The Radar and the Weight: Neuroception, Trauma and Empathy, and What the Research Shows

By Jordan Gold and Valery Gold

Illustration of a person seen from behind, seated in a wooden chair in a sunlit room, with soft concentric arcs and branching green and clay-colored filaments spreading across the wall like roots or river deltas. A houseplant stands at the left, a simple table and a tall window at the right, and an open doorway lets morning light through.
Illustration: a quiet room at morning light, the body at rest while the space is gently read.Illustration by Legends Only Racing (LOR). Illustrative and symbolic; not a real person, and not an anatomical or diagnostic image.

Walk into a room where two people have just stopped arguing. Nobody says a word. Your shoulders climb toward your ears, your stomach tightens, and some old, quick part of you has already decided how the next ten minutes are going to go. You couldn't explain it if someone asked.

A lot of people know that feeling, and some know it all day long. This article is about one name for it, neuroception, and about the question underneath: when a childhood goes wrong, does it leave a person more tuned in to other people, more tired from listening, or both?

It began with a message from Valery Gold about neuroception, trauma and empathy. The instinct at the heart of it is one we share and want to keep: a sensitive nervous system is something to understand with compassion, not something to be ashamed of. Then we did what we do with every research piece. We read the studies behind the claims. The picture that came back is richer and less tidy than the popular version. Some of it holds up well. Some of it is small, some is debated by qualified experts, and some we could not confirm. We say which is which, because the people most likely to read a piece like this deserve both kindness and accuracy.

A note before we start. This is general information, not medical advice, and it is not a substitute for a clinician who knows you. If any of this lands hard, or you are in distress, you are not alone. In the United States you can call or text 988, or chat at 988lifeline.org, any hour of any day. It is free and confidential [1].

How to read the labels. Study means a peer-reviewed report of data. Review means a paper that pools or surveys many studies. Proposal means a claim from a theory's author that others have not confirmed. Contested means qualified experts disagree in print. Our arithmetic and our reading mean we counted, converted or interpreted figures printed in the source. Opinion is ours and is marked. Every source was opened and checked on 15 November 2026 (US Central time).

What "neuroception" means

Stephen Porges introduced the word in the early 2000s. The earliest use we found is in the abstract of his 2003 paper in the Annals of the New York Academy of Sciences, which speaks of the nervous system matching behavior to "the neuroception of a safe, dangerous, or life-threatening environment" [2]. His 2004 article describes neuroception as the way neural circuits "distinguish whether situations or people are safe, dangerous, or life threatening" [3]. In his 2022 review he calls it "a neural process, distinct from perception," that shifts the body's state "without requiring conscious awareness" [4]. Both parts of the popular story check out: Porges did coin the term, and the idea that this happens below awareness, before thought, is his.

What does the system read? In Porges's words, the brain areas involved are sensitive to "voices, faces, gestures, and hand movements," and he describes neuroception as decoding "movements and sounds" [4]. He also says the body's side of the loop is "functionally equivalent to interoception," our sense of our own internal state, and that while we are "often unaware of the stimuli that trigger different neuroception responses," we are "generally aware of our body's reactions" [4]. So faces, tone of voice and inner bodily signals are well supported as part of the picture. Posture, proximity and light also appear on some popular lists. We did not find them named as neuroception inputs in the Porges papers we could read, so we leave them out rather than guess.

Illustration of a simple living room from a high angle, with a small faceless figure standing at the center and soft concentric rings and contour lines spreading across the floor toward a green sofa, a low table, a plant, and a sunlit doorway.
Illustration: the body quietly reading a room for safety.Illustration by Legends Only Racing (LOR). Illustrative and symbolic; not a real person, and not an anatomical or diagnostic image.

Porges describes three resulting states: social engagement, fight or flight, and shutdown [4]. "Freeze" is the popular word. Porges writes about immobilization, which he links to features such as "dissociation, withdrawal" [4].

One more thing worth saying plainly. In 2004 Porges wrote that "faulty neuroception might lie at the root of several psychiatric disorders" [3]. Note the word might. That was a proposal then, and it is a proposal now.

What holds up, and what is debated

Here an honest article has to slow down. Polyvagal Theory, the framework neuroception belongs to, is contested, and not gently.

Paul Grossman, emeritus research director at University Hospital Basel, published a 2023 review whose stated aim is to document how each of the theory's five basic premises "have been shown to be either untenable or highly implausible based on the available scientific literature" [5]. In the open author copy he goes further and writes that there is "no support" for the continuing assertion that one vagal brainstem region, the dorsal vagal motor nucleus, is responsible for "trauma-related dissociative or emotional freezing responses" [5]. He is one author arguing hard, and a review is an argument built from other studies, not a new experiment. A second group, Taylor and colleagues, surveyed the heart control of vertebrates and report that the other major vertebrate groups "all show both tonic and phasic control of heart rate" by the parasympathetic system. They suggest the mammalian pattern may have evolutionary roots in reptiles, amphibians and lungfish, which bears on the theory's account of how the system evolved [6]. That last step is our reading of their abstract.

Porges answers that critics misrepresent the theory. In 2023 he restated it, arguing that two distinct brainstem pathways to the heart resolve what he calls the "vagal paradox," and he describes the criticisms as "a classic strawman argument" [7].

We are not going to pretend to referee the anatomy. What we can say is this. The critiques we read aim at the theory's physiology: how the branches of the vagus nerve are organized, how the heart is controlled across species, how that evolved, and whether heart-rhythm measures such as respiratory sinus arrhythmia can serve as a readout. In the Grossman author copy and the Taylor full text, a text search finds no use of the word "neuroception," so those papers are not testing neuroception itself, and it would overstate them to say they have refuted it. (We read the Grossman open author copy and the published Taylor paper; we did not read the other critiques in full.) The two ideas are linked, though. Porges builds neuroception into the same framework, says it triggers the body's automatic "scripts," and describes it in his own 2023 restatement as "the hypothetical process through which these scripts are triggered" [7]. So the careful position is: neuroception is offered by its author as a hypothesis, the physiology around it is disputed, and neither side has shown that the experience people describe is not real.

Can it be measured? A questionnaire exists, the Neuroception of Psychological Safety Scale, and a 2025 paper validated it in 2,035 adults in the UK [8]. It is a 29-item self-report that asks how people felt over the past week, with statements such as "I felt valued" [8]. That makes it a measure of felt safety, which is useful, but not a direct measure of a process that by definition happens outside awareness (our reading). Porges is a coauthor on the paper [8]. In the set of sources we verified, we found no study that measures "neuroception" directly as its own mechanism. So here is how we hold it. Opinion: treat the word as a useful map of a real human experience, sensing safety or danger before you can say why, and hold the physiology underneath it loosely. A map can help you find your way without being the territory. Many people find that language helpful, and nothing in the debate takes the experience itself away.

Does trauma tune the threat detector?

Porges's own proposal is that a history of severe adversity can leave the nervous system "locked in states of defense" [4]. The support he gives in that paper is a report from foster parents of abused children [4]. That is a human observation worth listening to. It is not a controlled study, so we label it a proposal.

The independent evidence is narrower and sturdier. Pollak and colleagues studied children who had been physically abused. In one study of 40 children (23 physically abused, 17 not; mean age about 9), abused children treated the boundary between angry and other faces differently. They showed "alterations in their perceptual categories for anger," while identifying and discriminating other emotions "in the same manner as nonabused children" [9]. In another, abused children "accurately identified facial displays of anger on the basis of less sensory input than did controls" [10]. In a third, abused 8- to 11-year-olds were slower to shift their attention away from an angry face, and gained more from an angry face that correctly signalled where to look [11].

Read those carefully, because they correct a popular line. They are small studies of children who were physically abused, and the abused group in the first was recruited partly through a state psychiatric facility [9]. They show a tilt toward anger and threat cues, which makes good sense as a survival habit. They do not show that trauma survivors read all emotions "with high precision," and we found no study in our set that tested whether trauma makes safety slower to register. Quick to spot danger: supported in these children, for anger. The wider claims go past the data we could verify.

A newer and much larger review adds a note of caution without taking anything away from survivors. A 2026 meta-analysis of 40 studies and 6,369 adults looked at how accurately people recognize emotions on behavioral tests. It found a statistically significant but small difference in the opposite direction from "high precision": people with high childhood maltreatment exposure were slightly less accurate than people with low exposure (g = -0.13), with larger gaps for neutral (g = -0.27) and disgusted (g = -0.32) expressions [12]. The abstract also says that "no significant associations between childhood maltreatment and emotion recognition accuracy, nor effects for reaction times, were found" [12]. It does not spell out which analyses that sentence covers, and we could read only the abstract, so we do not lean on that line. In plain words: being quick to catch anger in some children is not the same as being more accurate across all emotions, and a small average gap in a pooled group says nothing about any one person. Many survivors read a room well, and nothing here says they do not. What the pooled data do not support is the blanket claim of "high precision" (our reading).

For scale, adverse childhood experiences are common. The CDC reports that three in four high school students had one or more, and one in five had four or more [13]. That figure is about teenagers, not adults, and "adverse experience" is a broad category. It is still a reminder that we are talking about a great many people, which is why getting this right matters.

Does trauma make people more empathic?

This is the claim the popular story rests on, and the study behind it is real. David Greenberg, Simon Baron-Cohen, Peter Fonagy and colleagues published Elevated empathy in adults following childhood trauma in PLOS ONE in 2018 [14]. Here is what it found, and how far it goes.

The finding. "On average, adults who reported experiencing a traumatic event in childhood had elevated empathy levels compared to adults who did not" [14]. The feeling-with-others side, affective empathy, was higher in both studies. The understanding-others side, cognitive empathy, was higher in the second study (the Perspective Taking scale) but "only approached significance" in the first [14]. So "both" needs a qualifier. (The first study prints a p-value of .74 beside that result. An F of 3.20 on 1 and 385 degrees of freedom works out to roughly p = .07, so the printed value looks like a dropped zero. That is our arithmetic, and it fits the authors' own words, so we do not quote the .74. The same paragraph prints F = 1.47 and p = .23 for the total EQ score, which is a second internal inconsistency; see "Checking Greenberg's numbers from the authors' own data" in Methods below.)

The sample. Two online surveys of adults recruited through Amazon Mechanical Turk, with 387 people analysed in Study 1 and 442 in Study 2 after exclusions. About 78 to 79 percent were White, and the samples were mostly female (about 70 percent in Study 1 and 62 percent in Study 2). Participants were not asked about clinical diagnoses [14]. The Study 2 text says 491 people completed the battery and 57 were excluded, which would leave 434, but it reports 442 analysed. We use the 442 the authors report, because the posted data file for Study 2 has 442 participants (348 with a trauma and 94 without), and we flag the mismatch rather than guess which number slipped. The Study 1 file has 387 (309 and 78), matching the paper. Roughly four in five counted as having a childhood trauma (309 of 387 and 348 of 442; our arithmetic). The checklist asked about events before and including age 17 and had five items: the death of a very close friend or family member, a major upheaval between parents such as divorce or separation, a traumatic sexual experience, being the victim of violence, and any other major upheaval [14]. So the "trauma" group is broad, and the authors themselves note that it outnumbered the comparison group and ask what people count as trauma [14].

The size. The significant differences were small. The authors report partial eta squared values between .012 and .026 [14], which we read as roughly 1 to 3 percent of the variation in scores. By the usual rules of thumb (.01 is "small," .06 is "medium"), that puts them between small and medium (our reading). Real, but modest.

Checking the numbers. The authors posted their data, so we reran the main comparisons. A few printed figures do not reproduce, including two Study 2 results that the text appears to transpose, and none of them changes the conclusion. The details and our method are in the Methods section below [14] (see Checking Greenberg's numbers).

Dot plot of effect sizes (partial eta squared) for empathy measures in two 2018 studies of adults with and without childhood trauma. Study 1, 387 adults: affective empathy .014 is significant; cognitive empathy .008 and social skills .004 are not. Study 2, 442 adults: global empathy .026, empathic concern .012, perspective taking .013 and fantasy .019 are significant; personal distress .001 is not. All lie below .06.
Source: Greenberg, D. M., Baron-Cohen, S., Rosenberg, N., Fonagy, P., and Rentfrow, P. J. (2018). Elevated empathy in adults following childhood trauma. PLOS ONE, 13(10), e0203886. Values are from the Results text of Study 1 (Empathy Quotient) and Study 2 (Interpersonal Reactivity Index): sex-adjusted ANCOVAs, F(1, 385) in Study 1 and F(1, 438) in Study 2, except where we note a recalculation from the authors’ posted data. Ns are people analysed after exclusions: Study 1, 410 completed, 23 excluded, 387 analysed; Study 2, 442 analysed, as the authors report (the text gives 491 completed and 57 excluded, which would leave 434; the authors’ posted data file contains 442 participants, so we use 442). The checklist asked about events prior to and including age 17. Partial eta squared: Study 1 Affective Empathy .014 (p ≈ .02; printed as .01), Cognitive Empathy .008 (not significant), Social Skills .004 (p = .20). Study 2 Global Empathy .026 (p < .001), Empathic Concern .012 (p ≈ .02), Perspective Taking .013 (p = .02), Fantasy .019 (p = .003), Personal Distress .001 (p = .45). The paper’s Results text lists the Empathic Concern and Fantasy statistics the other way round; the values here are our recalculation from the authors’ posted data (sex-adjusted analysis of covariance, item scoring from the authors’ script). Our reruns have degrees of freedom of 383 (Study 1) and 436 (Study 2), two lower than printed, because one Study 1 and three Study 2 participants have no sex recorded; this does not exactly account for the printed values, so printed figures are quoted where we did not recalculate. Both facets are significant under either assignment. The five significant effects (.012–.026) are small by convention, between the .01 small and .06 medium benchmarks. Conventional benchmarks often applied to partial eta squared are .01 (small), .06 (medium) and .14 (large); they are conventions, not thresholds from this paper. For Study 1 Cognitive Empathy, p ≈ .07, from the printed F and confirmed by our rerun of the authors’ posted data; the printed p = .74 appears to be a typo. The authors later write that this result “only approached significance.” The point is plotted as not significant, with a hollow marker. Study 1 total EQ is omitted because the paper prints no partial eta squared for it. The paper’s printed F = 1.47 (p = .23) does not reproduce; our rerun of the authors’ data gives F ≈ 3.16 (p ≈ .08). Neither is significant. No confidence intervals were reported. Samples were online Mechanical Turk surveys, self-report and largely correlational. Unadjusted standardized differences (trauma group minus comparison group, divided by the pooled standard deviation, not adjusted for sex), our calculation, from the authors’ posted data: Study 1 affective 0.32; Study 2 perspective taking 0.34, empathic concern 0.37, fantasy 0.41, global 0.49. These figures describe group differences on questionnaires; they are not a diagnosis or clinical advice.

The severity link. Among people with trauma, the authors report that severity was "positively linked to empathy levels" [14]. In the tables the picture is patchy. Some links were statistically significant and others were not, depending on the type of trauma and the empathy measure, and no correlation printed in either table was above .30 [14]. In Study 2, for example, overall empathy was significantly linked to severity only for a death (.24) and sexual abuse (.25). Greater severity goes with more empathy is true in the abstract and uneven in the detail.

The limits, in the authors' words. The study relied on self-report, and the authors raise the possibility that people who have been traumatized "may believe they have better empathy." It was "largely correlational," so it cannot say trauma caused the difference. Adults looked back on their own childhoods, and the authors call for performance tests and for studies of clinical groups, which they did not test [14]. The abstract says the findings "suggest that the experience of a childhood trauma increases a person's ability to take the perspective of another," which is more causal-sounding than a largely correlational design can support. We would read the result as a pattern worth taking seriously, not a cause established.

Has it been replicated? Study 2 replicates Study 1 inside the same paper, with a different questionnaire and a new sample. We did not find an independent, same-design replication. A later study, by Benz and colleagues in 2023, built on Greenberg's work and hypothesized a positive link between early adversity and empathy. Only one facet, personal distress, turned out to be related, as the next section describes [15].

The part that deepens the story

If Greenberg were the whole literature, we could end the section here. It is not, and the rest of the research does not erase his finding. It puts it in company.

A 2024 meta-analysis pooled 24 studies and 22,580 participants and found the opposite direction: child maltreatment was related to reduced empathy (r = -0.15, 95% confidence interval -0.17 to -0.13), with emotional neglect the strongest link (-0.18) and sexual abuse the weakest (-0.07) [16]. That is also a small correlation by the usual rules of thumb (our reading). The authors report that the results were not inflated by publication bias but did vary with the age of respondents, sample size, the empathy measure used and the type of maltreatment [16]. Differences like those may be part of why studies disagree, though we cannot say how much.

A 2023 study of 228 adults, 83 percent of them women, found that childhood abuse and neglect were tied to higher personal distress when other people suffer, while empathic concern, perspective taking and fantasy were "not related" [15]. That is a correlational, self-report study too. And a 2025 scoping review of 43 empirical papers tallied them like this: 15 reported that early adversity went with more empathy, 19 with less, and 12 found no significant association [17]. Those groups overlap. The three counts add up to 46, not 43, because many papers reported a mix of positive, negative and null findings [17] (our arithmetic). It is a tally of papers, not a pooled effect. The reviewers also summarize the pattern by age: in youth, early adversity went with higher affective but lower cognitive empathy, and in adulthood with higher personal distress [17].

Three horizontal bars showing a scoping-review tally from Kahhale, Byrd and Hanson, 2025: 15 articles suggesting more empathy, 19 suggesting less empathy, and 12 suggesting no difference. The category counts sum to 46 across 43 unique articles because many articles reported a mix of positive, negative and null findings. The bars are category counts, not a partition of 43. This is the authors’ tally, not a pooled effect.
Source: Kahhale, I., Byrd, A., and Hanson, J. (2025). Early life adversity and empathy: A scoping review of past research and recommendations for future directions. Clinical Child and Family Psychology Review, 28(2), 275–321. The abstract and the opening of the Results report 43 empirical articles, and the tally “15 articles suggested that ELA was associated with increased empathy, 19 that ELA was associated with decreased empathy, and 12 pointed to a null association.” The same Results section states that most manuscripts presented multiple associations and that “many articles reported a mix of positive/negative and null findings.” Category counts therefore sum to 46 across 43 unique articles. The authors also give these as 35%, 44% and 28% of 43, which add to more than 100% and are consistent with overlap. This is a count of articles, not a pooled effect size, and the studies mix ages and measures. It is not a diagnosis or clinical advice. The review covers 43 unique articles (its abstract, Results and Table 1). One passage in Section 4 says “41 manuscripts”; that looks like a count left over from before the authors’ January 2025 search update added 2 papers (our reading, not the authors’ statement). Table 1 lists 44 rows because one paper is split into two studies.

There is also the matter of naming your own feelings. A meta-analysis of 99 samples and 36,141 people found child maltreatment positively related to adult alexithymia, meaning difficulty identifying and describing one's emotions (r = .23, 95% confidence interval .19 to .27), most strongly after emotional abuse, emotional neglect and physical neglect [18]. Another meta-analysis, of 12 studies with 1,095 people who had PTSD, found a large association between PTSD and alexithymia (d = 0.80, 95% confidence interval 0.60 to 1.00, in the authors' posted manuscript) [19]. A person can be finely attuned to others and still find it hard to name their own inner weather. We did not find a study in our set that tested both at once, so that stays an open question.

Two panels of pooled correlations with confidence intervals around a zero line. Child maltreatment and empathy, 24 studies: overall r = −0.15, with all five maltreatment types between −0.18 and −0.07. Child maltreatment and alexithymia, 99 samples: overall r = .23, with the three listed types between .18 and .21. Different outcomes, not combined.
Source: Zhang, H., Gao, X., Liang, Y., Yao, Q., and Wei, Q. (2024). Does child maltreatment reduce or increase empathy? A systematic review and meta-analysis. Trauma, Violence, & Abuse, 25(1), 166–182. Values from the abstract: 24 studies, 22,580 participants; overall r = −0.15, 95% CI −0.17 to −0.13; emotional neglect −0.18 [−0.25, −0.12]; physical neglect −0.16 [−0.23, −0.09]; emotional abuse −0.15 [−0.21, −0.10]; physical abuse −0.12 [−0.17, −0.07]; sexual abuse −0.07 [−0.14, −0.01]. Zhang values are from the published abstract (confirmed on PubMed and the publisher page); we could not obtain the full text. Source: Ditzer, J., Wong, E. Y., Modi, R. N., Behnke, M., Gross, J. J., and Talmon, A. (2023). Child maltreatment and alexithymia: A meta-analytic review. Psychological Bulletin, 149(5–6), 311–329. Values from the abstract: 99 samples, 78 sources, 36,141 participants; overall alexithymia r = .23 [.19, .27]; emotional abuse .18 [.13, .23]; emotional neglect .21 [.16, .26]; physical neglect .18 [.15, .22]. Ditzer values are from the published abstract; we could not obtain the full article, but the authors’ public online supplement matches the overall, emotional abuse and emotional neglect estimates and gives physical neglect as .18 [.15, .21] against the abstract’s [.15, .22]. The two panels measure different things (empathy versus alexithymia) and are not pooled together. k and N by subtype were not given in the abstracts and are not shown. These figures are pooled correlations, not a diagnosis or clinical advice.

So what do we make of a field that points several ways? Greenberg and colleagues put it well: "there are likely varied pathways that lead toward or away from empathy after the experience of a trauma" [14]. Our reading is that trauma can make people more empathic is a fair description of some people in some samples, measured by self-report, and not a rule. What happened, when, with whom, and who helped afterward all seem to matter, and the evidence cannot yet sort that out. Survivors who feel deeply are not imagining it. Survivors who feel numb or find it hard to name feelings are not failing at something. Both show up in the research.

The double edge

The best point in the starting message is that this sensitivity cuts both ways. The research agrees, with some care about the words.

Abstract illustration of a continuous field of color and form, opening in warm cream and clay-amber curves on the left and gradually tightening into cooler, more angular blue-green contours on the right, with a pale ribbon winding through.
Illustration: a visual spectrum from open warmth to cooler, tighter ground.Illustration by Legends Only Racing (LOR). Illustrative and symbolic; not a real person, and not an anatomical or diagnostic image.

Empathy researchers draw a line between sharing a feeling and dissolving into it. Tania Singer and Olga Klimecki write that in empathy "one feels with someone, but one does not confuse oneself with the other," and that without that self-other distinction "we speak of emotion contagion" [20]. That is a useful way to talk about boundaries. When absorbing other people's moods feels automatic and exhausting, you may be dealing with contagion rather than a bigger heart. The adult data on personal distress [15] [17] point the same way. One wrinkle: Greenberg's second study found no difference in personal distress between groups [14], so even this is mixed. We offer the contagion idea as an interpretation (our reading), not a result those studies tested.

What about people-pleasing, blurred boundaries and difficulty knowing what you need? Many survivors and clinicians describe these as ways of staying safe around unpredictable people, and the description rings true to many. We looked for a study that tests it and did not find one we could verify. So we treat it as a lived description that deserves respect, not as a research finding. The same goes for the idea that clinicians often see this as a protective adaptation. It is plausible and widely said, and we could not tie it to a source.

None of this is a flaw in anyone's character. If you grew up scanning faces for weather, you got good at weather. The skill was real. So was the cost.

What helps: what we can and can't say

The starting message ends with healing through "co-regulation with trusted others, somatic practices, and environments that provide consistent cues of safety." Here we separate the theory from the evidence.

The theory is Porges's: that feeling safe, and calming alongside a trusted person (co-regulation), is how a nervous system learns to stand down [4]. It is a gentle idea, and it is a proposal. For transparency, Porges discloses in that paper that he holds patent rights and receives royalties for a listening program, the Safe and Sound Protocol [4]. We mention it because readers deserve to know, not because it settles anything. The CDC says something close in plainer language: "safe, stable, nurturing relationships and environments" prevent adversity and help "buffer the risk of harm when they do happen" [13]. Notice that the CDC is talking about children and prevention, but it points the same direction.

Illustration of two faceless silhouettes sitting toward each other, with synchronized wave lines and branching golden filaments meeting in a soft glow between them.
Illustration: two figures sharing a quiet, steady light.Illustration by Legends Only Racing (LOR). Illustrative and symbolic; not a real person, and not an anatomical or diagnostic image.

The evidence for treatments built on polyvagal ideas is thin so far. A 2021 systematic review of mindfulness and compassion practices through a polyvagal lens found only six eligible studies, one of them in people with OCD and two of the PTSD studies with no comparison group, and the authors wrote that "very limited research" exists and that the results are preliminary [21]. A 2024 review of 26 randomized trials of mind-body approaches for PTSD, not polyvagal-specific, rated the evidence as low for mindfulness-based and for meditation- and mantra-based approaches and moderate for movement-based ones, and noted that only two trials used objective outcome measures [22].

For people with diagnosed PTSD, the American Psychological Association's 2025 guideline (approved February 2025) recommends three first-line therapies, all forms of cognitive behavioral therapy: cognitive processing therapy, prolonged exposure and trauma-focused CBT. It suggests cognitive therapy, EMDR and narrative exposure therapy as second-line options [23]. The 2023 VA/DoD guideline recommends cognitive processing therapy, EMDR and prolonged exposure [24]. Neither is about neuroception, and both are written for diagnosed PTSD.

On body-oriented approaches the two guidelines land in the same place. APA states that "there is insufficient evidence for the Panel to recommend for or against" a list of body- and movement-oriented interventions that includes Somatic Experiencing and trauma-sensitive yoga [23]. VA/DoD says there is "insufficient evidence to recommend for or against" a list of mind-body interventions that includes somatic experiencing and yoga [24]. Please read that phrase for what it says. It is not "these do not work." APA's own footnote says the term refers to the quality standards used in the guideline and that for some therapies listed there is "a substantial amount of relevant evidence from other epistemologies" [23], and a 2021 scoping review of Somatic Experiencing called its early findings "promising" while saying they "require more support from unbiased RCT-research" [25]. It means the trials that would let a guideline panel recommend them are not there yet.

Polyvagal therapies were named in the APA panel's scope and search criteria, but the guideline makes no recommendation about them [23]. When we searched Europe PMC on 15 November 2026, we found no randomized trial of a polyvagal-informed therapy for trauma or PTSD, and none that tested co-regulation as a treatment for trauma. That is absence of evidence in our searches, not proof that none exists. It is a good reason to talk to a trained professional if symptoms are running your life.

Calm, safe, kind surroundings are good for nearly everyone, and nothing here argues against them. We just can't tell you that a specific polyvagal technique, or co-regulation as a treatment, works better than the alternatives, because we did not find the evidence. That is a statement about the research so far, not a verdict on anyone's experience of what helps.

What we take from it

The starting message closes with this line: "Recognizing neuroception as the underlying mechanism offers a more compassionate and accurate map of why some people feel so much, so quickly, and so deeply."

We love the compassion in that sentence and want to keep every bit of it. Opinion: a framework that says your sensitivity may be a nervous system doing its job is kinder than one that says you are too much. The one adjustment we would make is a small one in wording. The research supports calling neuroception a compassionate map. It does not yet support calling it the underlying mechanism, because neuroception has not been measured directly as a mechanism in the studies we checked, the theory around it is contested, and the research linking trauma to feeling more, rather than less, runs in several directions at once. So here is the version we can stand behind: neuroception is one compassionate framework for why some people feel so much, so quickly, and so deeply. It may fit some lives well. It is not yet an established explanation, and it is not the only one.

The steady version, the one worth keeping: your body keeps a record of what it had to survive. That record can make you quick to sense trouble, and sometimes quick to sense other people's pain. In some studies it also shows up as more empathy. In others it shows up as less, or as a loud inner life that is hard to name. The science does not hand us a single story, and we would rather say so than sell one.

What the science does support is modest and useful. Safety and steady relationships matter. Evidence-based care exists for trauma. And "sensitive" is not the same thing as "broken." If you carry some of this, you are in good company, and you are allowed to set it down sometimes.

If you are struggling right now, call or text 988 in the United States [1].

Methods, limits and corrections

What we checked. Every cited source was opened on 15 November 2026 (US Central), and each quoted passage or figure was checked against the text we retrieved. For the full texts we could open (the Porges 2022 and 2023 reviews, Greenberg 2018 and its posted data, Pollak and Kistler 2002, Benz 2023, Cogan 2025, Kahhale 2025, Kaplan 2024, Taylor 2022, the open author copy of Grossman 2023, the authors' posted manuscript of Frewen 2008, the APA and VA/DoD guideline documents, and the CDC, MDPI and 988 pages), we read the passages themselves. For several journal articles the publisher blocks automated readers, so we confirmed the abstract through the US National Library of Medicine's PubMed service, matched the citation details against Crossref, and claim only what the abstract says. That applies to the Porges 2003 and 2004 records, Pollak and Sinha, Pollak and Tolley-Schell, Zhang, Ditzer (where we also read the authors' posted supplement), Singer and Klimecki, the Fuchs meta-analysis and the Somatic Experiencing review. Our audit file marks which is which.

Labels and boundaries. We keep a theory's claims apart from tested findings, and we do not treat one study as the field. Effect sizes are as printed in the sources. The "1 to 3 percent," "four in five," "roughly p = .07" and "46 counts across 43 papers" figures are our arithmetic. The rerun of Greenberg's group comparisons, the standardized differences and the group means are our calculation from the authors' posted data, with the method stated under "Checking Greenberg's numbers" in this section.

Checking Greenberg's numbers from the authors' own data. The authors posted their data and analysis scripts on the Open Science Framework [14], so we reran the main comparisons. Our calculation: for each scale we compared the trauma and no-trauma groups with a one-factor analysis of covariance, with sex as the covariate as in the paper, and we computed the standardized difference (Cohen's d) as the difference in group means divided by the pooled standard deviation, unadjusted for sex. We rebuilt the Study 2 scales from the item responses using the item lists in the authors' own script, and the reliabilities matched the paper's. Three things came out of it. First, the paper's text appears to swap two Study 2 results. It prints Empathic Concern as F = 8.62 (ηp² = .019) and Fantasy as F = 5.42 (ηp² = .012). In the posted data it is the other way around: Fantasy gives F = 8.62 (ηp² = .019, p < .01) and Empathic Concern gives F = 5.42 (ηp² = .012, p ≈ .02). Both are significant either way, so the overall conclusion does not change, but we use the corrected assignment. Second, the printed total EQ result in Study 1 (F = 1.47, p = .23) does not reproduce. We get F(1, 383) = 3.16, p ≈ .08, using the authors' own total-score column, whose item scoring we could not rebuild independently. It is not significant by either figure. The same rerun reproduces the printed cognitive-empathy F of 3.20 (p ≈ .07). Third, in standardized terms the significant differences were about 0.3 to 0.5 standard deviations: Study 1 affective empathy 0.32, and in Study 2 Perspective Taking 0.34, Empathic Concern 0.37, Fantasy 0.41 and Global Empathy 0.49. For a feel for the scale, Study 1's affective score (0 to 10) averaged 5.6 in the trauma group and 4.8 in the comparison group (our calculation). By the usual rule of thumb (0.2 small, 0.5 medium) that is small to medium, which fits the eta-squared reading above. A smaller fourth point: the paper prints p = .01 for Study 1 affective empathy, while F = 5.48 on those degrees of freedom gives about .02, which looks like truncating rather than rounding (it prints .003 for a value near .0035). Our recomputation has two more caveats: our degrees of freedom are two lower than the printed ones (383 against 385 in Study 1, 436 against 438 in Study 2). One Study 1 and three Study 2 participants have no sex recorded, which explains our smaller samples, though not exactly the printed degrees of freedom, so we quote the paper's printed figures where we have not recalculated. It is a check of arithmetic, not a replication of the study.

What we could not confirm. The starting message and much popular writing describe posture, proximity and light as neuroception inputs, say trauma survivors read emotions "with high precision," say trauma makes the system "slower to register safety," say co-regulation heals trauma, describe people-pleasing and difficulty identifying one's needs as part of the pattern, say clinicians note a protective origin, and call neuroception "the underlying mechanism" and an "accurate map." We found no verified source for those statements in the form given, and we have worded them accordingly. We did not use any of the blog labels in the starting message as citations.

Limits. The empathy literature is mixed, mostly self-reported and mostly correlational. Several studies cited here are small or from convenience samples. Polyvagal Theory is contested, and this article does not settle it. We are a research and writing team, not clinicians, and nothing here is treatment advice.

Corrections. If you see an error, please use the corrections form on our site.

Sources

Numbers are in order of first citation. Every link was opened on 15 November 2026 (US Central), and each quoted passage was checked against the retrieved copy or, where noted, the abstract.

  1. 988 Suicide & Crisis Lifeline, home page. 988lifeline.org. Official service site.
  2. Porges, S. W. (2003). "Social engagement and attachment: a phylogenetic perspective." Annals of the New York Academy of Sciences, 1008, 31-47. doi.org/10.1196/annals.1301.004; PubMed 14998870. Proposal by the theory's author. Abstract read through PubMed and bibliographic details matched in Crossref; the full article was not read.
  3. Porges, S. W. (2004). "Neuroception: A Subconscious System for Detecting Threats and Safety." Zero to Three, 24(5), 19-24 (ERIC record with abstract). eric.ed.gov/?id=EJ938225. Proposal by the theory's author. Abstract-level record on ERIC; the full article was not read.
  4. Porges, S. W. (2022). "Polyvagal Theory: A Science of Safety." Frontiers in Integrative Neuroscience, 16, 871227. PMC9131189. Review by the theory's author; proposal, not independent confirmation. The author discloses patent rights and royalties for the Safe and Sound Protocol.
  5. Grossman, P. (2023). "Fundamental challenges and likely refutations of the five basic premises of the polyvagal theory." Biological Psychology, 180, 108589. doi.org/10.1016/j.biopsycho.2023.108589. Open author copy: Zenodo. Contested-debate source: a critical review by one author. The abstract quote is confirmed through PubMed; the "no support" passage is from the open Zenodo author copy and may differ from the published wording.
  6. Taylor, E. W., Wang, T., & Leite, C. A. C. (2022). "An overview of the phylogeny of cardiorespiratory control in vertebrates with some reflections on the 'Polyvagal Theory.'" Biological Psychology, 172, 108382. doi.org/10.1016/j.biopsycho.2022.108382. Review, critical of some polyvagal claims. Full text read (the published PDF, via the Aarhus University repository), and the abstract confirmed through PubMed.
  7. Porges, S. W. (2023). "The vagal paradox: A polyvagal solution." Comprehensive Psychoneuroendocrinology, 16, 100200. PMC10724739. Review and reply by the theory's author. A 2024 corrigendum (vol. 18, 100233) exists; it was not part of our checks.
  8. Cogan, N., Morton, L., Campbell, J., Irvine Fitzpatrick, L., et al., and Porges, S. (2025). "Neuroception of psychological safety scale (NPSS): validation with a UK based adult community sample." European Journal of Psychotraumatology, 16(1), 2490329. PMC12057785. Study; 2,035 UK adults, self-report scale validation. Porges is a coauthor, so this is not independent of the theory's originator.
  9. Pollak, S. D., & Kistler, D. J. (2002). "Early experience is associated with the development of categorical representations for facial expressions of emotion." PNAS, 99(13), 9072-9076. PMC124425. Study, 40 children (23 physically abused, 17 not); small sample.
  10. Pollak, S. D., & Sinha, P. (2002). "Effects of early experience on children's recognition of facial displays of emotion." Developmental Psychology, 38(5), 784-791. doi.org/10.1037/0012-1649.38.5.784. Study. Abstract verified through PubMed; the publisher page blocks automated readers, and we found no open full text, so the full article was not read.
  11. Pollak, S. D., & Tolley-Schell, S. A. (2003). "Selective attention to facial emotion in physically abused children." Journal of Abnormal Psychology, 112(3), 323-338. doi.org/10.1037/0021-843X.112.3.323. Study. Abstract verified through PubMed; the publisher page blocks automated readers, and we found no open full text, so the full article was not read.
  12. Fuchs, F. O., Konrad, A. C., Kanske, P., Geyer, L., Trautmann, S., et al. (2026). "The association between childhood maltreatment and emotion recognition: A systematic review and meta-analysis." Journal of Psychopathology and Clinical Science (advance online publication, 15 Oct. 2026). doi.org/10.1037/abn0001173; PubMed 42879220. Review; 40 studies, N = 6,369. Abstract read through PubMed; full text not read, and the abstract does not give volume or page numbers yet.
  13. Centers for Disease Control and Prevention, "About Adverse Childhood Experiences" (page updated Mar. 2, 2026). cdc.gov/aces/about. Government page. The prevalence figure is for high school students, not adults.
  14. Greenberg, D. M., Baron-Cohen, S., Rosenberg, N., Fonagy, P., & Rentfrow, P. J. (2018). "Elevated empathy in adults following childhood trauma." PLOS ONE, 13(10), e0203886. PMC6169872. Study; two online self-report surveys (MTurk), largely correlational. The authors' posted data and scripts (OSF) were also used for the recalculations described in the text. The paper has internal inconsistencies, described there.
  15. Benz, A. B. E., et al. (2023). "Increased empathic distress in adults is associated with higher levels of childhood maltreatment." Scientific Reports, 13, 4087. PMC10008534. Study; 228 adults (83% female), self-report, correlational.
  16. Zhang, H., Gao, X., Liang, Y., Yao, Q., & Wei, Q. (2024). "Does child maltreatment reduce or increase empathy? A systematic review and meta-analysis." Trauma, Violence, & Abuse, 25(1), 166-182. doi.org/10.1177/15248380221145734. Review. Abstract verified (PubMed and the publisher page abstract); full text not read.
  17. Kahhale, I., Byrd, A., & Hanson, J. (2025). "Early life adversity and empathy: A scoping review of past research and recommendations for future directions." Clinical Child and Family Psychology Review, 28(2), 275-321. PMC12340701. Scoping review of 43 articles; the counts are the authors' tally of papers (overlapping), not a pooled effect. Table 1 lists 43 unique manuscripts (44 rows, because one paper is split into two studies). One passage in Section 4 says "41 manuscripts"; that looks like a count left over from before the authors' January 2025 search update added 2 papers (our reading), and we use 43, as in the abstract and Table 1.
  18. Ditzer, J., Wong, E. Y., Modi, R. N., Behnke, M., Gross, J. J., & Talmon, A. (2023). "Child maltreatment and alexithymia: A meta-analytic review." Psychological Bulletin, 149(5-6), 311-329. doi.org/10.1037/bul0000391. Meta-analysis. Abstract verified through PubMed; the publisher page blocks automated readers. The authors' posted online supplement was also read, but the full article was not.
  19. Frewen, P. A., Dozois, D. J. A., Neufeld, R. W. J., & Lanius, R. A. (2008). "Meta-analysis of alexithymia in posttraumatic stress disorder." Journal of Traumatic Stress, 21(2), 243-246. doi.org/10.1002/jts.20320. Meta-analysis, 12 studies. Abstract verified (PubMed and the publisher page abstract). We also read the authors' posted manuscript from the University of Western Ontario repository, which is the pre-publication text and may differ from the journal version.
  20. Singer, T., & Klimecki, O. M. (2014). "Empathy and compassion." Current Biology, 24(18), R875-R878. doi.org/10.1016/j.cub.2014.06.054. Short review. Abstract verified through PubMed; the publisher page blocks automated readers, and we found no open full text, so the full article was not read.
  21. Poli, A., Gemignani, A., Soldani, F., & Miccoli, M. (2021). "A systematic review of a polyvagal perspective on embodied contemplative practices as promoters of cardiorespiratory coupling and traumatic stress recovery for PTSD and OCD." International Journal of Environmental Research and Public Health, 18(22), 11778. mdpi.com/1660-4601/18/22/11778. Systematic review of 6 small studies; the authors call the results preliminary.
  22. Kaplan, J., Somohano, V. C., Zaccari, B., et al. (2024). "Randomized controlled trials of mind-body interventions for posttraumatic stress disorder: a systematic review." Frontiers in Psychology, 14, 1219296. PMC10847595. Systematic review of 26 RCTs. Not specific to polyvagal approaches.
  23. American Psychological Association (2025). "APA Clinical Practice Guideline for the Treatment of Posttraumatic Stress Disorder (PTSD) in Adults" (approved by the APA Council of Representatives, February 2025). apa.org/ptsd-guideline/ptsd.pdf; summary page: apa.org/ptsd-guideline/treatments. Professional guideline; full guideline document read. It applies to adults with PTSD, not to neuroception.
  24. U.S. Department of Veterans Affairs and Department of Defense (2023). "VA/DoD Clinical Practice Guideline for Management of Posttraumatic Stress Disorder and Acute Stress Disorder," Version 4.0, Provider Summary, June 2023. healthquality.va.gov (PDF). Professional guideline; the provider summary was read, not the full guideline. Applies to PTSD and acute stress disorder.
  25. Kuhfuß, M., Maldei, T., Hetmanek, A., & Baumann, N. (2021). "Somatic experiencing - effectiveness and key factors of a body-oriented trauma therapy: a scoping literature review." European Journal of Psychotraumatology, 12(1), 1929023. PubMed 34290845. Scoping review of 16 studies (not a meta-analysis). Abstract read through PubMed; full text not read.
Next
Next

Tubes, Gyros and Flywheels: A History of the Other Ways to Move People