I have lost count of how many times I have heard the following question: where in the body do you feel this? I have heard it in supervisory contexts, in case consultations, in workshops, and in client transcripts shared by colleagues. It is one of the most commonly asked questions in contemporary psychotherapy. I have been skeptical of its efficacy towards incisive therapeutic inquiry. With respect to the question in question... I have my own questions:

What is "the body"?

What is "the score"?

How exactly is the latter kept in or by the former?

For some clients, the question is effective. For some clients, it produces nothing. For some clients, it makes things worse. And the therapists asking it cannot, when pressed, say why the question succeeds in one case and fails in another. It is asked because... it is what one asks. It is the move many somatic training courses teach trainees to make.

I want to write about this because something has shifted, recently, in the discourse inside which the question lives. In April of this year (2026), a paper appeared in Frontiers in Systems Neuroscience with a title designed to provoke: "The body does not keep the score: trauma, predictive coding, and the restoration of metastability."1 The authorship is striking. Karl Friston — the originator of the free energy principle, the most influential figure in contemporary computational neuroscience — sits on the byline, along with Steven Kotler, Michael Mannino, and Glenn Fox. The paper takes direct aim at Bessel van der Kolk's The Body Keeps the Score,2 the book that has shaped how a generation of therapists thinks about trauma.

The paper has begun to circulate and the analyses are following. EMDR International is hosting it. Psychology Today has a response up. Yoga teachers are writing blog posts about what it means for their work. The somatic-therapy world is, for the first time in more than a decade, being asked to consider that perhaps... the metaphor at the centre of its identity might be wrong.

I want to say what I think the paper has gotten right, what it has gotten half-right, and what it is missing entirely. And I want to say what I think this means for the question I started with.

The paper's argument

In ordinary language, Friston et al. argue the following: the brain does not store trauma in non-innervated tissue. There is no archive in the muscle, in the fascia, in the hip. The persistence of trauma is the persistence of a prediction loop. The brain, attempting to make sense of the world, comes to assign excessive confidence to threat: that is, it predicts danger, senses the body's arousal, and then takes that arousal as confirmation that danger is present, and predicts danger again. In this way, "the score" is not a score at all. It is, instead, an artefact of the manner in which a person's experience of badness is elevated by the experience of being in a loop.

Let me offer an example: think of phantom limb pain. The limb is really gone. The amputation has been performed. The loss is real. The body map has changed. But the pain the patient feels is not the lost limb sending signals, because the limb is not there to send anything. The pain is being produced by a brain whose model of the body has not updated, generating sensations as if the limb were still there and in distress. The fact that there is no limb to send signals to the brain means that the suffering — which is real! — is a production of one's brain and not one's arm. It is what the brain is doing now with what happened then. The Frontiers authors are arguing that post-traumatic suffering has this same structure. To recap:

  1. The events were real.
  2. The current suffering is real.
  3. But the suffering is not stored in the body or kept on any internal scoreboard.
  4. Suffering is what the brain is currently producing with the events of the past towards a prediction of the future.

The authors' answer to the loop is what they call metastability: i.e. the brain's capacity to switch fluidly among semi-stable states and to remain flexible enough that no single basin of fear can capture it for long. They propose that flow states (Kotler's territory) restore metastability through novel sensorimotor feedback. The brain, given new evidence, updates its model and recalibrates its predictions.

I take the empirical claim seriously. The neuroimaging evidence they cite is real and accumulating. The free energy framework that underlies it is the most sophisticated computational account of brain function we have. And the central empirical move — i.e. what looks like storage is, in fact, recurring prediction — is a substantive correction to a popular metaphor that has done real damage in clinics by encouraging therapists to look for buried material that, in many cases, was never buried in the first place. This is the foundational Freudian mistake, first made in *Studies on Hysteria in 1895... still alive and kicking today.

So far, so good. I am on board. I have been flogging this predictive horse for many years now.

The trouble is what it leaves untouched.

click on the hand There to submit an intake form

The metaphor remains alive and well

The paper claims the body does not keep the score. It does not say there is no score. It moves the score from the body to the brain's predictive apparatus. The score is now an artefact of circular inference rather than an inscription in tissue, but it is still the score: a definite, singular thing that someone is still said to be keeping.

This is not a small point. It is the spine of the critique I have been working on for the past number of years.

When van der Kolk wrote The Body Keeps the Score, he made two metaphysical decisions before any empirical question had been asked. He defined a thing called "the body," a gathering up of the actual, variable, individual, gendered, racialised, disabled, microbially-populated, environmentally-embedded bodies of his actual patients into a single unitary category, from a god's-eye vantage that none of us actually occupy. And he defined a thing called "the score", a definite, singular object that the body was said to be keeping. The book's central metaphor would not be sayable without those two prior definitisations.

When the Frontiers authors argue that the body does not keep the score, they retain both. They argue about where the score is kept and about who is doing the keeping. They do not contest that there is a definite thing called the score. They do not contest that "the body" is a category that gathers up actual bodies into a unitary Platonic object. They have performed a correction within a metaphysics they share with the object of their critique. They and van der Kolk are singing from the same hymn book.

This is what I want somatic therapists reading this to notice. The most sophisticated computational-neuroscience critique of the van der Kolk frame does not break the frame. It only changes the picture. The deeper structure — the metaphysics of presence,3 the assumption that the work of clinical inquiry is to identify what is there and where there is — survives the correction intact. This is normal. It is the way disciplines refine themselves while leaving their foundations untouched. Copernicus did not solve Ptolemy's cosmology by adding further epicycles; he recognised that the proliferation of new epicycles was itself the sign of a failing framework, and proposed a different one.

What is absent

Here is what I find most interesting about the Frontiers paper, and what I think is most important for working therapists to see.

The paper names five mechanisms by which the post-traumatic brain fails. Read carefully and you will find them:

  • Precision-weighting of danger priors becomes excessive — the brain assigns too much confidence to predictions of threat.
  • Sensory attenuation fails — the brain loses the capacity to downweight interoceptive signals that ought to be treated as noise.
  • Top-down regulation breaks — prefrontal systems can no longer suppress or update threat representations.
  • Executive precision collapses — working-memory-supported cognitive control gives way.
  • Metastability is lost — the brain cannot fluidly leave the basin it has fallen into.

These are five names for one function. Each describes a capacity to make signal absent: i.e. to downweight, to suppress, to attenuate, to dampen. Each describes what does not happen when the function fails. The function in question is what I have, in my own work, called inhibition. I use that term not in the popular sense (holding yourself back, being repressed, being uptight) but the technical sense: the active, energetically costly process by which an organism selectively dampens competing signals so that some processes can dominate while others recede.

What I have called inhibition or inhibitory tone is not incidental or optional. Literally everything depends on this capacity: every cognition, every instance of regulating one's emotions, every act of attending to something. To attend to one thing is to inhibit all others... all... which I have come to call infinity minus one. To predict one thing is to inhibit competing predictions. To feel one feeling is to inhibit the others. Inhibition is not a feature of mental life; it is the very precondition of it.

The Frontiers paper has named the mechanism in five vocabularies without recognising that the five names point to one function. This is not an accident. The discipline within which the paper sits — the natural-scientific tradition that animates predictive processing, computational neuroscience, and psychiatry alike — is constitutionally blind to absence. It can name presences (firing, activation, prediction, signal), and it can name absences only insofar as they show up as presences (a measurable attenuation, a quantifiable suppression). It cannot see absence as a unified family of phenomena, because its base epistemology cannot make absence into its own category of inquiry. The five names exist because each one snuck absence into the conversation under cover of describing something positive. The unification cannot happen within the frame because the frame is what prevents it.

What I am describing is the deep continuity with van der Kolk. Both van der Kolk and the Frontiers authors are working within a tradition that can only see what is there, i.e. presence. They disagree about which presences matter, i.e. bodily traces or neural predictions, but they share the unstated assumption that the work is to identify presences. No doubt they would not be thrilled with my assessment, but... what they share is bigger than the thing about which they disagree.

The question...

Which brings me back to where I started.

When a therapist asks a client, "Where in the body do you feel this?"... what, mechanistically, is happening?

For some clients, the question might be regulatory. Some people do not hear their bodies well: what we might call under-developed interoceptive attention. The question slows them down. It directs them to practice finer-grained sensory attention, which they can use to update over-generalised predictions ("I am bad," "I am unsafe," "I am wrong"). For these clients, the question is doing useful work, and the framework I am sketching here can give a perfectly clear account of why.

For other clients, the question might be dysregulatory. Some people hear their bodies too well: what we might call over-precise interoceptive precision-weighting. They are already over-attending to bodily signals. Every interoceptive shift is read as confirmation of catastrophe. To ask such a client to attend more closely to the racing heart is to deepen the loop that is the problem. The question can make things worse, and the framework I am sketching here can give a perfectly clear account of why.

These are different failure modes. They require different clinical responses. One client needs the precision dial turned up; the other needs it turned down. The same question, asked of two clients both seeking therapeutic support and insight, will help one and harm the other.

The current vocabulary of the somatic-therapy world cannot make this distinction. It has informal heuristics — "titrate," "stay within the window of tolerance," "go slow" — but it does not have a theoretical apparatus that names what the difference is between the client for whom interoceptive attention is regulatory and the client for whom it is not. The theory points everyone at the same thing, because the theory has a single (incorrect) answer for the question of trauma: i.e. it lives in the body, look for it in there. The question gets asked reflexively, universally, regardless of whether interoceptive attention is the right tool for this client at this moment.

This is what I have been frustrated about for many years. It is not that any individual somatic therapist is doing anything wrong. Most of them are kind, devoted, and clinically careful in ways that compensate for the theoretical paucity of the frame they have been handed. What I object to is the paucity itself. The seriousness of psychotherapy requires being able to ask the following question in each and every instance: what is actually happening when this person is sitting across from me, and what does it require of me? That question cannot be answered with a single reflex move. It requires a theoretical vocabulary that can make differential distinctions. The current vocabulary cannot.

The inhibitory hypothesis

What I have been calling the inhibitory hypothesis is, in its simplest form, the claim that the conditions we currently treat as separate psychopathologies — anxiety, OCD, ADHD, PTSD, and others — are expressions of a common underlying dynamic. That dynamic is the insufficiency of inhibitory function. Not too much of anything. Not buried material. Not a problem with what is there. A problem with what cannot be made absent.

This reframes, for me, what therapeutic work is. It is not excavation. It is not the recovery of buried tissue-memory. It is not the discharge of stored emotion. It is the cultivation of conditions under which inhibitory tone can develop and be sustained. Different clients need different cultivations. The framework gives a vocabulary for naming which cultivation is required for which client.

I have written this argument out at considerable length in a longer work, On Inhibition: Notes from Undisciplined Practice, which is currently out to readers for comment. The full story is beyond what I can offer here. This essay is an invitation to consider that the discussion currently taking place around the Frontiers paper might be more interesting and more clinically useful, if it ran the question all the way to the end.

The Frontiers authors have done useful work. They have shown that “the body” is not an archive. They have not yet shown — perhaps because their tradition prohibits it — that what they are pointing at, when they describe the post-traumatic brain, is inhibitory insufficiency. The score is not kept. There is no score! There is only the recursive loop of an organism whose inhibitory tone is unequal to the activity it is being asked to modulate... which is why trying harder is rarely the answer to these sorts of questions. Trying harder, in this light, ensconces a person ever more tightly in their loop.

When that is the question, the work changes. And the question one asks at the start of a session — where in the body do you feel this? — stops being a reflex and starts being a clinical judgement made fresh, each time, for this particular person, in this particular session, with this particular history. Which is what good therapy has always required, and what no theoretical apparatus has yet quite given the field permission to do.

If you would like to keep up with my work on this topic, please have a look on my site at my page dedicated to On Inhibition.

Footnotes

  1. Steven Kotler, Michael Mannino, Glenn Fox, and Karl Friston, "The body does not keep the score: trauma, predictive coding, and the restoration of metastability," Frontiers in Systems Neuroscience 20 (April 30, 2026): 1812957, https://doi.org/10.3389/fnsys.2026.1812957. ↩
  2. Bessel A. van der Kolk, The Body Keeps the Score: Brain, Mind, and Body in the Healing of Trauma (New York: Viking, 2014). ↩
  3. The phrase metaphysics of presence I take from a longer line of twentieth-century philosophical critique running through Heidegger, Derrida, and most usefully for my purposes, Bruno Latour. Latour's We Have Never Been Modern (Cambridge: Harvard University Press, 1993) and Pandora's Hope (Cambridge: Harvard University Press, 1999) describe the operation by which the natural sciences purify their objects — gather them up, definitise them, strip them of context — in order to make them tractable to investigation. The cost of the purification is that what gets investigated is no longer the thing one started with. The same operation, I argue in On Inhibition, structures the way psychiatry and clinical psychology have constructed their diagnostic objects. ↩
click on the hand There to submit an intake form