'The Body Keeps the Score': A Critical Reading

Bessel van der Kolk's *The Body Keeps the Score*, published in 2014, is one of the most widely read books on trauma in modern history. It synthesized decades of clinical observation and neuroscience into a single argument: that trauma is not merely a psychological event but a somatic one, encoded in the body and brain in ways that talk therapy alone cannot always reach. The book introduced millions of readers to concepts like the window of tolerance, the triune brain, and the role of subcortical structures in threat response. It championed somatic therapies, neurofeedback, EMDR, and yoga as legitimate interventions for post-traumatic stress.

A decade later, the book remains influential. It is assigned in graduate programs, cited in clinical settings, and recommended by therapists across disciplines. But influence is not the same as accuracy. Some of van der Kolk's central claims have been refined or contested by subsequent research. The triune brain model, for instance, is now considered an oversimplification. The emphasis on implicit memory storage has been clarified by more nuanced models of reconsolidation and prediction error. And the book's treatment of certain therapies has been critiqued for overstating efficacy or underrepresenting risk.

This article offers a critical reading. Not a dismissal, but an appraisal. What has held up under scrutiny. What has been revised. And how the Nervous System Intelligence framework reads the same material through a different lens—one that centers prediction, revision, and the intelligent adaptability of the system itself.

*The Body Keeps the Score* matters because it changed the conversation. Before van der Kolk, trauma was still largely understood as a disorder of memory—something to be processed, narrated, and resolved through insight. The book reframed trauma as a disorder of the nervous system, one that persists not because the person refuses to let go, but because the body continues to respond as if the threat is present. That reframing was clinically and culturally significant. It reduced shame. It expanded the therapeutic toolkit. It gave language to experiences that had been pathologized or dismissed.

But the book also matters because it has been widely misread. Clinicians and lay readers alike have extracted simplified takeaways—"trauma is stored in the body," "the reptilian brain takes over," "you have to do somatic work or you won't heal"—that do not reflect the complexity of the underlying science. Some of these ideas have become therapeutic dogma, repeated without scrutiny, applied without nuance, and sometimes used to justify interventions that lack robust evidence.

For clinicians, this creates a dilemma. The book opened doors. It legitimized body-based approaches in settings that had been skeptical. It gave practitioners permission to work with breath, movement, and sensation. But it also set expectations that are not always realistic, and it framed certain mechanisms in ways that newer research has complicated.

For patients, the stakes are different. Many people have found the book validating. It named what they felt. It offered hope that healing was possible outside the narrow confines of cognitive therapy. But others have been left confused or discouraged when their own experience did not match the narrative—when somatic therapy did not unlock a stored memory, when yoga did not resolve hypervigilance, when the body did not, in fact, seem to keep the score in the way the title promised.

A critical reading is not a rejection. It is an attempt to preserve what is valuable while updating what is incomplete. That is the work of science. And it is the work of clinical integrity.

Van der Kolk's central thesis rests on several pillars: that trauma disrupts normal memory encoding, that it dysregulates subcortical threat-detection systems, that it fragments the sense of self, and that it requires interventions that engage the body directly. Each of these claims has a basis in research, but each has also been refined or contested in the decade since publication.

**Memory and the Body**

The book argues that traumatic memories are stored differently than ordinary memories—encoded implicitly, somatically, and outside the reach of narrative recall. This draws on early work by LeDoux (1996) on fear conditioning and Brewin's dual representation theory (Brewin et al., 1996). The idea is compelling, but recent work has complicated it. Reconsolidation research suggests that even deeply encoded fear memories are not static; they are updated each time they are retrieved (Schiller et al., 2010). More recent meta-analyses have found that the distinction between implicit and explicit trauma memory is less categorical than once thought (Brewin, 2014; Krans et al., 2022). Traumatic memories can be fragmented, but they are not necessarily stored in a fundamentally different neural substrate. The body does not "keep" the memory in a literal sense; rather, the nervous system continues to predict threat based on prior learning, and those predictions shape perception, arousal, and behavior (Paulus & Stein, 2023, *Nature Reviews Neuroscience*).

**The Triune Brain**

Van der Kolk relies heavily on MacLean's triune brain model—the idea that the human brain is organized into reptilian, mammalian, and neocortical layers, and that trauma "shuts down" the cortex while activating the reptilian brain. This model has been abandoned by contemporary neuroscience. As Barrett and Simmons (2015) and Pessoa (2022, *Nature Reviews Neuroscience*) have shown, the brain does not operate in hierarchical layers; it functions as an integrated predictive system. The amygdala is not a "reptilian" structure, and the prefrontal cortex does not simply inhibit subcortical regions. Instead, cortical and subcortical systems work in concert to generate predictions, update models, and regulate arousal. The triune model is pedagogically useful, but it is not anatomically or functionally accurate (Cesario et al., 2020, *Neuroscience & Biobehavioral Reviews*).

**Somatic Therapies**

The book champions several body-based interventions: EMDR, neurofeedback, yoga, and sensorimotor psychotherapy. The evidence for these is mixed. EMDR has robust support for PTSD, comparable to trauma-focused CBT (Carlson et al., 2023, *JAMA Psychiatry*; Lewis et al., 2020, *Psychological Medicine*). Neurofeedback, by contrast, has limited high-quality evidence; a 2023 Cochrane review found insufficient data to recommend it as a first-line treatment (Micoulaud-Franchi et al., 2023). Yoga for trauma has shown promise in small trials (van der Kolk et al., 2014; Rhodes et al., 2022, *Journal of Traumatic Stress*), but effect sizes are modest and mechanisms remain unclear. Sensorimotor psychotherapy lacks large-scale RCTs (Payne et al., 2015, *Frontiers in Psychology*).

The issue is not that these therapies are ineffective, but that van der Kolk presents them with more certainty than the evidence warrants. He is correct that trauma treatment should not be limited to cognitive processing. But the claim that somatic work is necessary for healing is not supported by the literature. Many people recover through exposure-based or cognitive therapies alone (Cusack et al., 2023, *Lancet Psychiatry*).

**Polyvagal Theory**

Van der Kolk draws extensively on Porges' polyvagal theory, which posits that the vagus nerve has two branches—one associated with social engagement, the other with shutdown—and that trauma disrupts the balance between them. Polyvagal theory has been influential in trauma therapy, but it has also been critiqued on anatomical and physiological grounds. A 2023 review in *Biological Psychology* (Grossman & Taylor, 2023) found that many of polyvagal theory's core claims are not supported by human autonomic physiology. The vagal brake, the social engagement system, and the shutdown response are not as clearly demarcated as Porges suggests. This does not invalidate the clinical utility of vagal-focused interventions, but it does mean the theory should be held lightly, not taught as fact.

**What Has Held Up**

Despite these critiques, much of van der Kolk's work remains valuable. The emphasis on safety, regulation, and relationship in trauma treatment is well-supported (Cloitre et al., 2023, *American Journal of Psychiatry*). The recognition that trauma affects multiple systems—cognitive, emotional, somatic, relational—is now standard in phase-based models (Courtois & Ford, 2020). And the call to expand the therapeutic repertoire beyond talk therapy has been vindicated by the growth of integrative and embodied approaches (Payne et al., 2015; van der Kolk et al., 2014).

The Nervous System Intelligence framework reads *The Body Keeps the Score* as a transitional text—one that moved the field forward but stopped short of a fully predictive account. Van der Kolk was right that trauma is not just a problem of memory or cognition. But the mechanism is not that the body "stores" trauma. It is that the nervous system learns to predict threat in contexts where threat is no longer present, and those predictions become self-confirming.

This is not a metaphor. Prediction error—the mismatch between what the system expects and what actually occurs—is the engine of learning in the brain (Friston, 2010; Clark, 2013; Barrett, 2017). In trauma, the system has learned that certain cues predict danger. When those cues recur, the system generates a threat response—autonomic arousal, muscle tension, attentional narrowing—even if the actual danger is absent. The response itself then becomes evidence that the prediction was correct. The loop closes. The prediction is not revised.

This is where NSI diverges from van der Kolk's framing. The body does not keep the score. The nervous system keeps the model. And models are revisable.

The NIRVA Method's six movements—Notice, Interrupt, Identify, Regulate, Validate, Align—are designed to create the conditions under which prediction error can occur and be integrated. **Notice** brings awareness to the prediction itself: the sensation, the thought, the impulse. **Interrupt** creates a pause, a moment in which the automatic response is not enacted. **Identify** names the prediction and the context in which it was learned. **Regulate** stabilizes the system so that new information can be processed without overwhelming it. **Validate** acknowledges that the prediction was once adaptive. **Align** tests a new prediction—one that is more congruent with present reality.

This is not a rejection of somatic work. It is a reframing. Somatic interventions are valuable not because they release stored trauma, but because they provide the nervous system with new data—data that can update the model. A slow exhale signals safety. A grounded stance signals stability. A supported movement signals agency. These are not metaphors. They are prediction errors. And prediction errors, when integrated, revise the model.

Van der Kolk understood that trauma treatment must engage the body. NSI explains why: because the body is where predictions are tested. The nervous system is intelligent. It learns. And what has been learned can be revised.

For clinicians, *The Body Keeps the Score* remains a useful entry point—but it should not be the endpoint. The book's greatest contribution is its insistence that trauma treatment must be multimodal, that safety and regulation precede processing, and that the therapeutic relationship is itself a site of nervous system revision. These principles are well-supported and clinically essential (Cloitre et al., 2023, *American Journal of Psychiatry*; Courtois & Ford, 2020).

But clinicians should be cautious about adopting the book's models uncritically. The triune brain is not a valid framework for case formulation. Polyvagal theory should be presented as a heuristic, not a fact. And the claim that somatic therapies are necessary for trauma recovery is not supported by the evidence. Many patients respond well to exposure-based or cognitive therapies, and some may find body-focused work retraumatizing or inaccessible (Cusack et al., 2023, *Lancet Psychiatry*).

What matters is not the modality but the mechanism. Effective trauma treatment creates conditions under which the nervous system can revise its predictions. That can happen through narrative exposure, through somatic experiencing, through EMDR, through pharmacotherapy, or through relational attunement. The task is not to release stored trauma but to provide the system with new data—data that is safe enough, clear enough, and repeated enough to update the model.

Clinicians trained in NSI will recognize this as the work of the NIRVA Method. The movements are not techniques; they are conditions. **Notice** and **Interrupt** create the space for awareness without reactivity. **Identify** and **Regulate** stabilize the system. **Validate** and **Align** allow the prediction to be revised without invalidating the history that produced it. This can be done with breath, with movement, with language, with imagery, or with silence. The medium is less important than the mechanism.

Finally, clinicians should be mindful of the cultural weight the book carries. Many patients arrive in treatment having read it, and they may have internalized its language—"my body keeps the score," "I'm stuck in my reptilian brain," "I need to do somatic work." These beliefs are not inherently harmful, but they can become rigid. Part of the clinical work is to honor what the book has given the patient while gently expanding the frame. The body is not a vault. The nervous system is not broken. And healing is not about unlocking the past—it is about revising the future.

If you have read *The Body Keeps the Score* and found it helpful, you do not need to unread it. But you may benefit from reading it differently.

The book's central insight—that trauma lives in the body—is true in a limited sense. What is more accurate is that your nervous system has learned to predict threat in situations that resemble past danger, and those predictions shape what you feel, how you move, and what you notice. The good news is that predictions can be revised. You do not need to excavate the past. You need to provide your system with new data.

Start with **Notice**. Not in a therapeutic sense, but in a literal one. What does your body do when you feel unsafe? Does your breath shorten? Do your shoulders rise? Does your vision narrow? These are not symptoms. They are predictions. Your system is preparing for a threat it has learned to expect.

Next, **Interrupt**. Not to suppress the response, but to create a pause. A slow exhale. A deliberate unclenching of the jaw. A shift in posture. These are small acts, but they are also signals. They tell the system: this moment is different.

Then **Identify**. Not the trauma itself, but the prediction. What is your system expecting to happen? What is it preparing for? You do not need to relive the past to name the pattern.

**Regulate** is not about calming down. It is about stabilizing enough to take in new information. That might mean grounding your feet, softening your gaze, or simply noticing that you are still here.

**Validate** means acknowledging that the prediction was once accurate. Your system learned to respond this way because it needed to. That learning was intelligent. It kept you safe. It is not wrong. It is simply no longer necessary.

Finally, **Align**. Test a new prediction. Not through belief, but through action. A small movement. A different breath. A moment of stillness. See what happens. Let the system update.

This is not a cure. It is a practice. And it does not require a therapist, a yoga mat, or a neurofeedback device. It requires attention, patience, and a willingness to let the nervous system do what it does best: learn.