Brain scans show two routes to placebo pain relief are not the same

A pooled analysis of brain-scan data from 409 people finds that words alone and words paired with conditioning change pain-related brain activity in overlapping but partly different ways, a distinction researchers say matters for how future placebo-based protocols are designed and tested.

An individual-participant-data meta-analysis combining brain-scan results from 409 people across 16 studies has found that placebo pain relief produced by verbal suggestion alone and by verbal suggestion combined with classical conditioning relies on overlapping but partly distinct brain circuits. The study, published in Nature Communications on July 17, 2026, was led by Tamas Spisak, Ulrike Bingel and colleagues in the Placebo Imaging Consortium, the same group behind a 2021 meta-analysis of 603 healthy participants that first assembled much of this imaging data.

Both induction methods increased activity in the dorsolateral prefrontal and inferior parietal cortices during pain, and decreased activity in the insula, putamen and primary sensory areas, the researchers reported. Adding conditioning, in which a cue is repeatedly paired with reduced pain before any verbal suggestion is given, further engaged regions tied to context representation and pain modulation, and produced greater reductions in sensory and cerebellar activity. Conditioning also strengthened the link between reported pain relief and activity in the Neurologic Pain Signature, a brain-wide marker of nociceptive processing, and combining conditioning with instructions produced larger analgesic effects than instructions alone, mediated in part by increased ventromedial prefrontal and dorsal caudate activity.

The earlier 2021 analysis by the same group had found no clear difference between induction methods when comparisons were preliminary and less statistically powered; the new pooled dataset, built specifically to compare the two techniques within the same participants, was designed to settle that question. For families and clinicians following dementia research, the finding is relevant because any future attempt to harness placebo mechanisms for symptom management, including pain or distress in dementia, would need to specify which induction route, verbal, conditioned or both, is being used, since the study indicates the underlying brain mechanisms are not interchangeable.

No dementia patients were included in this research, and no clinical protocol was tested. The authors describe a difference in average brain activity across healthy volunteers undergoing brief, experimentally controlled pain, not a treatment, a diagnosis, or a recommendation for anyone living with a dementia.

Whether conditioning-based approaches could be adapted for people who have lost the ability to understand verbal suggestion, a question with obvious relevance to later-stage dementia, is not addressed by this study and would require its own trials in that population.