Neck Training: It Helps Neck Pain, It Is Not Proven to Prevent Concussion

Summary

Training the neck reliably makes the neck stronger, and for chronic neck pain that is a genuinely well-supported, cheap, low-harm intervention with a landmark year-long randomised trial and a Cochrane review behind it — but the far louder claim, that a stronger neck prevents concussion, rests on one much-quoted observational cohort in soccer, basketball and lacrosse players whose association does not survive pooling (the best meta-analysis, eight studies and 7,625 participants, found the relationship small, nonsignificant and very-low-certainty), a laboratory rig measuring head velocity rather

Why Moderate

Moderate overall, because the entry's most-asked question and its most-cited claim sit on different rungs and the entry is built to keep them apart. The neck-pain and cervicogenic-headache material would justify a higher tier on its own; the concussion material would sit lower on its own; and an entry whose headline question is "does neck training protect my brain" cannot honestly wear the tier earned by a different outcome.

NOT the tier above (Strong Evidence) because the concussion cluster — the reason most people arrive here — rests on a single much-quoted observational cohort whose association does not survive meta-analysis, a laboratory surrogate with a contradicting field literature, two small systematic reviews pulling in opposite directions with one of them now partly superseded, and one cluster trial whose randomised concussion comparison was null. The best synthesis of the association grades its own evidence very-low-certainty. International consensus does not list isolated neck strengthening as established prevention. Additionally, even the well-evidenced neck-pain side carries Cochrane's own verdict that no high-quality evidence exists in the field, and the programming guidance is almost entirely coaching convention that the best dosage review explicitly declines to endorse.

NOT the tier below (Emerging) because the neck-pain claim is genuinely well supported and it would be dishonest to drag it down: a 180-participant, twelve-month, examiner-blinded randomised trial with a large effect, an independent Cochrane review of 27 trials and 2,485 participants finding moderate-quality evidence for cervico-scapulothoracic strengthening, a four-arm cervicogenic headache trial with benefits maintained at twelve months, and a 13-trial meta-analysis in chronic whiplash. That is a real body of convergent evidence for a real clinical problem, and it is the part of neck training most people actually need.

The internal split — read this, not the headline:
• Neck strength or endurance training reduces chronic non-specific neck pain: Strong Evidence for direction and consistency; moderate-quality trials underneath, with attention-time and expectancy confounds, an active control arm, and Cochrane's no-high-quality-evidence caveat.
• Strength training and endurance training do not differ significantly for neck pain: Strong Evidence — a direct within-trial randomised comparison in which the strength arm gained roughly four times the strength and got no more symptom relief, and one of the least-quoted results in the field.
• Cervico-scapulothoracic strengthening beats stretching alone or general fitness alone: Strong Evidence for direction, moderate-quality underneath.
• Low-load specific neck exercise helps cervicogenic headache, with benefit at twelve months: Moderate.
• Neck and shoulder strength training does NOT help tension-type headache: Moderate — a genuine negative finding on small samples with heavy attrition, stated as a negative.
• Neck-specific exercise helps chronic whiplash grades 2–3: Moderate, with a serious independence caveat (many trials from the same research groups) and a pending check from an unread pragmatic trial.
• Craniocervical flexion training improves deep-flexor recruitment and coordination: Moderate — strong for the surrogate, weak-to-absent for load capacity, and carrying no concussion claim at all.
• Neck strengthening programmes increase neck strength: Moderate — the boring, robust part; three small trials, roughly 70 male participants, direction consistent, plus a 110% flexion gain in the year-long neck-pain trial.
• Greater neck strength is associated with lower concussion odds: Emerging — one large cohort found it; pooled across eight studies and 7,625 participants the relationship is small, nonsignificant and very-low-certainty, with extreme heterogeneity, and a professional rugby cohort found no predictive value. Downgraded from Moderate on that synthesis. Not an intervention effect at any tier.
• Neck strength and anticipatory activation reduce head velocity in a laboratory rig: Moderate for the lab finding, with a contradicting field literature stated generically.
• Neck strengthening reduces concussion incidence in humans: Emerging, and closer to inconclusive than to positive — the review that looked hardest concluded there is a lack of evidence.
• Loaded neck bridging is high-risk: Experimental — biomechanical and expert reasoning only; no measurement study, no case series, no registry, no injury statistic.
• Optimal neck-training dosage (sets, reps, frequency, tempo): not tiered as evidence — coaching consensus, and the best dosage review says the optimum is unknown.

Practical takeaway

The framing to hold: train the neck if you have neck pain, a cervicogenic headache pattern, a neck that fatigues under load, or a sport that loads it — and train it because those are real, addressable problems. Do not train it because you have been told it protects your brain, because that has not been shown. Almost every specific number below is labelled with what it rests on, because most of them rest on convention.

First: which of the two trainings do you actually need?
• If you have neck pain, a stiff or fatiguing neck, or a headache that starts in the neck — the target is capacity and control. Both loaded strength work and lighter endurance work have trial support here, and the landmark trial found no discernible difference between them on pain and disability, so pick the one you will actually do.
• If you have a documented deep-flexor control problem, or you are early in rehabilitation — low-load craniocervical flexion is the specific modality, and it is a coordination intervention, not a strength intervention. It will not build load capacity. Do not expect it to.
• If you play a sport that loads the neck — build the capacity, and drill the anticipation, and understand that the second is likely to matter as much as the first. Neck exercise inside a broader neuromuscular warm-up programme is a defensible recommendation, but be clear about why: the programme has cluster-trial support, and no one has shown that the neck component carries any of it. Isolated neck training sold as concussion protection is not defensible at all.
• If your goal is a thicker-looking neck — that is a normal thing to want, and this entry is not going to lecture about it. A stronger neck is a legitimate training target, the work is the same work, and appearance follows from the physical change rather than being the thing that gets trained. The platform's standing line applies: health drives appearance, appearance is a byproduct, and it is a poor target because it produces bad decisions about load and progression. See looksmaxxing_health_first_evidence.

What the trials actually used (trial-derived — this is the shortest list on the page).
• Manual isometric resistance, four directions. Flexion, extension, left lateral flexion, right lateral flexion. Three sets of ten-second holds per direction, twice weekly. A partner or your own hand supplies the resistance; you hold position and do not let the head move. This produced measurable strength gains across all planes in five weeks in the one trial that used it.
• Cable or machine isometrics in the same four directions, in a longer block with a maintenance phase — non-significant strength changes in the trial that used it, so treat it as an alternative delivery of the same idea rather than a superior one.
• A rotational device using self-generated centripetal force — six sets of 50 revolutions in each direction, twice weekly, produced a composite strength improvement over control in seven weeks.
• High-intensity isometric neck strength training using an elastic band, or lighter dynamic neck endurance training performed lying supine and prone, as delivered in the chronic neck pain trial — twelve months, both effective on pain and disability, neither discernibly better than the other. Note which is which: the elastic band belonged to the strength arm, and the endurance arm lifted the head against gravity from supine and prone. The endurance arm gained far less strength and got the same symptom relief.
• Low-load craniocervical flexion with pressure biofeedback — the standard rehabilitation protocol. You lie supine, perform a gentle nodding motion (a chin nod, not a chin retraction and not a neck curl) that flexes the upper cervical segments while the head stays heavy on the surface, and hold. The point is to produce the movement without the sternocleidomastoid and scalenes taking over — which is what the pressure cuff is for, and why the load must be low. This is the intervention with strong evidence for changing recruitment and weak-to-absent evidence for changing strength.

What is coaching consensus, clearly labelled as such (no trial support located for any of it):
• Isometrics before loaded dynamic work. Standard practice, and a reasonable way to establish tolerance before adding range under load. Not tested.
• Slow, controlled tempo; no ballistic movement. Consensus, and consistent with the one published caution that names jerky, rapid, abrupt-rotation and rapid flexion-extension movements as risk-relevant — though that caution was written for people with or at risk of arterial dissection, not as a general training rule.
• Two to three sessions per week, low volume. Convergent with what the trials happened to do (two of the three neck-strength protocols used twice weekly), which makes it a defensible convention rather than an established optimum. The dosage review says plainly that the optimum is unknown.
• Avoid loaded end-range extension. Reasoning only, and the reasoning is set out in the risks section below. There is no dosage study behind it.
• Progress load gradually. The one positive dosage signal — progressive load moderately more effective than fixed load for short-term pain in chronic neck pain patients — mildly cuts against an ultra-conservative "isometrics only, low volume forever" prescription. Progress, but slowly, and in the range you can control.
• Train all four directions, not just extension. Anatomically sound and universally recommended. The electromyography data support the specific version of this point: ordinary gym work (shrugs, upper-back work) already loads the extensor and trapezius complex substantially while doing nothing for the cervical flexors, so the flexor side is the genuine gap.
• Neck harnesses, plate-on-forehead work, and four-way neck machines. These dominate the instructional content and appear in none of the trials located. They are not thereby bad — absence of evidence is not evidence of harm — but nobody has shown they do anything the tested modalities do not, and the people demonstrating them mostly sell them.

Where neck training slots into a week. It is low-volume accessory work, not a session. Pair it with upper-body or general strength days rather than giving it its own slot, keep it away from anything requiring maximal effort immediately afterwards, and treat the first two weeks as tolerance-finding rather than progression. If you are also working on desk setup, screen height and time in a fixed posture — which is the other half of most people's neck complaint — that belongs to physical_counter_modern_postures, and if the complaint is worst on waking, sleeping position and pillow height belong to sleep_posture_and_sleep_quality.

Honest ceiling. Expect a stronger, more tolerant neck within roughly five to twelve weeks, and expect the chronic-neck-pain benefit to be real but partial and to take months rather than weeks — the trial that produced the largest pain reduction ran for a year, and its control group, which was doing aerobic exercise and stretching, improved meaningfully on its own. Do not expect a concussion outcome, because nobody has shown one.

Evidence detail

Why This Entry Exists

Neck training is a field where the volume of content is wildly out of proportion to the volume of evidence. The best systematic review of neck strength-and-conditioning protocols screened 2,462 articles and found three eligible studies totalling roughly seventy participants, all male. The best review of exercise dosage for chronic neck pain screened 3,990 citations, included twenty-six trials, and concluded in plain words that the optimal dosage is not known. Against that, there are neck harnesses, four-way neck machines, rotational devices, "build a bull neck" programmes, and a very large amount of instructional content — the overwhelming majority of it written by people who sell the equipment being demonstrated.

The entry has to hold two things at once that most content collapses into one. The first is that neck training genuinely works for something: chronic non-specific neck pain, where a twelve-month examiner-blinded randomised trial and an independent Cochrane review both support neck strength and endurance training, and where the intervention is cheap, self-administered and close to harmless. The second is that the claim doing nearly all the commercial and cultural work — that a stronger neck protects your brain — is not established. The mechanism is plausible, one large cohort did find an association, and when that association was tested across every available study it shrank to a small, statistically nonsignificant, very-low-certainty relationship. The step from plausible mechanism to demonstrated prevention has not been made, and the intermediate step — a dependable association — is weaker than this entry originally believed. Refusing to make either step is this entry's central discipline, in both directions: we will not sell prevention that has not been shown, we will not inflate an association that does not pool, and we will not throw away the neck-pain evidence because the concussion claim is inflated.

There is a third confusion worth naming up front. Two quite different practices share the name. One is loaded strength and size work for the neck — isometric holds, machine or cable resistance, harnesses, the "big neck" goal. The other is craniocervical flexion training, the low-load chin-nod work used in neck-pain rehabilitation, which deliberately targets the deep cervical flexors while keeping the superficial ones quiet. They aim at different deficits, have been studied in different populations, and the evidence for one is not evidence for the other. Notably, the concussion-side review that reported higher neck strength associated with lower risk explicitly did not find that association for deep neck flexor endurance — so the rehabilitation modality carries no concussion claim at all.

What bad advice this protects against, in all directions:
• "Every pound of neck strength cuts your concussion risk by five percent, so train your neck and protect your brain." → The number is real and correctly quoted from the study, but it is an odds ratio in an observational cohort, from a pilot whose stated purpose was to see whether a cheap screening measurement could flag higher-risk athletes — not to test whether training the neck lowers risk. The sports were boys' and girls' soccer, basketball and lacrosse. It has been re-labelled onto American football, rugby and combat sport by people citing it, which silently changes the population. A per-pound odds ratio does not extrapolate linearly: twenty pounds of added neck strength does not abolish risk. And most damagingly for the claim, when four prospective studies including this one were pooled, the relationship came out small and statistically nonsignificant on very-low-certainty evidence — so this is not even a settled association, let alone an effect.
• "A cluster trial showed a neck programme cut concussions by fifty-nine percent." → That figure comes from a schoolboy rugby trial whose intention-to-treat concussion result was a risk ratio of 0.71 with a confidence interval running from 0.48 to 1.05 — crossing one, so not significant. The fifty-nine percent appears only in the compliance-restricted subgroup, where teams that complied were not randomly assigned to comply. And the intervention was a whole-body movement-control warm-up including balance, resistance, plyometric and landing/cutting work. Even taking the subgroup at face value, the effect belongs to the programme, not to the neck.
• "Neck training is bro-science with nothing behind it." → The opposite over-correction, and the more costly one. Chronic neck pain is common and disabling, and neck strength and endurance training is one of the few musculoskeletal interventions that is cheap, self-administered, near-zero-harm and consistently supported across many trials, a Cochrane review, and the cervicogenic headache literature. Rejecting the whole field because the concussion marketing is inflated throws away the part that works.
• "Neck exercise fixes headaches." → Only one kind. Cervicogenic headache is supported by a four-arm randomised trial with benefits still present at twelve months, and by moderate-quality Cochrane evidence. Tension-type headache is not: a randomised trial of ten weeks of elastic-band neck and shoulder strength training found no between-group effect, and the control group improved more than the strength group. A second trial in working-age women with chronic headache was null on its primary outcome. Citing the cervicogenic evidence for tension-type headache or migraine is a category error.
• "You need heavy loaded neck work — light stuff is a waste of time." → The landmark neck-pain trial randomised women to intensive isometric strength training or lighter dynamic endurance training and found no statistically discernible difference between the two training arms on pain and disability — even though the strength arm got dramatically stronger and the endurance arm barely did. Both beat the control substantially. This is one of the most under-quoted results in the field, and it is inconvenient for anyone selling load.
• "Chin tucks are useless, just load the neck." → Deep cervical flexor training has strong evidence that it does the specific thing it is designed to do — improve neuromuscular coordination and shift recruitment away from the superficial flexors — in populations where that recruitment pattern is documented to be disordered. It also has little effect on strength or endurance at higher loads, so it is not a substitute for loading. Both modalities have a defensible lane; the error is claiming either covers the other's.
• "Shrugs cover your neck, you don't need separate work." → Partly wrong in the other direction. An electromyography study of elastic-band exercises found shrugs produced high activation not only in the upper trapezius but also in the upper neck extensors. So shrugs do more neck work than the "different musculature" claim admits. What they do not do is train the cervical flexors at all, train craniocervical flexion, or move the cervical spine through its own range under load — so they are not a substitute, particularly on the flexion side.
• "Neck bridges are how wrestlers built indestructible necks, so they're fine" and "neck bridges will wreck your spine." → Neither is evidenced. There is no trial, no case series, no biomechanical measurement study and no injury registry that quantifies cervical load during bridging or attributes injuries to it. A 2024 narrative review of cervical spine injuries in combat sports does not mention bridging at all. The caution in this entry is mechanistic reasoning and expert opinion, and is labelled as such — it is not backed by an injury statistic, and this entry will not invent one.
• "Lifting causes artery dissection, so heavy neck work is dangerous." → Do not import this. In the largest cervical artery dissection dataset examined, sport was a self-reported trigger in 6.3% of dissection patients versus 1.8% of healthy controls, and the review's own conclusion is that sport as a cause "seems uncommon". A risk-factor meta-analysis grades minor trauma at low certainty and most other candidates at very low. Resistance training does not appear as an established risk factor. The screening apparatus built around cervical manipulation belongs to cervical manipulation.

This entry owns the honest verdict on neck training — that it works for chronic neck pain and cervicogenic headache, that it reliably increases neck strength, that the concussion-prevention step is unproven and even the underlying association does not pool, and that the two modalities sharing the name are distinct — plus the technique and programming guidance with its evidence status labelled. It defers general chronic joint pain management to joint_pain_conservative_management, the headache differential and when a headache is not a neck problem to headache_migraine_cross_pillar_diagnostic, desk-posture and screen-position work to physical_counter_modern_postures, sleeping position and pillow questions to sleep_posture_and_sleep_quality, and the appearance-motive conversation to looksmaxxing_health_first_evidence. Where a question resolves to one of those, this entry hands off rather than re-arguing it.

Evidence

Read the tiers, not the thesis. The neck-pain and cervicogenic-headache claims are the firm end of this entry; the concussion claims are the thin end, and the gap between them is the single most important thing on this page.

Chronic neck pain: the well-evidenced side (Strong Evidence for direction, with an honest ceiling).

1. Neck strength training and neck endurance training both substantially reduce chronic neck pain, and neither beat the other (Strong Evidence). An examiner-blinded randomised controlled trial run between February 2000 and March 2002 recruited 180 working women aged 25 to 53 with constant or frequent neck pain of more than six months' duration through Finnish occupational health care, and randomised them in three arms of 60 to intensive isometric neck strength training performed with an elastic band, lighter dynamic neck endurance training performed lying supine and prone, or a control condition. The control arm was active, not idle — it performed aerobic exercise and stretching, which matters when reading its improvement. Training ran for twelve months. At twelve months both neck pain and disability had decreased in both training groups compared with the control group (P<.001), with reductions on the visual analogue scale of roughly 40 points in the strength group, 35 in the endurance group and 16 in the control group — the widely quoted "73%, 59% and 21%" falls are the corresponding percentage changes from baseline reported in the paper's results, not figures that appear in the abstract. The arms separated sharply on strength and not on symptoms: maximal isometric strength improved 110% in flexion, 76% in rotation and 69% in extension in the strength group, versus 28%, 29% and 16% in the endurance group and 10%, 10% and 7% in the control — and yet there was no statistically discernible difference between the two training groups on pain and disability. (Ylinen J, et al. "Active neck muscle training in the treatment of chronic neck pain in women: a randomized controlled trial." JAMA 2003;289(19):2509–16. PMID 12759322. Strong Evidence — a large, long, examiner-blinded randomised trial. Cui bono note: the 73% figure is quoted constantly in physiotherapy and equipment marketing, almost always without the control group's improvement and without noting that the control arm was exercising too; the strength-versus-endurance null is quoted almost never, because a fourfold difference in strength gain producing no difference in pain relief is ruinous for anyone selling load.)

2. Cervico-scapulothoracic strengthening produces moderate-to-large pain benefit for mechanical neck disorders, while stretching alone and general fitness alone do not (Strong Evidence for direction, moderate-quality underneath). The Cochrane review of exercises for mechanical neck disorders included 27 randomised trials with 2,485 analysed participants, pooling selectively rather than meta-analysing all 27 together. It found moderate-quality evidence that cervico-scapulothoracic strengthening gives moderate-to-large pain benefit immediately after treatment and at follow-up; that combined strengthening plus stretching benefits pain and function; and that strengthening and stabilisation work improves pain and function at intermediate-term follow-up. Stretching alone, general fitness alone and breathing exercises alone showed minimal or no effect on low-quality evidence. Cochrane's own caveat must travel with the finding: no high-quality evidence was found, so uncertainty about the effectiveness of exercise for neck pain remains. (Gross A, Kay TM, Paquin JP, Blanchette S, Lalonde P, Christie T, et al. "Exercises for mechanical neck disorders." Cochrane Database Syst Rev 2015;1:CD004250. PMID 25629215. Strong Evidence for consistency and independence; the underlying trial quality is moderate at best. Trial and participant counts carried from the review's abstract, not recounted from the included-studies table in this pass. Cui bono note: independent Cochrane review with no product to sell — and it is notably unkind to the cheap-to-market options, finding stretching and general fitness near-useless in isolation.)

3. Neck-specific exercise helps chronic whiplash-associated disorder, with a real independence caveat (Moderate). A systematic review and meta-analysis of 13 randomised trials and 2,427 participants aged 18 to 63 with chronic whiplash grades 2 to 3 concluded that neck-specific exercise, with or without an added behavioural approach, might help reduce pain and disability — and that adding the behavioural layer bought nothing measurable over the exercise alone (pain at 6 to 12 months, p=0.71; Neck Disability Index at 6 to 12 months, p=0.97). The reviewers themselves say the findings should be viewed cautiously because there are few studies, many originate from the same research groups, and protocols are heterogeneous. (Correia L, Carvalho P, Amaral L, et al. "The Effect of Neck-Specific Exercise with or Without a Behavioral Approach in Chronic Whiplash-Associated Disorders: A Systematic Review and Meta-Analysis." Muscles 2025;4(4):49. Moderate — thirteen trials concentrated in a small number of labs is closer to one research programme replicating itself than to independent replication. Cui bono note: whiplash rehabilitation is a high-volume market funded substantially through insurance and litigation, and providers benefit directly from evidence favouring long multi-session programmes over a single advice consultation. A pragmatic randomised trial published in the Lancet in 2014 — the PROMISE trial, Michaleff ZA et al. — tested exactly that comparison in chronic whiplash; we have confirmed its existence, journal and year but could not read its result, so we are naming it as the strongest available check on "more exercise is better" and explicitly NOT stating its conclusion.)

Headache: split it or get it wrong (Moderate, and it cuts both ways).

4. Cervicogenic headache responds to low-load specific neck exercise, and the benefit holds at twelve months (Moderate). A multicentre randomised trial with unblinded treatment and blinded outcome assessment allocated 200 participants with cervicogenic headache to manipulative therapy alone, low-load specific exercise alone, both, or a control, with six weeks of treatment and follow-up at three, six and twelve months. At twelve months both manipulative therapy and specific exercise had significantly reduced headache frequency and intensity and reduced neck pain, with effects maintained. Combined therapy was not significantly superior to either alone, though roughly 10% more patients gained relief with the combination. Cochrane independently supports this, finding moderate-quality evidence that static-dynamic cervico-scapulothoracic strengthening and endurance exercise including pressure biofeedback improves pain, function and global perceived effect at long-term follow-up in chronic cervicogenic headache. (Jull G, Trott P, Potter H, Zito G, Niere K, Shirley D, Emberson J, Marschner I, Richardson C. "A randomized controlled trial of exercise and manipulative therapy for cervicogenic headache." Spine 2002;27(17):1835–43. PMID 12221344; corroborated by Gross A, et al., Cochrane 2015;1:CD004250. Moderate — treatment could not be blinded, and cervicogenic headache is itself a contested diagnosis with imperfect criteria, so the population definition is soft. Cui bono note: manual therapy and physiotherapy providers benefit from a broad "neck work fixes headaches" claim that quietly absorbs tension-type headache and migraine into a finding that holds only for the cervicogenic subgroup.)

5. Tension-type headache does NOT respond to neck and shoulder strength training on primary outcomes — a genuine negative finding (Moderate). A randomised controlled study randomised 60 tension-type headache patients to ten weeks of elastic-band neck and shoulder strength training or to ergonomic and posture correction, with follow-up at 19 to 22 weeks. There was no between-group effect. Within groups, the strength arm improved less (frequency −11%, P=0.041; duration −10%, P=0.036) than the control arm (frequency −24%, P=0.0033; duration −27%, P=0.041), and the authors concluded the within-group effects did not reach clinical significance. Attrition was substantial and should be stated as a number: of the 60 randomised, 23 completed the strength intervention and 21 the control. Separately, a trial of 116 working-age women with chronic headache randomised to a six-month progressive neck-shoulder home programme versus placebo-dosed TENS plus stretching found no between-group difference on the primary outcome of pain intensity; headache frequency, a secondary outcome, fell more in the exercise group (4.5 to 2.4 days per week versus 4.4 to 3.0). (Madsen BK, Søgaard K, Andersen LL, Tornøe B, Jensen RH. "Efficacy of strength training on tension-type headache: A randomised controlled study." Cephalalgia 2018;38(6):1071–1080. PMID 28750588; Rinne M, Garam S, Kukkonen-Harjula K, et al. "Neck-Shoulder Region Training for Chronic Headache in Women: A Randomized Controlled Trial." Clin Rehabil 2023;37(10):1322–1331. Moderate, with small samples and heavy attrition in the first trial. Cui bono note: a null primary with a favourable secondary is exactly the pattern that gets reported as a win. We are declining to report it that way. The dismissive side has its own bias — headache medicine has historically been sceptical of cervical explanations, and that scepticism should not be allowed to erase the cervicogenic finding above.)

Deep cervical flexor training: a different intervention with a narrower claim (Moderate).

6. Craniocervical flexion training improves the recruitment pattern it targets, and does not build load capacity (Moderate). People with neck pain show a documented pattern of reduced deep cervical flexor activity, compensatory increased superficial flexor activity, and reduced endurance — a control deficit rather than a strength deficit. A systematic review found strong evidence that deep cervical flexor training improves neuromuscular coordination, with no or only small effects on strength and endurance at higher contraction loads, and improvement in head and cervical posture; evidence for other outcomes was limited or contradictory. A separate systematic review with meta-analysis found low-load craniocervical flexion effective specifically for deep-flexor impairments in chronic neck pain. ("Effects of deep cervical flexor training on impaired physiological functions associated with chronic neck pain: a systematic review." BMC Musculoskeletal Disorders 2018;19:373. PMC6263552 — cited by title and journal because we did not verify an author line. Supporting: a systematic review and meta-analysis of cranio-cervical flexion versus other treatments for non-specific chronic neck pain, Musculoskeletal Science and Practice 2019 — identified but not read in full, treated as indicative. Moderate — the outcomes are heavily surrogate-loaded: electromyography recruitment patterns, muscle thickness on ultrasound, pressure-biofeedback scores and posture measures rather than pain and function. Cui bono note: the posture-correction and "text neck" industry leans on this literature to sell a posture-causes-pain story it supports only weakly; the loaded-training camp dismisses it while ignoring that the documented deficit in neck-pain populations is a control problem.)

Concussion: the contested centre (Emerging for the association, Emerging for prevention, and the gap between them is the point).

7. One large observational cohort of high-school athletes found greater neck strength associated with lower concussion odds — this is the number the whole field quotes (Moderate for the cohort itself; see claim 8 before using it). A prospective observational cohort nested in a multi-phase feasibility and pilot study followed 6,704 high-school athletes across 51 schools in 25 states during 2010–2011, measuring neck strength with a purpose-built hand-held tension scale validated against a dynamometer. Smaller mean neck circumference, smaller neck-to-head circumference ratio and weaker overall neck strength were each significantly associated with concussion, and each additional pound of neck strength was associated with an odds ratio of 0.95 (95% CI 0.92–0.98) — roughly 5% lower odds per pound. The sports were boys' and girls' soccer, basketball and lacrosse. Not American football, not rugby, not ice hockey, not combat sport. (Collins CL, Fletcher EN, Fields SK, Kluchurosky L, Rohrkemper MK, Comstock RD, Cantu RC. "Neck strength: a protective factor reducing risk for concussion in high school sports." J Prim Prev 2014;35(5):309–19. PMID 24930131. Moderate as a single well-conducted cohort, but observational: neck strength travels with body mass, sex, sport, training age, pubertal maturation and general athleticism, any of which could carry the effect; stronger-necked athletes may also be better positioned and more experienced; concussion was ascertained by school athletic trainers, so diagnostic sensitivity varies by site; and the study's stated aim was to see whether a cheap screening measurement could flag higher-risk athletes, not to test training. Cui bono note: this single odds ratio is the most-cited number in the entire neck-training sales funnel, usually stripped of the word "observational" and usually re-labelled onto populations the study never touched. The authors also had a legitimate interest in a cheap validated screening device, which is part of why the paper's own title says "protective factor" — that phrase is the authors', not a demonstrated effect.)

8. When every available prospective study is pooled, the neck-strength/concussion association is small, nonsignificant and very-low-certainty — and a professional cohort found it has no predictive value at all (Emerging, and this supersedes the single-cohort reading). A systematic review with meta-analysis identified eight studies totalling 7,625 participants and meta-analysed four prospective longitudinal studies. Its finding, quoted directly: "There was very low-certainty evidence suggesting a small, nonsignificant relationship between greater neck strength and a lower risk of sustaining a SRC." Pooled effects were r = 0.08 to 0.14 with substantial heterogeneity (I²>90%), which the reviewers attribute to studies with vastly different participant ages, playing levels and sports. Independently, a prospective cohort of 136 male professional rugby players, in which 40 players sustained 51 concussions over a season, measured peak isometric flexion and extension strength plus endurance and found neck muscle strength similar between players who were concussed and players who were not (peak isometric extension odds ratio ≈ 1.01); prior concussion history, not neck strength, was the dominant predictor, carrying over twice the odds. The authors' conclusion is that neck-strength assessment offers no clinical value for predicting concussion. Read this against claim 7 and not separately. The correct statement is not "a stronger neck is associated with fewer concussions" but "one large cohort in adolescent field sports found that association, the field has tested it since, and it does not reliably reproduce." (Garrett JM, Mastrorocco M, Peek K, van den Hoek DJ, McGuckian TB. "The Relationship Between Neck Strength and Sports-Related Concussion in Team Sports: A Systematic Review With Meta-analysis." J Orthop Sports Phys Ther 2023;53(10):585–593. PMID 37428807. Liston, Leckey, Whale, van Dyk. "Neck Strength Assessment Offers No Clinical Value in Predicting Concussion in Male Professional Rugby Players: A Prospective Cohort Study." J Orthop Sports Phys Ther 2023, PMID 37017931, doi 10.2519/jospt.2023.11723 — author initials not independently verified. Emerging — very-low-certainty evidence by the reviewers' own GRADE assessment, and extreme heterogeneity means the pooled estimate is fragile in both directions; a nonsignificant pooled result is not proof of no association, it is failure to demonstrate one. Cui bono note: this is the result that should have replaced the 2014 odds ratio in every piece of marketing and has replaced it in almost none. Note also that the honest sceptical reading has limits — heterogeneity that severe means the true relationship could differ by sport, sex and level, which is exactly what a later population-split review has been reported to suggest.)

9. In a laboratory rig, both neck strength and anticipatory muscle activation independently reduce head velocity on impulsive loading — but the field data do not reproduce it cleanly (Moderate for the lab finding). A descriptive laboratory study of 46 male and female contact-sport athletes aged 8 to 30 measured maximum isometric neck strength in each anatomical plane, then peak head linear and angular velocity under controlled impulsive loading, with and without anticipatory cervical muscle activation. Both greater isometric strength and anticipatory activation were independently associated with lower head linear and angular velocity across all planes (P<.001), with correlations from r = 0.417 to 0.657. The outcome was head velocity in a rig, not concussion, and not even on-field impact. A counter-literature reports weak or absent associations between cervical strength and head accelerations in real impacts, including in youth ice hockey and instrumented head-impact work, and biomechanical modelling argues the head-neck coupling is weak enough that the added effective mass from a stronger neck is far smaller than the theory assumes. We state that kinematics counter-finding generically, without a citation, because we did not read any single one of those papers in full — the named counter-evidence in this entry (claim 8) is about concussion incidence, not head kinematics, and the two must not be conflated. The consistent thread across honest readings is that anticipation and bracing do more measurable work than static maximal strength — and unanticipated impacts are exactly the ones that concuss people. (Eckner JT, Oh YK, Joshi MS, Richardson JK, Ashton-Miller JA. "Effect of neck muscle strength and anticipatory cervical muscle activation on the kinematic response of the head to impulsive loads." Am J Sports Med 2014;42(3):566–76. PMID 24488820. Moderate — cross-sectional and correlational; strength was measured, not trained, so it cannot show that training changes kinematics; the age band mixes maturation with strength; and head velocity is a surrogate whose link to concussion is itself unproven. Model-to-human overreach flag: a large slice of the "neck protects the brain" literature runs on cadavers, anthropomorphic test dummies and finite-element head models and then gets reported as if demonstrated in living athletes. That is the dominant overreach class in this cluster. Cui bono note: mechanism is the equipment industry's favourite currency because it is visual and testable in a gym while the outcome that matters is rare and invisible. Note also that the anticipation half of this finding is commercially inconvenient — you cannot sell a bracing reflex in a box — which is part of why the strength half gets amplified and the anticipation half gets dropped.)

10. Systematic review: there is a lack of evidence that neck strengthening reduces impact injury risk in adult sport (Emerging — this is the honesty check). A systematic review screened 2,462 articles and found only three eligible studies in adult collision-sport athletes, all male. Its conclusion is quoted verbatim: "This review has shown that there is currently a lack of evidence to support the use of neck strengthening interventions in reducing impact injury risk in adult populations who participate in sport." The review says outright that it was written because strength and conditioning coaches routinely cite neck strengthening as concussion mitigation — that is, because coaching culture had outrun the data. Two of the three studies found significant isometric strength gains; none reported any effect on cervical spine injuries or concussions. (Daly E, Pearce AJ, Ryan L. "A Systematic Review of Strength and Conditioning Protocols for Improving Neck Strength and Reducing Concussion Incidence and Impact Injury Risk in Collision Sports; Is There Evidence?" J Funct Morphol Kinesiol 2021;6(1):8. PMC7838928. Participant-count note: the review's abstract states 68 participants; the three included samples as described (25, 27 and 18) sum to 70. We report the discrepancy rather than silently picking one — the number is "roughly seventy, all male" either way. Emerging — three studies, short follow-up, and the review's own stated limitation is lack of statistical power. Cui bono note: this negative finding is the single most commercially inconvenient result in the field, which is a reason to state it plainly rather than bury it in a caveat.)

11. A more permissive review supports neck exercise inside multi-component programmes — and the reducing agent is the programme (Emerging, with one component now superseded). A second systematic review of six studies concluded that higher neck strength, but explicitly not deep neck flexor endurance, was linked to reduced concussion risk in observational work, and that injury-reduction programmes that included neck exercises can reduce head and neck injury incidence including concussion, recommending neck exercises be incorporated into such programmes. Two things must travel with this. First, read the programme claim precisely: the agent it identifies is a multi-component injury-reduction programme, and the effect cannot be decomposed and handed to the neck. Second, its observational-association component has since been superseded — the larger meta-analysis in claim 8, with eight studies and 7,625 participants, found that same association small and nonsignificant on very-low-certainty evidence. Only the multi-component-programme half of this review survives; do not cite its strength-association half as current. (Elliott J, Heron N, Versteegh T, Gilchrist IA, Webb M, Archbold P, Hart ND, Peek K. "Injury Reduction Programs for Reducing the Incidence of Sport-Related Head and Neck Injuries Including Concussion: A Systematic Review." Sports Med 2021;51(11):2373–2388. PMID 34143411. Emerging — six studies, non-identifiable component effects, and one core finding overtaken by later synthesis. The internal check worth keeping: deep neck flexor endurance was NOT associated with reduced risk, which cuts against "any neck work protects".)

12. The load-bearing trial's randomised concussion result did not reach significance (Moderate for the trial, and it does not say what it is quoted as saying). A cluster randomised controlled trial in schoolboy rugby covered 40 independent schools, 118 teams and 3,188 players aged 14 to 18, comparing a pre-activity movement-control programme against a reference warm-up, with 441 time-loss match injuries over 15,938 match exposure-hours. The intention-to-treat concussion result was a risk ratio of 0.71 with a 95% confidence interval of 0.48 to 1.05 — crossing one. Only in the adherence-restricted comparison, among teams averaging three or more sessions per week, did the reduction become clear: risk ratio 0.41 (95% CI 0.17–0.99), which is the roughly 59% reduction quoted almost everywhere. The intervention was whole-body neuromuscular training — balance work, whole-body resistance training, plyometrics, and controlled rehearsal of landing and cutting. One disclosure so a reader checking the source is not confused: the paper's own abstract describes the intention-to-treat concussion result as showing "clear reductions" despite the interval crossing one. We report it as not reaching significance because that is what the interval says; the authors framed it more favourably, and you should know that before you read them. (Hislop MD, Stokes KA, Williams S, et al. "Reducing musculoskeletal injury and concussion risk in schoolboy rugby players with a pre-activity movement control exercise programme: a cluster randomised controlled trial." Br J Sports Med 2017;51(15):1140–1146. PMC5530334. Moderate — a genuine cluster trial with a dose-response pattern in the right direction, but the headline number is a compliance-restricted subgroup in which compliance was self-selected, not randomised; teams that comply differ from teams that do not. Cui bono note: "59% fewer concussions" is the most-laundered number in neck-training marketing, repeatedly attached to neck training specifically and sometimes printed beside a harness. Governing bodies also benefit from promoting a cheap warm-up as their concussion answer. The dismissive read is wrong too — warm-up neuromuscular programmes are cheap, broadly beneficial for musculoskeletal injury, and essentially harmless.)

13. International consensus does not list isolated neck strengthening as an established preventive measure (Emerging, and read with a stated verification limit). The 6th International Conference on Concussion in Sport consensus statement's prevention list covers mouthguards in ice hockey, policy disallowing bodychecking in ice hockey, and neuromuscular training warm-up programmes in adolescent rugby. Isolated neck strengthening is not on it. Where the neck appears with confidence in that consensus is treatment rather than prevention: cervicovestibular rehabilitation for people with dizziness, neck pain or headache persisting beyond roughly ten days after concussion. (Patricios JS, Schneider KJ, Dvorak J, et al. "Consensus statement on concussion in sport: the 6th International Conference on Concussion in Sport — Amsterdam, October 2022." Br J Sports Med 2023;57(11):695–711. PMID 37316210. Verification limit, stated openly: we verified the statement's identity and citation directly, but could not access the full text; the prevention content above comes from the abstract plus reputable secondary summaries. Do not quote its prevention section verbatim without reading it. A 2025 systematic review reportedly splits by population — cervical strengthening embedded in standard training may reduce concussion in male rugby players, with insufficient evidence in male or female football players or female rugby players — confirmed only at search-snippet level and treated as indicative, not cited as established. A 2025 systematic review and Delphi consensus reportedly recommending systematic inclusion of neck strengthening was seen referenced but not retrieved, and is deliberately not cited here.)

What the trials actually did, which is not what the content shows (Emerging).

14. The three trial protocols used manual isometric resistance, cable and scrum-machine isometrics, and a rotational device — no harness, no plate-on-forehead, no four-way machine (Emerging). Within the three-study systematic review: Geary and colleagues (2014) trained 25 male professional and semi-professional rugby union players with isometric manual resistance applied by a strength coach, three sets of ten-second holds in four directions (flexion, extension, left and right lateral flexion), twice weekly for five weeks, reporting increased isometric strength in all planes (flexion 334.45 ± 39.31 to 396.05 ± 75.55 N; extension 606.19 ± 97.34 to 733.88 ± 127.16 N). Naish and colleagues (2013) trained 27 male professional rugby union players with isometric exercise on a cable machine and a scrum machine over 26 weeks split into a 13-week strengthening phase and a 13-week maintenance phase, reporting non-significant increases in all directions. Versteegh and colleagues (2019) trained 18 male college American football players (8 intervention, 10 control) on a rotational device driven by self-generated centripetal force, six sets of 50 revolutions in each direction, twice weekly for seven weeks, reporting a composite strength improvement favouring the intervention (mean change 32 N, 95% CI 13–50, versus 12 N in control). (Protocols as described in Daly E, Pearce AJ, Ryan L, J Funct Morphol Kinesiol 2021;6(1):8. Emerging — three studies, roughly 70 male collision-sport athletes, all-male samples. Absence of harness studies is absence of evidence, not evidence that harnesses are worse. Cui bono note: the equipment market publishes the overwhelming majority of technique content and has direct commercial interest in modality-specific claims that no trial supports.)

15. Optimal dosage for neck exercise is not known, and the source says so (Moderate). A systematic review with narrative synthesis screened 3,990 citations and included 26 trials on exercise dosage for chronic non-specific neck pain. Its conclusion: the optimal dosage to improve pain and disability in a chronic neck pain population is not known, and optimal motor-control plus segmental exercise variables and dosage are unknown and require clarification. Evidence was low to moderate for motor control plus segmental exercise at short-term follow-up, with no high-quality trials investigating long-term outcomes, and pervasive high risk of bias and small samples. Two limited signals emerged: higher frequency may improve motor-control effectiveness (limited evidence), and progressive load was moderately more effective than fixed load for short-term pain. (Price J, Rushton A, Tyros I, Tyros V, Heneghan NR. "Effectiveness and optimal dosage of exercise training for chronic non-specific neck pain: A systematic review with a narrative synthesis." PLOS ONE 2020;15(6):e0234511. Moderate. Population note: this is a chronic neck pain population, not healthy people training for neck strength — applying its dosage conclusions to healthy trainees is an extrapolation, and we are flagging it as one. Cui bono note: coaching content needs specific numbers to be useful and sellable, and "nobody knows the optimal dose" makes a poor article. That pressure is precisely how consensus hardens into pseudo-evidence — and a product card that needs a set-and-rep prescription is under the same pressure.)

16. Shrugs load the upper neck extensors substantially, but do nothing for the flexors (Moderate). An electromyography study of elastic band exercises in 11 healthy males, mean age 25.9, across six exercises at 12RM and 20RM with bilateral recording normalised to maximal voluntary activation, found that shrugs produced high activation in both the upper trapezius (100.3 ± 29.8 %MVE) and the upper neck extensors (61.9 ± 16.8 %MVE). The authors concluded shrugs and reverse flyes produce high-intensity activation of both. ("Electromyographic Evaluation of Specific Elastic Band Exercises Targeting Neck and Shoulder Muscle Activation." Applied Sciences 2020;10(3):756, doi 10.3390/app10030756 — cited by title and DOI only, because we did not confirm the author list. Moderate. Eleven healthy young males, elastic band resistance, single session — this is an activation study, not a training study, and electromyography amplitude does not predict hypertrophy or strength adaptation. "Upper neck extensor" is a surface electrode site over overlapping muscles, not a clean single-muscle measurement. A separately circulating figure of "67.6 ± 29.8 %MVE for cervical extension" carries a standard deviation identical to the shrug upper-trapezius value, which suggests two table rows were welded together somewhere in secondary reporting — we have discarded it and it should not be used. Cui bono note: "shrugs aren't neck training" is load-bearing for equipment vendors, since it establishes a gap only their product fills, and the data partly undercut it; general coaches who don't want to add neck work benefit from the opposite overclaim, which the flexor gap clearly refutes.)

Mechanism

This entry owns enough mechanism to make the neck-pain claim, the concussion non-claim, and the risk guidance intelligible — not the full cervical anatomy.

Why the neck is a control problem before it is a strength problem, in people who hurt. The cervical spine carries a heavy, mobile head on a small, highly mobile column, and it is stabilised by two layers doing different jobs. The deep layer — longus capitis and longus colli in front, the deep segmental extensors behind — provides low-force, continuous segmental support and positional control. The superficial layer — sternocleidomastoid, the anterior scalenes, upper trapezius, splenius, semispinalis — produces gross force and gross movement. In people with neck pain, the documented pattern is reduced deep flexor activity, compensatory increased superficial flexor activity, and reduced endurance. That is why the rehabilitation intervention is low-load and precision-targeted rather than heavy: the deficit being addressed is which muscles are recruited and when, not how much force the neck can produce. The causal direction is not settled — the altered activation may be a consequence of pain rather than its cause — and the analogous "core stability" model in the lower back has been substantially deflated over the past decade without the cervical version facing the same scrutiny. Hold it as a real, reproducible, trainable observation, not as a proven cause of pain.

Why loading the neck is a separate stimulus with a separate rationale. The neck responds to progressive resistance the way other skeletal muscle does, and the load it can tolerate is a real, measurable, trainable capacity. The neck-pain trials show that building that capacity reduces pain — but they also show that a lighter endurance stimulus did just as well on symptoms while producing a fraction of the strength gain, which suggests the effective ingredient is repeated, tolerable, direction-specific loading of the region rather than maximal force production. That distinction matters because it is exactly where "you must load heavy" claims lose their footing. See resistance_training_and_body_composition for the general adaptation picture and specificity_of_adaptation for why a strength gain does not automatically become a real-world outcome.

The concussion mechanism, stated honestly and then stopped. The theory is head-neck coupling. A concussion is driven by rapid linear and rotational acceleration of the brain inside the skull. If the head and neck act as a coupled system, then a stiffer, stronger, pre-activated neck adds effective mass and resistance to that acceleration, so the same impulse produces less head motion. The laboratory data are partly consistent with this: in a rig, both isometric strength and anticipatory activation independently reduced head linear and angular velocity. That is where the mechanism stops being able to carry weight. Four things break the chain. First, effective mass depends on the neck actually being coupled at the moment of impact, and biomechanical modelling argues the coupling is much weaker than the theory assumes. Second, the studies that measure real impacts on real fields frequently fail to find the strength-kinematics association that the rig produces. Third, the outcome the theory is meant to predict does not track it either: pooled across prospective studies, stronger necks are not reliably associated with fewer concussions, and in a professional rugby cohort neck strength did not differ between concussed and non-concussed players at all. Fourth, and most tellingly, the anticipation effect is at least as large as the strength effect — and the impacts that concuss people are disproportionately the ones nobody saw coming, when there is nothing to anticipate with. A mechanism that is plausible in a rig, inconsistent on the field, unsupported at the outcome, and dependent on a condition that fails precisely in the injurious case is not a mechanism you can convert into a prevention claim.

Why the surrogate is not the outcome. Head velocity in a laboratory is a surrogate for head kinematics in sport, which is itself a surrogate for concussion. Each step has a break in it. A great deal of the "neck protects the brain" literature also runs on cadavers, crash-test dummies and finite-element head models and then reaches the public as if it were demonstrated in living athletes. That model-to-human jump is the dominant overreach in this field, and recognising it is most of what it takes to read the literature correctly.

Why anticipation matters more than it sells. Pre-activation — bracing the neck because you know contact is coming — appears in the data as an independent contributor, and it is trainable through sport-specific drilling and situational awareness rather than through a device. It is the part of the mechanism with the best internal support and the least commercial appeal, which is roughly why you have heard less about it.

Risks And Contraindications

• RED FLAGS — stop training the neck and get assessed. These are not negotiable and they are not a wait-and-see. Radiating pain, numbness, tingling or weakness down an arm. Loss of hand dexterity — dropping things, fumbling buttons, shoelaces, keys, handwriting deteriorating. Gait disturbance or unexplained falls, or a wide-based unsteady walk. New upper-limb weakness. New bladder or bowel disturbance — urgency, hesitancy, retention or incontinence. Lhermitte's sign — an electric-shock sensation running down the spine or into the limbs when you bend your neck forward. These are the symptoms of degenerative cervical myelopathy — compression of the spinal cord itself — with the hand, gait and neck-pain items drawn from a 2025 scoping review of case-control studies, and myelopathy progresses and benefits from timely assessment. Three honesty notes attached: these signs identify myelopathy, they do not establish that neck training caused it; case-control diagnostic-accuracy figures systematically overstate real-world performance; and the bladder/bowel and Lhermitte's items are added here as standard clinical teaching on myelopathy rather than as items we traced to that specific review's sensitivity table. The stop-training instruction itself is clinical prudence and coaching consensus, not a trial-derived or guideline-derived rule — no source we located tests or recommends stopping neck resistance training on these symptoms. We hold it anyway, because the cost of stopping unnecessarily is a fortnight of lost training and the cost of not stopping is a spinal cord.
• SEPARATE RED FLAG — sudden severe neck or occipital pain, sudden severe headache, sudden dizziness, visual disturbance, slurred speech, one-sided facial or limb numbness, or drop attacks during or shortly after neck work is an urgent medical presentation, not a training question. State this without a citation, because we could not source a specific symptom list for it: the general picture is consistent with the cervical artery dissection literature — in which sudden neck or head pain is typically the earliest and most common feature — but the widely circulated screening mnemonic originates in manual-therapy pre-manipulation screening and we did not verify it. Distinguish this from ordinary post-set light-headedness, which is overwhelmingly benign — orthostatic, vasovagal, dehydration, or the aftermath of a breath-hold. Sudden onset plus a neurological deficit, or sudden onset of the worst neck or head pain of your life, is the pattern that matters; mild wooziness after a hard set is not.
• Neck bridges, especially loaded, get a clear caution — and the caution is reasoning, not data. Bridging places body mass through the head in end-range cervical extension, combining axial compression, extreme extension and shear in the position where the cervical spine has least passive protection. That is biomechanically coherent and consistent with what is known about cervical injury mechanisms generally. What does not exist: any trial, case series, biomechanical measurement study or injury registry that quantifies cervical load during bridging or attributes an injury to it. A 2024 narrative review of cervical spine injuries in combat sports does not mention bridging, neck bridge drills, or neck strengthening at all; its injury mechanisms are direct blows, forced flexion, takedowns and submissions. Every caution we could trace runs back to coaching blogs and equipment vendors, several of whom sell the recommended alternative. There is no bridging injury statistic and this entry will not invent one. The wrestling spine-injury rate sometimes quoted alongside this topic (0.71 per 1,000 athlete-exposures, from collegiate injury surveillance, carried from that narrative review and not independently verified in this pass) is all wrestling spine injury from competition and practice in a contact sport — it says nothing about bridging as a drill and must never be presented as a bridging injury rate. Verdict: treat loaded bridging as high-risk-by-mechanism, unnecessary given that tested alternatives exist, and a poor first choice for anyone without a coach and a specific reason. Absence of published injury reports is weak evidence of safety here, because nobody surveils it and wrestlers who bridge are also being thrown.
• Do NOT import cervical-manipulation dissection risk onto ordinary resistance training. In the largest dissection dataset reviewed, sport was a self-reported trigger in 61 of 966 cervical artery dissection patients (6.3%) versus 0.8% of non-dissection stroke patients and 1.8% of healthy controls, and the authors conclude that sport as a cause "seems uncommon". A risk-factor systematic review with meta-analysis grades the evidence as moderate certainty for migraine and MTHFR TT homozygosity, low for minor trauma, and very low for everything else — resistance training does not appear as an established risk factor. Many dissections occur during entirely ordinary activities, which argues that temporal association is often coincidental. The trigger data are self-reported and retrospective, and the 6.3%-versus-1.8% comparison is raw, not an adjusted risk estimate; do not quote it as a risk ratio. These percentages are carried from the cited narrative review and were not independently re-derived from the source dataset in this pass.
• The one published movement caution that does exist, quoted in its correct scope. A review of cervical artery dissection and sport names jerky rapid movements, abrupt rotation, triggered reflexive reactions, long-lasting hyperextension, and rapid flexion-extension of the neck as patterns to avoid as much as possible. This is mechanism-based expert recommendation within a narrative review, written for people who have had a dissection or are at elevated risk — not as a general population training rule. It is the strongest published warrant for the ballistic-and-end-range caution above, and it is being extended beyond its stated audience here, which we are flagging rather than hiding.
• Get clinical clearance before loaded neck training if you have known cervical pathology. Disc herniation with radicular symptoms, radiculopathy, myelopathy, significant symptomatic degenerative change, prior cervical fusion, or an inflammatory arthropathy affecting the neck. This list is clinical prudence with no citation attached, and we are saying so. No paper, guideline or consensus statement we located enumerates contraindications to neck resistance training. Neck training in these populations is understudied, not proven harmful — and a blanket "significant degenerative change" exclusion would sweep in a large asymptomatic population, since incidental degenerative findings on cervical imaging are extremely common in people with no symptoms at all. The honest instruction is to have a clinician who knows your neck make the call, not to self-exclude on the basis of a scan report.
• Heavy isometrics with breath-holding produce large acute blood pressure rises, and the caution belongs to people with vascular risk. Connective tissue disorders and prior arterial dissection are the clearest cases — the dissection review recommends that people with connective tissue disorders permanently avoid combat sports and heavy isometric exercise, and sets out a staged return-to-sport timeline after dissection. That recommendation is expert opinion within a narrative review, and it does not distinguish vascular from hypermobile Ehlers-Danlos, which carry very different risk profiles; the safe handling is clinical clearance rather than a blanket ban. Uncontrolled hypertension and known aneurysm belong on the caution list as standard clinical exercise-screening practice, which we did not verify against a named guideline and are therefore stating generically. A number that must not be attached to neck training: the famous intra-arterial recordings of 320/250 mmHg (with one subject exceeding 480/350) come from five trained male bodybuilders performing the double-leg press to failure. There is no intra-arterial blood pressure data for neck isometrics, and the pressor response scales with active muscle mass, so neck work would plausibly produce a far smaller rise. Anyone quoting the leg-press figure beside a neck exercise has welded two studies together.
• Atlantoaxial instability in Down syndrome: a genuine consideration, commonly overstated. Radiographic instability occurs in 6.8% to 27% of the Down syndrome population, but fewer than 1–2% are thought to develop symptomatic instability, and neurological injury during sport is described as extremely rare. Routine radiographic screening before sport is no longer the American Academy of Pediatrics recommendation, though Special Olympics still requires radiological examination including flexion and extension views for named sports (equestrian, gymnastics, diving, diving starts and butterfly swimming, pentathlon, high jump, football, alpine skiing). The cited work recommends a short symptom screen plus neurological and neck-control assessment rather than routine imaging, and explicitly notes the benefits of sport participation. Scope flag: that paper is about sport preparticipation screening, not resistance training, and cannot be cited as if it addresses neck training. Loaded neck extension and bridging are arguably closer to the flagged sports than to general exercise — but the paper does not say so, and we are not putting words in it.
• The imbalance concern is plausible and undemonstrated. The worry that neck training overdevelops the superficial muscles at the expense of the deep stabilisers is anatomically reasonable — the electromyography data show that common gym work loads the extensor and trapezius complex heavily while doing nothing for the deep flexors — and it is the premise of the whole craniocervical flexion literature. But no trial tests whether neck training produces harmful imbalance, the causal direction between altered activation and pain is unestablished, and the analogous lumbar "core stability" imbalance model has been substantially deflated. Imbalance narratives sell corrective-exercise services and certifications, and have a poor track record when tested. Train both directions because the anatomy justifies it, not because you have been sold a dysfunction.
• On soreness and progression rate: we found no evidence at all on neck-specific delayed-onset muscle soreness. Any statement about how sore a neck should get, or how fast to progress it, is observational practice. Common sense applies — a sore neck interferes with sleep, driving and head-turning in a way a sore quadriceps does not, which is a practical argument for conservative early progression regardless of what the literature does not say. General soreness principles are at muscle_soreness_recovery_doms.

Controversy

Nature: an intervention with genuine, guideline-adjacent support for one outcome (neck pain) and a far louder, far weaker claim for a different outcome (concussion prevention), where the weaker claim is the one funding the research, selling the equipment, and shaping the coaching culture. Error is possible at both poles: selling unproven brain protection, and dismissing a cheap effective treatment for a common painful condition because its marketing is dishonest.

Position A — "Train the neck, it protects your brain." The prevention take.
• Best evidence: a large prospective cohort in 6,704 high-school athletes found roughly 5% lower concussion odds per pound of neck strength; a laboratory study found both isometric strength and anticipatory activation independently reduced head linear and angular velocity; a systematic review found neck exercises within injury-reduction programmes can reduce head and neck injury incidence; a cluster randomised trial of a neuromuscular warm-up showed a dose-responsive concussion reduction in its compliant subgroup; and the mechanism — head-neck coupling and effective mass — is intuitive and partially supported.
• Where it overreaches: the cohort is observational, from a pilot designed for screening, in soccer, basketball and lacrosse athletes, and gets re-labelled onto footballers and fighters — and, decisively, when four prospective studies were pooled the relationship came out small, nonsignificant and very-low-certainty, while a professional rugby cohort found neck strength no different between concussed and non-concussed players. The laboratory outcome is head velocity in a rig, and field studies frequently fail to reproduce the association. The multi-component programme's effect cannot be decomposed and handed to the neck. The trial's randomised concussion comparison was not significant. And the most-quoted numbers in this position — "5% per pound" and "59% fewer concussions" — are systematically stripped of the words "observational" and "subgroup" respectively.

Position B — "Neck training is fitness theatre." The dismissive take.
• Best evidence: the best synthesis of the association itself — eight studies, 7,625 participants — found it small, nonsignificant and very-low-certainty, and the best-powered professional cohort found no predictive value at all; the best systematic review of neck strength-and-conditioning protocols found three eligible studies in roughly 70 male participants and concluded there is a lack of evidence that neck strengthening reduces impact injury risk in adults; international consensus does not list isolated neck strengthening among established preventive measures; deep neck flexor endurance specifically was not associated with reduced concussion risk; and biomechanical modelling suggests the added effective mass from a stronger neck is much smaller than the theory assumes.
• Where it overreaches: it uses a null on concussion to dismiss the entire practice, and it treats a nonsignificant pooled estimate with I²>90% as if it were a demonstrated absence of effect, which it is not — heterogeneity that severe means the relationship could still be real in specific sports, sexes or levels. More importantly, the chronic neck pain evidence is a twelve-month examiner-blinded randomised trial plus an independent Cochrane review of 27 trials and 2,485 participants, all pointing the same direction; cervicogenic headache has a four-arm trial with benefits at twelve months; chronic whiplash has a 13-trial meta-analysis. And a stronger neck is a real trainable capacity with essentially no harm profile — dismissing a cheap, self-administered, low-risk intervention because its most commercial claim is inflated is throwing away the part that works.

The funding and bias dimension — cui bono, both ways. Toward overclaim, the incentives are heavy and converge. Neck harness and neck machine manufacturers publish the overwhelming majority of technique content, and the "stop doing bridges, buy this instead" framing is a direct commercial position, not a finding. The entire academic literature runs on concussion-prevention funding, which selects for male collision-sport athletes and away from general-population neck health — so the evidence base is shaped by who pays for it, not by what most people need. Sporting bodies and leagues benefit enormously from "train the neck, prevent concussion", because a conditioning intervention is far cheaper and less disruptive than rule changes or reduced contact exposure, which is the intervention class with the better evidence. Parent-facing youth-sport marketing is the sharpest edge: selling protection against paediatric brain injury on a single observational odds ratio that has since failed to pool. And note which half of the mechanism gets amplified — the strength half, which a device can raise, rather than the anticipation half, which no device can sell. Toward under-recommendation, the incentives are quieter but real: physiotherapy and rehabilitation clinics benefit from framing loaded work as dangerous and clinician-supervised work as necessary; equipment vendors selling the "safe alternative" have direct interest in bridging being seen as reckless; and this platform's own bias runs toward over-warning, because caution costs nothing to publish and offloads liability from us to a clinician. That last one is ours and we are naming it rather than indulging it — which is why the bridging caution is labelled as mechanism-only, why the contraindication list is labelled as prudence with no citation, and why the dissection literature is kept in its correct scope instead of being borrowed to make the section sound more serious.

Realised Position: Train the neck for the neck. For chronic neck pain, neck strength or endurance training is one of the better-supported, cheaper, lower-harm interventions available, it works over months rather than weeks, and lighter endurance work performed just as well on symptoms as loaded work in the trial that compared them despite gaining a fraction of the strength — so pick the version you will keep doing. For cervicogenic headache, low-load specific neck exercise has trial support that held at twelve months; for tension-type headache and migraine it does not, and that distinction is not optional. Deep-flexor chin-nod work and loaded "big neck" work are two different interventions with two different jobs, and neither substitutes for the other. On concussion, hold the line exactly here: neck training reliably makes the neck stronger; one large observational cohort in soccer, basketball and lacrosse athletes found stronger necks associated with lower concussion odds, and when that association was pooled across every available prospective study it came out small, statistically nonsignificant and very-low-certainty, with a professional rugby cohort finding no predictive value at all; the trial everyone quotes did not reach significance in its randomised comparison; international consensus does not list isolated neck strengthening as established prevention. Therefore neck training is a reasonable component of a broader neuromuscular programme and a dishonest thing to sell as brain protection. Wanting a stronger, thicker neck is a fine reason to train it. Believing it will save your brain is not a reason we will give you.

Cross-Pillar Connections

Neck training sits at the junction of strength work, a common pain condition, a headache subtype, and an appearance motive — so its connections run across all four.
• Physical (physical_counter_modern_postures): the other half of most people's neck complaint. Screen height, desk setup, sustained fixed postures and movement breaks belong there; this entry owns the training intervention and defers the environmental one rather than re-arguing it.
• Cross-pillar (headache_migraine_cross_pillar_diagnostic): the headache differential, and the boundary that must not be blurred. Neck exercise has trial support for cervicogenic headache and does not have it for tension-type headache or migraine — route the "what kind of headache is this" question there before applying anything from this entry.
• Cross-pillar (joint_pain_conservative_management): the general conservative-management frame for musculoskeletal pain — graded loading, expectation setting, when to escalate. This entry owns the neck-specific evidence and defers the general approach.
• Cross-pillar (low_back_pain_evidence_and_management): the instructive parallel. The lumbar "core stability" and deep-stabiliser model has been substantially deflated over the past decade; the cervical version of that model underpins deep-flexor training and has not faced the same scrutiny. Read the two together to see why we hold the cervical control model as a reproducible observation rather than a proven cause.
• Physical (resistance_training_and_body_composition): the general adaptation picture the neck follows — progressive resistance, tolerable load, adaptation over months. Neck training is accessory resistance training, not a separate discipline.
• Physical (specificity_of_adaptation): the entry's central discipline in general form. A trained strength gain in one measured capacity does not automatically produce an outcome in a different domain — which is exactly the step the concussion claim needs and has not made, and the neck-pain trial is the cleanest illustration on this page: four times the strength gain, the same symptom relief.
• Cross-pillar (looksmaxxing_health_first_evidence): the appearance motive, handled honestly. A thicker neck is a common and legitimate aesthetic goal; the standing line is that health drives appearance and appearance is a byproduct rather than the target, because appearance-as-target produces bad decisions about load and progression. Route the aesthetic framing question there rather than moralising here.
• Physical (grip_strength_loaded_carries): the closest analogue for how a small, specific, easily-measured strength quality gets over-interpreted as a general health lever — useful as a template for reading "strong neck equals protected brain" claims sceptically.
• Sleep (sleep_posture_and_sleep_quality): if the neck complaint is worst on waking, sleeping position and pillow height are the more likely lever than training volume; that question belongs there.
• Physical (muscle_soreness_recovery_doms): the general soreness and recovery frame. Worth noting that no evidence at all exists on neck-specific soreness, so the general principles are all there is — and a sore neck interferes with sleep, driving and head-turning in a way a sore leg does not, which argues for conservative early progression.

What would change our mind

Industry bias note

Sources (26)

• Ylinen J, et al. "Active neck muscle training in the treatment of chronic neck pain in women: a randomized controlled trial." JAMA 2003;289(19):2509–16. pubmed.ncbi.nlm.nih.gov/12759322↗/" target="_blank" rel="noopener">PMID 12759322↗. (Occupational-health recruited; no product interest. Author list beyond the first author not independently verified — cited as "et al." deliberately.) — examiner-blinded RCT run February 2000 to March 2002; 180 working women aged 25–53 with chronic non-specific neck pain over six months, randomised in three arms of 60 to high-intensity isometric strength training with an elastic band, lighter dynamic endurance training from supine and prone, or an active control performing aerobic exercise and stretching; twelve months. Pain and disability fell in both training groups versus control (P<.001), with visual-analogue reductions of roughly 40, 35 and 16 points respectively (commonly quoted as 73%, 59% and 21% falls from baseline — those percentages come from the results tables, not the abstract). Maximal isometric strength gains: strength arm 110% flexion, 76% rotation, 69% extension; endurance arm 28%, 29%, 16%; control 10%, 10%, 7%. No statistically discernible difference between the two training groups on pain and disability.
• Gross A, Kay TM, Paquin JP, Blanchette S, Lalonde P, Christie T, et al. "Exercises for mechanical neck disorders." Cochrane Database Syst Rev 2015;1:CD004250. pubmed.ncbi.nlm.nih.gov/25629215↗/" target="_blank" rel="noopener">PMID 25629215↗. (Independent Cochrane review; reaches conclusions unhelpful to easily-marketed options. Trial and participant counts carried from the abstract.) — 27 randomised trials, 2,485 analysed participants, selective rather than wholesale pooling; moderate-quality evidence for cervico-scapulothoracic strengthening (moderate-to-large pain benefit); minimal or no effect for stretching alone, general fitness alone, breathing exercises alone; explicit statement that no high-quality evidence was found; separately, moderate-quality support for static-dynamic cervico-scapulothoracic work including pressure biofeedback in chronic cervicogenic headache.
• Jull G, Trott P, Potter H, Zito G, Niere K, Shirley D, Emberson J, Marschner I, Richardson C. "A randomized controlled trial of exercise and manipulative therapy for cervicogenic headache." Spine 2002;27(17):1835–43. pubmed.ncbi.nlm.nih.gov/12221344↗/" target="_blank" rel="noopener">PMID 12221344↗. (Academic; unblinded treatment with blinded outcome assessment.) — 200 participants, four arms, six weeks of treatment; at twelve months both manipulative therapy and low-load specific exercise significantly reduced headache frequency and intensity and neck pain; combined therapy not significantly superior, roughly 10% more patients gaining relief.
• Madsen BK, Søgaard K, Andersen LL, Tornøe B, Jensen RH. "Efficacy of strength training on tension-type headache: A randomised controlled study." Cephalalgia 2018;38(6):1071–1080. pubmed.ncbi.nlm.nih.gov/28750588↗/" target="_blank" rel="noopener">PMID 28750588↗. (Negative finding, reported as negative.) — 60 tension-type headache patients randomised to ten weeks of elastic-band neck/shoulder strength training versus ergonomic and posture correction, follow-up at 19–22 weeks; no between-group effect; strength arm improved less (frequency −11%, P=0.041; duration −10%, P=0.036) than control (−24%, P=0.0033; −27%, P=0.041); authors state within-group effects did not reach clinical significance; only 23 and 21 participants respectively completed.
• Rinne M, Garam S, Kukkonen-Harjula K, et al. "Neck-Shoulder Region Training for Chronic Headache in Women: A Randomized Controlled Trial." Clin Rehabil 2023;37(10):1322–1331. — 116 working-age women, six-month progressive neck-shoulder home programme versus placebo-dosed TENS plus stretching; null on the primary outcome (pain intensity); secondary outcome of headache frequency fell more with exercise (4.5→2.4 days/week versus 4.4→3.0).↗
• Correia L, Carvalho P, Amaral L, et al. "The Effect of Neck-Specific Exercise with or Without a Behavioral Approach in Chronic Whiplash-Associated Disorders: A Systematic Review and Meta-Analysis." Muscles 2025;4(4):49. (Reviewers' own caveat: few studies, many from the same research groups, heterogeneous protocols.) — 13 randomised trials, 2,427 participants aged 18–63, chronic whiplash grades 2–3; neck-specific exercise with or without a behavioural approach may reduce pain and disability; no significant difference between the two (pain at 6–12 months p=0.71; Neck Disability Index p=0.97).↗
• Michaleff ZA, et al. "Comprehensive physiotherapy exercise programme or advice for chronic whiplash (PROMISE): a pragmatic randomised controlled trial." Lancet 2014. — RESULT NOT VERIFIED. Title, journal, year and design confirmed from an indexed listing; full text inaccessible. Listed here because it is the strongest available check on whether a comprehensive programme beats good advice in chronic whiplash. Its conclusion is deliberately not stated anywhere in this entry.↗
• "Effects of deep cervical flexor training on impaired physiological functions associated with chronic neck pain: a systematic review." BMC Musculoskeletal Disorders 2018;19:373. PMC6263552. (Cited by title and journal — author line not verified.) — strong evidence that deep cervical flexor training improves neuromuscular coordination; no or only small effects on strength and endurance at higher contraction loads; improved head and cervical posture; limited or contradictory evidence for other outcomes. Supporting and indicative only: a systematic review with meta-analysis of cranio-cervical flexion versus other treatments for non-specific chronic neck pain, Musculoskeletal Science and Practice 2019 — identified, not read in full.↗
• "Effect of Pressure biofeedback training on deep cervical flexors endurance in patients with mechanical neck pain: A randomized controlled trial." PMC7931293. — cited for the standard method of dosing and monitoring craniocervical flexion; the protocol concept is attributed generically to Jull and colleagues, as no specific primary paper of theirs was retrieved in this pass.↗
• Collins CL, Fletcher EN, Fields SK, Kluchurosky L, Rohrkemper MK, Comstock RD, Cantu RC. "Neck strength: a protective factor reducing risk for concussion in high school sports." J Prim Prev 2014;35(5):309–19. pubmed.ncbi.nlm.nih.gov/24930131↗/" target="_blank" rel="noopener">PMID 24930131↗. (Observational pilot with a screening-tool interest; the framing "protective factor" is the paper's own title, not a demonstrated effect. Design, N, schools, states, years, sports and the odds ratio all verified directly against the abstract.) — 6,704 high-school athletes, 51 schools, 25 states, 2010–2011; neck strength by validated hand-held tension scale; smaller neck circumference, smaller neck-to-head ratio and weaker neck strength each significantly associated with concussion; OR 0.95 per pound (95% CI 0.92–0.98); concussion ascertained by school athletic trainers; sports were boys' and girls' soccer, basketball and lacrosse. Read alongside Garrett 2023 and Liston 2023 below — this cohort's association does not survive pooling.
• Garrett JM, Mastrorocco M, Peek K, van den Hoek DJ, McGuckian TB. "The Relationship Between Neck Strength and Sports-Related Concussion in Team Sports: A Systematic Review With Meta-analysis." J Orthop Sports Phys Ther 2023;53(10):585–593. pubmed.ncbi.nlm.nih.gov/37428807↗/" target="_blank" rel="noopener">PMID 37428807↗. (Independent academic synthesis; the result is commercially inconvenient in every direction.) — eight studies, 7,625 participants; four prospective longitudinal studies meta-analysed; pooled effects small (r = 0.08–0.14) and nonsignificant, heterogeneity I²>90% attributed to differing ages, playing levels and sports; conclusion verbatim: "There was very low-certainty evidence suggesting a small, nonsignificant relationship between greater neck strength and a lower risk of sustaining a SRC." This supersedes single-cohort readings of the strength-concussion association, including the Elliott 2021 review's observational component.
• Liston, Leckey, Whale, van Dyk. "Neck Strength Assessment Offers No Clinical Value in Predicting Concussion in Male Professional Rugby Players: A Prospective Cohort Study." J Orthop Sports Phys Ther 2023. pubmed.ncbi.nlm.nih.gov/37017931↗/" target="_blank" rel="noopener">PMID 37017931↗, doi 10.2519/jospt.2023.11723.↗ (Author initials and volume/pages not independently verified.) — 136 male professional rugby players, 40 players sustaining 51 concussions over a season; peak isometric flexion and extension strength plus endurance measured; neck muscle strength similar between concussed and non-concussed players (peak isometric extension OR ≈ 1.01); prior concussion history was the dominant predictor, carrying over twice the odds; authors conclude neck-strength assessment has no clinical value for concussion prediction.
• Eckner JT, Oh YK, Joshi MS, Richardson JK, Ashton-Miller JA. "Effect of neck muscle strength and anticipatory cervical muscle activation on the kinematic response of the head to impulsive loads." Am J Sports Med 2014;42(3):566–76. pubmed.ncbi.nlm.nih.gov/24488820↗/" target="_blank" rel="noopener">PMID 24488820↗. (Academic laboratory study.) — 46 male and female contact-sport athletes aged 8–30; maximum isometric neck strength and anticipatory activation each independently associated with lower head linear and angular velocity across all planes (P<.001), r = 0.417–0.657; outcome is head velocity in a rig, not concussion. Counter-literature reporting weak or absent strength-kinematics associations in real impacts and in modelling work is stated generically in this entry, with no citation attached, because no individual counter-paper was read in full. That generic kinematics counter-claim is separate from the named incidence counter-evidence above and must not be conflated with it.
• Daly E, Pearce AJ, Ryan L. "A Systematic Review of Strength and Conditioning Protocols for Improving Neck Strength and Reducing Concussion Incidence and Impact Injury Risk in Collision Sports; Is There Evidence?" J Funct Morphol Kinesiol 2021;6(1):8. PMC7838928. (The most commercially inconvenient conclusion in the field.) — 2,462 articles screened, 3 eligible studies, all male; participant count discrepancy noted openly: the abstract states 68 while the three included samples as described (25, 27, 18) sum to 70 — "roughly seventy" either way; verbatim conclusion of a lack of evidence to support neck strengthening interventions for reducing impact injury risk in adult sport; two of three studies found significant isometric strength gains and none reported any effect on cervical spine injuries or concussions. Protocols: Geary et al. 2014, 25 male professional/semi-professional rugby union players, manual isometric resistance 3×10s holds, four directions, twice weekly, five weeks (flexion 334.45±39.31→396.05±75.55 N; extension 606.19±97.34→733.88±127.16 N); Naish et al. 2013, 27 male professional rugby union players, cable and scrum-machine isometrics, 26 weeks split 13 strengthening + 13 maintenance, non-significant increases; Versteegh et al. 2019, 18 male college American football players (8 intervention, 10 control), rotational device, 6×50 revolutions each direction, twice weekly, seven weeks, mean change 32 N (95% CI 13–50) versus 12 N. No harness, plate-on-forehead or four-way-machine study appeared.↗
• Elliott J, Heron N, Versteegh T, Gilchrist IA, Webb M, Archbold P, Hart ND, Peek K. "Injury Reduction Programs for Reducing the Incidence of Sport-Related Head and Neck Injuries Including Concussion: A Systematic Review." Sports Med 2021;51(11):2373–2388. pubmed.ncbi.nlm.nih.gov/34143411↗/" target="_blank" rel="noopener">PMID 34143411↗. — six studies; higher neck strength but not deep neck flexor endurance associated with reduced concussion risk in observational work; injury-reduction programmes including neck exercises can reduce head and neck injury incidence; the identified agent is the multi-component programme. Its observational strength-association component is superseded by Garrett 2023 and should not be cited as current; only the multi-component-programme finding stands.
• Hislop MD, Stokes KA, Williams S, et al. "Reducing musculoskeletal injury and concussion risk in schoolboy rugby players with a pre-activity movement control exercise programme: a cluster randomised controlled trial." Br J Sports Med 2017;51(15):1140–1146. PMC5530334. (Trial figures verified directly.) — 40 schools, 118 teams, 3,188 players aged 14–18, 441 time-loss match injuries over 15,938 match exposure-hours; intention-to-treat concussion RR 0.71 (95% CI 0.48–1.05); adherence-restricted (≥3 sessions/week) RR 0.41 (95% CI 0.17–0.99); intervention was whole-body neuromuscular training (balance, resistance, plyometrics, landing and cutting rehearsal). Framing disclosure: the paper's own abstract describes the intention-to-treat concussion result as showing "clear reductions" although the interval crosses one; we report it as not reaching significance.↗
• Patricios JS, Schneider KJ, Dvorak J, et al. "Consensus statement on concussion in sport: the 6th International Conference on Concussion in Sport — Amsterdam, October 2022." Br J Sports Med 2023;57(11):695–711. pubmed.ncbi.nlm.nih.gov/37316210↗/" target="_blank" rel="noopener">PMID 37316210↗. (Verification limit: identity and citation verified directly; full text inaccessible, so the prevention content used here comes from the abstract plus reputable secondary summaries. Do not quote its prevention wording verbatim without reading it.) — prevention measures identified: mouthguards in ice hockey, policy disallowing bodychecking in ice hockey, neuromuscular training warm-up programmes in adolescent rugby; isolated neck strengthening not among them; cervicovestibular rehabilitation appears as treatment for dizziness, neck pain or headache persisting beyond roughly ten days after concussion.
• Price J, Rushton A, Tyros I, Tyros V, Heneghan NR. "Effectiveness and optimal dosage of exercise training for chronic non-specific neck pain: A systematic review with a narrative synthesis." PLOS ONE 2020;15(6):e0234511. — 3,990 citations screened, 26 trials; optimal dosage explicitly not known; low-to-moderate evidence for motor control plus segmental exercise short-term; no high-quality long-term trials; limited evidence that higher frequency may improve motor-control effectiveness, and that progressive load is moderately more effective than fixed load for short-term pain. Population is chronic neck pain patients, not healthy trainees.↗
• "Electromyographic Evaluation of Specific Elastic Band Exercises Targeting Neck and Shoulder Muscle Activation." Applied Sciences 2020;10(3):756, doi 10.3390/app10030756.↗ (Cited by title and DOI only — author list not verified.) — 11 healthy males, mean age 25.9, six exercises at 12RM and 20RM, bilateral electromyography normalised to maximal voluntary activation; shrugs produced 100.3 ± 29.8 %MVE upper trapezius and 61.9 ± 16.8 %MVE upper neck extensors. A separately circulating "67.6 ± 29.8 %MVE for cervical extension" figure carries a standard deviation identical to the shrug upper-trapezius value and appears to be two table rows merged in secondary reporting — discarded, not used.
• Bakirtzis et al. "A Narrative Review of Combat Sports Injuries With a Particular Focus on Cervical Spine Injuries." Cureus 2024;16(12):e74980. (Cited primarily for what it does NOT contain.) — injury mechanisms covered are direct blows, forced flexion and whiplash, takedowns and throws, and submissions; prevention section covers rules, equipment and concussion education; neck bridging, neck bridge drills and neck strengthening are not mentioned at all. Also the source of collegiate wrestling spine-injury context (57 recorded spine injuries, estimated 2,040 nationally, 0.71 per 1,000 athlete-exposures — carried from this review, not independently verified) — this is all wrestling spine injury in a contact sport and is not a bridging-attributable rate.↗
• "Cervical Artery Dissection and Sports." Frontiers in Neurology 2021;12:663830. (Narrative review; vascular-neurology authorship sees the rare catastrophic case and not the uneventful denominator, which biases toward caution. Percentages carried from this review, not re-derived from the source dataset.) — reporting CADISP data, sport self-reported as a trigger in 61 of 966 dissection patients (6.3%) versus 0.8% of non-dissection stroke patients and 1.8% of healthy controls, with the conclusion that sport as a cause "seems uncommon"; many dissections occur during ordinary daily activities; movement patterns named as risk-relevant are jerky rapid movements, abrupt rotation, triggered reflexive reactions, long-lasting hyperextension and rapid flexion-extension; recommends permanent avoidance of combat sports and heavy isometric exercise in connective tissue disorders, and a staged return-to-sport timeline after dissection. Addressed to people with or at elevated risk of dissection, not to the general population.↗
• "Risk factors for cervical artery dissection: a systematic review with meta-analysis." pubmed.ncbi.nlm.nih.gov/40866078↗/" target="_blank" rel="noopener">PMID 40866078↗. — moderate certainty for migraine and MTHFR TT homozygosity; low for minor trauma; very low for everything else; resistance training does not appear as an established risk factor.
• MacDougall JD, Tuxen D, Sale DG, Moroz JR, Sutton JR. "Arterial blood pressure response to heavy resistance exercise." J Appl Physiol 1985;58(3):785–790. Fidelity warning — do not attach these numbers to neck training. — five experienced male bodybuilders, direct intra-arterial recording, exercises to failure at 80/90/95/100% of max; peak pressures highest during the double-leg press, group mean 320/250 mmHg, one subject exceeding 480/350 mmHg. There is no intra-arterial blood pressure data for neck isometrics, and the pressor response scales with active muscle mass.↗
• "Clinical signs and symptoms for degenerative cervical myelopathy: a scoping review of case-control studies to facilitate early diagnosis." Spinal Cord 2025, doi 10.1038/s41393-025-01065-1↗; PMC11906348. Corroborated by "Degenerative Cervical Myelopathy: Recognition and Management." Am Fam Physician 2020;102(12):740. (Spinal surgery has a documented interest in earlier myelopathy detection; the review frames itself around facilitating early diagnosis, so the urgency register is not neutral — though myelopathy genuinely progresses and genuinely benefits from timely decompression.) — highly sensitive symptoms: hand numbness, hand paraesthesia, loss of dexterity, gait abnormality, neck pain; examination sign specificities Babinski 93–100%, Tromner 79–100%, clonus 96–99%, inverted supinator 78–99%; most sensitive signs Tromner 93–97% and hyperreflexia 15–85%; case-control designs inflate accuracy relative to primary care. The bladder/bowel and Lhermitte's items in this entry's red-flag list are standard clinical teaching on myelopathy, added for completeness, and are not traced to this review's sensitivity table.
• "Sport Preparticipation Screening for Asymptomatic Atlantoaxial Instability in Patients With Down Syndrome." Clin J Sport Med 2018, pubmed.ncbi.nlm.nih.gov/30119085↗/" target="_blank" rel="noopener">PMID 30119085↗, doi 10.1097/JSM.0000000000000642.↗ (Scope flag: about sport preparticipation screening, not resistance training.) — radiographic atlantoaxial instability in 6.8–27% of the Down syndrome population, fewer than 1–2% symptomatic, neurological injury during sport extremely rare; routine radiographic screening no longer the AAP recommendation; Special Olympics still requires radiological examination for named sports; recommends a short symptom screen plus neurological and neck-control assessment over routine imaging, and notes the benefits of participation.
• Funding notation: the entry's firmest claims — that neck strength and endurance training reduces chronic neck pain, that both training modes work about equally despite a fourfold difference in strength gain, that cervico-scapulothoracic strengthening beats stretching alone, that the pooled strength-concussion association is small and nonsignificant, and that there is a lack of evidence for neck strengthening reducing impact injury risk — come from an occupational-health-recruited randomised trial, an independent Cochrane review, an independent academic meta-analysis, and a systematic review whose conclusion is directly damaging to the equipment market. The claims running the other way are the ones with money behind them: concussion prevention funds essentially the whole neck-training research programme and selects its populations; harness and machine manufacturers write the technique content and have direct interest in modality-specific claims no trial supports; and sporting bodies benefit from a conditioning answer that costs less than rule change. Four specific fidelity hazards are flagged in place rather than buried — the intra-arterial blood pressure figures belong to the leg press and not the neck, the Down syndrome screening paper covers sport participation and not resistance training, one circulating electromyography figure appears to be two table rows welded together and has been discarded, and the Ylinen trial's elastic band belonged to its strength arm rather than its endurance arm. Two sources are cited with explicit verification limits (the consensus statement's prevention wording, the PROMISE trial's result), one carries an open participant-count discrepancy (68 in the abstract versus 70 across the included samples), and several claims are stated generically with no citation attached because no verified source was located — the field-based counter-literature on head kinematics, the vertebrobasilar symptom list, the contraindication list, and neck-specific soreness.*↗