The Science of Hypnosis

For most of its history, hypnosis has been debated in philosophical terms. Is it real? Is it a special state? Is the hypnotist doing something to the subject, or is the subject doing it to themselves? Those questions persisted for more than a century because nobody had the tools to look inside the brain during the process. That changed in the 1990s, and what researchers found surprised nearly everyone.

From Monoideism to Neuroimaging

The scientific study of hypnosis begins with James Braid, a Scottish surgeon who coined the term in 1843. Braid rejected the prevailing theory of "animal magnetism," the idea that hypnotists projected a physical force onto their subjects, and proposed something more grounded. He called it monoideism: the concentration of the mind on a single idea, sustained long enough to suppress competing sensory input. It was the first model that treated hypnosis as a psychological process rather than a mystical one.

Braid got the broad strokes right, but his explanation was incomplete. Over the next century, the field split into two camps that would argue for decades. On one side, "state" theorists like Ernest Hilgard believed that hypnosis produced a genuinely altered state of consciousness, something measurably different from ordinary waking awareness. On the other side, "non-state" theorists like Theodore Barber and Nicholas Spanos argued that hypnotic behavior could be fully explained by motivation, expectation, and social context. No altered state required. Just people doing what they believed was expected of them.

The state-versus-non-state debate dominated the field from the 1960s through the 1990s. It generated enormous volumes of research, but it also created a kind of paralysis. Without a way to observe the brain in real time during hypnosis, both sides could point to behavioral evidence that supported their position. It took functional neuroimaging, specifically fMRI and PET scans, to break the stalemate.

What the Brain Actually Does

In 1997, neuroscientist Pierre Rainville published a landmark study using PET scans to observe what happened in the brain when hypnotized subjects were given suggestions about pain. The experiment was elegant. Subjects immersed their hands in painfully hot water. Under hypnosis, some were told the pain would feel less unpleasant (targeting the emotional dimension), while others were told the sensation itself would diminish (targeting the sensory dimension).

Rainville found something remarkable. Suggestions targeting unpleasantness produced measurable changes in the anterior cingulate cortex, the brain region that processes the emotional component of pain, while leaving the primary somatosensory cortex largely unchanged. Suggestions targeting intensity did the opposite: they modulated the somatosensory cortex while leaving the emotional processing intact. This "double dissociation" demonstrated that hypnotic suggestion could selectively target different components of perception. The brain was not simply ignoring the pain. It was reorganizing how it processed the experience, component by component.

Nearly two decades later, David Spiegel and his team at Stanford published fMRI research that identified three distinct neural signatures of the hypnotic state in highly hypnotizable individuals. First, decreased activity in the dorsal anterior cingulate cortex, a key node in the brain's salience network that decides what deserves attention. Second, increased connectivity between the prefrontal cortex and the insula, strengthening the brain-body connection. Third, reduced connectivity between the prefrontal cortex and the default mode network, the system responsible for self-referential thinking and mind-wandering. In practical terms, the hypnotized brain was paying less attention to peripheral stimuli, more attention to internal experience, and less attention to self-conscious evaluation. It was not asleep. It was not unconscious. It was reorganized.

The Predictive Coding Framework

Contemporary neuroscience has moved beyond the state-versus-non-state debate entirely. The most compelling current model frames hypnosis through predictive coding, the theory that the brain continuously generates predictions about incoming sensory data and then updates those predictions based on what it actually receives.

Under this framework, hypnotic suggestion works by altering the brain's "priors," its baseline expectations about what it is about to experience. When a hypnotist suggests that a subject's arm is becoming lighter, the suggestion does not bypass the subject's perception. It changes the prediction. The brain begins expecting lightness, and sensory data that would normally contradict that expectation gets down-weighted. The result is a genuine perceptual shift, not a performance.

This model explains one of the most striking findings in hypnosis research. In 2000, Kosslyn and Spiegel demonstrated that when highly hypnotizable subjects were told to perceive a gray image as colored, the color-processing regions of their brains (the fusiform and lingual gyri) actually activated. The brain was not pretending to see color. It was seeing color, because the suggestion had altered the predictive model that governed visual processing. The hallucination was neurophysiologically indistinguishable from actual perception.

Measuring Who Responds

One of the most consistent findings in hypnosis research is that people vary in their responsiveness, and that variation is remarkably stable over time. The Stanford Hypnotic Susceptibility Scales, developed by Hilgard and Weitzenhoffer in the late 1950s and early 1960s, remain the gold standard for measurement. Scores on these scales follow a roughly normal distribution: about 10 to 15 percent of people are highly responsive, 10 to 20 percent are low responders, and the majority fall somewhere in the middle.

What makes this finding significant is its stability. Longitudinal research has shown that an individual's hypnotizability score remains consistent over periods of 10, 15, and even 25 years, with test-retest correlations around 0.7. That level of stability puts hypnotizability in the same category as major personality traits. It is not a mood. It is not a skill that improves dramatically with practice. It is a relatively stable cognitive characteristic.

Crucially, hypnotizability does not correlate with the traits that popular culture associates with it. It has no meaningful relationship with submissiveness, gullibility, or low intelligence. If anything, the research points in the opposite direction. Hypnotizability correlates positively with absorption, the capacity to become deeply engaged in sensory or imaginative experience, and with certain measures of cognitive flexibility. The most responsive hypnotic subjects are not the most compliant. They are the most imaginatively engaged.

What Hypnosis Is Not

The neurobiological evidence is equally clear about what hypnosis is not. EEG and MEG studies consistently show that the hypnotic state does not produce the delta wave patterns characteristic of sleep. The brain during hypnosis shows increased theta and alpha oscillations, markers of relaxed, focused internal attention, but it remains demonstrably awake. Research by Banyai and Hilgard showed that subjects could achieve full hypnotic responsiveness while pedaling a stationary bicycle, confirming that physical relaxation and stillness are not requirements of the state.

The evidence also contradicts the "mind control" narrative. Martin Orne's research in the mid-twentieth century demonstrated that when hypnotized subjects appeared to comply with dangerous instructions (reaching for a venomous snake, throwing acid at a researcher), unhypnotized control subjects who were simply asked to "act as if" they were hypnotized complied at the same rate. The subjects were not being controlled. They were trusting the experimental context. When interviewed afterward, they reported deducing that the situation was safe because it was taking place at a university. They were active problem-solvers, not passive automatons.

Similarly, hypnosis does not function as a "truth serum" or a reliable memory enhancer. Research has consistently shown that while hypnosis can increase the volume of information a subject reports, it also increases the rate of confabulation, the unintentional production of false memories. The subject becomes more productive but not more accurate, and critically, their confidence in the accuracy of those memories increases regardless of whether the memories are real. This finding has had significant legal implications and is one reason why most jurisdictions restrict or exclude hypnotically refreshed testimony.

Where the Science Stands

The scientific picture of hypnosis in 2026 is substantially clearer than it was even twenty years ago. Neuroimaging has confirmed that hypnosis produces measurable, replicable changes in brain activity and connectivity. The predictive coding framework provides a coherent mechanism for how suggestion alters perception. Psychometric research has established hypnotizability as a stable trait with known correlates. The clinical evidence for hypnotic analgesia, the use of hypnosis for pain management, is strong enough that it is recognized by both the American Psychological Association and the National Institutes of Health.

What remains contested is narrower than it used to be. Researchers continue to investigate the genetic basis of hypnotizability, with some studies pointing to the COMT gene as a potential factor, though results have been inconsistent. The mechanisms of post-hypnotic amnesia, whether it represents a genuine dissociative barrier or a strategic attentional choice, remain debated. And the precise relationship between hypnosis and other focused-attention states like meditation is still being mapped.

But the core question that dominated the field for a century has been answered by the imaging data. The brain under hypnosis behaves differently from the brain in ordinary waking consciousness. The changes are specific, measurable, and reproducible. The question now is not whether it works, but how it works, and how that understanding can be applied.