In 1978, a dental surgeon at the University of California, San Francisco, published a four-page study in The Lancet (pages 654–657) that rewired how medicine understands its own tools — and provided pioneering experimental evidence linking expectation-based pain modulation to endogenous neurochemistry. Following surgical third-molar extractions conducted under pre-procedural calming premedication, local anesthesia (mepivacaine), and nitrous oxide, participants were monitored for post-operative sensory feedback using a Visual Analogue Scale (VAS). Jon Levine, Howard Fields, and Newton Gordon initially administered an intravenous placebo presented as a painkiller; a substantial subgroup of participants reported genuine pain relief. In the subsequent experimental phase, researchers administered receptor blockers—a competitive opioid receptor antagonist that binds with high affinity to mu-opioid receptors, blocking both endogenous and exogenous opioids without intrinsic activating efficacy. When placebo responders received receptor blockers, their pain relief was promptly reversed, whereas control participants who received receptor blockers without prior expectation-driven sensory gating also reported heightened pain. This seminal four-page trial provided experimental evidence consistent with endogenous endorphin involvement in expectation-driven sensory gating, demonstrating that opioid receptor blockade reverses expectation-driven pain relief without directly assaying endogenous endorphin peptide concentrations in biological fluids.
That conclusion collapsed a century of comfortable assumption. The placebo effect was not imagination or measurement noise; Levine et al.’ findings demonstrated that in acute post-operative pain, expectation could mobilize endogenous endorphin pain reduction that was pharmacologically reversible. In experimental research, scientists carefully distinguish between the placebo response (the aggregate improvement observed across an entire placebo control group, which includes natural physiological baseline, spontaneous recovery, and statistical regression to the mean) and the true placebo effect (the specific psychobiological outcome directly driven by psychological expectation, perceptual conditioning, and the symbolic context of care). While later research established that true placebo effects recruit diverse neurochemical pathways—including striatal dopaminergic reward prediction and endocannabinoid signaling depending on conditioning history—the Levine trial proved that expectation-driven sensory gating can be pharmacologically blocked by an opioid antagonist, confirming that expectation mobilizes endogenous biochemical cascades rather than imagination or measurement artifact. Four decades of neuroscience have since mapped the machinery , endorphins and enkephalins onto mu-opioid receptors, dopamine surging through the striatum, dose-response curves and blocking agents. What follows is an anatomy of that machinery: the learning systems that arm it, the circuit that fires it, the toxic twin that runs it in reverse, and the honest sugar pill that triggers it anyway.
Belief is not the only shortcut to the body’s natural painkillers; ice-water lab trials show how swearing dulls acute pain through a sudden fight-or-flight surge.
The Receptor blockers Experiment: An Early Demonstration of Endogenous Endorphin Pathways
Methodological Clarification: controlled studies draw a rigorous distinction between the true placebo effect (endogenous psychobiological pain reduction driven by neurotransmitter release) and the broader placebo response, which encompasses spontaneous remission, regression to the mean, and reporting bias. Crucially, expectation-driven sensory gating modulates subjective symptom perception and discomfort; it does not eliminate underlying microbial pathogens, reverse tissue damage, or replace structured behavioral adjustments.
This landmark 1978 study provided an early experimental demonstration supporting an endogenous endorphin mechanism in expectation-driven sensory gating. Levine et al. treated post-surgical dental pain with a placebo analgesic; when responders received receptor blockers, their reported pain relief diminished, providing evidence consistent with the proposed endorphin-mediated mechanism. In the trial’s control observations, participants administered receptor blockers without receiving an effective placebo experienced higher reported pain than placebo responders, an outcome consistent with the pharmacological blockade of endogenous endorphin activity, although the trial did not directly measure baseline circulating or central opioid peptide concentrations.
These findings provided strong experimental support for the hypothesis that expectation-driven pain reduction is mediated, in significant part, by endogenous endorphin signaling , endorphins and enkephalins acting upon mu-opioid receptors. While subsequent research demonstrated that non-endorphin mechanisms (such as endocannabinoids and dopamine) also contribute, receptor blockers antagonism demonstrated that psychological expectation recruits measurable biological pathways. Twenty-seven years later, Jon-Kar Zubieta’s laboratory made the invisible visible: PET scans showed expectation of relief measurably activating the endogenous endorphin system in the prefrontal cortex, the anterior cingulate, and deep nuclei feeding the brainstem’s pain-control circuitry. The placebo effect had been photographed at its receptor.
Two Engines: Pavlovian Conditioning and Conscious Expectation
What arms the placebo effect? Neuroscience distinguishes two engines.
The first is conditioning, the associative engine Ivan Pavlov described a century ago. Pair a bell with food often enough and the bell alone provokes salivation; pair a capsule with relief often enough and the capsule alone begins to mobilize physiology. Crucially, this half of the placebo effect operates below consciousness and reaches systems expectation struggles to touch: immune suppression, hormone secretion, bronchoconstriction and its reversal.
Switching participants silently from an active solution to a placebo, for instance, partially reproduces the drug’s physiological signature: the body has learned the routine. The white coat and the injector are the bell.
The second engine is expectation: a conscious, top-down prediction about what the intervention will do, generated in the prefrontal cortex, the same predictive machinery that lets the nervous system anticipate the body’s needs, from the reflexes that wrinkle your fingers in water to the forecasts that guide healing.
Neuroimaging work by Tor Wager and colleagues showed that placebo instructions increase activity in the dorsolateral prefrontal cortex (dlPFC) and the rostral anterior cingulate cortex (rACC) (regions that hold expectations online) and that the strength of this engagement predicts how much relief each person reports. The engines are also separable: expectation dominates conscious outcomes such as pain and nausea, while conditioning governs autonomic and immunological parameters.
The Open-Label Paradox: Honest Placebos That Still Work
Here the story turns on its head. If expectation requires belief, an openly labeled inert capsule should be self-defeating: tell participants plainly that the capsules contain inert substance, and endogenous endorphin and dopamine release should go silent. Ted Kaptchuk’s program at Harvard Medical School tested that assumption, and the assumption lost.
In a 2010 randomized controlled trial, individuals with irritable digestive sensations (IBS) received pill bottles labeled, in print, “Placebo Pills”, sugar pills with no active pharmacological agents, plus a frank explanation that placebos can work through mind-body self-healing processes. Three weeks later, the open-label group reported significantly greater global improvement and symptom relief than passive control groups. The result was evaluated in chronic low back pain by Carvalho et al. (2016) in a randomized controlled trial of 97 participants: adding an open-label placebo to Standard Protocol (TAU) produced significantly greater reductions in both pain and disability scores compared with Standard Protocol alone (with both groups maintaining their normal daily routines), and pilot studies report early signals in migraine. Honesty, it turns out, is not a contraindication.
The leading interpretation dissolves the paradox: the active ingredient is not belief in the pill but the embodied ritual of care, taking the capsule, performing the act of receiving help, embedded in a supportive interpersonal context. Conditioning supplies the physiology; the ceremony supplies the cue. Conscious conviction is optional equipment: the body learns by association, and the ritual is the teacher.
The Biological Boundaries: What Placebos Cannot Do
Every powerful idea invites overreach, so the line must be drawn. The placebo effect is a modulator of subjective and autonomic outcomes, and its influence has hard biological limits.
Placebos powerfully modulate how the mind reads the body; pain, nausea, fatigue, cough, bronchoconstriction, tremor, depressive symptoms. They do not destroy physical cellular lesions, clear systemic bacterial infections, or regenerate articular cartilage.
That boundary is the mechanism read in reverse. Everything on the “can” list routes through the nervous system’s control rooms, pain through the descending pain modulatory pathway, nausea through brainstem vomiting centers, airway tone through autonomic outflow, and the deep fatigue the body raises before collapse, as we covered in how your body signals fatigue before you collapse. The placebo effect reaches exactly those control rooms; everything on the “cannot” list is a problem of tissue or pathogen.
The boundary explains a subtler truth: placebos change the experience of illness more than its course, and relief that lets an individual sleep, move, and eat is not an imaginary benefit. The honest position is precision about the territory the placebo effect can reach.
The Descending Circuit: How the Brain Closes the Gate on Pain
How does a placebo reach a wound in the toe from a thought in the head? The answer is the descending pain modulatory pathway, a dedicated neural highway that adjusts the intensity of spinal pain traffic.
The chain reads like a relay race. Expectation-related activity in the rACC projects to the periaqueductal gray (PAG), a midbrain hub densely packed with opioid receptors. The PAG then drives the rostral ventromedial medulla (RVM); brainstem territory that also runs the autonomic surprises of sleep onset, including the jerks that wake you as you drift off, whose neurons descend to the dorsal horn, the landing zone where nociceptive fibers first synapse before any signal ascends toward consciousness.
From this descending terminal, serotonin and norepinephrine release onto inhibitory interneurons in the dorsal horn, and the gate on nociceptive traffic physically narrows. Pain signals do not get weaker at the wound; they get filtered before the brain finishes reading them.
In 2009, Falk Eippert and colleagues imaged the entire circuit under expectation-driven sensory gating and showed that receptor blockers undid it, collapsing both the behavioral relief and the coupling along rACC, PAG, and RVM. The descending pain modulatory pathway is the physical infrastructure of the placebo effect: a prediction from the cortex, implemented as spinal pharmacology.
And expectancy-driven physiology has a currency of its own. Placebo anticipation releases dopamine in the ventral striatum and nucleus accumbens, the reward infrastructure that also fires for food, money, and music, tying pain reduction to reward anticipation.
The Predictive Brain: How Expectations Shape Physiological Responses
Zoom out, and this top-down neural modulation becomes a statement about what kind of organ a brain is. The classical view treats perception as passive reception. A growing body of neuroscience argues for the opposite architecture: the brain is a prediction machine that generates a model of the body’s state in advance, then uses incoming signals mainly to correct it; the principle we examined in what neuroscience reveals about human adaptation. Pain, in this framework, is not a measurement but a conclusion.
Seen through that lens, this expectancy response is not an anomaly but the prediction machine operating as designed. A credible prediction of safety recruits the endogenous endorphin system, tilts the dorsal horn’s gate, quietens nociceptive traffic, and the pain reduction engages measurable descending inhibitory circuitry within the central nervous system. The nocebo effect is the same machinery running on catastrophic priors. What medicine calls context, the white coat, the reassurance, the ritual, is, to a predictive brain, evidence about what comes next.
The discovery refuses to stay a laboratory curiosity. Open-label placebos are entering pilot protocols; trial designers now weigh placebo-controlled procedures; researchers are learning that words are dosed like drugs. If expectation is a biological signal, the ritual of care is a cue anyone can recognize: the nurse who explains instead of dismisses, the researcher whose tone carries the expectation as surely as the tablet.
The 1978 receptor blockers findings initiated a fundamental shift in neurobiological understanding, illustrating how expectation and conditioning alter physiological symptom perception. Fake medicine works because nothing about it is fake, the pill is inert, but the brain it enters is not, and the brain has a pharmacopoeia of its own.
The Nocebo Effect: When Expectation Turns Toxic
The machinery has an evil twin. If the brain can synthesize pain reduction on the strength of a prediction, it can synthesize the opposite, the nocebo effect, the placebo effect running in reverse run in reverse: negative expectation producing measurable hyperalgesia, nausea, and symptom amplification from suggestion alone.
In Fabrizio Benedetti’s Turin laboratory, a nocebo suggestion, telling study participants that a sensory stimulus would intensify, reliably amplified their reported pain and anxiety. Mitigating anticipatory anxiety with calming premedication attenuated both the heightened hypothalamic-pituitary-adrenal (HPA) axis response and the nocebo-induced hyperalgesia, highlighting anxiety’s catalytic role. More striking still, in Fabrizio Benedetti’s landmark trials, the non-specific cholecystokinin (CCK) antagonist proglumide completely blocked the nocebo hyperalgesia without suppressing HPA axis hyperactivity, demonstrating that cholecystokinin (CCK) plays a key neurochemical role in mediating anticipatory anxiety-induced hyperalgesia, alongside other cognitive and neurochemical pathways that shape nocebo responses. In laboratory models of anticipatory hyperalgesia, cholecystokinin (CCK) functions as an identified biochemical mediator facilitating pain transmission during anxiety-induced expectation, operating alongside other cognitive and autonomic pathways rather than serving as the sole universal transmitter across all nocebo phenomena. Negative expectation is not the absence of relief but an actively manufactured opposite.
No villainy is required; ordinary supportive framing suffices. Reading a drug’s side-effect leaflet measurably increases the rate of those exact side effects, and so does a researcher’s tone; nocebo suggestions even raise heart rate and cortisol as part of the anticipatory stress response. The circuitry that lets a warm word blunt a headache lets a warning label manufacture one.
When Surgery Itself Was the Placebo
The most uncomfortable experiments involved scalpels. In 2002, J. Bruce Moseley and colleagues reported in the New England Journal of Medicine that 180 participants with knee discomfort, randomized to arthroscopic debridement, arthroscopic lavage, or a sham procedure in which surgeons made real incisions, simulated the arthroscopy, and completed the procedure for the participant; improved equally. Over two years, the sham surgery group matched the participants who received the actual procedure on pain and function; the placebo effect had matched the scalpel, point for point.
A decade and a half later, cardiology ran its own version. In the ORBITA trial, 200 participants with stable chest discomfort were randomized to a stent, or to a sham procedure: real catheterization, a stent placed and then withdrawn, six weeks of blinded follow-up. The comparison found no significant difference between stent and sham in exercise time or angina frequency; this anticipation reproduced most of the perceptual benefit on its own.
Neither trial says surgery is theater. They say something more precise: the procedural ritual engages the same expectation circuitry as any capsule, and the resulting relief can be large enough to mimic hardware, which is why placebo-controlled surgical designs are now considered ethically achievable.
Why Placebos Work Even When You Know They Are Fake
Does the placebo effect mean my symptoms were fake?
No, and the receptor blockers experiments provide strong evidence supporting the mechanism. If expectation-driven sensory gating can be reversed by an opioid antagonist, it was built from real neurochemistry, not imagination. The placebo response changes measurable physiology: endogenous endorphins release, dopamine rises in the striatum, spinal nociceptive processing dampens. Only the trigger was inert, the relief was not. Self-limiting illness can also mislead, which is why the last question matters.
Why do two fake pills work better than one, and injections better than pills?
Because placebos are dosed by symbolism, not milligrams. The placebo effect scales with the perceived intensity of the ritual: capsules outperform tablets, two capsules outperform one, injections outperform capsules, and sham surgery produces the largest responses of all. Price works the same way, placebos presented as expensive brand-name tablets relieved pain more than identical pills presented as discounted generics. The effect tracks what the ritual teaches the body to expect.
Can someone develop tolerance or addiction to a placebo?
Tolerance, yes in a limited sense: this psychobiological healing response can wane with repetition as the ritual’s novelty fades, and in one study of postoperative recovery, reducing chemical strength while keeping pill-taking behavior intact let the placebo response substantially replace the drug dose. Addiction, no: placebo responses involve no external substance, no receptor-level dependence, and no withdrawal discomfort when they stop, placebo conditioning is even being studied to ease opioid tapering.
How do scientists separate the expectation effect from natural physiological recovery?
With control groups, not intuition. Regression to the mean, natural remission, and symptom fluctuation affect untreated groups too, so a proper trial compares a placebo arm against a passive control arm in parallel. The difference isolates the placebo effect; everything both arms share is regression, remission, or attention. When placebo and passive arms genuinely diverge, as in the IBS and open-label back-pain trials, the gap is the effect, not the condition’s own timeline.
Is taking a placebo dishonest medicine?
The open-label trials answer this directly: placebos disclosed as placebos still outperformed no intervention in IBS and chronic low back pain. Expectancy-mediated pain reduction there ran on the embodied ritual plus the interpersonal context, not deception. What remains ethically charged is the covert use of placebos while claiming they are real drugs; the idea that a placebo must be a lie no longer is.
Peer-Reviewed Studies & Sources
- Levine JD, Gordon NC, Fields HL. The mechanism of expectation-driven sensory gating. Lancet. 1978;2(8091):654–657. PMID 80579
- Zubieta JK, Bueller JA, Jackson LR, et al. Placebo effects mediated by endogenous endorphin activity on mu-opioid receptors. J Neurosci. 2005;25(34):7754–7762. PMID 16120776
- Wager TD, Rilling JK, Smith EE, et al. Placebo-induced changes in fMRI in the anticipation and experience of pain. Science. 2004;303(5661):1162–1167. PMID 14976306
- Eippert F, Bingel U, Schoell ED, et al. Activation of the endorphinergic descending pain control system underlies expectation-driven sensory gating. Neuron. 2009;63(4):533–543. PMID 19709634
- Scott DJ, Stohler CS, Egnatuk CM, et al. Placebo and nocebo effects are defined by opposite opioid and dopaminergic responses. Arch Gen Psychiatry. 2008;65(2):220–231. PMID 18250260
- Benedetti F, Amanzio M, Casadio C, Oliaro A, Maggi G. Blockade of nocebo hyperalgesia by the cholecystokinin antagonist proglumide. Pain. 1997;71(2):135–140. PMID 9211474
- Kaptchuk TJ, Friedlander E, Kelley JM, et al. Placebos without deception: a randomized controlled trial in gut-brain sensitivity. PLoS One. 2010;5(12):e15591. PMID 21203519
- Carvalho C, Caetano JM, Cunha L, Rebouta P, Kaptchuk TJ, Kirsch I. Open-label placebo protocols in chronic low back pain: a randomized controlled trial. Pain. 2016;157(12):2766–2772. PMID 27755279
- Moseley JB, O’Malley K, Petersen NJ, et al. A controlled trial of arthroscopic surgery for osteoarthritis of the knee. N Engl J Med. 2002;347(2):81–88. PMID 12110735
- Al-Lamee R, Thompson D, Malik N, et al. Percutaneous coronary intervention in stable angina (ORBITA): a double-blind, randomised controlled trial. Lancet. 2018;391(10115):31–40. PMID 29103656
- Waber RL, Shiv B, Carmon Z, Ariely D. Commercial features of placebo and perceptual efficacy. JAMA. 2008;299(9):1016–1017. PMID 18319411
- Benedetti F, Carlino E, Pollo A. How expectation shapes neural sensory processing. Neuropsychopharmacology. 2011;36(1):339–354. PMID 20592717
