Can Hypnosis Recover Lost Memories? What the Science Says

The movie version: a hypnotist counts down, the fog clears, and the buried memory surfaces — intact, accurate, admissible.

Memory doesn’t work that way. Neither does hypnosis.

No — hypnosis cannot reliably recover lost memories. What it reliably does is increase the amount you recall and your confidence in it, not its accuracy. Under hypnosis people produce more material, but much of the new material is wrong, and they believe the false details as firmly as the real ones. That’s why the American Medical Association warned against it four decades ago and why courts restrict hypnotically refreshed testimony. Here’s the evidence.

The Experiment That Broke the Myth

The cleanest test came from Jane Dywan and Kenneth Bowers, published in Science in 1983. Subjects studied 60 pictures, then spent a week trying to recall them. Then half were hypnotized and encouraged to remember more. They did remember more — that part of the myth is true. The problem: most of the newly recalled material was incorrect, and the effect was strongest in highly hypnotizable subjects under hypnosis.

That’s the pattern that repeats across this entire literature. Hypnosis boosts output, not accuracy. It lowers your threshold for reporting something as a memory, so more comes out — real fragments, plausible guesses, and pure confabulation, all delivered with the same feeling of remembering.

What 24 Studies Found: More Errors, More Confidence

A 1994 meta-analysis (Steblay & Bothwell, Law and Human Behavior) pooled 24 studies comparing hypnotized and non-hypnotized eyewitnesses. The findings:

Hypnotized subjects produced more intrusion errors and higher levels of pseudomemory — false memories of things that never happened.
Hypnotized subjects were more confident in their recall than non-hypnotized subjects — regardless of whether they were right.
Recall was less accurate for hypnotized subjects at delays under 24 hours (d = −.29) and at one week (d = −.24). The one condition favoring hypnosis — non-leading questions at a 1–2 day delay — didn’t change the authors’ conclusion: caution against relying on hypnotically refreshed memory.

That second bullet is the finding most coverage misses, and it’s the dangerous one. If hypnosis merely produced errors, you could filter them. But it produces errors plus inflated confidence, which strips away the honest “I’m not sure” that normally flags a shaky memory. The error and the conviction arrive welded together.

It gets worse when suggestion enters the room. A 2002 experiment (Scoboria, Mazzoni, Kirsch & Milling) tested 111 students with misleading or neutral questions, in or out of hypnosis. Both hypnosis and misleading questions significantly increased memory errors, and the two effects were additive — misleading questions asked during hypnosis produced the most errors of all. Notably, the effect didn’t depend on how suggestible the person was. A well-meaning practitioner asking “and then what did he do to you?” inside hypnosis is running the highest-error condition in the study.

Why Courts Don’t Trust It

This isn’t just a lab concern. In 1985, the AMA’s Council on Scientific Affairs reviewed the evidence and concluded that recollections obtained during hypnosis “can involve confabulations and pseudomemories and not only fail to be more accurate, but actually appear to be less reliable than nonhypnotic recall” — and warned that using hypnosis on witnesses could have serious consequences for the legal process.

The U.S. Supreme Court engaged with the same science in Rock v. Arkansas (1987). The opinion lays out three problems with hypnotic recall: the subject becomes suggestible and may try to please the hypnotist; the subject tends to confabulate, filling gaps with imagination; and the subject experiences “memory hardening” — great confidence in both true and false memories, making effective cross-examination more difficult. The Court noted that a number of states had responded by excluding hypnotically refreshed witness testimony outright as unreliable. (The narrow holding of Rock was that a blanket ban couldn’t be applied to a criminal defendant’s own testimony — the Court still declined to endorse hypnosis as an investigative tool.)

When a technique’s output is treated as presumptively unreliable by both the AMA and much of the court system, “hypnotic memory recovery” is not a fringe skepticism position. It’s the mainstream scientific one.

The “Repressed Memory” Problem

The recovered-memory pitch usually leans on a second claim: that trauma gets repressed — locked away intact, waiting for hypnosis to unlock it. That model is scientifically contested at best. A 2019 review (Otgaar, Howe, Lynn, Loftus and colleagues, Perspectives on Psychological Science) found that belief in repressed memories persists on a nontrivial scale — 58% — and appears to have increased among clinical psychologists since the 1990s, even as the research community documented how therapeutic techniques aimed at recovering such memories can carry adverse effects. The authors’ warning is blunt: the “memory wars” never ended, and the consequences play out in clinics and courtrooms.

Put the two findings together and the risk is obvious. A technique that inflates confidence in false details, applied in search of memories that may never have been stored the way the model assumes, is a machine for manufacturing certainty — not truth.

What This Means in Practice

Here’s the distinction that gets lost: hypnosis failing at memory retrieval says nothing about hypnosis failing in general. The same suggestibility that corrupts recall is exactly what makes hypnosis clinically useful for anxiety, sleep, and habit change — there, suggestion is the mechanism, pointed at how you respond going forward, not at reconstructing your past. The evidence for those uses is a separate question with a very different answer — see is hypnosis just a placebo? and what happens in your brain during hypnosis.

So the honest rules are simple:

Treat any “recovered” memory from hypnosis as unverified. It may contain truth, but the feeling of certainty attached to it is not evidence.
Avoid regression-style “memory recovery” work. It sits in the highest-error condition the research has identified: suggestion plus hypnosis plus a gap the mind wants to fill.
Judge clinical hypnosis on its own evidence base — future-directed suggestion for measurable outcomes, which is precisely how modern clinical hypnosis (including Oriamind’s sessions) is built. No excavation of the past required.

Hypnosis is a real tool with real evidence behind it. Memory recovery is the one job it demonstrably cannot do — and knowing that is what separates clinical use from pseudoscience. For more myths put against the data, see can you get stuck in hypnosis?

This post is part of our science of clinical hypnosis series.