Is the placebo effect just people imagining they feel better?

No. Placebo responses involve measurable neurochemical changes — endogenous opioid release, dopamine activity, and in some contexts endocannabinoid signalling. These are real physiological events, not fabricated reports. The response is shaped by expectation, conditioning, and context, but the underlying biology is concrete.

What is the nocebo effect and how does it differ from placebo?

The nocebo effect is the negative counterpart to placebo: an inert substance or procedure produces an adverse response because of negative expectation. It involves different neurochemical mediators — notably cholecystokinin (CCK) — and is just as biologically real. Some reported side effects in clinical trials may be nocebo-driven rather than pharmacologically caused.

Why do clinical trials need a placebo arm?

Without a placebo control, it's impossible to separate the pharmacological effect of an active compound from the response driven by expectation, the clinical setting, and prior conditioning. The placebo arm establishes a baseline that the active treatment must exceed to demonstrate a genuine drug effect.

Does the route of administration affect the size of the placebo response?

Research evidence suggests yes. Injections tend to produce larger placebo responses than oral dosing, likely because the ritual and perceived invasiveness of the procedure amplify expectation effects. This is one reason route of administration matters in clinical trial design beyond pharmacokinetic considerations.

Do these effects have any relevance to cannabinoid or peptide research?

Yes. Much early-phase research on cannabinoids and peptides uses open-label or small-scale designs that are particularly vulnerable to placebo and nocebo effects. Properly interpreting that literature requires understanding how large expectation-driven responses can be, and why blinded, adequately powered trials are the methodological standard.

The Placebo and Nocebo Effects

By Priya Nandakumar · 2 June 2026 · 7 min read

A person in white gloves is looking through a microscope

When "nothing" does something

About 30–50% of participants in placebo-controlled clinical trials show measurable physiological responses to inert treatments. That's not a rounding error. That's a core pharmacological phenomenon, and honestly, I think it's one of the most underrated topics in all of medicine. We spend so much time talking about molecular targets and receptor kinetics , and rightly so , but the machinery that generates placebo and nocebo responses is just as real, and just as worth understanding.

Let me be clear from the start: none of what follows is a claim that believing hard enough will treat any condition. This is pharmacology education. What I want to walk through is the neurobiology, the endogenous signalling, and the research methodology that explains why these effects exist and why every clinical trial has to account for them.

Defining the terms

A placebo effect occurs when a person receives an inert substance or sham procedure and shows an objective physiological or subjective response that goes in the direction of benefit. A nocebo effect is the mirror image: an inert substance or procedure produces an adverse response, driven by negative expectation.

Neither is "just psychological." That framing dismisses the mechanism. Both involve measurable changes in neurochemistry, autonomic tone, and sometimes immune signalling. The brain is not floating separately from the body's pharmacology.

The endogenous opioid system is heavily involved

The most thoroughly characterised mechanistic pathway underlying placebo analgesia runs through endogenous opioid peptides , primarily endorphins and enkephalins. Fabrizio Benedetti's group at the University of Turin has shown, in work published across multiple decades, that naloxone (an opioid antagonist) can significantly attenuate placebo analgesic responses. That's not a subjective finding; it's receptor pharmacology demonstrating that the response depends on actual opioid signalling.

Conditioning matters here too. When a person has previously received an active opioid, the body can "replay" a portion of that response when they subsequently receive an inert substance in the same context. The conditioned stimulus; the act of taking a pill, the clinical setting, even the colour of a capsule, activates circuits that previously co-fired with the pharmacological effect. This is Pavlovian conditioning at a neurochemical level, not imagination.

And it links, interestingly, to the endocannabinoid system. Benedetti's lab published evidence as early as 2011 in the Journal of Neuroscience suggesting that cannabinoid CB1 receptors (see our overview of the CB1 receptor) are recruited in some placebo analgesic responses, especially when opioidergic pathways are blocked. Endocannabinoid tone, including fluctuations in anandamide; appears to contribute to the broader picture of expectation-driven modulation. That's a live research area, not a settled question, but the signalling overlap is genuinely interesting from a pharmacology perspective.

Dopamine and the expectation-reward loop

Placebo effects aren't confined to pain. In Parkinson's research, placebo responses in motor function have been associated with measurable dopamine release in the striatum, as mapped via PET imaging. The expectation of receiving an active dopaminergic agent appears to trigger anticipatory dopamine activity in circuits that were supposed to be waiting for the drug.

This is not trivial. Dopamine release in these contexts shows up in the nucleus accumbens, the same region central to reward processing and motivation. The pharmacokinetics of endogenous signalling can, under certain conditions, mirror a portion of what an exogenous compound would do. Understanding pharmacokinetics in that broader sense, how substances, including endogenous ones, move through and act on biological systems, helps contextualise why placebo responses are dose-shaped, context-shaped, and subject to extinction over time.

The nocebo: expectation in the other direction

The nocebo effect deserves at least as much attention as placebo, but gets far less of it. Which is a problem, honestly, because it has direct implications for how we communicate with patients.

Cholesterol-lowering statin trials have been a useful natural experiment here. The SAMSON trial (published in the New England Journal of Medicine in 2020) used an n-of-1 design across 60 participants and found that the vast majority of muscle-related symptoms attributed to statins in that cohort were nocebo responses; generated by the knowledge of taking the drug rather than its pharmacology. The symptoms were real. The participants weren't faking. But when they were given identical-looking inert capsules with no disclosed content, the rates of symptom reporting tracked expectation, not pharmacology.

At the neurochemical level, nocebo responses have been associated with cholecystokinin (CCK) release. CCK is an anxiogenic neuropeptide, and Benedetti's group demonstrated that CCK antagonists can partially block nocebo hyperalgesia. So again: a specific molecular target, a specific endogenous signal. Not mysticism.

Why clinical trial design depends on this

Every properly designed clinical trial uses a placebo arm precisely because of this. But the nuances run deeper than most people realise. The response magnitude varies with:

the ritual surrounding administration (more elaborate procedures tend to produce larger placebo responses)
the warmth and confidence of the clinician
the prior conditioning history of the participant
the route of administration, injections often produce larger placebo responses than oral dosing, which is something worth noting if you're reading about subcutaneous injection protocols in any context
open-label versus blinded conditions (open-label placebo trials, where participants know they're taking inert substances, still show response, challenging the simple "deception is required" model)

From a research design perspective, this is why the nocebo effect creates an ethical bind. Fully informed consent; required under Australian and international research ethics frameworks, may itself generate nocebo responses by priming participants to expect adverse effects. The nocebo effect is partly iatrogenic, embedded in the communication of risk.

What this means for interpreting research on any compound

I spend a lot of time reading pharmacology literature on cannabinoids, peptides, and plant-based extracts. And one thing I keep coming back to is how poorly the placebo response is characterised in many trials outside of mainstream pharmaceutical research. Smaller trial sizes, shorter durations, and open-label designs, common in early-phase research on compounds like cannabidiol; make it genuinely hard to parse drug effect from expectation-driven response.

That's not an argument against studying these compounds. It's an argument for methodological rigour, for adequately powered trials, and for not overclaiming from preliminary data. I'd argue researchers and science writers have an obligation here, present the mechanism, present the limitations, and let the reader hold the nuance.

Between long runs I listen to a lot of true crime podcasts, I know, I know, very cliché; and the thing that makes a good crime documentary isn't the conclusion, it's the quality of the evidence chain. Pharmacology is the same. A compelling story about a molecule means nothing if the evidence trail doesn't hold up to scrutiny.

Open-label placebo: a newer research direction

One of the more surprising developments in this space is the open-label placebo (OLP) trial design, pioneered substantially by Ted Kaptchuk's group at Harvard. In OLP trials, participants are explicitly told they are receiving an inert pill with no active ingredient, and yet measurable responses persist above no-treatment controls in some populations and contexts.

The proposed mechanism shifts from expectation of pharmacological effect to conditioned physiological response and the engagement of self-healing rituals around clinical contact. This is not well-enough understood to make clinical recommendations from, but it does fundamentally challenge the assumption that deception is a prerequisite for placebo response. The research is genuinely interesting. It's also methodologically young, and should be read with appropriate scepticism.

Sources

, Priya Nandakumar, Pharmacology writer; peptides & PK

Common questions

Is the placebo effect just people imagining they feel better?: No. Placebo responses involve measurable neurochemical changes — endogenous opioid release, dopamine activity, and in some contexts endocannabinoid signalling. These are real physiological events, not fabricated reports. The response is shaped by expectation, conditioning, and context, but the underlying biology is concrete.
What is the nocebo effect and how does it differ from placebo?: The nocebo effect is the negative counterpart to placebo: an inert substance or procedure produces an adverse response because of negative expectation. It involves different neurochemical mediators — notably cholecystokinin (CCK) — and is just as biologically real. Some reported side effects in clinical trials may be nocebo-driven rather than pharmacologically caused.
Why do clinical trials need a placebo arm?: Without a placebo control, it's impossible to separate the pharmacological effect of an active compound from the response driven by expectation, the clinical setting, and prior conditioning. The placebo arm establishes a baseline that the active treatment must exceed to demonstrate a genuine drug effect.
Does the route of administration affect the size of the placebo response?: Research evidence suggests yes. Injections tend to produce larger placebo responses than oral dosing, likely because the ritual and perceived invasiveness of the procedure amplify expectation effects. This is one reason route of administration matters in clinical trial design beyond pharmacokinetic considerations.
Do these effects have any relevance to cannabinoid or peptide research?: Yes. Much early-phase research on cannabinoids and peptides uses open-label or small-scale designs that are particularly vulnerable to placebo and nocebo effects. Properly interpreting that literature requires understanding how large expectation-driven responses can be, and why blinded, adequately powered trials are the methodological standard.