That a sham treatment can relieve pain is well established: when the brain expects relief, it releases opioids and messengers such as dopamine, and the effect is measurable. Less clear was whether the two routes clinicians use to trigger it do the same thing inside the brain. A new analysis suggests they only partly do.
Two techniques are used in practice. In the first, the clinician builds an expectation with words — describing a drug's benefits in detail, for instance. In the second, the patient is conditioned: during a learning phase, painful stimuli delivered under a sham treatment — a cream, an inert preparation — are secretly weakened, sometimes with real painkillers added, until the association holds.
"Whether verbal suggestion and conditioning rest on the same neural pathways has so far been unclear," writes a team led by Tamas Spisak of the University of Duisburg-Essen. The reason is methodological: most studies used only one of the two techniques, which ruled out direct comparison.
A shared core, plus an extra route
Spisak and colleagues from the Placebo Imaging Consortium therefore pooled individual participant data from 16 placebo neuroimaging studies — 409 people whose brain activity had been recorded both with and without sham treatment. In five studies verbal instruction acted alone; in eleven, conditioning was added. The work appeared in Nature Communications in July 2026.
Both techniques share a core: during pain, activity rises in the dorsolateral prefrontal and inferior parietal cortices while falling in the insula, the putamen and primary sensory areas. Conditioning then adds something on top. It draws more heavily on regions that represent context and modulate pain, and it pushes down activity in pain-processing areas more firmly. Measured against the Neurologic Pain Signature — an established pattern for the processing of painful stimuli — it strengthens the link between the relief people report and dampened nociceptive activity.
In practical terms: combining explanation with conditioning produces the strongest relief, and it does so partly through different circuits than talking alone. That is more than an academic nicety. Anyone designing pain treatment, or reading a clinical trial in which the placebo arm improves substantially, has been working with the assumption of a single unified effect. The data argue that there are two levers — and that they can be combined deliberately, without adding any active drug.