Is Trauma Stored in the Body? What the Evidence Supports, and What It Does Not
Evidence Review · The Reading Room, Vol. 10 · 11 min read · What the storage claim rests on, what it does not, and why the practices can still be worth doing
The phrase comes from a book that has sold millions of copies. A 2025 evaluation in the BJPsych Bulletin tested its central claims against the evidence.
The sentence has become furniture. It appears in therapy intake forms, in yoga class, in the explanation a friend offers over coffee for why your shoulders have been up around your ears since 2019. Trauma is stored in the body. The body keeps the score.
It is a compelling idea, it is repeated with great confidence, and almost nobody who repeats it has read the evidence underneath it. This review examines what that evidence establishes. It is a reading of published psychiatry and epidemiology; we are not clinicians, the findings belong to the researchers named below, and only the plain-English rendering is ours.
The short answer: the broad claim — that chronic stress and adversity have real, measurable physical consequences — is well supported. The specific claim — that traumatic experience is retained in bodily tissue as a kind of memory, which must be discharged through the body before recovery is possible — is not. A 2025 evaluation in the BJPsych Bulletin examined the neurobiological and treatment claims in The Body Keeps the Score and found several of the central ones unsupported by causal evidence.¹
There is a better name for what you are actually carrying, and it is not a deposit waiting to be excavated: the running cost. Years of staying switched on have a physiological price that accumulates in ordinary ways — sleep, immune function, cardiovascular strain, a body that reacts before you have decided anything. That price is real, it is documented, and it explains the tight shoulders better than the vault does. It also implies something entirely different about what to do next.
- Three prospective reviews, the latest identifying 25 studies. Where brain differences have been tracked before and after trauma exposure rather than only afterwards, many appear to precede the trauma — consistent with pre-existing vulnerability rather than damage.¹
- 8,056 adults, a real dose-response, and no causal claim. The original adverse childhood experiences cohort found smoking rising from 7% to 17% and severe obesity from 5% to 12% across the score range — from a retrospective, cross-sectional design its own authors say cannot establish causation.²
- 64 women, 10 weeks: the practice worked anyway. Trauma-informed yoga beat an active comparison group, with 52% versus 21% no longer meeting PTSD criteria.³
- Trauma-focused talking therapy still has the most robust outcomes. The 2025 evaluation concludes that structured trauma-focused CBT produces the most durable results across meta-analyses, and that the claim only body-based approaches can process trauma is unsupported.¹
- The mechanism claims are the weakest part. Epigenetic explanations are characterised by small effects, unclear mechanisms and replication failures; stress-hormone explanations by early findings that proved difficult to replicate.¹
The source for this review
Prof. Michael S. Scheeringa, MD
Professor of Psychiatry at Tulane University School of Medicine in New Orleans, and a long-standing researcher on post-traumatic stress — including the diagnostic criteria used for it in young children. He is not an outsider taking aim at a popular book. He is a trauma researcher assessing claims made in his own field, which is what gives the evaluation its weight and also what makes his own position on the debate worth knowing about.
What we read: Scheeringa, BJPsych Bulletin (2025), alongside the Felitti and Anda adverse childhood experiences cohort and van der Kolk’s own randomised trial of trauma-informed yoga.¹²³
Where to read his work: Tulane School of Medicine · ResearchGate · Psychology Today
Scope and method of this review
This review covers the claim that traumatic experience is physically stored in the body and must be released through the body. It examines the neurobiological evidence offered for that claim, the treatment evidence for body-based approaches, and the distinction between the two.
Sources were selected as follows: the 2025 evaluation, which is to date the most direct published assessment of the claims; the primary cohort study most often cited as evidence for the adversity-and-health link; and the randomised trial of body-based treatment conducted by the author of the book himself, on the grounds that a claim is best tested against its strongest available support rather than its weakest.
What this review does not cover. It does not assess whether trauma is real or whether it has consequences — it is and it does. It does not evaluate PTSD treatment guidelines or recommend between therapies, which is a clinician’s decision made with a person in the room. It does not cover children, where the evidence base and the developmental questions are different. It does not assess any individual practitioner, method or training programme. And three of the rows in the table below are characterised through the 2025 evaluation rather than extracted independently from the underlying studies, which is stated in each limitation cell.
Two claims wearing one sentence
“Trauma is stored in the body” is doing two jobs at once, and separating them settles most of the argument.
The first claim is about consequences. Prolonged adversity and chronic stress leave physical marks — on sleep, on the immune system, on cardiovascular risk, on how readily the body mounts a stress response. This claim is well supported and is not seriously disputed. The original adverse childhood experiences cohort of 8,056 adults found a graded relationship between the number of childhood adversities reported and adult health outcomes, with smoking rising from 7% among those reporting none to 17% among those reporting four or more, and severe obesity from 5% to 12%.²
The second claim is about mechanism and location. It says the experience itself is retained in tissue — in fascia, in muscle, in the nervous system as an encoded record — and that this record must be physically discharged before recovery can happen. This is the claim that fails.
Definition: allostatic load is the cumulative physiological cost of repeatedly mounting a stress response and not fully standing down between times. It is the mechanism doing the honest work that the storage metaphor is doing dishonestly: a running cost rather than a sealed deposit.
The 2025 evaluation is direct about the second claim. Cross-sectional neuroimaging studies do show differences between people with and without PTSD, but such designs cannot establish which came first. Where prospective designs exist — three literature reviews, the most recent identifying 25 such studies — the differences frequently appear to precede trauma exposure, which points towards pre-existing vulnerability rather than damage inflicted by the event.¹ The claim that trauma directly causes obesity, heart disease and autoimmune conditions is not supported by causal evidence. The epigenetic account is characterised by small effect sizes, unclear mechanisms and replication failures. The stress-hormone account is described as inconclusive, with early findings proving difficult to replicate.¹
None of that means nothing happened to you. It means the physical marks of what happened are the ordinary marks of prolonged strain, not a sealed deposit of the event itself waiting to be extracted.
If the honest answer is that the load has been too heavy for too long, the Quiet Audit is ten private minutes on exactly that. No programme, no advice, nobody sees your answers. Start here.
The evidence table
| Source | Sample / scope | Finding | Limitation |
|---|---|---|---|
| Scheeringa, 2025¹ | Evaluation of the neurobiological and treatment claims in The Body Keeps the Score | Several central claims unsupported. Trauma-as-damage, trauma-causes-disease, and body-based-only-processing all judged unsupported by causal evidence | An evaluation by one author in a journal’s bulletin, not a systematic review or meta-analysis with a registered protocol |
| Prospective neuroimaging reviews¹ | 3 reviews; the most recent identifying 25 prospective studies | Differences often precede trauma. Consistent with pre-existing vulnerability rather than acquired damage | Characterised via the 2025 evaluation rather than extracted here; prospective studies remain a small minority of this literature |
| Felitti et al., 1998² | 8,056 adults (from 13,494 surveyed), aged 19–92 | Graded dose-response. Smoking 7% to 17%, severe obesity 5% to 12% across the range | Retrospective self-report is prone to recall bias; sample 79% white and 43% college-educated; the design cannot establish causation |
| van der Kolk et al., 2014³ | 64 women (60 completers), mean age 42.9, chronic treatment-resistant PTSD | Yoga outperformed the control. CAPS d = 1.07 vs 0.66; 52% (16/31) vs 21% (6/29) no longer met criteria | Well-educated women with treatment-resistant PTSD only; 10 weeks with no follow-up; components not dismantled, so the active ingredient is unidentified |
| Epigenetic mechanism literature¹ | Field-level characterisation | Weak. Small effect sizes, unclear mechanisms, replication failures | Characterised via the 2025 evaluation; no pooled estimate is offered and none is claimed here |
| Stress-hormone mechanism literature¹ | Field-level characterisation | Inconclusive. Early findings proved difficult to replicate | As above. See also the separate review of cortisol measurement, where seven studies point in three directions |
Read the finding column downward. The consequences are real. The storage mechanism is not established. Those two statements are compatible, and holding both of them is the accurate position.
Key terms
Cross-sectional — a study that measures a group at one point in time. It can show that two things occur together. It cannot show which came first, which is why almost every finding in this area carries an asterisk.
Prospective — a study that measures people before an event and follows them afterwards. Far more expensive, far rarer, and the only design that can tell you whether a brain difference is a consequence of trauma or was there beforehand.
Effect size (Cohen’s d) — how large a difference is, expressed independently of sample size. Roughly: 0.2 small, 0.5 medium, 0.8 large.
Somatic — simply “of the body”. In therapy contexts it describes approaches that work through physical sensation and movement rather than through talking alone.
Why the practices work anyway
This is the part that gets lost when the mechanism is questioned, and it matters more than the questioning.
The strongest single trial of a body-based approach was run by the author of The Body Keeps the Score. Sixty-four women with chronic, treatment-resistant PTSD, mean age 42.9, were randomised to ten weekly hours of trauma-informed yoga or to a supportive women’s health education group.³ This was an active comparison, not a waiting list, which is the harder test. The yoga group showed a large effect on clinician-assessed symptoms (d = 1.07); the comparison group showed a medium-to-large one (d = 0.66). By the end, 52% of the yoga group and 21% of the comparison group no longer met diagnostic criteria.
That is a real result, and it is worth taking seriously alongside its limitations: ten weeks, no follow-up, a narrow and well-educated sample, and no dismantling study to identify which component did the work.
Notice what it does not demonstrate. It does not show that anything was released from tissue. Regulated breathing, predictable movement, an hour of attention that is not a demand, and the ordinary effect of being in a room with people in the same situation are all plausible mechanisms, and none of them require the storage claim. The comparison group improved too — considerably. Something about structured weekly attention helps, and yoga added to it.
The practical consequence is straightforward. If the practice helps you, the finding that its explanation is shaky is not a reason to stop. It is a reason not to pay a premium for the explanation.
How strong is this evidence?
Contested. The literature does not agree with itself, and the disagreement is partly about evidence and partly about what the words are claiming.
Well established: that prolonged adversity is associated with worse adult physical health in a graded way,² and that at least one body-based intervention outperformed an active control in a randomised trial.³
Not established: that traumatic experience is retained in bodily tissue; that trauma causes the physical diseases attributed to it in a demonstrated causal sense; that epigenetic or stress-hormone mechanisms explain the effects; or that body-based approaches are the only ones capable of processing trauma.¹
Also worth saying: the 2025 evaluation is a single-author assessment in a bulletin, not a registered systematic review, and its author has published a book of his own arguing a related case. That does not make it wrong. It does mean it should be read as a serious argument from inside the field rather than as a settled verdict, and this review treats it that way.
The limitation that matters most for the reader of this page: almost none of this work was done on women in their late forties, fifties and early sixties carrying ordinary accumulated strain rather than diagnosed PTSD. The yoga trial recruited women with chronic treatment-resistant PTSD. If your question is whether twenty years of over-functioning is “stored” somewhere, this literature was not built to answer it, and anyone who tells you it was is extrapolating.
Five things that look like this and are not this
The sections above protect the literature. This one is for the reader, and it exists because attributing an unexamined physical symptom to stored trauma is one of the more reliable ways a treatable condition goes unfound.
Five ordinary explanations produce the tight shoulders, the racing heart and the exhaustion that the storage story is usually invoked for.
- An overactive thyroid. A fast or irregular heartbeat, heat intolerance, tremor, weight loss. One blood test, and it is misread as anxiety constantly.
- Iron deficiency. Breathlessness on stairs and a flatness that a good weekend does not shift. Common in the years before the final period.
- Obstructive sleep apnoea. A body that never fully stands down at night produces a body that never fully stands down in the day. Somebody else has usually noticed the breathing first.
- An anxiety disorder or depression. Both are treatable, both are common, and neither requires a trauma history to explain it — which is worth saying, because “something must have happened to me” is a heavy and often unnecessary search.
- An unexamined physical symptom. Persistent pain, palpitations or gut symptoms that have not been assessed belong with a clinician before they belong with a framework of any kind.
None of these is excluded by a difficult history, and a difficult history does not exclude any of them. They are questions to raise in order, not a list anyone can clear for themselves.
Who else has assessed this
A single evaluation is a single evaluation. Three other bodies of work bear on the same question from different directions.
Vincent Felitti (site · faculty bio) and Robert Anda (site · LinkedIn), working with Kaiser Permanente and the US Centers for Disease Control, produced the cohort that established the association everybody cites — and stated in the paper itself that the retrospective, cross-sectional design cannot establish causation.² The finding is genuinely important. It has been overstated relentlessly since, usually by people who did not read that sentence.
Bessel van der Kolk (Wikipedia · site · Amazon author page) himself ran the randomised trial that gives body-based treatment its best evidence.³ It is a real result against an active comparison, and it is the strongest card in the hand of the position this review is questioning — which is exactly why it is here rather than omitted.
The prospective neuroimaging literature, summarised across three reviews with the most recent identifying 25 studies, is the part that does the actual damage to the storage claim, because it is the only design that can order the events.¹ The finding that differences often precede exposure is the single most important line in this review.
An epidemiological cohort, a treatment trial, a prospective imaging literature and one critical evaluation. They converge on a split verdict: the consequences hold, the mechanism does not.
The marks are real. The vault is a metaphor.
What this actually changes
Not whether something happened. Not whether it cost you. Neither of those was ever in question here.
What it changes is what you are looking for. Under the storage model, recovery is an excavation: there is a thing inside you, it must be found, and if you are still tired then you have not found it yet. That framing quietly converts exhaustion into a personal failure of technique, which is the last thing a woman who has been over-functioning for twenty years needs added to the pile.
Under the running-cost model, there is nothing to find. There is a bill that has been accruing, and the useful questions are what is still charging it and what has been paid down. That is duller. It is also actionable this week, and it does not require you to produce a memory to justify how you feel.
When this belongs with a doctor
A framework is not a substitute for an assessment. Speak to a clinician rather than reading further if intrusive memories, nightmares or flashbacks are interrupting daily life; if avoidance is narrowing where you go or what you do; if there are persistent physical symptoms — pain, palpitations, gut symptoms — that have not been examined; or if low mood is new or worsening.
Effective treatments for trauma exist and are well studied. Which one suits a particular person is a clinical decision, not one to be settled by a book, a class, or an evidence review. If you are having thoughts of harming yourself, that is a reason to speak to someone today rather than to keep reading — a GP, an emergency service, or a crisis line in your country.
What to say at the appointment, and what you will probably hear back
The difficulty here is unusual: the vocabulary available to describe this is mostly borrowed from a book. Three lines that are not.
Short enough to read off a phone screen
Say: “I have been running at capacity for years and my body has stopped standing down. I would like the physical side checked and a view on whether a talking therapy is appropriate.”
If you hear “have you tried yoga”: “I might, and I would still like thyroid and ferritin checked first, so we are not treating a deficiency with breathing.”
If a therapy is being chosen: “What does the evidence support for what I am describing, and what would tell us in eight weeks that it is not working?”
If what you want to bring into that appointment is a clear account of what the last few years have actually contained, the Quiet Audit is ten private minutes that produce exactly that page. Nobody sees your answers. Start here.
The one move
The one move
Before starting any body-based practice, ask three questions in order. Is it sold as a practice or as a treatment? What was it compared against — nothing, a waiting list, or another group doing something structured for the same hour? And what would tell me in eight weeks that it is not working?
A practice with no failure condition is not being evaluated. It is being believed.
And a smaller move, for the sentence itself. When you next hear that your body is keeping the score, try the more accurate version: the strain of the last several years has had physical costs, and those costs are real whether or not anyone has named them. It is less quotable. It is also true, it does not require you to excavate anything, and it does not imply that the reason you are still tired is that you have failed to release something correctly.
Questions this review is asked
Is trauma literally stored in the body?
No evidence establishes that traumatic experience is retained in bodily tissue as a recoverable record. The 2025 evaluation found the neurobiological claims behind that idea unsupported by causal evidence, with prospective studies suggesting brain differences often precede trauma exposure rather than follow it.¹
So is The Body Keeps the Score wrong?
Partly. Its central observation — that adversity has physical consequences and that treatment should attend to the body — is defensible. Its specific neurobiological and mechanism claims, and the claim that only body-based approaches can process trauma, were judged unsupported.¹
Does that mean somatic therapy and trauma-informed yoga do not work?
No. A randomised trial of 64 women found trauma-informed yoga outperformed an active comparison group, with 52% versus 21% no longer meeting PTSD criteria after ten weeks.³ The practice can work while the explanation offered for it remains unproven. Those are separate questions.
What about the ACE study — does that not prove trauma causes illness?
It shows a graded association between reported childhood adversity and adult health outcomes across 8,056 adults.² It is retrospective, cross-sectional and self-reported, and its own authors note the design cannot establish causation.
Why does the phrase feel so true if the mechanism is not established?
Because the experience it describes is real. Sustained strain does produce tight shoulders, disrupted sleep and a body that reacts before you have decided anything. The metaphor names something accurately. It is the literal mechanism attached to it that outruns the evidence.
If it is not stored, why am I still exhausted years later?
Usually because the load has not actually reduced, only changed shape. That is a duller explanation than storage and considerably more actionable — distinguishing a life cause from a medical one is the useful next question.
Informational, not medical advice. Trauma symptoms that interfere with daily functioning warrant a clinician.
Where to go next
In order, and each for a reason.
1. The responsiveness test — how to tell a life cause from a medical one, which is the question underneath most of this.
2. Then what a standard panel checks and what it never looks at, because four of the five alternatives above live there.
3. Then what a cortisol test can and cannot tell you, since the stress-hormone story is the mechanism most often offered next.
You do not have to excavate anything. You have to put something down.
References
1. Scheeringa, M. S. (2025). Evaluating evidence behind popular trauma narratives: neurobiological and treatment claims in The Body Keeps the Score. BJPsych Bulletin. Read the evaluation
2. Felitti, V. J., Anda, R. F., Nordenberg, D., et al. (1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study. American Journal of Preventive Medicine, 14(4), 245–258. Read the paper
3. van der Kolk, B. A., Stone, L., West, J., et al. (2014). Yoga as an adjunctive treatment for posttraumatic stress disorder: a randomized controlled trial. Journal of Clinical Psychiatry, 75(6), e559–e565. Read the trial
4. Frontiers in Systems Neuroscience (2026). The body does not keep the score: trauma, predictive coding, and the restoration of metastability. Read the paper — cited as an example of an alternative model now being proposed, not as established evidence.
On the researcher. Michael Scheeringa is a psychiatrist and trauma researcher at Tulane, not your clinician, and nothing in his published work is medical advice for an individual. He has written a book of his own arguing a related case about how trauma is understood, which is a relevant interest and is disclosed here rather than left for you to discover. This review cites him for what his 2025 evaluation states about specific studies, each of which can be checked at the links above.
Sourcing, disclosure and disclaimers
This is not medical advice. Blue Leaf Journal publishes general information for a general readership. Nothing here is a diagnosis, a treatment recommendation, or a substitute for care from a clinician who knows your history. Nothing here should be used to decide whether to start or stop a therapy.
We are not clinicians. Blue Leaf Journal is an independent publication. We read published research and translate it. The findings belong to the researchers and institutions named and linked above; the plain-English rendering is ours, and so is any error in it.
No affiliation and no endorsement. Blue Leaf Journal is not affiliated with Michael Scheeringa, Bessel van der Kolk, Vincent Felitti, Robert Anda, Tulane University, the Centers for Disease Control, Kaiser Permanente or any trauma-therapy training body. None of them has reviewed, approved or endorsed this article, and none of them is responsible for it.
No commercial relationship. Nobody named above paid for or was paid for this coverage. There are no affiliate links, no sponsored placements and no gifted products in this article. We do not sell therapy, training or courses in any modality discussed here, and we earn nothing from any of them.
How this was checked. Every figure above is drawn from the primary sources, each linked in the references, and can be verified there. Where a claim is characterised through the 2025 evaluation rather than extracted from the underlying studies, the table says so in the limitation cell. Sources were checked on 27 August 2026. If you find something we have got wrong, write to miriamalderton@blueleafjournal.com and we will correct it and say that we did.
Miriam Alderton is Research Editor at Blue Leaf Journal. She reads the methods section first and the abstract last, and every figure in this piece is linked to its source above.
Written by Miriam Alderton, Research Editor, for Blue Leaf Journal. Updated: 27 August 2026.
