Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 18 of the 20 references it cites were found in a registry or an index; 6 of the 18 sources its claims cite were checked for support.

The body does not keep the score: trauma, predictive coding, and the restoration of metastability

Open access
Kotler, Steven, Mannino, Michael, Fox, Glenn, Friston, Karl · 2026 · Frontiers in Systems Neuroscience

Abstract

The Body Does Not Keep the Score: Trauma, Predictive Coding, and the Restoration of Metastability By Steven Kotler, Michael Mannino, Glenn Fox, Karl Friston For nearly a decade, the idea that "the body keeps the score" has shaped public and clinical understanding of trauma. It is an enticing metaphor-implying that experience is literally inscribed in flesh, that the body bears the scars of what the mind cannot face. Yet recent advances in computational and systems neuroscience reveal that this image, while emotionally compelling, is biologically inaccurate. The body does not store trauma; the brain dynamically reenacts it through maladaptive inference. What endures after trauma is not a memory lodged in tissue but a collapse of flexibility-a loss of metastability, the brain's ability to fluidly switch among semi-stable network states. The traumatic memory is real, but it is entrenched in deep, defensive ravines in the landscape of our beliefs and thoughts.The landscape in question is a free energy landscape where every (Bayesian) belief is equipped with a measure of its plausibility. On this view, making sense of the worldand our bodies -entails a process of (Bayesian) belief updating that can be read as minimizing free energy or, in the vernacular, finding apt explanations for our sensations that are the least surprising, or the most plausible.When trauma strikes, the brain's Bayesian mechanics, as described by the Free Energy Principle and Active Inference -the continuous anticipation and minimization of surprise-can lock into a narrow regime of threat expectation. Functional imaging studies show this concretely: in post-traumatic stress disorder (PTSD), neural networks become hypersynchronous, dominated by recurrent loops between the amygdala, hippocampus, and medial prefrontal cortex. Signal variability drops, connectivity patterns harden, and the brain's dynamic repertoire shrinks.In computational terms, trauma overweights the precision of danger priors: the brain assigns excessive confidence to threat predictions, constraining inference based on the prior premise of enduring and ever present danger. The result is hypervigilance, flashbacks, and avoidance-symptoms of a system caught in self-confirming predictions. Mathematically, this overconfidence means one cannot escape local minima -in the free energy landscape -that become deeply and precisely engrained (i.e., trapped in a ravine with steep sides, where precision corresponds to the local curvature or steepness). This rigidity contrasts with the healthy brain's metastable attractor landscape, where neuronal networks continually integrate and segregate in response to context; allowing neuronal dynamics to explore multiple (unstable) minima, and implicit Bayesian beliefs. Mišić and colleagues describe metastability as the hallmark of cognitive flexibility: the capacity for neural coalitions to assemble transiently, enabling rapid adaptation. Trauma erodes this fluidity, trapping the brain in narrow attractor basins of fear and defensive salience. To restore mental health is therefore not to "release" stored emotion but to reestablish dynamic equilibrium-to recover the brain's ability to move with graceful agility over a landscape of beliefs, commitments and intentions.From this view, trauma is a disorder of prediction, not storage. Predictive coding reframes perception as active inference: the brain does not passively register the world but actively predicts it, adjusting only when errors arise -or acting to resolve such errors. For example, a reflexive movement fulfils the brains predictions that our limbs should be in a particular place; thereby minimizing prediction error (and free energy). However, the brain can also resolve prediction errors by affording them less precision. A process known as sensory attenuation. In trauma, or ability to attenuate sensory precision is lost and prediction errors are mis-weighted. Internal threat expectations dominate the search for -and attention to -sensory evidence of danger; unattenuated interoceptive signals-racing heart, tight chest-are interpreted as confirmation of danger rather than imprecise noise. The "score" the body appears to keep is thus an artifact of circular inference: the brain predicts pain, senses arousal, and takes that arousal as proof that pain persists. The body participates in trauma, but as messenger, not archive. This dynamic interpretation aligns with the broader field of embodied cognition. The body and environment are extensions of the brain's predictive loop, scaffolding thought through action. Yet embodiment is active and transient-it does not imply storage. Trauma-related somatic symptoms are better understood as mis-calibrated feedback between prediction, action, and sensation, not as remnants of the past frozen in muscle or fascia. The distinction matters. Where the storage model leads to metaphors of exorcism-finding and purging what was buried-the inference model leads to tra

Overview

CoChat Summary

TL;DR

This theoretical paper argues that trauma is not stored in the body but reflects rigid, over-precise predictive inference and a loss of brain metastability. It proposes flow and other flexibility-restoring interventions as ways to recalibrate precision and restore metastable dynamics, while noting direct PTSD metastability evidence is still lacking.

Key findings

  • Trauma is framed as a disorder of prediction, not storage: over-weighted precision of danger priors traps inference in local minima/free-energy ravines, reducing metastability and cognitive flexibility.
  • PTSD neuroimaging cited shows amygdala hyperactivation with medial prefrontal hypoactivation, with mPFC responsivity inversely associated with symptom severity (Shin et al., 2006); MEG threat processing showed hyperconnectivity in posterior cingulate/right parietal regions and increased node strength/clustering in amygdala and mPFC (Engdahl et al., 2010).
  • Resilience data: a systematic review of 54 studies found the resilient trajectory in approximately 65.7% of cases (Bonanno et al., 2015), and most trauma-exposed individuals do not develop chronic PTSD.
  • Flow hypothesis: Walter et al. (2023) randomized 96 active duty service members to 6 weeks of surf or hike therapy; both groups improved in depression, surf therapy participants were more likely to achieve MDD remission at three-month follow-up, but groups did not differ significantly at post-program. Direct confirmation that flow restores metastability using metrics such as the Kuramoto order parameter has not yet been achieved.

Methods

  • Theoretical/perspective synthesis; no new primary data. It integrates predictive coding/active inference, the free-energy principle, metastability, the somatic marker hypothesis and convergence-divergence zones, resilience research, and flow neurodynamics.
  • Draws on cited empirical literature, including PTSD neuroimaging and MEG studies, TBI metastability work (Hellyer et al., 2015), a resilience review, an RCT of surf/hike therapy, and an analog executive-function study.
  • Main analysis is conceptual argument, not statistical or experimental.

Limitations

  • Stated: the core claim that PTSD specifically involves reduced metastability has not yet been directly tested using established metastability metrics in PTSD populations; Hellyer et al. (2015) pertains to traumatic brain injury, a neurologically distinct condition, so conflation would be unwarranted.
  • Stated: flow-metastability restoration is a theoretical prediction; direct empirical confirmation using established metrics has not been achieved, flow literature is methodologically heterogeneous, neurochemical measurement during naturalistic flow remains limited, and the Walter et al. RCT cannot isolate flow-specific effects from exercise, nature exposure, and social connection.
  • Evident: the paper is conceptual and relies on indirect evidence; the Bomyea and Amir (2011) finding is from an analog sample and, as the text states, awaits replication in clinical PTSD populations.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3389/fnsys.2026.1812957

Cites

18 of 20 references verified0 citation errors · 2 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

6 Valid · 1 Not valid · 13 Not checked · 43 Own results and interpretations

21 more claims

Create a free account to see all 63 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 29
  • Probably not a substantial claim. 13

Create a free account to see these claims.

Citations

Cites

Cited by

See all 5

Metadata

Authors
Kotler, S., Mannino, M., Fox, G., & Friston, K.
Affiliations
  • Florida Atlantic University (US)
  • Florida Atlantic University
  • University of Southern California (US)
  • University College London (GB)
Venue
Frontiers in Systems Neuroscience
Published
2026-04-30
Volume
20
Pages
1812957-1812957
ISSN
1662-5137 (ISSN-L)
OpenAlex
W7159854024
Cited by
5
References
19
Citations by year
2026: 5
FWCI
20.65
Type
journal article
Language
en
Open access
Locations
  • Frontiers in Systems Neuroscience (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Embodied and Extended Cognition
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Memory and Neural Mechanisms
  • Social Sciences › Psychology › Clinical Psychology › Posttraumatic Stress Disorder Research
Keywords
trauma, predictive coding, metastability, body, embodiment, neuroplasticity

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-01 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.