Concussions in Combat Athletes: What the Evidence Says About Risk, Recognition, and Recovery

Key Takeaways: Concussions in Combat Athletes

• The most practical way to reduce harm is to recognize concussion early, avoid unnecessary head impacts, use structured return-to-competition protocols, and take persistent symptoms seriously.
• A concussion does not require loss of consciousness. Many cases may go unrecognized when athletes, coaches, or officials look only for a visible knockout.
• Boxing appears to carry a higher concussion risk than several other combat disciplines, although reported rates vary widely depending on how concussion is defined and detected.
• A normal acute CT scan can rule out major bleeding, but it does not exclude concussion or every form of traumatic brain injury. Persistent symptoms may still warrant neurological follow-up and, in selected cases, MRI.
• Repeated concussive and subconcussive head impacts are associated with measurable neurological changes and may increase long-term neurodegenerative risk, although the exact individual risk remains uncertain.
• Brain-bank studies of CTE are affected by strong selection bias, so their prevalence figures cannot be applied directly to all fighters. Even so, the broader evidence supports a relationship between cumulative head-impact exposure and later neurological disease.
• Previous knockout losses may indicate increasing vulnerability to future concussion, but no fixed number of knockouts has been validated as a universal retirement threshold.
• Headgear may reduce some acute head impacts and improve short-term cognitive outcomes, but it should not be treated as complete protection against concussion or long-term neurodegenerative disease.
• Fixed suspension periods are only minimum safeguards. No calendar-based waiting period can replace an individualized medical assessment before returning to sparring or competition.
• Both the winner and the loser of a bout may sustain a concussion and should be assessed when symptoms or concerning signs are present.
• Combat sports are not uniformly dangerous or uniformly safe. Recreational training, controlled sparring, grappling-based disciplines, and high-level competition can involve very different levels of head-impact exposure.
• Neurological risk is only one part of the athlete’s overall health profile. Combat-sport athletes may still benefit from high levels of physical activity, cardiovascular fitness, strength, and long-term exercise participation.
• Better overall fitness or lower all-cause mortality does not rule out an elevated risk from one specific cause of death, such as neurodegenerative disease. A population can be healthier than average in many respects while still carrying a meaningful sport-specific neurological risk.

Introduction: Concussions in Combat Athletes

Combat sports differ from many other contact sports because deliberate blows and, in some disciplines, knocking out the opponent are accepted parts of competition. A systematic review covering MMA, boxing, taekwondo, karate, and kickboxing reported that, among the head injuries it classified, 72.8% were concussions and 27.2% were categorized as traumatic brain injuries [1].

During my own years in combat sports, my coach used to say that after three knockouts, a fighter should seriously consider ending their career. Similar beliefs circulate widely in gyms: that athletes who have been knocked out too many times begin to lose speed, sharpness, or reaction time, and that some veteran fighters appear noticeably slower after years of repeated head impacts. These stories may reflect genuine concern, but they are still anecdotes rather than scientific evidence.

The public discussion has also been shaped by high-profile examples such as Muhammad Ali, who later developed Parkinson’s disease. His illness contributed to the broader concern that repeated blows to the head may increase the risk of long-term neurological and neurodegenerative disease. But the central question is not whether head impacts can be harmful. It is how often serious brain injuries actually occur in combat sports, what repeated knockouts and subconcussive impacts may mean over time, and how confident we can be about those risks based on the available evidence.

From a clinical perspective, it is also important not to reduce the entire health profile of a combat athlete to neurological risk alone. A fighter may be exposed to repeated head impacts while also maintaining far higher levels of physical activity, cardiovascular fitness, strength, and metabolic health than a sedentary person. The more useful question is therefore not whether combat sports are simply healthy or unhealthy, but how their neurological risks, physical benefits, and other sport-specific harms interact over the course of an athlete’s life. This article examines what the evidence actually says about concussion risk, recognition, recovery, and long-term health in combat athletes — and where uncertainty still remains.

How Common Are Concussions in Combat Athletes?

Reported concussion rates in combat athletes vary widely by sport and by how “concussion” is defined and detected. A meta-analysis of head trauma in boxing found concussion rates per 100 participants ranging from 14.0 to 41.5 across boxing studies, with rates below 3 per 100 participants in taekwondo and karate [3]. A separate direct comparison within the same analysis found professional boxing bouts carrying a higher concussion rate than amateur bouts (2.62 vs. 0.33 per 100 bouts, respectively) [3]. The same analysis found that boxers had a significantly elevated risk of concussion compared with combat athletes in other disciplines [3].

Video-based research tells a similar story from a different angle. A study reviewing 60 professional boxing and MMA matches found that physician reviewers judged that a concussion had likely occurred in 47 of them, with concussions occurring at a mean rate of 0.061 per minute of fight time (0.047 for boxers, 0.085 for MMA) [4]. Notably, the fighter who sustained the first concussion in a bout went on to lose 98% of the time, and the same reviewers judged that 40% of the fights they assessed should have been stopped earlier than they were [4].

In MMA specifically, injury surveillance places concussions at somewhere between 3.8% and 20.4% of competition injuries, with the head and face accounting for two-thirds to nearly four-fifths of all injuries sustained [9]. For comparison, one large database of amateur boxers followed over five years found an incidence of roughly 0.53 concussions per 1,000 hours of competing and training, while a 16-year surveillance study of professional kickboxing reported 19.2 concussions per 1,000 fight participations [8].

In clinical practice, one of the main challenges is that concussion is often recognized in combat-sport settings only when a fighter is visibly knocked out. Loss of consciousness, however, is not required for a concussion to occur. Many patients with concussion remain awake throughout the event, and the diagnosis depends on the broader clinical picture rather than on whether they collapsed or blacked out.

This means that some combat athletes may have sustained concussions without recognizing them as such at the time. In my clinical experience, this possibility often comes up when reviewing a history of repeated head impacts, especially when an athlete describes episodes of confusion, slowed thinking, headache, balance disturbance, or memory gaps that were never formally evaluated. The absence of a knockout does not by itself rule out concussion.

What Happens in the Brain

Among combat athletes, a concussion is currently defined as a form of mild traumatic brain injury caused by a direct blow to the head or an impulsive force transmitted to the head from elsewhere on the body, producing a rapid but usually short-lived disturbance of neurological function rather than a structural injury visible on standard imaging [8]. What makes combat sports distinct is not just the concussions that are formally diagnosed, but the burden of subconcussive impacts — blows that don’t produce recognized clinical symptoms but may still produce measurable biomarker changes consistent with neuronal or glial stress or injury.

Research in amateur boxers has shown this directly. Cerebrospinal fluid levels of brain-injury biomarkers — neurofilament light protein, GFAP, total-tau, and S-100B — were elevated in more than 80% of boxers within one to six days after a bout, even in fighters who showed no loss of consciousness or concussion symptoms [5]. After a two-week rest period, most of these biomarkers normalized, but roughly 20% of boxers still had significantly elevated neurofilament light and GFAP levels compared with non-boxing controls [5]. In a longitudinal cohort of active and retired professional fighters, baseline plasma GFAP concentrations were higher in retired boxers than in MMA fighters, while active boxers had higher plasma neurofilament light concentrations than MMA fighters [6]. Separately, a controlled study of elite male boxers found that a single match significantly raised inflammatory markers including interleukin-6, along with biochemical markers that can reflect exercise-related tissue stress, including AST and ALT, reflecting the broader physiological toll of a fight beyond the brain itself; creatinine did not change significantly in that same study [15].

Repeated head trauma is a central exposure of concern in chronic traumatic encephalopathy, although the exact prevalence and progression of CTE remain uncertain and require separate neuropathological sources. In the heterogeneous boxer populations summarized by one systematic review and meta-analysis, dementia or amnesia was reported in 46 of 71 boxers (61.79%), cognitive disorders in 36 of 70 (51.43%), and abnormal CT or EEG findings in 57 of 109 (52.29%) [3]. The same review reported cavum septum pellucidum in 147 of 631 amateur and professional boxers (23.30%) and some form of brain atrophy in 125 of 411 (30.41%) [3]. Because these figures came from different, selected study populations with different denominators, they should not be interpreted as population-wide prevalence estimates for boxers in general.

After a knockout — or even without loss of consciousness when the clinical features are consistent with concussion — the immediate concern is to rule out a more serious intracranial injury, particularly bleeding. In practice, the threshold for brain imaging is often low when a patient has lost consciousness, developed focal neurological symptoms, shown prolonged confusion, or otherwise presents with features that raise concern for structural injury. A normal acute CT scan is reassuring in terms of major bleeding, but it does not exclude every form of traumatic brain injury.

If symptoms persist despite an initially normal CT scan, the patient may be referred to a neurologist or a dedicated brain-injury clinic for further assessment. Depending on the clinical picture, brain MRI may then be considered because it can reveal more subtle abnormalities that are not visible on CT, including signs of traumatic axonal injury. However, such findings are not present in every patient, and a normal MRI does not rule out concussion.

From a clinical perspective, these are often among the most difficult cases to manage. A patient may have no acute bleeding, no obvious structural lesion, and still continue to experience headaches, dizziness, cognitive slowing, fatigue, memory difficulties, balance problems, or other persistent post-concussive symptoms. These symptoms can be functionally significant even when conventional imaging remains normal, which is why ongoing neurological evaluation may become relevant.

The important point is that the absence of an acute bleed does not mean that the brain has necessarily escaped injury. At the same time, it would be too strong to describe every concussion as permanently disabling or completely irreversible. Many patients recover well, while others are left with residual symptoms that can persist for a prolonged period. Concussion should therefore be understood as a genuine brain injury with a variable course rather than as either a harmless event or an inevitably permanent condition.

Lessons from American Football: Correcting for Selection Bias in CTE Research

Because CTE research specific to concussions in combat athletes is still relatively sparse, it is worth looking at what a much larger body of work in American football has established — with the caveat that football and combat sports differ in impact mechanics and cannot be assumed to carry identical risk. The largest case series to date, published in JAMA, examined the donated brains of 202 former football players and found neuropathological evidence of CTE in 87% of the sample overall, rising to 99% (110 of 111) among those who had played in the NFL [10].

That headline figure carries an important caveat that the study’s own authors raised: brain donation programs draw disproportionately from families who already suspected a link between their loved one’s symptoms and head trauma, which inflates the apparent prevalence relative to the true population of former players [10]. To correct for this, a separate analysis published in Neurology used the total number of deceased NFL players eligible for brain donation over the same period as a denominator, and estimated a minimum prevalence at death of 9.6% among professional football players — explicitly a lower bound rather than a central prevalence estimate, since the true figure could plausibly be considerably higher depending on the prevalence among the players whose brains were never donated [12]. A later analysis using the same brain bank cohort, statistically adjusting for the selection process itself, found that college-level players had 2.38 times and professional players 2.47 times the odds of a CTE diagnosis compared with players whose football exposure was limited to high school [13]. This dose-response pattern, persisting after statistical adjustment for the modeled selection process, strengthens — but does not by itself prove — the case for a causal relationship between cumulative head-impact exposure and CTE risk.

The methodological lesson for combat sports is direct: the boxer autopsy series cited earlier in this article rely on similarly donated or clinically selected brains, and are therefore also likely to overstate population-wide CTE prevalence in the same way the initial NFL findings did before correction. Selection-bias-corrected prevalence estimates specific to combat sports do not yet exist at the scale seen in football — a research gap worth naming rather than glossing over.

Muhammad Ali is probably the best-known example of how a single famous case can shape public perception. He was an exceptional boxer with extraordinary career exposure, and he later developed Parkinson’s disease. Whether the disease was caused by boxing, accelerated by repeated head trauma, or would have developed regardless is something we cannot determine from his case alone.

In my view, Ali’s story became so influential precisely because it was so memorable. When a globally recognized fighter later develops a neurological disease, people naturally connect the two events. That association may be understandable and may even point toward a real biological concern, but one individual case cannot establish causation. The scientific question has to be answered from broader patterns across many athletes rather than from the emotional power of one famous example.

Performance and Cognitive Impact

Beyond long-term neurodegenerative risk, concussion has measurable acute effects on how fighters think and perform. A systematic review focused specifically on executive function in combat athletes — pooling boxing and MMA cohorts totaling 1,130 fighters — found that roughly 82% of studies reported impairment of executive function following concussion in combat athletes, with memory affected most often (36.5% of reported impacts) followed by inhibitory control (27.5%) [2].

This cognitive vulnerability compounds with repeated exposure. Fighters who had experienced a concussion in their most recent match had a significantly higher average number of previous knockout losses (2.9) than fighters who had not been concussed [14] — a knockout-loss history rather than a formally diagnosed concussion history, but one broadly consistent with the well-documented phenomenon that a prior concussion increases susceptibility to subsequent ones. Headgear appears to offer partial, acute protection: in a controlled study of amateur Olympic boxers, fighters wearing headgear absorbed significantly fewer punches to the head (38.5 vs. 51.1) and performed better on post-bout automatic and controlled executive function testing than the same fighters sparring without headgear [7]. Whether this acute protective effect meaningfully changes long-term outcomes such as CTE risk remains an open, actively debated question in the literature, particularly given recent headgear bans in some Olympic boxing competitions [3].

Interestingly, these findings are not far removed from something my own coach used to say: after three knockouts, it may be time to consider ending a fighting career. That rule was never presented to me as a scientific threshold, but rather as a practical judgment based on years of observing fighters.

The study above does not prove that three knockout losses should be treated as a universal retirement cutoff. However, the fact that recently concussed fighters had an average of 2.9 previous knockout losses suggests that this kind of empirical gym wisdom may not be entirely arbitrary. In my view, the important point is not that “three knockouts” is a validated medical rule, but that repeated knockout exposure may signal increasing neurological vulnerability and should prompt a more careful individual assessment of whether continued competition remains reasonable.

Recognizing and Assessing Concussions in Combat Athletes

Assessing concussions in combat athletes presents a genuine diagnostic challenge that most contact sports don’t face: distinguishing a concussion from the normal disorientation of absorbing a hard, legal strike, in real time, under a ten-count. The Association of Ringside Physicians’ consensus statement lists the core signs to watch for — somatic symptoms such as headache and nausea, cognitive symptoms such as a foggy feeling or slowed reaction time, balance impairment, behavioral changes, and any loss of consciousness — and notes that when a knockout results from strikes to the head, a concussion should be assumed present [8]. A technical knockout, by contrast, can be awarded for reasons unrelated to concussion, such as a cut or a corner stoppage, so not every stopped bout implies brain injury, and not every concussion produces a stoppage at all [8].

Postbout evaluation should apply to both the winner and the loser of a bout, since the winning fighter can also sustain a concussion during the contest, and signs can be delayed rather than immediately obvious ringside [8]. The Association of Ringside Physicians recommends a first evaluation immediately postbout and a second, more thorough evaluation later in a quiet, distraction-free setting such as the locker room [8].

In Finland, there does not appear to be a single, nationally standardized post-bout concussion pathway that automatically directs every combat athlete for medical assessment after a match. In practice, the approach may depend on the individual sport federation, and the resulting rules can vary considerably in both detail and medical basis.

From my own experience in Finnish combat sports, the guidance has not always been equally developed across disciplines. In Taekwon-Do (ITF), for example, the practical rule was essentially that an athlete who had been knocked out needed to see a doctor, without a broader structured return-to-competition protocol. Finnish kickboxing federation, by contrast, had a more formal policy that required brain imaging after a knockout, although the rule did not clearly distinguish whether CT or MRI was required.

These kinds of federation-level rules may still improve safety by creating a minimum threshold for follow-up. However, they should not automatically be interpreted as evidence-based medical protocols unless they are explicitly grounded in specialist consensus or established concussion guidelines. In my view, the wider problem is not that federations have no rules at all, but that the quality, rationale, and clinical consistency of those rules may differ substantially between sports.

Managing Concussions in Combat Athletes: Suspension and Return-to-Competition Protocols

Unlike most contact sports, commissions overseeing combat athletes typically use fixed minimum suspension periods for concussions rather than a purely symptom-based return-to-play timeline, in recognition of the fact that head contact will resume as soon as a fighter is cleared to compete again. Under USA Boxing’s framework, a TKO or KO without loss of consciousness carries a minimum 30-day suspension, a KO with loss of consciousness under one minute carries 90 days, and a KO with loss of consciousness over one minute carries 180 days. Suspensions escalate further with repeat knockouts: a second TKO/KO within 90 days of the first raises the minimum to 90–360 days depending on severity, and a third TKO/KO within 365 days raises it to 12 months, or 18 months for a third knockout involving loss of consciousness [8]. The Association of Ringside Physicians’ own guidelines converge on similar minimums: 30 days after a head-strike TKO, 60 days after a KO without loss of consciousness, and 90 days after a KO with loss of consciousness, with sparring restricted for the same period [8].

Critically, these fixed suspension windows are treated as a floor, not a substitute for clinical clearance. The Association of Ringside Physicians recommends that suspension continue until a specialist trained in concussion management — a neurologist, neurosurgeon, or primary care sports medicine physician — clears the athlete, and that all fighters undergo a validated neuropsychological baseline test, repeated annually; vestibular/ocular and balance baseline testing is also recommended as an adjunct where feasible, though the annual-repeat recommendation is specified for the neuropsychological test [8]. The proposed return-to-fight pathway itself unfolds over three supervised phases: a return to general fitness, a return to non-contact fight-specific drills, and finally a graded return to sparring and contact. Non-contact training, beginning with light aerobic activity, may start one week after the concussion or head-strike TKO/KO provided symptoms are improving, and the fighter should progress no more than one stage per day, waiting a minimum of 24 hours and confirming symptoms have not recurred or worsened before advancing to the next step [8].

In Finnish amateur kickboxing, a knockout currently results in a one-month suspension from competition. In my view, this is at least a meaningful minimum safeguard, even if it may be somewhat less conservative than the graduated suspension periods used in frameworks such as USA Boxing, where the duration can increase depending on factors such as loss of consciousness and repeated knockouts.

A fixed one-month suspension does not, by itself, guarantee that an athlete has fully recovered or is ready to return to head contact. Nevertheless, having a defined mandatory restriction is preferable to leaving the decision entirely to the athlete or coach. The suspension provides a minimum recovery window, although individual symptoms and clinical assessment may warrant a longer period before returning to competition.

In any case, I would recommend that every suspected concussion be assessed clinically by a physician. No fixed waiting period can replace a proper medical evaluation, because recovery cannot be judged reliably from the calendar alone.

Are Combat Athletes Healthier Overall? What the Longevity Data Shows

It is worth stepping back from concussion-specific risk to ask a broader question: do combat athletes, on balance, live longer or shorter lives than the general population? A meta-analysis of 165,033 former elite athletes found lower overall mortality than in the general population, as well as lower cardiovascular and cancer mortality among male athletes [11].

These benefits were not distributed evenly across sport categories. Endurance and team-sport athletes had significantly lower cardiovascular mortality, whereas the broad power-sport category — which combined boxing and wrestling with weightlifting and throwing events — showed no significant all-cause mortality advantage (pooled SMR 1.04) and no significant cardiovascular mortality advantage (SMR 1.10) [11]. The all-cause analysis included 2,826 power athletes, while the cardiovascular analysis included only 1,885, making both estimates relatively imprecise and unsuitable as combat-sport-specific mortality figures.

The meta-analysis did not establish why the apparent survival advantage was absent in power athletes. Exposures such as repetitive head trauma and rapid weight-cutting are possible explanations in some combat athletes, but this is a clinical hypothesis rather than a conclusion directly supported by these mortality data.

American football offers a useful and cautionary illustration of why overall mortality and cause-specific mortality can tell very different stories. A 2026 population-based retrospective cohort study included 19,824 current and former NFL players who debuted between 1960 and 2019, played at least one regular-season or postseason game, and were followed for mortality through 2023 [16]. NFL players had lower all-cause mortality than the general population (SMR 0.70), alongside lower mortality from several major causes.

Neurodegenerative mortality, however, was markedly elevated (SMR 3.94), including all-cause dementia (SMR 3.80), Parkinson’s disease (SMR 3.88), and amyotrophic lateral sclerosis (SMR 4.55) [16]. Among players who died before age 60, neurodegenerative mortality was 12.43 times the expected rate. Players with careers lasting five or more seasons also had significantly greater neurodegenerative mortality than those who played for one to four seasons, supporting an exposure-response pattern [16].

In other words, the same population can have lower mortality from most major causes and still carry a markedly elevated risk of neurodegenerative death. This is a large observational cohort, not a randomized comparison, so it cannot rule out residual confounding or establish individual-level causation — but the study covered the full enumerated population of qualifying players rather than a self-selected or donated-brain sample, drew cause of death from National Death Index records, and found a dose-response relationship with career length; the elevated neurodegenerative mortality also remained roughly threefold in sensitivity analysis. Taken together, the pattern is consistent with a relationship between cumulative repetitive head-impact exposure and later neurodegenerative disease. It also underscores a broader methodological point for interpreting the smaller and more heterogeneous power-sport data above: a single all-cause mortality figure does not, by itself, reveal what is happening to any one specific cause of death within that population.

Overall, I do not think concern about head trauma or neurodegenerative disease should automatically prevent someone from taking up a combat sport. Combat sports can be practised in many different ways, and the level of exposure varies substantially depending on the discipline, the training environment, and whether the athlete competes.

From a clinical perspective, it is important to separate recreational training from repeated competitive head contact. Every combat sport carries some risk of injury, but sensible coaching, appropriate protective equipment, controlled sparring, and avoiding unnecessary head impacts may reduce that risk. For those who are uncomfortable with striking, grappling-based disciplines may offer many of the physical and psychological benefits of combat sports without the same degree of repeated direct head impact.

This is a question I am sometimes asked by the parents of younger athletes: is it safe to allow a child to participate? In my view, recreational participation is not automatically unreasonable simply because the sport belongs to the combat-sport category. Competition is a separate decision, because exposure to hard contact and head impacts may increase. That decision needs to be individualized, and I cannot make it on behalf of another family. What I can say is that the neurological risks deserve to be taken seriously, but they are not, by themselves, an absolute reason to avoid all combat-sport training.

Conclusion: Concussions in Combat Athletes

Concussion is an inherent risk in striking-based combat sports, but the risk is neither uniform nor fully captured by knockout statistics alone. Many concussions occur without loss of consciousness, some injuries leave no visible abnormality on acute imaging, and repeated exposure may have consequences that only become apparent over time. At the same time, the available evidence does not justify treating every combat athlete as inevitably destined for long-term neurological disease.

In my view, the most important practical distinction is between accepting that risk exists and becoming careless about how it is managed. A knockout, suspected concussion, or persistent neurological symptom warrants proper clinical assessment, and no fixed suspension period can replace an individualized medical evaluation. Return-to-training decisions should take symptoms, recovery, previous injuries, and the nature of the athlete’s exposure into account rather than relying on the calendar alone.

The long-term evidence remains imperfect, particularly because much of the CTE literature is affected by selection bias and because combat-sport-specific mortality data are still limited. Even so, the broader pattern is difficult to dismiss: cumulative exposure to repeated head impacts appears to matter, and all-cause mortality statistics can conceal important cause-specific neurological risks.

At the same time, neurological risk is only one part of the athlete’s overall health profile. Combat-sport training can involve high levels of physical activity, cardiovascular fitness, strength, coordination, and long-term engagement in exercise. The available mortality research on elite athletes more broadly suggests that athletic populations may have lower mortality from several major causes, even when a specific sport-related risk remains elevated. In other words, a combat athlete may be healthier than the general population in many respects while still carrying a meaningful neurological risk from repeated head impacts.

That does not mean combat sports should be avoided altogether. Recreational training, controlled sparring, grappling-based disciplines, appropriate protection, and sensible coaching may offer a very different risk profile from repeated high-level competitive head contact.

Ultimately, the goal should not be to portray combat sports as either harmless or inherently unacceptable. The more useful approach is to understand where the risk comes from, recognize concussion early, reduce unnecessary head-impact exposure, and make return-to-competition decisions with the same seriousness that would be applied to any other genuine brain injury.

References

1. https://www.researchgate.net/publication/388054766_Determinants_and_key_indicators_of_concussion_in_combat_sport_athletes_a_systematic_review
2. https://www.tandfonline.com/doi/full/10.1080/02640414.2024.2433902
3. https://doi.org/10.1097/JSM.0000000000001195
4. https://www.tandfonline.com/doi/abs/10.1080/00913847.2020.1856631
5. https://pmc.ncbi.nlm.nih.gov/articles/PMC3829937/
6. https://link.springer.com/article/10.1186/s13195-023-01310-w
7. https://revistaretos.org/index.php/retos/article/download/106894/79759/435567
8. https://ringsidearp.org/wp-content/uploads/2020/01/Concussion-Combat-Sports-2018.pdf
9. https://pmc.ncbi.nlm.nih.gov/articles/PMC5315252/
10. https://pubmed.ncbi.nlm.nih.gov/28742910/
11. https://pmc.ncbi.nlm.nih.gov/articles/PMC7846545/
12. https://www.neurology.org/doi/10.1212/WNL.0000000000006699
13. https://pmc.ncbi.nlm.nih.gov/articles/PMC9989358/
14. https://www.jstage.jst.go.jp/article/jpfsm/11/4/11_255/_article/-char/ja/
15. https://pmc.ncbi.nlm.nih.gov/articles/PMC6630693/
16. https://www.thelancet.com/journals/eclinm/article/PIIS2589-5370(26)00304-4/fulltext