Is Polyvagal Theory scientifically valid?
The wiki met Polyvagal Theory (PVT) secondhand, as the physiological scaffolding of Somatic Experiencing, and flagged it as “contested in the wider literature” with no critical source on hand. The Polyvagal ingest supplies both sides first-hand — Porges (2023) restating the theory and Grossman (2023) arguing it is refuted — so the dispute can finally be stated properly.
What is actually being disputed
Not the clinical value of the framework (both sides note its popularity), but a chain of physiological premises. PVT’s distinctiveness — the “poly” — is the claim that the cardiac vagus has two functionally opposite brainstem sources. If that is false, the three-circuit hierarchy, the dorsal-vagal account of shutdown/freeze, and the uniquely-mammalian social-engagement story lose their mechanism.
| claim | Porges | Grossman/Taylor |
|---|---|---|
| Does the dorsal vagus (DMNX) cause massive bradycardia? | Yes, esp. in immature/compromised organisms | No — DMNX contributes ~nothing to heart rate; NA does it all |
| Is the ventral/myelinated cardiac vagus uniquely mammalian? | Yes (a mammalian repurposing) | No — present across vertebrates |
| Is RSA uniquely mammalian and = cardiac vagal tone? | A mammalian cardiorespiratory form; a valid ventral-vagal index | RSA-like coupling exists in non-mammals; RSA ≠ vagal tone (category mistake) |
| Are the critiques fair? | No — a strawman that redefines RSA | Yes — premises quoted verbatim from Porges (2011) |
Where the evidence points, and where it doesn’t
The anatomical/physiological evidence runs against PVT’s core. Selective ablation, pharmacogenetic inhibition and optogenetic stimulation converge on the nucleus ambiguus as the seat of cardiac vagal control, with the DMNX contributing negligibly — a modern consensus (Grossman cites Veerakumar et al. 2022, Nature) that directly contradicts Premise 1, on which the other four depend. Comparative biology likewise undercuts the “uniquely mammalian” claims. On these points the wiki weights Grossman heavily.
Two things keep the debate formally open rather than closed:
- The parties are not always answering the same question. Porges insists PVT is mammal-centric and that observing cardiorespiratory coupling in a rattlesnake does not touch a theory about mammalian modifications — while Grossman insists the premises, as written, make universal-sounding anatomical claims that the data falsify. Some of the disagreement is about what PVT actually asserts, which a debate cannot resolve by evidence alone.
- Clinical utility is a separate axis. Even if the anatomy is wrong, autonomic state clearly shapes felt safety and threat, and interventions that move RSA can help — see Weng and MABT. The wiki has already recorded that trauma-relevant vagal findings arise without invoking PVT (Bonaz, Weng), which is evidence about the framework’s dispensability, not a verdict on its truth. Kerdabadi et al. (2024) now supplies the same lesson from the pro-PVT side: their meta-analysis of “polyvagal interventions” pools a small but robust benefit (g ≈ 0.26, p < .001) out of a bundle that is mostly generic relaxation — massage, yoga postures, slow breathing, HRV biofeedback, positive psychotherapy — with only the Safe and Sound Protocol actually Porges-designed. Interventions that raise vagal tone helping does not require PVT’s premises to be true; the construct inflation is itself an exhibit for Grossman’s dispensability point, and the paper overstates its own small effect as “medium”/“large.”
The pending evidence
This debate is about mechanism; a parallel question is whether RSA — whatever it indexes — actually predicts the outcomes PVT and the wider vagal literature claim. Three RSA meta-analyses were queued in raw/papers/Polyvagal/ (children’s adaptive functioning, self-control, childhood adversity); all three are now in, and each lands on this evidential-robustness axis, not on mechanism.
Wesarg et al. (2022), pooling ~27,000 participants, finds no overall association between childhood adversity (a canonical developmental risk factor) and either baseline vagal activity or vagal reactivity, with small effects (r ≈ −.05 to −.11) surviving only in clinical samples and for direct/maltreatment adversity. It is a null within the mainstream vmHRV paradigm (authored from the Neurovisceral Integration tradition, Thayer on the byline; it does not test PVT’s dorsal/ventral anatomy), so it neither confirms nor refutes the disputed premises — but it shows RSA’s real-world associations are thinner than the frameworks resting on it assume.
Graziano & Derefinko (2013) is the more pointed of the two for PVT specifically. Across 44 studies and ~4996 children, RSA withdrawal predicts fewer externalizing, internalizing and cognitive/academic problems (each r ≈ −.16, small and adaptive) — but is unrelated to social functioning (r = −.06), which they call their “most surprising” result and flag as “counter to what one would expect based on Polyvagal theory and the Social Engagement System.” Social engagement is the outcome the ventral-vagal story is built to explain, and the RSA-W↔social effect is not only absent overall but flips maladaptive in clinical/at-risk samples (M = +.16 vs −.14 in community samples). Like Wesarg it is a null within the vagal-reactivity paradigm, not a test of the brainstem anatomy — but it removes an empirical prop the social-engagement claim has leaned on, and the authors conclude PVT’s phylogenetic social-behaviour account “has little empirical support for individuals with other significant clinical impairment.”
Zahn et al. (2016) completes the trio from the most cognitive angle: does resting HRV predict laboratory self-control, as the CAN/inhibitory-capacity account (shared by PVT and Thayer’s Neurovisceral Integration Model) requires? Across 26 studies the pooled effect is small (r = 0.15) and, tellingly, drops to a nonsignificant r = 0.04 once publication bias is corrected — so “there might not be an association.” This one speaks more to the Neurovisceral Integration Model than to PVT proper, because Zahn et al. deliberately restricted HRV to Task-Force metrics and excluded Porges–Bohrer RSA, conceding the results “may only partially be generalizable to Polyvagal Theory” — the mirror image of Grossman’s own complaint about what RSA is allowed to mean. All three meta-analyses are nulls within the vagal paradigm, not tests of the brainstem anatomy, but together they thin the predictive base the frameworks rest on.
The intervention-efficacy meta-analysis in that folder is also in — Kerdabadi et al. (2024) — and lands on the clinical axis above: a small pooled benefit for a heterogeneous set of vagal-tone techniques, of no evidential weight for the disputed premises. Filed open: the physiological premises are under strong, expert challenge; Porges disputes that the premises say what the critics quote; and the clinical framework’s usefulness is not settled by either.