Theory 16/Psychology & Behaviour /Cognitive theory/3 min
Placebo Is Precision, Not Belief
Two patients believe the treatment will work. One responds. The difference may be how confident the brain is in its own prediction.
The size of a placebo response is set by the precision of the prior, not by the strength of the belief: manipulating the reliability of the cue changes the response, while changing the stated expectation at fixed reliability does not.
01The evidence
- Placebo analgesia has been reformulated as a predictive-coding phenomenon in which expectations act as priors integrated with sensory evidence, which makes precision the operative variable rather than valence.Neuron, predictive coding perspective
- Work on how meaning shapes placebo effects has extended the Bayesian account, showing that the content of the expectation matters as well as its presence, which complicates a simple strength model.Trends in Neurosciences, 2026
02The tension
Expectation-strength models predict that stronger belief produces larger responses, and the correlation is real but weak. Many high-expectation patients do not respond, and some low-expectation patients do. The unexplained variance is large enough that the model cannot be used to predict who will respond.
03The theory
In a predictive-coding system, a prior is weighted by its precision, the inverse of its variance. A belief held confidently but vaguely has low precision and is easily overridden by sensory input. A belief held with a narrow distribution, even a less extreme one, dominates the percept.
Placebo response should therefore scale with precision rather than with the mean of the expectation. This is measurable: precision is reflected in reaction-time variability, in the sharpness of conditioning, and in the reliability of the cues that established the expectation.
The theory predicts a dissociation. Increasing the reliability of the treatment cue, without changing what the patient says they expect, should increase the response. Increasing the stated expectation while keeping cue reliability fixed should not.
It also explains nocebo. A precise negative prior is more powerful than a strong but vague one, which predicts that clear, consistent warnings produce more side effects than emphatic but inconsistent ones, a testable claim about informed consent.
04What this is not
Those models treat the response as a function of how strongly the patient expects benefit. This treats it as a function of how precisely the expectation is held, and predicts that strength and precision can be dissociated experimentally.
05Predictions
- P1Increasing cue reliability at fixed stated expectation should increase placebo analgesia.
- P2Increasing stated expectation at fixed cue reliability should leave the response unchanged.
- P3Responders should show lower sensory noise and less variable priors than non-responders at matched expectation.
06What would refute it
A response that scales with expectation strength at fixed precision would restore the simpler account and leave precision as a correlate.
07What it would change
It would make placebo response predictable from a measurable cognitive parameter, and change how consent and side-effect warnings are worded.
08Evidence base
- Placebo analgesia: a predictive coding perspective — Neuronhttps://www.sciencedirect.com/science/article/pii/S0896627314001925
- Beyond Bayesian integration: how meaning shapes placebo effects — Trends in Neuroscienceshttps://www.cell.com/trends/neurosciences/fulltext/S0166-2236%2826%2900143-8