The placebo effect has a reputation problem. Tell someone their improvement might be “just placebo” and you have, in three words, dismissed their experience as imaginary. But the neuroscience of the last three decades tells a different story. The placebo effect is not a failure of perception or a trick of gullibility. It is a measurable neurobiological event involving endogenous opioidsNatural painkillers produced by the brain and body, such as endorphins, that are released in response to expectation of treatment and block pain signals., dopamine release, and changes in brain activity visible on functional MRIA neuroimaging technique that measures brain activity by tracking blood flow patterns, allowing visualization of which brain regions are active during specific tasks or mental states.. And in one of the more quietly strange findings in modern medicine, it works even when patients know they are taking a sugar pill.
One of our editors wanted to know what the science actually shows. Here is what it shows: the placebo effect is real, it is getting stronger in US clinical trials, and the pharmaceutical industry has spent decades trying to eliminate something it cannot fully explain.
What the Placebo Effect Actually Is (and Is Not)
First, a clarification that matters. The placebo effect is not “getting better because you think you will.” That framing is both too simple and subtly wrong. More precisely, it is a set of neurobiological responses triggered by the context of a therapeutic encounter: the ritual of taking a pill, the authority of a clinician, the expectation of relief, and the conditioning built from prior treatment experiences.
This distinction is important because it separates the placebo effect from two things it is often confused with. One is natural remission, where symptoms improve on their own regardless of treatment. The other is regression to the meanA statistical phenomenon where extreme values on one measure tend to be less extreme on a repeated measure. When two variables are imperfectly correlated, outliers move closer to average., a statistical phenomenon where extreme symptoms tend to move toward average levels over time. Neither of those involves active neurobiological processes. The placebo effect does.
It also is not uniform. Placebo responses vary dramatically by condition. They are strong in pain, depression, nausea, and irritable bowel syndrome. They are weak to nonexistent in cancer tumor shrinkage, bone fracture healing, or lowering cholesterol. The placebo effect changes how you experience symptoms. It does not, in general, alter the underlying disease pathology.
Your Brain on Placebo: The Neurobiology
The evidence that the placebo effect involves real neurochemistry dates to 1978, when researchers demonstrated that the opioid antagonist naloxone could block placebo analgesia. Naloxone is a drug that occupies opioid receptors without activating them, preventing both external opioids (like morphine) and your body’s own endorphins from working. When researchers gave naloxone to patients experiencing placebo pain relief, the relief disappeared. This meant the brain was not merely “imagining” less pain. It was actively releasing endogenous opioids, its own painkiller molecules, in response to the expectation of treatment.
Fabrizio Benedetti, a neuroscientist at the University of Turin and one of the most cited researchers in the field, expanded this work significantly. His team showed that placebo responses activated through expectation cues were opioid-dependent and could be fully blocked by naloxone, while placebo responses generated through prior conditioning with non-opioid drugs used entirely different neurochemical pathways and were naloxone-insensitive. In other words, the brain does not have one placebo mechanism. It has several, and which one activates depends on how the expectation was formed.
Brain imaging studies have added further layers. Functional MRI research has shown that placebo analgesia reduces activity in pain-processing regions including the thalamus, anterior cingulate cortex, and insula. A 2018 meta-analysis of individual participant fMRI data found that placebos produced small but consistent reductions in the “neurologic pain signature,” a brain activation pattern that reliably tracks physical pain intensity. The brain is not just reporting less pain. It is processing less pain.
The neurotransmitter list extends beyond opioids. Placebo responses have been linked to dopamine release in the striatum (particularly relevant in Parkinson’s disease, where dopamine is precisely what patients lack), endocannabinoid activation, serotonin modulation, and changes in cholecystokinin (CCK), a peptide that appears to mediate nocebo effects, the harmful opposite of placebo.
Open-Label Placebos: The Part That Should Not Work
If the placebo effect depends on expectation, then telling patients they are receiving a placebo should destroy it. For decades, this assumption was so obvious that nobody bothered to test it seriously. Then Ted Kaptchuk did.
Kaptchuk, a professor at Harvard Medical School and director of the Program in Placebo Studies and the Therapeutic Encounter, ran a landmark 2010 trial on patients with irritable bowel syndrome. Participants were randomized to either open-label placeboA treatment where patients receive a placebo while being explicitly informed it is a placebo, yet still experience therapeutic benefits through mind-body self-healing mechanisms. pills or no treatment. The placebo group was told, explicitly, that they were receiving “placebo pills made of an inert substance, like sugar pills, that have been shown in clinical studies to produce significant improvement in IBS symptoms through mind-body self-healing processes.” No deception. No hidden active ingredients.
The open-label placebo group improved significantly more than the no-treatment group, on both global improvement scores and symptom severity. The effect sizeA standardized measure of the magnitude of difference between groups in a study, independent of sample size. was not trivial. It was comparable to the effects seen with some active IBS medications.
This was not a one-off result. Subsequent studies have found open-label placebo effects in chronic low back pain, cancer-related fatigue (a 2018 trial showed 29% improvement in fatigue severity), migraines, and knee osteoarthritis. A 2024 meta-analysis in the European Journal of Pain confirmed a small but consistent positive effect across multiple conditions, with larger effects in clinical populations compared to non-clinical samples. A 2023 non-inferiority trial went further, providing the first evidence that open-label placebos may be as effective as deceptive placebos for pain relief.
Perhaps most revealing: a 2022 study published in Pain found that open-label placebo analgesia could also be blocked by naloxone. The same opioid-mediated mechanism that drives “traditional” placebo pain relief also drives the open-label version. Even when patients know they are taking nothing, their brains still release endorphins.
Nobody fully understands why. The leading hypotheses involve conditioned responses (your body has learned to associate pill-taking with relief), the therapeutic ritual itself (the structured encounter with a clinician activates healing expectations regardless of what you consciously believe), and embodied cognition (the physical act of swallowing a pill engages neurological pathways that bypass conscious skepticism). These are not mutually exclusive.
The Pharmaceutical Industry’s Growing Placebo Problem
The placebo effect is not just a curiosity for neuroscientists. It is a structural problem for drug development, and it is getting worse.
A 2015 meta-analysis led by Alexander Tuttle and Jeffrey Mogil at McGill University examined randomized controlled trials for neuropathic pain drugs from 1990 to 2013. They found that placebo responses had increased steadily over this period, reaching an average 30% decrease in pain levels by 2013. Drug responses, meanwhile, remained stable. The treatment advantage, the gap between drug and placebo that determines whether a drug gets approved, had been cut by two-thirds: from 27% in 1996 to just 9% in 2013.
The finding that makes this genuinely strange: the increase occurred only in trials conducted in the United States. European and Asian trials showed no change in placebo response rates over the same period. The McGill team identified two structural factors. US trials had grown larger and longer over time, and both trial size and duration were independently associated with stronger placebo responses. But the US-only pattern also raises questions about the role of direct-to-consumer pharmaceutical advertising, which is legal only in the United States and New Zealand, and which may inflate patient expectations of drug efficacy.
The consequences are concrete. Over the past decade, more than 90% of potential drugs for neuropathic and cancer pain have failed in advanced clinical trials. Not necessarily because the drugs do not work, but because the placebo arm keeps getting better, making it statistically impossible to demonstrate superiority. Pharmaceutical companies have responded by attempting to screen out “placebo responders” during run-in periods before trials begin, a practice that raises its own methodological and ethical questions.
Why “Just Placebo” Is the Wrong Framework
The phrase “just placebo” encodes an assumption: that if an effect is placebo-mediated, it is not real. The neuroscience reviewed above makes this position difficult to maintain. If a treatment produces endogenous opioid release, measurable changes in brain activation, and clinically significant symptom improvement, the mechanism being “placebo” does not make the outcome less real for the patient.
This does not mean the placebo effect can replace medicine. It cannot shrink tumors, cure infections, or mend fractures. Its domain is primarily subjective symptom experience: pain, fatigue, nausea, mood. But for conditions where subjective experience is the primary clinical outcome, dismissing placebo as “not real” misunderstands what “real” means in this context.
It also does not mean placebos are free of complications. The nocebo effectThe harmful opposite of the placebo effect, where negative expectations produce negative outcomes through the same neurobiological mechanisms as placebo benefits., where negative expectations produce negative outcomes, operates through the same mechanisms in reverse. Patients warned about side effects in clinical trials report those side effects at significantly higher rates, even in the placebo arm. The brain’s pharmacy dispenses in both directions.
What Remains Genuinely Unknown
The placebo effect is better understood than it was twenty years ago, but several fundamental questions remain open.
Individual variation is enormous and poorly explained. Some people are consistent placebo responders; others show no response regardless of context. Genetic factors appear to play a role: a 2012 study identified a variant in the catechol-O-methyltransferase (COMT) gene, which regulates dopamine metabolism, as associated with stronger placebo responses in IBS patients. But genetics explains only a fraction of the variance. Personality traits, prior treatment history, the quality of the patient-clinician relationship, and cultural factors all contribute, and nobody has a unified model of how they interact.
The mechanism behind open-label placebos remains genuinely puzzling. The conditioned-response hypothesis is plausible but does not fully explain why patients with no prior positive treatment history also respond. The therapeutic-ritual hypothesis is elegant but hard to test in isolation. And the field is still grappling with a basic epistemological question: if a placebo works through the same neurobiological pathways as an active drug, at what point does the distinction between “placebo” and “treatment” become a matter of convention rather than science?
That last question is not rhetorical. It is the question the clinical trial industry is structured to avoid, because the entire regulatory framework for drug approval depends on a clean separation between “drug effect” and “placebo effect.” If that boundary is blurrier than assumed, the implications extend well beyond any single trial.
This article covers a health-related topic. It is intended for informational purposes only and does not constitute medical advice. Consult a qualified healthcare provider for diagnosis and treatment decisions.
Sources
- Benedetti, F. (2005). Neurobiological Mechanisms of the Placebo Effect. Journal of Neuroscience, 25(45), 10390-10402.
- Benedetti, F. et al. (1999). Neuropharmacological Dissection of Placebo Analgesia. Journal of Neuroscience, 19(1), 484-494.
- Kaptchuk, T.J. et al. (2010). Placebos without Deception: A Randomized Controlled Trial in Irritable Bowel Syndrome. PLOS ONE, 5(12).
- Tuttle, A.H. et al. (2015). Increasing placebo responses over time in U.S. clinical trials of neuropathic pain. Pain, 156(12), 2616-2626.
- Bąbel, P. (2024). Rethinking placebo: Exploring the effectiveness of open-label placebos. European Journal of Pain.
- Von Wernsdorff, M. et al. (2021). Effects of open-label placebos in clinical trials: a systematic review and meta-analysis. Scientific Reports, 11, 3855.
- Kleine-Borgmann, J. et al. (2022). Open-label nondeceptive placebo analgesia is blocked by the opioid antagonist naloxone. Pain.
- Hoenemeyer, T.W. et al. (2018). Open-Label Placebo Treatment for Cancer-Related Fatigue: A Randomized-Controlled Clinical Trial. Scientific Reports, 8, 2784.
- Benedetti, F. (2010). How Placebos Change the Patient’s Brain. Neuropsychopharmacology, 36, 339-354.
