Experimental Mental

Tension and Trauma Releasing Exercises (TRE): what the tremor is and what it has been shown to do

Summary

The shaking that Tension and Trauma Releasing Exercises reliably produce is real and is almost certainly ordinary fatigue-enhanced physiological tremor in deliberately tired hip and thigh muscles; the practice is a cheap, plausible relaxation routine with a built-in stop switch, but the story that the tremor discharges stored trauma has never been measured, the clinical evidence is a handful of small pilots, two tiny peer-reviewed trials and one larger trial that never reached a journal, and that larger trial, which tested tremoring against the same exercises without tremoring, found no differ

Why Experimental

TRE as a treatment for any named condition: Experimental. The peer-reviewed randomised evidence is forty-six people across two trials, one null and one positive only after covariate adjustment against usual care. The largest trial exists only as a grant report and is null against a tremor-free placebo. Everything else is uncontrolled, dissertation-level or case material, substantially author-conflicted. There is no systematic review or meta-analysis. This is the definition of Experimental.

TRE for fatigue and self-reported pain in multiple sclerosis: approaching Emerging, held at Experimental. One independent, peer-reviewed RCT with a between-group signal, preceded by a consistent pilot. Held back by n=28, no blinding, usual-care control, adjusted-only significance and no located registration. An independent replication would move it.

TRE as a relaxation practice for sleep and anxiety: Experimental in its own right; Moderate by analogy to progressive muscle relaxation. The ingredients are shared and the one head-to-head found equivalence in nine people, but TRE has not itself been tested against a relaxation control at scale. The analogy is stated as an analogy.

The tremor as ordinary enhanced physiological tremor: Foundational physiology, applied by inference. The underlying tremor physiology is established. Its application to TRE is the parsimonious inference from mechanism, not a measurement, because nobody has recorded the TRE tremor.

The discharge, stored-trauma, neurogenic and autonomic-regulation claims: no tier, because there is no evidence to tier. Stated as absence. The only measurements bearing on them, the tremor-free controls and the controlled HRV null, point away.

The anecdotal claims of height change and spontaneous release: no tier. No located evidence of any kind. Ordinary physiology accounts for the felt experiences.

Safety: Experimental. Small studies, one trial that counted adverse events without describing them, self-selected samples. The practice's own exclusion list is practitioner guidance.

Practical takeaway

If you want to try it as a relaxation practice. Learn the sequence properly, ideally once from a competent instructor or a good instructional source, since the leg positions matter for producing the tremor comfortably. Practise on a mat, knees bent, feet together, with a wall or cushion to hand. Let the tremor run for no more than ten to fifteen minutes, which is the common practitioner convention, then straighten the legs and lie still for a few minutes more. Two or three times a week is what the pilots used. Treat it as you would progressive muscle relaxation or a body scan: a way of unwinding, useful before bed, with a realistic expectation of feeling looser and calmer afterwards and no expectation of anything leaving your body.

Use the stop switch, and use it early. Straightening the legs ends the tremor. The practitioner convention is "less is more": stop while you still feel in control rather than pushing for a bigger shake. If you notice rising distress, a sense of unreality, intrusive images or strong emotion you did not choose, stop, put your feet flat, open your eyes, orient to the room, and do not restart that day.

What to expect from the sensations. Feeling taller or straighter afterwards is real as posture and disc height and is not skeletal growth. Tremors, twitches or emotional waves arriving outside sessions in the first weeks are consistent with a learned motor pattern and with heightened body attention; if they are occasional and mild they are not a sign of anything, and if they are intrusive, uncontrollable or paired with destabilising feelings they are the signal to stop the practice and speak to a clinician.

What it is not for. TRE is not an evidenced treatment for PTSD, complex trauma, anxiety disorder, chronic pain or insomnia. If you have any of those, the evidenced routes are trauma-focused cognitive behavioural therapy, EMDR or prolonged exposure for PTSD, cognitive behavioural therapy for insomnia for sleep, and the standard psychological and pharmacological options for anxiety. TRE can sit alongside those as a relaxation adjunct with your clinician's knowledge. It should not replace them, and a person with a trauma history should not practise it alone in the expectation that it will process the trauma.

Compared with the alternatives. Progressive muscle relaxation delivers the same ingredients with a larger evidence base and no need to produce a tremor. Slow breathing practices, body-scan meditation and gentle stretching before bed are the near neighbours. Choose TRE over these only because you prefer it or find it more engaging; there is no evidence it does more.

Cost. Nothing once learned. Introductory classes and one-to-one sessions are priced like any bodywork session. Certification as a provider is a multi-thousand-pound training business and is not something a person practising for their own wellbeing needs.

Evidence detail

Why This Entry Exists

Tension and Trauma Releasing Exercises, usually shortened to TRE, is a short routine developed by David Berceli. You tire the legs with a wall sit, calf raises and hip stretches, then lie on your back with knees bent and feet together and let the legs slowly close. Within a minute or two the thighs and pelvis begin to shake, and the shaking is allowed to run for ten to fifteen minutes before the legs are straightened to end it. The practice has spread widely because it works in the narrowest sense: nearly everyone who follows the instructions shakes. What has spread with it is a large explanatory story. The tremor is said to be a "neurogenic" reflex that all mammals use to discharge the charge of a threat, which humans have learned to suppress, and which, once reactivated, releases tension and trauma stored in the muscles, restores the nervous system to a state of safety, and completes an unfinished stress cycle.

This entry separates the phenomenon from the narrative, because they have very different evidence behind them. The phenomenon is a reliably producible tremor with a well-understood physiology that nobody in the TRE literature has ever recorded. The narrative is a set of claims about mammals, muscles and memory that has never been tested and in several places contradicts what has been measured. Between them sits a set of anecdotes that travel with the practice: better sleep, less anxiety, pain relief, "processing trauma", shaking or crying arriving days later outside any session, and the recurring report that a person came out of a course standing, or even measuring, taller. Each of those gets handled on its own terms below, because some have a real and ordinary mechanism underneath and some have none at all.

The entry also examines the two frames the practice leans on. The first is yoga and its neighbours: kundalini kriyas, spontaneous movement, trance and "energy release", offered as evidence that induced shaking is an ancient and therefore validated route to healing. The second is "nervous system safety", the vocabulary of polyvagal theory, in which the tremor moves a person from a defensive state into a safe one. Both frames contain something worth taking seriously and both are used to smuggle in conclusions the evidence does not carry. The corpus has already ruled that polyvagal theory is contested at the level of theory and that Realised entries describe autonomic mechanism plainly rather than through its branding; this entry follows that ruling and defers the critique itself.

A note on generosity, because the tone here matters. TRE is free once learned. It needs no equipment, takes a quarter of an hour, and sells nothing at the moment of practice. It plausibly delivers a relaxation effect through the same ingredients as progressive muscle relaxation, which does have evidence. A reader who finds that it helps them unwind and sleep is not deluded, and this entry does not tell them to stop. What it corrects is the explanation, and the one place where the explanation does harm: the idea that a person with a trauma history can treat this as therapy, alone, in place of treatments that have been shown to work.

What bad advice this protects against, in all directions:
• "The tremor is your body releasing stored trauma." Wrong because no study has shown that muscles store trauma, that the tremor carries information about a person's history, or that anything leaves the body during it. The tremor is what tired, loaded postural muscles do. A person with no trauma history at all will shake just as reliably, which is itself the strongest argument that the shaking is about the muscles, not the biography.
• "TRE is a proven treatment for PTSD, anxiety and insomnia." Wrong because the clinical literature is a few small pilots, two peer-reviewed randomised trials totalling forty-six people, and one larger trial that exists only as a grant report. For PTSD specifically, the largest trial ever run, with a placebo arm doing the same exercises without tremoring, found that tremoring added nothing. Evidenced treatments for PTSD exist and TRE is not among them in any clinical guideline.
• "All mammals shake after a threat to discharge it; humans have forgotten how." Wrong as a statement of fact because no animal study supporting it has been located. Animals do tremble after fright, which is adrenaline acting on muscle, and shiver when cold, which is thermoregulation. That either of these "discharges" anything, or that an animal prevented from shaking is harmed, has not been shown. The one systematic study of shaking in a mammal found it clustered at transitions between activities, not after stress.
• "It's just fatigue tremor, so the whole thing is worthless." Wrong as a dismissal. The fatigue-tremor account explains the shaking; it does not abolish the fifteen minutes of lying still, attending to the body and letting muscle tone fall afterwards, which is exactly the recipe of progressive muscle relaxation and has a real evidence base for sleep and anxiety. A skeptic who throws out the practice with the story has overreached in the other direction.
• "Shaking or crying that arrives outside a session days later is a sign the healing is deepening." Wrong, and this one can hurt. A tremor pattern practised daily lowers the threshold for producing it, and intense attention to body sensation raises emotional lability in the days after. Neither is progress. The practitioner community itself treats uncontrolled tremoring as something to damp, not encourage, and intrusive or destabilising episodes are the signal to stop and get support.
• "It gets rid of the nervous system's stored charge, so you can safely do as much as you like." Wrong because practitioner guidance commonly caps tremoring at ten to fifteen minutes (the one large trial allowed fifteen to thirty) and lists a long set of people who should not do it alone, including anyone with dissociation, active PTSD, psychosis or seizures. Those guardrails are a record of what facilitators have seen go wrong, and they only make sense if the tremor is a strong body-attention stimulus rather than a release valve.

What this entry OWNS: the physiology of the TRE tremor; the complete located clinical literature on TRE, described study by study as exactly what each is; the anecdotal claims attached to the practice; the "safety" and "discharge" framing as it applies to TRE; the practical and safety guidance for someone considering it as self-help.

What this entry DEFERS: the polyvagal critique itself, to breath_movement_coupling, which carries the primary citations and the corpus ruling; kundalini and kriya practice, breathwork-induced states and intensive embodied protocols, to intensive_embodied_practice; progressive muscle relaxation and mindfulness as comparators, to anxiety_lifestyle_levers, behavioral_sleep_onset_protocol, mental_mindfulness_meditation and meditation_styles_compared; heart-rate-variability measurement and its confounds, to hrv_training_for_stress_resilience and vagal_tone_practices; evidence-reading method, to rct_vs_observational_evidence and reading_a_study_before_its_abstract.

Evidence

Read the tiers, not the thesis. The physiology of tremor is Foundational and is stated first because it is the only strongly established science in this entry. Everything downstream, meaning every outcome claim made for TRE, sits at Experimental, with a single small signal in multiple sclerosis approaching Emerging. Several of the claims that travel with the practice have no located evidence at all, and where that is the case it is stated as absence rather than softened to "limited". As of this writing there is no peer-reviewed systematic review or meta-analysis of TRE anywhere; the only review is a 2023 master's dissertation.

1. What the tremor is. Physiology (established), Foundational; applied to TRE by inference, because no one has recorded the TRE tremor.

Every person has a physiological tremor, an oscillation of roughly 8 to 12 Hz made up of a mechanical-reflex component (the limb's own resonance plus the stretch-reflex loop) and central components (Deuschl, Raethjen, Lindemann & Krack 2001, Muscle & Nerve; Elble 2003, Clinical Neurophysiology, with an EMG peak at 9 to 12 Hz; Elble 2026 review). Its amplitude rises with muscle fatigue, isometric loading, adrenaline, anxiety and cold. Enhanced physiological tremor is a named category in the Movement Disorder Society consensus classification (Deuschl, Bain & Brin 1998). Directly relevant to TRE, Ebenbichler et al. 2000 (Clinical Neurophysiology, n=18) recorded the quadriceps during sustained isometric knee extension at 30, 50 and 70 percent of maximum and found tremor power in the 6 to 20 Hz band changing with fatigue in a load-dependent way, with all loads converging on the same tremor level at exhaustion. Wu et al. 2026 (European Journal of Sport Science, 22 trained men) found 8 to 12 Hz tremor rose after fatiguing exercise in the endurance-trained half of the sample but not the strength-trained half, a reminder that the size of the fatigue effect varies with training history even though the direction is the same. The TRE protocol pre-fatigues the quadriceps, adductors and hip flexors, then holds the legs in a partly loaded butterfly position and slowly closes the knees. That is a textbook recipe for enhanced physiological tremor and reflex-driven oscillation in a loaded posture. The practitioner literature's own mechanical account, an alternating cycle between the Golgi tendon organ and the stretch reflex, describes ordinary spinal reflex physiology.

The word "neurogenic" needs unpacking. In neurology it means centrally generated oscillation, and nobody has shown that for TRE. In everyday reading it implies a special nervous-system reflex distinct from fatigue tremor. All tremor is generated by the nervous system, so the label adds nothing, and the distinctness it implies has never been demonstrated: no EMG, accelerometry or frequency characterisation of the TRE tremor exists in the indexed literature, on the certifying body's research list, or in the 175-page scoping review (zero hits for the relevant terms). The parsimonious account has never been tested against the practice's account by the simple step of recording the tremor. Cui bono: the certifying body and providers benefit from the tremor being a special reflex rather than a fatigue by-product; the physiology itself has no owner.

2. Post-fright shaking is adrenaline acting on muscle, not a discharge valve. Physiology (established), Strong.

The shaking people notice after an exam, an accident or a fright is catecholamine enhancement of physiological tremor: circulating adrenaline raises tremor amplitude through beta-2 adrenergic receptors, and the standard pharmacological account, built on decades of work with beta-agonists and beta-blockers, places most of that effect on receptors in the muscle and its spindles, which is why a beta-2 agonist inhaler makes the hands shake and why propranolol is used for the tremor of stage fright. Baker & Baker 2012 (PLoS ONE, eight healthy volunteers) is the cleanest modern demonstration: a single dose of salbutamol, a beta-2 agonist, increased tremor amplitude, and propranolol reduced tremor in the 6 to 12 Hz band, while their corticomuscular coherence data led them to suggest that a central site contributes as well. Whatever the exact split between muscle and brain, the point stands: adrenaline shaking is a side-effect of adrenaline on the tremor system. Shivering is a separate hypothalamic thermoregulatory system. Neither is a valve that empties anything.

3. Stress does raise resting muscle tone; muscles do not store memory. Physiology (established) for the first half, no located evidence for the second.

The honest kernel under "the psoas stores trauma" is that psychological stress measurably raises resting muscle activity. Lundberg et al. 1994 (International Journal of Behavioral Medicine, n=62 women) found that mental arithmetic, a Stroop task and cold pressor raised trapezius EMG alongside blood pressure, heart rate, catecholamines and cortisol, more so with concurrent physical load. Stress-related hypertonicity is real and reverses with relaxation, which is the opposite of storage. No psoas-specific study under stress was located, and there is no described mechanism by which a muscle encodes a memory. "Stores trauma" has no located evidence of any kind.

4. The Danish multiple sclerosis work: the strongest evidence for TRE, and independent. Human uncontrolled pre-post, then a small HUMAN RCT. Emerging for MS fatigue only.

Lynning, Svane, Westergaard, Bergien, Gunnersen & Skovgaard 2021 (Journal of Traditional and Complementary Medicine, PMID 34522632) ran a nine-week uncontrolled pilot. Eleven people with MS were recruited through the Danish MS Society's mailing list, nine completed, with weekly sessions from a certified TRE instructor and daily home practice. Fatigue on the Modified Fatigue Impact Scale fell from 43.7 to 22.0 (p=0.0014), eight of nine daily-rated symptoms fell, stress fell 1.2 points and sleep quality rose 0.4 on a five-point scale. The one objective measure, ankle plantar-flexor reflex torque by a portable spasticity device, did not change significantly (right p=0.057, left p=0.406) while self-reported spasticity fell. That dissociation between what people felt and what the instrument measured is the most informative finding in the whole TRE literature. Funded by the Danish MS Society, authored by its researchers and a University of Copenhagen neuroscientist, with no TRE-organisation funding, though delivered by a certified provider.

Skovgaard, Trénel, Hanehøj & Lynning 2025 (Advances in Mind-Body Medicine 39(1):4–12, PMID 40064004) followed with the only peer-reviewed randomised trial of TRE to report a between-group benefit. Twenty-eight people with MS were randomised to eight TRE sessions plus daily home practice or to usual care, with no active or attention control and no blinding. Unadjusted analyses showed no significant between-group differences. After adjustment for covariate imbalance, the primary fatigue outcome reached P=.03 (physical subscale P=.06). Daily self-rated spasticity and pain fell across the intervention period and a five-U-turn gait test improved; sleep measured by Fitbit showed no significant result in the abstract. Per-arm numbers, dropouts, adverse events, funding and conflicts were not retrievable from the abstract, the full text is paywalled, and no trial registration was located. The authors themselves say the effects were seen "when adjusting for covariate imbalance", conclude only that a real effect on fatigue "cannot be ruled out", and call for further research. Cui bono: independent patient-society research team, no TRE-organisation authorship; this is the least conflicted work in the field and it deserves its due. Its honest limit is that it supports "a supervised TRE course may modestly improve self-reported fatigue and pain in MS compared with doing nothing". It cannot separate tremoring from eight sessions of attention, gentle stretching, a daily routine and expectancy, and an adjusted-only result in twenty-eight unblinded people is fragile.

5. The complete list of randomised trials, and what the two with a tremor-free control found. HUMAN RCT ×2 peer-reviewed, ×1 grant report, ×3 dissertation. Experimental.

Harrison, Keating & Morgan 2018 (Journal of the American Board of Family Medicine 31(5):783–794, PMID 30201675) is the only other peer-reviewed RCT. Eighteen people with restless legs syndrome were randomised to six weekly sessions of "exercises to stretch and fatigue lower limb muscles and invoke therapeutic tremors" plus home practice, or to a discussion group. At six weeks there was no significant between-group difference on any outcome: restless-legs severity, global severity, stress, the Pittsburgh Insomnia Rating Scale, or depression. Both groups improved similarly, and the authors invoke repeated assessment, regression to the mean and the group environment. Independent physiotherapy authors at Monash. The discussion group doing as well as TRE is the clearest demonstration in the literature that the improvements people report are non-specific.

Kent 2018 is the largest and best-designed TRE trial ever run, and it has never been published in a peer-reviewed journal. It exists as a 26-page final grant report to the US Army Medical Research and Materiel Command (July 2018), archived by the Defense Technical Information Center and hosted in full on the certifying body's research page, so its numbers can be read directly. Funded by a US Department of Defense award of $494,859 to the Carl T. Hayden Medical Research Foundation at the Phoenix VA, with Martha Kent as principal investigator, it screened 140 veterans, required a PTSD diagnosis confirmed on the CAPS-5, excluded anyone in concurrent prolonged exposure, cognitive processing or other trauma-focused therapy, and randomised in waves to three arms: standard TRE with tremoring; the same exercises without the lying-down tremor phase, presented to participants as "placebo yoga"; or a wait list. Sessions were twice weekly for four weeks, sixty minutes, in groups of ten, all delivered by the same trainer; the people administering the tests were blind to allocation; treatment fidelity was rated by Berceli himself, who also trained the trainers. Ninety-two completed pre- and post-testing (TRE 30, placebo 32, wait list 30) and 83 completed the six-month follow-up. On the primary outcomes, both the TRE and placebo groups improved on the PTSD Symptom Scale and held the gain to six months while the wait list did not, but the group-by-time interaction was not significant; on depression only the placebo group improved within itself, again with no between-group difference. Affective pain and insomnia improved in the sample as a whole with no significant difference by arm (insomnia p<.09, both active arms drifting down and the wait list flat), wellbeing did not move in anyone, and physical health did not change. The only results favouring TRE were cognitive: attention improved more in TRE than in the wait list, and digit span improved more in TRE than in the placebo group (interaction p=.06), while the placebo group improved more than TRE on a category-inhibition task. In plain words, with a tremor-free placebo, tremoring added nothing measurable to PTSD, depression, pain or sleep. Twenty-eight adverse events were logged across the three years under the ethics-board procedure, but the report does not say what they were, which arm they occurred in or whether any was judged related to the exercises. An abstract was accepted as a poster at the International Society for Traumatic Stress Studies meeting in November 2018; the report says papers were planned, and none has appeared. A null result at that cost sitting in a grant archive rather than a journal is itself a signal about the field. Cui bono: the founder trained the trainers, rated fidelity and co-authored the poster; the result is adverse to TRE's core claim, which is a point in the report's favour.

Three dissertation RCTs exist and are grey literature. Berceli 2009 (Arizona State University, n=61 students, two weeks) compared TRE with the same exercises without tremor: state anxiety improved in the TRE group, another arousal scale did not change, and the heart-rate stress response fell slightly more in the tremor-free control. Atterbury 2019 (Stellenbosch, n=36 idiopathic Parkinson's, nine weeks) compared TRE with an exercise control and found gait and mood scores favouring TRE. Swann 2019 (Saybrook, n=9, PTSD with chronic low back pain, four weeks) compared TRE with progressive muscle relaxation and found it "at least as effective", with low baseline symptoms. Johnson & Naidoo 2017 (South African Journal of Psychology, n=63 teachers, self-assigned) found personal burnout reduced against a waitlist and work burnout unchanged. Across every randomised design the total analysed is roughly 244 participants, of whom only forty-six are in peer-reviewed trials. Two trials isolated the tremor with a tremor-free exercise control, Berceli 2009 and Kent 2018, and neither showed a tremor-specific advantage on its stress or PTSD outcomes.

6. The founder's pilot and the provider surveys. Human uncontrolled pre-post, founder- or provider-authored. Experimental.

Berceli, Salmon, Bonifas & Ndefo 2014 (Global Advances in Health and Medicine 3(5):45–48, PMID 25568824) is a single-arm ten-week pilot in twenty-one staff at a children's village in Cape Town. A wellness quality-of-life questionnaire rose from 3.81 to 4.35 (P<.05) across five domains; participant diaries could not be analysed for literacy reasons; there was no control, no blinding and no objective measure. The first author developed TRE and disclosed receiving royalties; the second author is a TRE trainer. Exactly what it is: a twenty-one-person uncontrolled pilot run by the method's founder, in a stressed workplace where ten weeks of weekly group attention would move any wellbeing questionnaire.

Beattie & Berceli 2021, an unpublished "global case study" hosted on the TRE research page, recruited clients and trainees through certified providers. Fifty-one completed baseline, twenty-two the four-week survey, seventeen the twelve-week survey and fourteen all three, an attrition of 57 to 73 percent. Fifty-nine percent were already doing other therapies. Perceived stress fell at four and twelve weeks in the completers; flourishing was significant only at twelve weeks; pain self-efficacy was non-significant throughout; the test across all three timepoints in the fourteen who completed everything was non-significant for stress. The authors' own recommendations were objective HRV measures, an active control, and "evaluation required from researchers outside of the TRE community". Cui bono: founder co-author, distributed by the certification body; the completers are the satisfied.

7. Anxiety. Human uncontrolled pre-post ×2, dissertation RCT ×1 mixed, grey null ×1. Experimental.

Berceli 2009 (above) is positive on one of two anxiety scales with the heart-rate measure favouring the control. Oh & Shin 2021 (Asia-Pacific Journal of Convergent Research Interchange 7(11)) measured state anxiety before and after a single three-hour training in twenty-five graduate students with no control; the authors are affiliated to a brain-education institution. McCann 2011 (University of Cape Town master's, n=50 course participants) found no significant change in state anxiety or the SF-36 after training. Botha 2022 (South African dissertation, n=11) found state anxiety fell after one session on an uncontrolled measure. Net: no peer-reviewed controlled evidence for anxiety. State anxiety falls immediately after almost any three-hour supervised body practice.

8. Sleep. HUMAN RCT ×2 null between groups, small uncontrolled positive ×1. Experimental.

Sleep is one of the commonest testimonials and every controlled test so far shows the control arm improving just as much. Harrison 2018: insomnia ratings improved equally in the TRE and discussion-group arms. Kent 2018: the Insomnia Severity Index improved in the whole sample with no significant difference against the tremor-free placebo (the group-by-time test was p<.09, with both active arms drifting down and the wait list flat). Skovgaard 2025: Fitbit sleep, no significant result in the abstract. Lynning 2021: a 0.4-point rise on a five-point daily rating, uncontrolled. Heath & Beattie 2019: one man's self-report. A registered Cairo trial has sleep as a co-primary outcome but had not started recruiting at the time of writing.

9. Pain and spasticity. HUMAN RCT ×1 small positive on a subjective secondary outcome, ×2 null, grey nulls. Experimental.

Skovgaard 2025's daily self-rated pain and spasticity fell against usual care, unblinded and subjective. Lynning 2021's objective reflex torque did not change while self-reported spasticity fell. Kent 2018's affective pain improved in the whole sample with no between-group difference. Swann 2019 found TRE no better than progressive muscle relaxation. Beattie & Berceli 2021's pain self-efficacy was non-significant at both timepoints. A Slovenian fibromyalgia master's thesis was not retrieved.

10. PTSD and "processing trauma". HUMAN RCT ×1 null, never peer-reviewed; human controlled non-randomised ×1 with unverifiable details; case reports. Experimental.

The one adequately powered placebo-controlled test, Kent 2018, was null on PTSD and depression, and its full report is public. Heath & Beattie 2019 (Journal of Military and Veterans' Health 27(3):35–40) is a single case: a 48-year-old former Australian soldier with PTSD after a brain injury attended a two-day TRE training, self-practised, and his Perceived Stress Scale fell from 28 to 18 at one month and 16 at four months while he continued specialist medical and psychological care; the authors note the PSS was the only validated measure, and both authors are TRE providers. Parker, Shook, Washington, English & Tatum 2024 (Psychology 15(1), published by SCIRP, an author-pays publisher on Beall's list) reports eight weekly group sessions for East African refugee women against a delayed-treatment control, with reductions on the Harvard Trauma Questionnaire; the full text was inaccessible, so per-arm numbers, whether allocation was randomised, attrition and effect sizes are unverifiable, and the design as described is a small non-randomised controlled pre-post in which weekly group contact in a refugee community is itself an active ingredient. Swann 2019 (n=9) found TRE roughly equal to progressive muscle relaxation. Unpublished case material on firefighters and Ukrainian psychotherapists exists and adds nothing controlled. TRE appears in no PTSD clinical guideline.

11. Has anyone measured what TRE does to the autonomic nervous system? Absence, with three weak HRV measurements. Experimental at best.

No cortisol study exists. Heart-rate variability has been measured three times, all weakly. Berceli 2009 (n=61, dissertation): the heart-rate stress response fell slightly more in the tremor-free control. Almeida & Rodrigues 2021 (unpublished, hosted by the certifying body, first author a TRE provider): eight women, a single session, five-minute windows before, during and after; only mean heart rate and RR interval changed significantly, every frequency-domain index was non-significant (LF p=0.401, HF p=0.123, LF/HF p=0.362), the "during" window was recorded while participants were shaking, and the facilitator instructed deeper breathing during it, which alone raises HRV. The paper's text nonetheless describes an increase in vagal activity. Botha 2022 (dissertation, n=11 people with epilepsy, one thirty-minute session against a control): no effect on heart rate, blood pressure or HRV. So the claim that "tremoring activates parasympathetic pathways" rests on no supporting measurement and one controlled null. A Beijing protocol (Zhao et al. 2024, BMC Complementary Medicine and Therapies 24:403, ChiCTR2100044553) plans HRV, cortisol, EEG and fMRI in 140 adolescent–mother dyads as an add-on to pharmacotherapy; it is independently funded, its background could find nothing stronger to cite than Berceli 2009 and the Lynning pilot, and as of this writing no results have been published.

12. The mammalian discharge premise. Practitioner claim; animal observation ×1 pointing the other way. No located supporting evidence.

"All mammals shake after a threat to discharge activation; humans suppress it" traces to Peter Levine's wildlife observations in Somatic Experiencing and is repeated in TRE materials without a primary citation: Kent's funding abstract asserts it uncited, other TRE authors cite Berceli's books. The tonic-immobility literature (Carli & Farabollini 2022, Progress in Brain Research; Woodruff 2025, Frontiers in Psychology) describes the animal freeze ending under risk appraisal and access to refuge, and nowhere documents post-threat shaking as a discharge with a protective function. The only systematic study of post-arousal shaking in a mammal, Bryce, Nurkin & Horowitz 2024 (Animals 14(22):3248; 96 dogs observed at a public dog run and a private daycare), found shake-offs clustered after greetings, interruptions and other transitions between activities, with posture and proximity stable afterwards, and the authors found no support for reading them as stress signals, proposing a behavioural-transition function instead. Nobody has shown that animals prevented from trembling develop persistent dysregulation. "Completing the stress cycle" is a popular framing with no located test. Cui bono: the premise underwrites both the TRE and the Somatic Experiencing training economies; the latter's own scoping review (Kuhfuß et al. 2021) found sixteen mixed-quality studies and called for unbiased RCTs.

13. Getting taller. Physiology (established) for the ordinary account; no located evidence of any kind for the claim. Anecdote.

No case report, survey item or study mentions height. Adult stature varies through the day by about 19 mm, 1.1 percent of height, with 54 percent of the loss in the first hour after rising and about 70 percent regained in the first half of the night (Tyrrell, Reilly & Troup 1985, Spine, n=8; 17.6 to 17.9 mm in Healey et al. 2011, Clinical Biomechanics, n=16). Lying supine for fifteen to twenty minutes regains millimetres through disc rehydration. Reduced hip-flexor and adductor tone after fatigue shifts pelvic tilt and felt uprightness. Self-measured height carries error of the same order as the effect. A felt "taller" is disc height plus posture plus attention, inside normal daily variation. No adult skeletal height change from any exercise exists.

14. Spontaneous release outside sessions. Anecdote; no located data on frequency.

Nothing in the literature counts it. Ordinary accounts exist: a rehearsed tremor pattern acquires a lowered threshold and can be triggered by cue or fatigue, and the fact that almost everyone produces the tremor within minutes of first instruction is itself the plainest evidence that it is a learnable motor pattern rather than a buried reflex; heightened attention notices normal hypnic jerks and fasciculations; and emotional lability after intense interoceptive practice under a trauma-release expectancy is well described across relaxation and yoga practices and is not specific to tremoring. Practitioner guidance commonly caps tremor at ten to fifteen minutes and instructs stopping if overwhelmed, which means the practitioner community treats uncontrolled release as an adverse event to be damped, not a benefit.

15. The cross-cultural shaking frame. Anthropological review, Moderate for the route-to-state claim; nothing for a discharge function.

Seligman & Kirmayer 2008 (Culture, Medicine and Psychiatry) frame trance, possession and ritual dissociation as neurobiologically grounded capacities shaped and interpreted by culture, not pathological when culturally consonant. Induced shaking recurs in the San healing dance, Pentecostal worship, Balinese trance, dynamic meditation and kundalini kriyas, and the recipe is shared: rhythm, fatigue or hyperventilation, expectancy and a group container. Ubiquity is good evidence that humans reliably find shaking states meaningful and often relieving. It is not evidence of a discharge mechanism, and the ethnographies locate the healing in community, music and meaning, which solo TRE strips out. The same literature carries a warning: Lindahl et al. 2017 (PLoS ONE, sixty Western Buddhist practitioners recruited for meditation-related difficulties) documented involuntary movements and "somatic energy" alongside serious destabilisation, and case series of meditation-associated psychosis exist (Sharma et al. 2019, twenty-eight cases). Those samples were selected for difficulty and describe far higher doses than a fifteen-minute TRE session; the relevance is that involuntary-movement states have a documented adverse tail in vulnerable people, and TRE has never been examined for the same tail.

16. Safety data. Small-study adverse-event reporting; the one trial that counted events never described them. Experimental.

No located study describes the adverse events it collected. The one trial with a formal adverse-event procedure, Kent 2018, logged twenty-eight adverse events across three years in a study of 92 completers, but the report gives no description, arm, severity or judgement of relatedness, so those events can be read neither as harms of TRE nor as reassurance, and that trial excluded veterans with active suicidality, psychosis or substance dependence. Lynning 2021 reported none in nine people, Botha 2022 no adverse cardiac autonomic response in eleven, Harrison 2018 none, Skovgaard 2025 nothing in the abstract, and the Zhao protocol does not state an adverse-event plan in its abstract. The only side-effect list found is secondary and uncited: mild nausea, headache and muscle aches (Hall 2020). No case report of flooding, dissociation or re-traumatisation attributable to TRE was located in indexed literature, and none would be expected from a field in which the one study that counted events did not say what they were. "No reported harms" means nobody has described what was counted, not shown safe, and it says nothing about PTSD or dissociative populations practising alone.

17. Reviews. Grey scoping review ×1; no peer-reviewed synthesis.

Roos 2023 (University of Pretoria master's dissertation in occupational therapy, JBI framework, English-only) found eighteen studies and twenty records totalling 660 participants: six RCT records (four dissertations, Harrison, Kent), one quasi-experimental, seven single-group pre-post, one survey, two case studies, three qualitative; all non-probability sampling; ten of eighteen recommending controlled trials and five recommending objective physiological measures. Its conclusion: results "should be considered as preliminary" and "tentative due to variability in methodological quality". A Canadian rapid-response report from 2015 exists by title but its content could not be retrieved and is unverifiable. Independent critique: Hall 2020 (Science-Based Medicine) reads the tremor as fatigue tremor, the testimonials as unreliable, relaxation alternatives as likely equivalent, and the practice as "probably safe to try". The certifying body's own research page describes the null restless-legs trial as showing "therapeutic value" and the unpublished eight-person HRV paper as showing TRE "regulates the autonomic nervous system".

Mechanism

What happens in a TRE session, described plainly and in order.

The tremor. The wall sit, calf raises and hip stretches fatigue the quadriceps, adductors and hip flexors. Lying supine with the soles together and knees apart, then slowly closing the knees, puts those tired muscles under a sustained low-grade isometric load at a length where the stretch reflex is active. Fatigue recruits higher-threshold motor units and raises the gain of the reflex loop; the limb's own mechanical resonance and the stretch-reflex cycle interact; and the result is a visible oscillation in the usual physiological band, amplified further if the person is anxious or has adrenaline circulating. Practitioners describe the shaking "travelling up the spine" into the pelvis and trunk as the session proceeds; whether that spread reflects an innate programme, or instruction, postural change and expectation, has not been tested. Straightening the legs removes the load that sustains the oscillation and ends the tremor within seconds. That stop switch is coherent under the fatigue-tremor account. It is much less coherent under a discharge account, where interrupting a release halfway ought to be either impossible or harmful, and it is neither.

Why people feel better afterwards. Nothing here requires the tremor to be special. The session contains tense-then-release across large postural muscles, ten to twenty minutes lying still, interoceptive attention, slow breathing that facilitators often cue, post-exertion parasympathetic rebound, expectancy, and in group settings a supportive social container. That is the ingredient list of progressive muscle relaxation, which has a real if heterogeneous evidence base for sleep and anxiety (Donato et al. 2026 meta-analysis of thirty-one RCTs, 2,277 participants; Muhammad Khir et al. 2024 systematic review). Stress-related muscle hypertonicity is real and falls with relaxation, so "I feel less braced" after a session is exactly what you would expect. The one trial that compared TRE with progressive muscle relaxation (Swann 2019, n=9) found them roughly equal, the trial that compared it with a discussion group (Harrison 2018) found the same, and the trial that compared it with the identical exercises minus the tremor (Kent 2018) found the same again.

What has not been shown. That the tremor differs from ordinary enhanced physiological tremor (never recorded). That muscles store trauma or that anything "leaves" during shaking (no evidence, no proposed mechanism). That memory reconsolidation or any trauma-memory change occurs during or after tremoring (no study). That cortisol, HRV or any autonomic marker moves in a TRE-specific way (no cortisol study; the HRV measurements are one confounded uncontrolled study, one controlled null, and one dissertation where the control did slightly better). That any mammal discharges threat through shaking (no located evidence; the one systematic mammal study favours a transition function).

On "safety". What TRE practitioners mean operationally by a nervous system moving into safety is a supervised, low-demand, supine, eyes-open session in which involuntary movement is reframed as benign and can be ended at will. Controllability, predictability and permission are ordinary conditions for reduced sympathetic arousal and would apply equally to any relaxation practice. The felt state, calmer, slower, less braced, is real and reportable. The measured autonomic change is essentially unshown. The explanatory apparatus that the practice borrows from polyvagal theory, in which the tremor shifts a person from a "dorsal vagal" defensive state into a "ventral vagal" safe one through "neuroception", is the practice's vocabulary, not established mechanism; the theory is contested at its foundations and the corpus critique lives in breath_movement_coupling. The parts of that vocabulary that do useful work, slow breathing, attention to body state, a safe relationship and permission to stop, do not need the theory.

Risks And Contraindications

Physical. For a healthy adult the physical dose is small. No musculoskeletal case report was located. Plausible minor risks from the mechanics are delayed-onset soreness in the thighs and adductors, cramp, knee-cap load during the wall sit, and lumbar or sacroiliac irritation from sustained pelvic oscillation in someone with existing instability or hypermobility. Anyone with recent surgery, a serious injury in the last three months, or an unstable spine or joint should not do it without clearance. This is reasoning from mechanism plus absence, not evidence of safety.

Psychological, and this is where the risk lives. The practice combines a strong involuntary body stimulus with sustained interoceptive attention and an expectancy that trauma is surfacing. In people with a trauma history, dissociation or a fragile sense of body ownership that combination can produce flooding: overwhelming affect, intrusive memories, depersonalisation or a dissociative episode. The self-regulation the practice depends on, noticing rising distress in time to straighten the legs, is exactly the capacity impaired in dissociation. No study has quantified these events, because the one trial that counted adverse events did not describe them and no other study collected them. Neighbouring intensive practices that produce involuntary movement have a documented adverse tail in vulnerable people at high dose.

The practice's own exclusion list, offered as practitioner guidance rather than evidence. Provider bodies commonly exclude or gate: pregnancy; epilepsy or seizure-like episodes; psychosis, bipolar disorder, suicidal ideation, self-harm or personality disorder without the agreement of a GP or mental-health team; severe or active PTSD or complex PTSD; dissociative disorders or frequent dissociation; significant cardiovascular disease; surgery or serious injury within three months; and "significant difficulty self-regulating". The certifying body's own page is softer, calling TRE "generally safe and effective for most people" while directing those with a complex trauma history to a certified provider and stating it is "not a substitute for medical or psychological trauma recovery procedures". Read the longer list as a record of what facilitators have watched go wrong.

Who should not do this alone. Anyone with PTSD, complex trauma, a dissociative disorder, psychosis or bipolar disorder. If such a person wants to try it, that is a conversation with their treating clinician and a practice done with a facilitator who knows their history, as an adjunct to evidenced treatment.

Adverse events were never described. Only one trial counted them, and it logged twenty-eight events in three years without saying what they were, in which arm, or whether any was related to the exercises; no other located study reports a plan for capturing harms. The absence of described harm across roughly 660 mostly self-selected participants is weak reassurance and not evidence of safety in clinical populations.

Misattribution. A person who starts to attribute unrelated tremor, twitching or neurological symptoms to "release" may delay seeking assessment. New tremor that appears at rest, is asymmetric, persists outside practice, or comes with weakness, stiffness or gait change warrants a medical appointment, not more sessions.

Controversy

Nature of the controversy. This is not a dispute about whether the tremor happens; it happens. It is a dispute about what the tremor means, whether it does anything a comparably long relaxation practice would not, and whether the framing around it is harmless colour or a claim that misdirects vulnerable people.

Position A: the practice's account. The tremor is an innate neurogenic mechanism shared with all mammals for discharging the activation of threat. Humans suppress it. Reactivating it under safe conditions releases chronically held tension and stored traumatic charge from deep postural muscles, particularly the psoas, restores the nervous system to a regulated, safe state, and allows unprocessed stress to complete. Its adherents point to a growing literature, to the MS trial, to widespread testimonials, and to the practice's ancient cross-cultural cousins.

Position B: the skeptical account. The tremor is enhanced physiological tremor in fatigued, loaded muscles, which is why it appears in everyone regardless of history and stops when the load is removed. The mammalian discharge story is an anecdote with no ethological evidence. "Stored trauma in muscle" has no mechanism and no evidence. The literature is small, mostly uncontrolled, substantially founder- or provider-authored, and its two tremor-versus-no-tremor tests were null. Whatever people gain is the relaxation they would gain from progressive muscle relaxation, and the trauma framing is at best unnecessary and at worst a reason for someone who needs treatment to substitute a mat exercise for it.

The funding/bias dimension: cui bono, both ways. The practice is trademarked and copyrighted, and only authorised trainers may certify providers; full certification runs into the low thousands of pounds or dollars, with paid modules, directories, books and video. The certifying body's research page inflates: a null trial is described as showing therapeutic value and an unpublished eight-person HRV study as showing autonomic regulation. The founder is first author on the one published pilot, trained the trainers and rated treatment fidelity for the largest trial and co-authored its poster, and several of the positive uncontrolled studies are authored by trainers. That is standard proponent behaviour and it is worth naming. On the other side, nobody profits from TRE being dismissed. There is no pharmaceutical or device incumbent it displaces; the skeptic's stake is reputational, and the most prominent skeptical review still concluded it was probably safe to try. The best evidence in the field comes from a patient society with no stake in the method, and it is modestly positive. The "sells nothing once learned" argument genuinely cuts in the practice's favour as a practice, while the certification layer behaves like any wellness-training business.

Realised Position: The phenomenon is real and ordinary; the narrative is unsupported; the experience people report is not imaginary. TRE is a free, low-dose relaxation exercise with a built-in stop switch that plausibly delivers the benefits of progressive muscle relaxation through the same ingredients. Treated that way, it is a reasonable thing to try. The explanatory story, mammalian discharge, stored trauma, neurogenic release and a nervous system moved into safety, has never been measured, contradicts the two trials that isolated the tremor, and should be understood as the practice's vocabulary rather than as mechanism. The one place the story does harm is when it licenses a person with a trauma history to treat this as therapy, alone, instead of the treatments that have been shown to work. The entry describes the practice accurately, corrects the story, and holds the line on that use case.

Cross-Pillar Connections

Mental. TRE's real relationship is to the relaxation and stress-management literature. The comparators that matter are progressive muscle relaxation and body-scan practice (anxiety_lifestyle_levers, mental_mindfulness_meditation, meditation_styles_compared); the stress physiology it invokes is handled in chronic_stress_management and stress_response_is_adaptive_but_context_dependent, which already make the point that acute arousal is adaptive and does not need "discharging". The polyvagal question is owned by breath_movement_coupling; HRV measurement and its confounds by hrv_training_for_stress_resilience and vagal_tone_practices; the plain autonomic account by autonomic_nervous_system_balance.

Physical. The tremor is a movement-physiology phenomenon and the felt loosening afterwards is a change in postural muscle tone. movement_fascia_mobility and physical_counter_modern_postures cover the hip-flexor and pelvic-tilt mechanics behind "standing taller"; low_back_pain_evidence_and_management is the entry to read before recommending sustained pelvic oscillation to anyone with an unstable or painful back.

Sleep. The plausible sleep benefit is the plausible sleep benefit of any pre-bed relaxation routine; behavioral_sleep_onset_protocol is where the evidenced version lives.

Cross-pillar and method. intensive_embodied_practice carries a "neurogenic tremoring" pathway written before this entry existed; this entry now owns TRE, and where the two disagree, this one reflects the located evidence. rct_vs_observational_evidence and reading_a_study_before_its_abstract are the tools for reading a field made almost entirely of uncontrolled pre-post pilots and one adjusted-only trial.

What would change our mind

Falsifiability: explicit upgrade/downgrade criteria from source

Toward a higher tier for TRE as a relaxation practice: a randomised trial with a progressive-muscle-relaxation or attention-matched control, pre-registered, blinded outcome assessment, at least a hundred participants, showing TRE outperforms the comparator on sleep or anxiety. Even a well-run equivalence trial would earn it a place alongside PMR at PMR's tier for those outcomes.

Toward taking the tremor-specific claim seriously: a recording study characterising the TRE tremor by EMG and accelerometry and showing it differs in frequency, distribution or central coherence from enhanced physiological tremor in matched fatigue conditions; and a replication of the tremor-versus-no-tremor design of Kent 2018 with a published, pre-registered, positive result. Peer-reviewed publication of the Kent trial itself, with the arm-level data its grant report already contains, would move the field's credibility whatever it shows.

Toward the MS signal graduating to Emerging with confidence: an independent replication of Skovgaard 2025 with an active control, a pre-specified primary analysis that is significant unadjusted, and objective outcomes moving alongside self-report.

Toward accepting an autonomic mechanism: a controlled study with HRV or cortisol recorded at rest, before and after, with breathing controlled and movement artefact excluded, showing a TRE-specific change that a matched relaxation control does not produce.

Toward accepting the mammalian premise: any ethological or experimental study showing that post-threat trembling predicts recovery, or that preventing it produces persistent dysregulation, in any mammal.

Toward a lower assessment: systematic adverse-event collection with described events revealing a meaningful rate of dissociative or flooding episodes in general users, or case reports of harm in self-help practitioners with trauma histories.

What would NOT move us: more uncontrolled pre-post pilots, however many; provider-recruited surveys; testimonials in any volume; studies whose comparator is usual care or a waitlist; single-session state-anxiety drops; HRV recorded during shaking or with breathing instruction; inclusion of TRE in a polyvagal-branded textbook; or the argument that a practice's cross-cultural age is evidence of its mechanism.

Industry bias note

Structural incentives the evidence base may reflect

Who benefits from TRE's claims being true. The trademark and copyright holders, the network of certification trainers and the certified providers who sell sessions and courses. The certification pathway is a multi-module, multi-thousand-pound training; provider directories, books and video sit around it. The founder authored the first pilot, trained the trainers and rated fidelity for the largest trial, co-authored its poster and the largest survey, and a meaningful share of the positive uncontrolled literature is written by trainers. Two structural features of the literature follow from this: a null trial funded at roughly half a million dollars that reached a grant archive and a conference poster but never a journal, and a research page that presents a null trial and an unpublished eight-person study as supporting evidence. The practice also sits inside a wider somatic-therapy economy, Somatic Experiencing and polyvagal-informed training among them, that shares the discharge premise and the safety vocabulary and benefits from their repetition across modalities.

Who benefits from TRE's claims being false. Almost nobody with money at stake. TRE displaces no drug, device or billable clinical procedure, and it is too cheap to threaten competing modalities materially. The skeptical position is held by clinicians and science writers whose stake is reputational, and the field's leading skeptical review still called the practice probably safe. Evidence-based trauma therapies are not competing with TRE for the same patients in any meaningful commercial sense.

Who has done the honest work. The Danish MS Society research team, with no TRE affiliation and a patient-society mandate, produced the only independent pilot and the only positive peer-reviewed trial, and reported their own limitations plainly, including the disagreement between what participants felt and what the instrument measured. The Monash physiotherapy group ran a fair trial and published a null. The Phoenix VA team ran the only placebo-controlled trial, reported its null result in full in a public grant report, and did not spin it. The Pretoria scoping reviewer mapped the field and called it preliminary. A Beijing hospital has registered the only adequately powered independent trial in progress. That is where the credible signal, such as it is, comes from.

How the bias shapes what a reader sees. Testimonials are the primary marketing content of provider sites; "I got taller" and "I released something" are promotional artefacts of a community that reinforces reports of dramatic experience. The polyvagal vocabulary functions as shared marketing infrastructure across somatic modalities and lends the practice a scientific register it has not earned. The plain-language corrective, that this is fatigue tremor and a relaxation practice, has no marketing budget and travels only through skeptical media.

Sources (61)

Related entries

HRV Training: Do the Breathing, Don't Just Read the NumberModerateAnxiety: The Levers That Work Are Behavioural, Not PillsStrongAutonomic Nervous System BalanceStrongBehavioral Sleep Onset ProtocolStrong
High on the physiology, low on every outcome claimLast updated August 2026
Open in the Library: search, filter, every entry →

We set no cookies and run no ad trackers. We count visits with Cloudflare's cookieless, privacy-first analytics. The only thing stored on your device is which example you last viewed.

Realised is the trading name of Realised AI Ltd, a company registered in England and Wales. Company number 17279960.