced similarly by both types of expectations were the rostral anterior cingulate and dorsomedial prefrontal cortex. These results indicate that cue-based expectations about stimulus intensity and placebo-based expectations about treatment outcomes are distinct, and that pain researchers should differentiate between sources of expectations. Furthermore, cue-based expectations were associated with more consistent effects than treatment-based expectations, suggesting that clinicians should be particularly mindful of how they present information about impending pain.