The Body Keeps the Score: What the Science Actually Says in 2026
By Dr. Anindo Mitra · MBBS, MD Psychiatry (JIPMER) · Consultant Psychiatrist, Athena Behavioural Health, Gurugram
TL;DR
Bessel van der Kolk's 2014 book The Body Keeps the Score argued that trauma is stored in the body, which is why it drove therapies like EMDR, yoga, and movement work.
A 2026 paper by Kotler, Mannino, Fox, and Friston reframes trauma as a prediction problem rather than a storage problem.
The brain gets stuck assigning too much confidence to the belief that danger is still present, and it reads ordinary body sensations as proof that belief is correct.
The neuroscience van der Kolk leaned on, particularly the amygdala and prefrontal cortex findings, still holds up well.
This is a theoretical reframe built from existing evidence, not a study that overturns the book. Both models end up supporting many of the same treatments.
For nearly a decade, one idea has shaped how millions of people think about trauma: that it lives in the body, waiting to be released. That idea comes from Bessel van der Kolk's bestselling book, and it has quietly become common knowledge in therapy rooms and Instagram captions alike. In 2026, a group of neuroscientists published a paper with a title built to get attention: the body does not keep the score. This piece walks through what the book actually claimed, what the new paper argues instead, and where the two views agree more than the headlines suggest.
What Did The Body Keeps the Score Actually Argue?
Van der Kolk's central claim is that trauma disconnects the brain's rational systems from its emotional ones during a traumatic event, leaving behind fragments rather than a coherent memory. What remains are sensations and physiological reactions that resurface long after the danger has passed, which is why he argued that talk therapy alone often falls short and body-based approaches are necessary.
Published in 2014, the book built its case around a specific idea: that overwhelming experience gets held somatically rather than narrated normally, and that the amygdala keeps reacting to danger that is, objectively, over. This is the foundation for treatments like EMDR, yoga, and movement-based therapy, all of which aim to work through the body rather than only through conversation. The book's popularity turned a clinical hypothesis into a cultural shorthand, one that a 2026 opinion piece in Frontiers in Systems Neuroscience directly engages with as it has shaped public and clinical understanding of trauma for nearly a decade.
Why Are Scientists Pushing Back Now?
A 2026 paper by psychologist Steven Kotler, along with Michael Mannino, Glenn Fox, and neuroscientist Karl Friston, argues that trauma is better explained as a disorder of prediction than a disorder of storage. The paper does not deny that trauma affects the body. It argues the mechanism connecting mind and body has been misdescribed for a decade.
The paper was published as an Opinion piece in Frontiers in Systems Neuroscience in April 2026 , which matters for how it should be read. An Opinion piece is a theoretical synthesis of existing research, not a new clinical trial. The authors are proposing a different lens on evidence that already exists, built on Karl Friston's free energy principle, a framework he has spent decades developing in computational neuroscience.
The Free Energy Principle, Explained Simply
The free energy principle treats the brain as a prediction machine that is constantly generating beliefs about the world and updating them to reduce the gap between what it expects and what it actually senses. Under this model, perception is not passive recording. It is active guessing, corrected in real time.
Applied to everyday life, this explains why a familiar hallway feels unremarkable but an unfamiliar dark alley feels charged with alertness even before anything happens. The brain has already generated a prediction, and it is scanning incoming signals to check whether that prediction holds. Most of the time, this system runs so smoothly that people never notice it is happening at all.
How Predictive Coding Reframes Trauma
Under the predictive coding model, trauma is what happens when the brain assigns excessive, almost unshakeable confidence to the prediction that danger is still present, long after the original threat has ended. The paper's authors use an image from the free energy landscape itself: a ball rolling into a steep ravine that becomes progressively harder to climb out of the longer it stays there.
This reframing changes where the "problem" is located. It is not that the trauma is trapped in the muscles or the nervous system waiting to be released. It is that one prediction, the belief that threat is ongoing, has become so dominant that it overrides more accurate, more current information. The nervous system behaves like a control system that keeps comparing incoming signals to that old prediction and acting to close the gap , which is a fundamentally different starting point than storage.
The Body as Messenger, Not Archive
A racing heart or a tight chest after a trauma reminder is not proof that the trauma is stored inside that tissue. According to this model, it is the brain reading an ordinary stress response as confirmation that its danger prediction is correct, a kind of circular reasoning that keeps the prediction locked in place.
This is a subtle but important shift. The body still reacts, and those reactions are real and often distressing. What changes is the explanation for why they keep happening. The body is described in the paper as a messenger carrying signals the brain has already decided to interpret as danger, not an archive holding onto the original event.
What Van der Kolk Got Right
The neuroimaging findings van der Kolk built much of his argument around remain well supported. People with PTSD reliably show heightened amygdala activity alongside reduced activity in the medial prefrontal cortex, the region that normally helps regulate fear responses, and this pattern has been replicated across multiple studies since it was first reviewed nearly two decades ago .
This matters because it means the new paper is not dismantling the clinical picture van der Kolk described. People with trauma histories genuinely do show a nervous system that stays on alert. Where the disagreement lies is narrower and more technical: is that alert state the result of trauma being stored somewhere in the body, or is it the downstream output of a brain that has locked onto an inaccurate prediction and keeps defending it. The symptoms are not in dispute. The mechanism is.
Where the Critique Lands
Beyond the storage-versus-prediction disagreement, the 2026 paper also points out that van der Kolk's book underengages with research on false memory and doesn't fully account for how memory recall can be reconstructive rather than a literal replay of stored events. It also revisits Antonio Damasio's convergence-divergence zones, a concept van der Kolk himself cites, which describes neural hubs that store the pattern of a memory or emotion and can replay it contextually.
That last point is worth sitting with. Even the brain-based model of emotional memory that van der Kolk draws on is a model of pattern storage in neural tissue, not evidence that memories live outside the brain in muscles or organs. The critique, in other words, is not that trauma has no bodily footprint. It is that the mechanism connecting brain and body has likely been oversimplified in translation from research to bestseller.
What Does "Restoring Metastability" Even Mean?
The paper's title includes a specific term worth unpacking: metastability, the brain's normal ability to flexibly shift between different network states depending on context. Trauma, in this model, locks the brain into one rigid state instead of allowing that flexibility.
This isn't a purely speculative claim. Researchers studying traumatic brain injury have shown that damage to structural brain connectivity reduces metastability, and that this reduction tracks closely with reduced cognitive flexibility . The 2026 paper borrows this concept and applies it to trauma more broadly, arguing that a trauma-affected brain isn't damaged in the same physical sense but behaves as though it is stuck in a narrow, repetitive pattern of threat detection.
So Is the Body Keeping the Score or Not?
Both models agree the body reacts strongly to trauma reminders and that this reaction is real, not imagined. Where they differ is on what that reaction means: van der Kolk frames it as stored trauma resurfacing, while the predictive coding model frames it as a brain defending an outdated threat prediction using the body's signals as evidence.
The distinction changes what treatment is aiming at. If trauma is an entrenched prediction rather than trapped material, then treatment is less about releasing something and more about giving the brain new, safer evidence it can actually update on. That reframing is where the more hopeful, actionable part of this research lives, and it points toward interventions that were not the primary focus of the original book.
Frequently Asked Questions
Does this new research mean The Body Keeps the Score was wrong?
Not exactly. The 2026 paper agrees with much of what van der Kolk observed clinically and with the neuroimaging evidence he cited. What it challenges is the explanation for why the body reacts the way it does, proposing a prediction-based mechanism instead of a storage-based one.
Is the 2026 Frontiers paper a new clinical study?
No. It is an Opinion piece, a theoretical synthesis built on existing research rather than a new experiment or clinical trial. It proposes a framework for interpreting evidence that already exists, using Karl Friston's free energy principle.
Does this change how trauma should be treated?
Not immediately. The paper argues its framework may explain why very different therapies, including EMDR, exposure therapy, and mindfulness, all show benefit, because they may work by helping the brain update an outdated danger prediction rather than by releasing stored material.
What does it mean to say the body is a "messenger, not an archive"?
It means physical sensations like a racing heart are treated as signals the brain is currently generating and interpreting, not as evidence of a stored memory sitting inside body tissue. The sensations are real; the explanation for their origin is what has shifted.
Should I stop body-based trauma therapies like yoga or EMDR based on this paper?
No, and the paper's authors do not argue for that either. These therapies remain evidence-supported. The paper offers a different explanation for why they may work, not a reason to discontinue them. Always discuss any change to your treatment plan with your own clinician.
About the author: Dr. Anindo Mitra is a Consultant Psychiatrist at Athena Behavioural Health, Gurugram, holding an MBBS and MD in Psychiatry from JIPMER, Puducherry. His clinical practice covers trauma-related disorders, anxiety, depression, and rational psychopharmacology, and he writes on evidence-based psychiatry at dranindomitra.com.
This article is for general education and does not constitute individualised medical advice. If you are struggling with symptoms of trauma or PTSD, please consult a qualified mental health professional.
