Background article for impactfeaturespace.com. Narrative review; limits of the evidence are stated in the text and in the evidence notes.
Key points
Concussion is defined clinically. Standard CT and conventional MRI are normal by definition. The proposed sequence is mechanical deformation of the brain, mainly in shear, then axonal and membrane disturbance, then an ionic, glutamatergic and energy disturbance that unfolds over hours to weeks.
Much of the cellular sequence is known from animal and in vitro work. Direct measurement of strain during injurious impacts in living humans is not possible, and the links from cellular events to symptoms are largely hypotheses.
Most athletes recover within weeks. The persisting-symptom minority is heterogeneous. It combines injury-related biology with physiological subtypes, comorbidity and perception effects.
Continuing to play, delayed care and an early second injury are associated with longer recovery. Whether they worsen the injury itself is not established, and the catastrophic form (second impact syndrome) is contested.
No peer-reviewed serial FDG-PET study of sport concussion was identified in the sources reviewed for this article. The mechanism of knock-out is unresolved, and the role of mechanoporation in concussion is contested.
Long-term risk after repetitive head impacts is unresolved. The best available studies are limited by selection bias and confounding.
1. The event and its definition
The Amsterdam 2022 consensus defines sport-related concussion as a traumatic brain injury caused by a direct blow to the head, neck or body that transmits an impulsive force to the brain. It initiates a neurotransmitter and metabolic cascade, with possible axonal injury, blood flow change and inflammation. Symptoms and signs may appear immediately or evolve over minutes or hours, commonly resolve within days, and may be prolonged. Standard structural imaging (CT; T1- and T2-weighted MRI) shows no abnormality, though research imaging may show functional, blood flow or metabolic changes [1].
Three qualifications matter for experts:
The definition was accepted by 78.6% of the panel, below the prespecified 80% consensus threshold. The voting figure records a dissent arguing that concussion should be defined by pathological, not clinical, constructs because symptoms are non-specific [1].
It provides no diagnostic criteria. The ACRM operational criteria for mild TBI were developed in parallel, and a unified conceptual and operational definition remains an aim [1].
Terminology (concussion, mild TBI, Kommotion, Kontusion) is treated in the companion article on terms. Here, "concussion" is the clinical syndrome defined above and "mild TBI" the broader operational category.
2. From head motion to tissue deformation
Linear and rotational acceleration occur in nearly every concussion [2]. What the tissue experiences is deformation. Brain tissue resists volume change far more than shear: Kleiven put the bulk modulus five to six orders of magnitude above the shear modulus, and argued on that basis that most traumatic brain injuries are driven by shear and therefore by rotational kinematics [3].
Direct measurements in living people exist only for non-injurious loading, and the authors of one of those studies note that experiments on in vivo deformation during injury-level impacts cannot be performed [4]. Tagged MRI of volunteers has shown:
Brain deformation during mild head deceleration (20–30 m/s², about 40 ms), consistent with contrecoup patterns [5].
Radial-circumferential shear strain patterns during mild angular acceleration [6].
Three-dimensional strain tensors after a mild head impact, with maximum principal strain of 0.019–0.053 correlating strongly with peak angular velocity, patterns that differed between neck rotation and extension, and the first in vivo estimates of axonal fibre stretch [4].
Head acceleration measured by a sensor is therefore an indirect stand-in for tissue deformation. How well kinematics predict concussion is covered in the companion article on linear and rotational acceleration.
3. The axon and the membrane
3.1 Axonal injury
Axons in white matter appear especially vulnerable to mechanical loading. Diffuse axonal injury is found across all severities of TBI and may be a key pathological substrate of mild TBI. It covers a spectrum from mechanical breaking of the cytoskeleton, through interrupted transport, swelling and proteolysis, to secondary changes [7]. Giza and Hovda summarise the experimental picture [8]:
Calcium entry collapses neurofilament side-arms and activates proteases that digest cytoskeletal proteins.
Microtubule breakage interferes with two-way axonal transport and can isolate the synapse.
Disconnection does not necessarily kill the neuron. Perisomatic axotomy has been followed by shrinkage without cell death in animals.
In one fluid percussion comparison, unmyelinated axons were more vulnerable than myelinated ones.
Little pathological data exists for human concussion, and mild-injury models often show cognitive impairment with little cell death.
3.2 Mechanoporation and its competitor
The mechanoporation hypothesis proposes that mechanical force creates small sublethal defects in the lipid membrane. Potassium leaves, sodium and calcium enter, and glutamate is released indiscriminately [8]. Altered axolemmal permeability after injury was shown early with tracers [9].
In a rat diffuse injury model, a normally excluded 10 kDa fluorescent dextran was infused before injury and a second tracer at 2 or 6 hours afterwards. At 4 and 8 hours, 55% of tracer-flooded neurons contained both tracers (enduring leakage, many necrotic). About 39% (4 h) and 24% (8 h) contained only the pre-injury tracer, consistent with resealing. Delayed new perturbation occurred in 5.5% and 20.9%, and the changes were linked to calpain-mediated spectrin proteolysis [10].
In an in vitro axonal stretch model, calcium rose immediately but was absent in calcium-free medium, abolished by tetrodotoxin or sodium replacement, and reduced by a voltage-gated calcium channel blocker. The authors found no evidence of calcium entry through mechanically produced pores, and proposed that deformation opens mechanically sensitive sodium channels, which trigger calcium entry through voltage-gated calcium channels and reversal of the sodium–calcium exchanger [11]. Sodium influx then led to proteolysis of the channel's α-subunit, a feed-forward process [12].
Assessment: both processes may operate, at different strains, strain rates or cell compartments. Neither has been shown in human concussion. The tracer method detects defects large enough for a 10 kDa molecule, so small ion-permeable defects would not be seen (inference). Resealing fits the reversibility of concussion, but that is interpretation. Integrin-mediated cytoskeletal anchor points have been proposed as a primary molecular target in neurons and vascular cells [8].
4. The chemical and energy disturbance
4.1 The neurometabolic cascade
The cascade as summarised by Giza and Hovda [8, 13, 14]:
1. Ion flux and glutamate release. Potassium efflux and sodium and calcium influx, with indiscriminate glutamate release. This may create a diffuse spreading-depression-like state, a possible substrate of the very acute impairments [8].
2. Energy crisis. ATP-dependent pumps work overdrive, causing hyperglycolysis and depletion of energy reserves. Early on this occurs with normal or reduced cerebral blood flow, so supply and demand are uncoupled [8].
3. Mitochondrial strain. Calcium is sequestered in mitochondria, which can impair oxidative metabolism. Oxidative stress alters the redox state [8].
4. Later hypometabolism. In adult animals impaired glucose metabolism lasted up to 7–10 days, and about 3 days in juveniles [8].
5. Altered neurotransmission in excitatory and inhibitory systems in animals [8].
6. Inflammation. Cytokine and inflammatory gene changes after mild TBI in animals, less studied acutely than in severe injury [8].
Metabolic changes can exist in head-injured patients with a normal Glasgow Coma Scale [13].
4.2 The human evidence
Human data come mainly from proton spectroscopy. N-acetylaspartate was reduced in frontal white matter after concussion, normalised over about 30 days, and took about 45 days after a second concussion. The reduction did not correlate well with symptoms [8]. Giza and Hovda expect human recovery to take longer than in animal models [8].
4.3 Serial FDG-PET
No peer-reviewed serial FDG-PET study in sport concussion was identified in the sources reviewed, and the existence of one cannot be excluded. Serial FDG-PET does exist in mixed-severity human TBI (13 of 54 patients were scanned acutely and again later), where the reduced-metabolism phase began to resolve at about one month regardless of severity [15]. What was found:
Two conference abstracts from one PET/MRI programme. Collegiate athletes were scanned acutely and again after more than three months, because baseline scans on all athletes were impractical. The 2018 abstract reports whole-brain uptake up to 97% above a control in two concussed athletes, falling with time, with MRI-derived grey matter oxygen metabolism up to 45% lower. It interprets this as raised cortical glucose metabolism with depressed oxidative metabolism [16]. The 2019 abstract reports seven acute scans (31–94 h) and three follow-ups. Whole-brain uptake was lower in the acute phase than at follow-up (p=0.02, unpaired), with a trend toward acute hypometabolism and regional findings that did not survive familywise correction [17]. The two interpretations differ. Both are tiny, unreviewed abstracts.
Mixed-severity human TBI. FDG-PET within a month of injury showed reduced cerebral glucose metabolic rate that was not tightly coupled to level of consciousness [18]. The patient numbers given here for that study (42 patients, 84% affected) come from a news summary and were not checked against the paper. The dynamic profile begins to resolve at about one month regardless of severity [15]. In severe TBI (19 patients, mean GCS 6), combined PET and microdialysis at about 36 hours showed a metabolic crisis not explained by ischaemia [19]. Serial FDG-microPET in rats after fluid percussion injury showed metabolic recovery that correlated with behavioural recovery [20]. None of these are concussion populations.
A critical review of FDG-PET in mild TBI [21]. Only its introduction and technical sections were reviewed. It notes that FDG reflects uptake and phosphorylation, which is not total metabolism, and that clinical PET spatial resolution is about 5–10 mm.
Reasons serial FDG-PET is hard in concussion (inference): radiation dose in healthy young athletes, the need for an acute scan within hours to days, the absence of baseline, and the fact that FDG cannot separate glycolytic from oxidative metabolism without an additional measure of oxygen use. The energy-crisis hypothesis therefore rests on animal work, human spectroscopy and the indirect clinical rationale for delaying return to contact.
4.4 The window of vulnerability
In animals, the worst metabolic disturbance followed injuries three days apart. Injuries seven days apart did not compound. In juvenile rats a second injury on day 1 worsened outcome, and one on day 5 acted as a separate injury [8]. In a cell-culture model, stretch below 5% strain caused no overt change, but axons showed increased sodium channel expression by 24 hours. A second identical mild injury then produced pathological calcium rises and axon degeneration [22]. Clinically, prior concussion is a strong risk factor for another, with the greatest risk in the first 10 days [8]. The length of the window probably varies between individuals, and no biomarker can currently measure it [8].
| Time after impact | Proposed events | Main evidence |
|---|---|---|
| Milliseconds | Skull accelerates, tissue deforms | Models; volunteer MRI at mild loads [4–6] |
| Seconds to minutes | Membrane disturbance, ion flux, glutamate release | Animals, in vitro [8, 10, 11] |
| Minutes to hours | Energy crisis, flow–metabolism mismatch, transport disruption | Animals [8] |
| Hours to days | Hypometabolism, symptoms, inflammatory response | Animals; clinical observation [8] |
| Days to weeks | Symptom recovery in most; spectroscopic normalisation (about 30 days) | Human clinical studies [1, 8] |
| Months and longer | Persisting symptoms in a minority; long-term outcome uncertain | Clinical and epidemiological studies [1] |
5. From cellular events to symptoms
Giza and Hovda propose links between processes and symptoms, and state that studies proving the associations remain to be done [8]:
| Process | Proposed correlate |
|---|---|
| Ion flux | Migraine-like headache, photophobia, phonophobia |
| Energy crisis | Vulnerability to a second injury |
| Axonal injury, impaired neurotransmission | Impaired cognition, slowed processing and reaction time |
| Protease activation, cytoskeletal change, cell death | Chronic atrophy, persistent impairment (hypothesis for repeated injury) |
Headache history and prior migraine are reported risk factors for prolonged recovery [8]. A family with a CACNA1A calcium channel mutation and familial hemiplegic migraine showed delayed cerebral oedema after minor head trauma [23]. Giza and Hovda suggest ion-channel genetics may underlie both "glass jaw" susceptibility and malignant swelling [8]. This is a case-family observation and not a population risk estimate.
6. Inflammation and what is not seen
Cell death is minimal in mild injury models even with cognitive impairment [8]. In a Glasgow archive of TBI cases of mixed severity, reactive microglia and ongoing white matter degeneration were present in 28% of cases with survival beyond one year, up to 18 years after a single injury [24]. Those were not sports concussions, so the finding illustrates a possible pathway and does not describe concussion.
7. Diagnostic tools and their limits
DTI gives variable results after mild injury. Fractional anisotropy was increased in some acute studies (for example corpus callosum at 6 days in adolescents, correlating with symptom scale) and decreased in others, depending on timing, age, severity, region and method. Routine use for individual patients is premature [8]. A German expert-assessment guideline reaches a similar conclusion for individual proof of brain injury [25].
fMRI has shown abnormal activation patterns days to weeks after concussion in some studies, with mixed findings [8].
GFAP and UCH-L1. In 1959 adults with suspected head injury (GCS 9–15), within 12 hours of injury the combined test had sensitivity 0.976 and negative predictive value 0.996 for CT-detected intracranial injury. It is a way to avoid CT, not a test for concussion [26].
Neurofilament light was raised in cerebrospinal fluid, and amyloid β lowered, in 16 professional hockey players with post-concussion syndrome for more than 3 months [27]. In a larger cohort, serum NfL separated players with persisting symptoms from preseason controls (AUROC 0.97), with Class III evidence [28].
Amsterdam 2022 concludes that advanced imaging, fluid biomarkers and genetic testing are valuable research tools but not yet suited to routine clinical use. It adds that physiological change may outlast clinical recovery, and that it is unknown whether residual changes are pathological, adaptive or benign [1].
8. Knock-out and impact convulsions
Brief loss of consciousness is a hallmark of some concussions but is not required [1]. Its mechanism remains unexplained.
Swine and primate rotation experiments.
In swine, non-impact head rotation in the axial plane (transverse to the brainstem) produced immediate prolonged coma, and coronal rotation did not. Diffuse axonal injury occurred in both planes. Coma severity correlated with the extent of brainstem axonal damage (p<0.01) and not with damage elsewhere [29].
A later swine mild-injury study found 10–35 minutes of loss of consciousness after axial rotation (n=3) and none sustained after coronal (n=2). Extensive axonal pathology was present at 7 days in both, even without sustained loss of consciousness [30].
In primates, coronal head motion produced the longest comas and most axonal injury: coma over 6 hours in 20 of 26 animals with full lateral motion [31].
The rotation plane that matters is therefore reversed between species, and Smith and colleagues note differing brainstem and callosal axonal damage between swine and primates [29].
Candidate mechanisms.
Brainstem disruption. The centripetal hypothesis proposes that rotational shear affects the surface first and reaches the upper brainstem with increasing severity [32]. A teaching resource summarises criticisms: it requires high energies, and patients who lose consciousness often lack more structural damage (secondary source) [33].
Diffuse depolarisation or "convulsive" explanations. Shaw (2002), as cited by Bayly, proposed that rapid deformation causes diffuse mechanically induced depolarisation of cortical neurons [5]. A chapter abstract by Shaw lists five hypotheses (vascular, reticular, centripetal, pontine cholinergic, convulsive) and concludes that only the convulsive theory is readily compatible with the neurophysiological data; the 2002 paper itself was not reviewed. Giza and Hovda describe a spreading-depression-like state as a possible substrate [8]. A pontine cholinergic mechanism was proposed in cats [34]. In cats, microinjection of a cholinergic agonist into the pontine reticular formation produced a reversible flaccid coma without epileptic EEG discharges or convulsive movements [35]. The inference drawn here is that this sits uneasily with a purely convulsive account and fits the pontine cholinergic hypothesis.
Historical vascular and autonomic view. Denny-Brown and Russell distinguished acceleration concussion, in which cerebral blood flow was greatly increased through stimulation of nerves at the foramen magnum, from compression concussion. They considered failure of the "veno-pressor system" important in death shortly after experimental concussion [36].
Glucose metabolism of deep structures after TBI. In 23 patients, thalamic, brainstem and cerebellar glucose metabolic rates correlated with the level of consciousness at the time of PET, and the authors interpret this as functional disconnection from subcortical white matter damage [37]. These were comatose and non-comatose TBI patients, not concussions.
Concussive convulsions. In 22 elite football cases, convulsions began within 2 seconds of impact, with tonic stiffening then myoclonic jerks up to 150 seconds. Recovery of consciousness was rapid, no structural or permanent injury was found, no player developed epilepsy over a mean 3.5 years, and the authors proposed a transient traumatic functional decerebration [38]. In 102 videotaped cases, tonic posturing occurred in 25 and was predicted only by loss of consciousness [39]. German literature calls the phenomenon kommotionelle Konvulsion or Aufprallanfall [40].
Synthesis: prolonged coma in animals is tied to brainstem axonal injury. Brief knock-out may reflect transient brainstem or network dysfunction without structural damage. Loss of consciousness is a poor guide to axonal injury [30]. No tested mechanism explains why the critical plane of rotation differs by species, and no source was found that tests a link between ion flux or mechanoporation and loss of consciousness.
9. Recovery and persisting symptoms
Typical course. Unrestricted return to sport typically occurs within one month, with a pooled mean of 19.8 days (95% CI 18.8–20.7 across 57 studies, with very high heterogeneity). 93% returned to learning without added support by 10 days. Early symptoms remain the most robust predictor of recovery. Continuing to play and delayed access to a clinician were associated with longer recovery, and sleep disturbance in the first 10 days with persisting symptoms [1].
Persisting symptoms are defined as lasting more than four weeks in all age groups. They may be pre-existing, concussion-related or both [1]. A conservative estimate of prolonged post-concussive symptoms is 10–15% of mild TBI, with psychogenic, cervicogenic and neurophysiological contributors [41]. Rates depend on population and definition:
In a Canadian emergency-department cohort of children and adolescents with acute head injury, 31.0% had persistent post-concussion symptoms at 28 days [42].
In a US adult trauma-centre cohort, 53% of mild TBI patients and 38% of orthopaedic-injury controls reported injury-related impairment at 12 months on the Glasgow Outcome Scale Extended. Among mild TBI patients, 61% with a positive admission CT and 49% with a negative CT were impaired (relative risk 1.24, 95% CI 1.08–1.43) [43].
Neither cohort is sporting, and the outcome measure is non-specific. The control result matters: a large share of people with an unrelated peripheral injury also report impairment.
Candidate contributors.
Injury-related biology. Early symptom severity is the most consistent predictor of slower recovery in a systematic review of 101 studies, and subacute headache or depression likely raise the risk of symptoms beyond a month [44]. A study of adults found higher post-injury symptom burden predicted symptoms beyond 28 days [45]. In 135 mild TBI patients, 27% of CT-normal cases had an abnormal early MRI. A contusion on MRI (odds ratio 4.5) and four or more foci of haemorrhagic axonal injury (3.2) were independently associated with worse 3-month outcome [46]. Serum NfL discriminated hockey players with persisting symptoms [28].
Physiological subtypes. One classification separates a physiological disorder (global metabolic impairment), a vestibulo-ocular disorder and a cervicogenic disorder, distinguished by history, examination and exercise testing [47]. Symptoms alone did not distinguish physiologic concussion from cervical or vestibular injury [48].
Comorbidity. Prior mental health problems appear to raise risk. ADHD and learning disabilities do not appear to raise it substantially. The teenage years, particularly girls in high school, may be most vulnerable [44]. Migraine history, prior concussion with symptoms longer than a week and female sex were components of the 12-point risk score in children [42].
Perception and reporting. In a healthy sample, 35.9–75.7% reported experiencing a given post-concussion-like symptom in the past two weeks, and 2.9–15.5% reported more severe symptoms, correlating with depressive symptoms [49]. After mild head injury, early beliefs in serious lasting consequences predicted enduring symptoms at 3 months, and adding injury severity, post-traumatic stress, anxiety and depression did not improve prediction [50]. In a non-clinical sample, the diagnostic label for an identical injury description changed expected symptoms and illness perceptions [51].
Management. In adolescent athletes, those who continued to play took longer to recover (44.4 vs 22.0 days) and were 8.8 times more likely to have recovery of 21 days or more (unadjusted analysis). An adjusted regression in the same report gave an odds ratio of 14.27 [52]. The two figures come from different analyses with different group sizes, and I took them from abstracts and slides, not from the paper's tables. The study is observational.
Can persisting symptoms be recognised beforehand? Partly. The 5P risk score for children performed better than physician judgement but had modest discrimination and required external validation [42]. Amsterdam states that evidence-based recommendations for specific tests to diagnose persisting symptoms are not possible on existing research [1].
Is the concussion definition too wide, or too narrow? Assessment:
The data support "too heterogeneous" more than either option as posed. Diagnosis rests on non-specific symptoms [48, 49], and the persisting group probably combines concussion-related physiological dysfunction, cervical, vestibular and oculomotor disorders, migraine, sleep and mood disorders, pre-existing conditions and perception effects.
A structural subgroup exists inside "mild" [46], which the "complicated" versus "uncomplicated" distinction and the German commotio/contusio split try to capture. The split does not map cleanly onto outcome: impairment at 12 months was common in CT-negative patients (49%) [43]. Brief or absent loss of consciousness does not exclude substantial axonal pathology [30].
"Too narrow" has support for exposure and not for the clinical criterion. Some imaging studies report changes in athletes with repeated head impacts and no diagnosed concussion [8, 53].
A multi-axis characterisation (mechanism, structural findings, physiology, comorbidity) fits better than moving a threshold. A review of mild TBI recommends a probabilistic diagnostic and prognostic framework [54].
10. Can external factors make a mild case severe?
"Severe" has four meanings here.
A longer or worse clinical course. Continued play, delayed clinician access and sleep disturbance are associated with longer recovery [1, 52]. These are observational findings. Athletes with worse injuries may be the ones who keep playing or present late, so a causal effect is not established.
A second injury inside the window. See section 4.4 [8, 22].
A catastrophic outcome. Second impact syndrome (malignant swelling after a second blow) is discussed controversially [25], and the genetic observation in section 5 is a family report [23].
A missed intracranial complication. The 2015 German adult guideline treats coagulopathy and anticoagulation as CT indications, and a patient who was initially lucid and then deteriorates as a sign of worsening intracranial injury. It notes that older patients may develop chronic subdural haematoma after relatively mild trauma [55].
11. Repetitive head impacts and long-term outcome
Three terms need separating:
A concussion is a clinical event.
Repetitive head impacts is an exposure, including impacts that never cause a diagnosed concussion.
CTE neuropathological change is a post-mortem finding. Traumatic encephalopathy syndrome (TES) is a proposed clinical syndrome whose research criteria require substantial exposure to repetitive head impacts, core cognitive or behavioural features, a progressive course, and no full explanation by other conditions [56]. CTE-NC is not a clinical diagnosis [1].
Brain-donation studies. In 202 deceased American football players, 177 (87%) had CTE, including 110 of 111 former NFL players, in a convenience sample [57]. In 152 deceased contact-sport athletes younger than 30, 63 (41.4%) had CTE, 95.2% mild (stages I–II). The donors were highly symptomatic regardless of CTE status, and there were no differences in cause of death or symptoms by CTE status [58]. Amsterdam states that such studies are not cohort studies and cannot quantify risk, that CTE-NC is uncommon in community samples with strict criteria, and that prevalence, whether it causes specific problems, and whether it is inevitably progressive are unknown [1].
Cohort and case-control studies. A systematic review of 28 studies found no increased risk of mental health or neurological disease in former amateur athletes. Some studies of former professionals suggest higher risk of dementia and ALS, but most had high risk of bias and little control for confounding [59]. In 7676 former Scottish professional footballers and 23,028 matched controls, neurodegenerative disease was the primary cause of death in 1.7% of former players versus 0.5% of controls (subhazard ratio 3.45; 95% CI 2.11–5.62). With neurodegenerative disease as a primary or contributory cause, the hazard ratio was highest for Alzheimer's disease (5.07; 2.92–8.82) and lowest for Parkinson's disease (2.15; 1.17–3.96). All-cause mortality was lower in players up to age 70 and higher after, with lower ischaemic heart disease (hazard ratio 0.80) and lung cancer mortality (0.53). Mortality did not differ significantly between goalkeepers and outfield players. The authors call for prospective confirmation [60].
Heading. A systematic review of 30 studies (1691 players) concluded that methodological shortcomings limit the evidence for persistent effects of football on brain structure and function, including poor assessment of heading frequency and inadequate control of statistical error and controls [61]. An imaging study associated heading with white matter microstructural and cognitive abnormalities [53].
Mechanism. Giza and Hovda note that no unifying biological mechanism yet explains how an acute, usually recoverable pathophysiology could become a chronic progressive disorder. Hypotheses include repeated injury before recovery, impaired protein clearance with tau accumulation, and persistent inflammation or axonal degeneration [8, 24, 27].
12. Practice and prevention (consensus)
Amsterdam 2022 recommends removal from play when concussion is suspected and serial re-evaluation. It advises relative rest for the first 24–48 hours, then gradual return to light activity. It also states that strict rest is not beneficial, and that sub-symptom-threshold aerobic exercise within 2–10 days reduces persisting symptoms. Cervicovestibular rehabilitation is advised if dizziness, neck pain or headache persists beyond 10 days. Prevention evidence it reports:
A ban on body-checking in children's ice hockey: 58% reduction in game concussion rates.
Limits on contact practice in American football: 64% reduction in practice-related concussions.
Mouthguards in ice hockey: 28% reduction.
Neuromuscular warm-ups in rugby are recommended.
Evidence in children aged 5–12, female athletes and para athletes is limited [1].
13. What remains unknown
The causal mix behind persisting symptoms in an individual athlete.
Whether the energy crisis and its recovery can be measured in individual humans.
What produces brief loss of consciousness, and why the critical rotation plane differs between species.
Whether mechanoporation, mechanosensitive channels or cytoskeletal coupling dominates, and at which strains.
Whether subconcussive exposure alone produces lasting harm.
Evidence notes
This is a narrative review written for a general scientific readership. Where the data are weak, the text says so. The following limits apply to how the sources were checked.
Much of the cellular sequence rests on animal and in vitro work. Statements about humans are limited to what the cited human studies measured.
Two PET findings in section 4.3 come from unreviewed meeting abstracts. Patient numbers for Bergsneider 2000 come from a news summary and were not checked against the paper.
The Elbin 2016 figures were taken from abstracts and slides, not the paper's tables.
The criticism of the centripetal theory in section 8 comes from a teaching website (a secondary source). Shaw 2002 is cited through Bayly 2005 and a chapter abstract.
The Vespa 2005 and Moore 2000 volume, page and author details and the Katayama 1986 authors were not verified against the bibliographic records.
Section 9 uses cohorts that are not sporting, and symptom outcomes are non-specific. The comparisons are illustrative and not estimates for athletes.
Statements marked “assessment”, “inference” or “synthesis” are interpretations by the authors of this article and not findings of the cited studies.
Note on interests. The author originated the ACT concept and is a senior advisor to Northern Sports Insight and Intelligence Oy, with no compensation or ownership.