How does hypnosis work? What we know, and what we still don't

What "hypnosis" actually refers to

Division 30 of the American Psychological Association describes hypnosis as a state of consciousness involving focused attention and reduced peripheral awareness, characterised by an enhanced capacity for response to suggestion. Its authors deliberately chose neutral wording: it is a shared vocabulary, not proof that a single mechanism has been discovered.

Two notions help. The induction is the entry procedure — fixing a point, following the breath, a countdown, images. It sets up a frame; it does not act like a substance that "puts" someone under. A suggestion is an invitation to experience a perception, an idea or an action in a certain way: "your hand may start to feel lighter". The person understands it, interprets it and responds to it in varying ways. That role of suggestion is what sets hypnosis apart from simple relaxation.

Attention, absorption and suggestion

Attention is selective: focusing on your breathing or a sensation pushes everything else into the background without erasing it. A noise can come back to the foreground, a thought can interrupt an image. That isn't failure: attention stays alive, it moves and comes back.

Absorption is the sense of being so taken up by an experience that the rest fades into the background. It exists outside hypnosis — a gripping book, music, a meticulous task. Imagination isn't limited to "seeing" either: some people think mostly in words, sensations or movements. Imagining a lighter arm doesn't change gravity, but it can change the attention paid to the arm and how a movement is experienced.

These dimensions overlap and help describe what happens. They are neither a scale of success nor a formula for effectiveness: a vivid image or strong absorption does not guarantee any benefit.

What changes — and what doesn't — during a session

Some people describe narrowed attention, time that seems to pass differently, a very present image, or a movement that seems to happen "by itself". Others remain very aware of the room and feel nothing dramatic. These accounts are real as experiences, but on their own they don't tell us which mechanism is at work.

What doesn't change matters just as much. People generally keep hearing, understanding, thinking about the suggestions and remembering the session. They are not asleep. Feeling no clear break from ordinary awareness doesn't mean nothing happened, and an intense experience doesn't prove a lasting effect.

Why a suggestion can influence experience or behaviour

A suggestion doesn't work like a physical command. It goes through understanding, imagination and expectation. Irving Kirsch proposed that response expectancies — what you anticipate will happen — account for a large part of hypnotic phenomena. If you expect heaviness, you may notice your points of contact more. The theory has been very influential without becoming the only explanation.

Context matters too: how credible the setting is, the time given, the relationship with a practitioner. Kirsch even described clinical hypnosis as a "nondeceptive placebo"; that is a debated thesis, not a verdict that hypnosis is "just placebo". Saying "it's all in your head" explains nothing: every psychological experience involves mental and brain processes.

What research actually supports

The strongest data concern specific settings, notably anxiety and pain around medical care and procedures, where meta-analyses of randomised trials exist. Other uses remain promising but uncertain, and a result obtained in a clinical setting doesn't automatically carry over to personal practice or to a recording.

Format matters as much as the label: hypnotherapy led by a professional, a guided session and listening to audio are not the same intervention. Reading a study means knowing what was actually tested, in whom, and against what.

The brain: what we know and what we don't yet

Imaging studies have observed measurable changes. In a small study of highly suggestible volunteers, suggestions changed how unpleasant pain felt without changing its intensity, with a correlate in the anterior cingulate cortex. Another, using positron emission tomography in eleven volunteers, found correlates distinct from simple mental imagery. A functional MRI study of people selected at the extremes of hypnotisability reported differences in activity and connectivity in the most responsive group.

Yet a systematic review of the brain correlates of hypnosis found few patterns that replicated from one study to the next, as protocols differed widely. There is no identified "hypnosis centre". Associated brain activity is not proof of causation: every expectation, task and sensation has brain correlates. And the lack of a single signature doesn't prove that nothing changes either.

Awareness and control: the person isn't "switched off"

The feeling that a movement happens on its own — subjective automaticity — is one possible experience. It doesn't mean the ability to choose has gone: the person understands the situation, interprets the requests and can generally refuse, change or stop.

The opposite would also be a myth: saying no isn't always easy when perceived authority, social pressure or an audience come into play. These influences aren't unique to hypnosis, but they are a reminder of the importance of consent, clear limits and the right to stop.

Responsiveness: why people react differently

Hypnotic suggestibility is measured with standardised scales that present a series of suggestions and score the responses. A score summarises responses to specific tasks on a given day; it measures neither intelligence, nor willpower, nor gullibility.

Scores show some stability without being fixed: one study observed lower scores on retest, partly explained by expectations and engagement. "I can't be hypnotised" therefore turns a local result into an identity; it would be more accurate to say "I didn't respond to that suggestion, under those conditions". Conversely, a high score confers no special status.

Common myths

Safety, limits and context of use

Hypnosis should be described neither as "risk-free" nor as inherently dangerous. Unwanted effects have been reported — usually brief, sometimes more serious — but the data don't allow any general risk rate to be stated. Context changes a lot: a stage show, supervised care and personal practice don't offer the same safeguards.

It replaces neither diagnosis nor treatment. In a mental health crisis, or with a history of psychosis, epilepsy or significant dissociative episodes, individual professional advice comes before any practice. And it should never be used to "recover" what supposedly happened.

Going further

If you'd like to move from understanding to practice, the self-hypnosis guide explains how to begin simply and carefully. The AI hypnosis and personalized hypnosis pages describe what artificial intelligence does inside HypnoAI, its limits, and what can vary from one session to another.

To explore each of these questions in depth, the HypnoAI book "Hypnosis — Understand Before You Practice" devotes its first part to definitions, theories and myths, then examines the evidence, memory and safety.