Background & Aims
Placebo hypoalgesia and nocebo hyperalgesia are striking examples of the impact of expectations on pain perception [2]. These effects have recently been conceptualised as Bayesian inferential processes, which propose that pain arises from the integration of sensory data (e.g., noxious input) and prior predictions (e.g., expectations) with consideration to precision levels [2]. However, direct empirical evidence in support of Bayesian accounts of placebo hypoalgesia and nocebo hyperalgesia is lacking [4]. Here, we explore for the first time whether these phenomena can be unified within the same Bayesian framework by testing whether pain is predicted not only by expectations but also by their level of precision, both measured at the metacognitive level. While expectation precision has been measured at the metacognitive level in cue-based paradigms [1,3], this is the first time this has been implemented in placebo/nocebo research.
Methods
Sixty healthy volunteers were administered a sham treatment and received two blocks of noxious stimulations, each comprising eight electrical stimuli. They were randomised into three groups, either receiving hypoalgesic (Placebo), hyperalgesic (Nocebo) or neutral (Control) verbal suggestions associated with the sham treatment. Trial-by-trial expectations magnitude, their precision, and perceived pain were measured. Skin Conductance Response (SCR) was also recorded as a marker of autonomic arousal. Bayesian Linear Mixed Models (BLMM) analyses were used to test for the successful induction of placebo and nocebo responses and to investigate whether our data is well described from a Bayesian inferential perspective according to which pain is predicted not only by expectations but also by their level of precision.
Results
BLMM analyses revealed that verbal suggestions effectively elicited placebo hypoalgesia and nocebo hyperalgesia. On this premise, the main BLMM analyses provided support for a Bayesian conceptualisation of these phenomena. First, by revealing that pain is predicted by the interaction between expectations and their precision. Specifically, with a 95% confidence level, for each unit increase in expectation precision, the relationship between expectations and pain changes between 0.01 and 0.06. Secondly, there was a strong and consistent main effect of expectation precision on DeltaPain, representing the difference between expected and perceived pain. With 95% confidence, for each unit increase in expectation precision, DeltaPain changes between -0.34 and -0.14. This indicates that the discrepancy between expected and perceived pain decreases with the increase in expectation precision. A main effect of time was reported for SCR, suggesting a habituation effect for this response.
Conclusions
Our data provide the first direct evidence indicating that both placebo hypoalgesia and nocebo hyperalgesia can be unified within the same Bayesian framework in which not only expectations but also their level of precision are key determinants of pain perception, also when measured at the metacognitive level. The identification of expectation precision as a predictor of pain suggests a novel target for potential pain treatment interventions. Further clinical studies are warranted to extend this model to chronic pain patients.
References
[1] Brown CA, Seymour B, El-Deredy W, Jones AKP. Confidence in beliefs about pain predicts expectancy effects on pain perception and anticipatory processing in right anterior insula. Pain 2008;139:324–332. doi:10.1016/j.pain.2008.04.028.
[2] Büchel C, Geuter S, Sprenger C, Eippert F. Placebo Analgesia: A Predictive Coding Perspective. Neuron 2014;81:1223–1239. doi:10.1016/j.neuron.2014.02.042.
[3] Mancini F, Zhang S, Seymour B. Learning the statistics of pain?: computational and neural mechanisms. bioRxiv 2021.
[4] Milde C, Brinskelle LS, Glombiewski JA. Does Active Inference Provide a Comprehensive Theory of Placebo Analgesia? Biol Psychiatry Cogn Neurosci Neuroimaging 2023. doi:10.1016/j.bpsc.2023.08.007.