Concussion and the Nervous System

A normal scan is part of the diagnosis

The 2023 diagnostic criteria for mild traumatic brain injury state outright that normal imaging does not exclude the diagnosis. CT and MRI photograph what the brain is made of, and a concussion changes what the brain is doing.

A player goes down, gets helped off, and says she is fine. By evening she is in an emergency department with a headache and a strange slowness in her thinking. The scan is done and the scan is normal. Someone tells her that is good news, and she goes home confused about why she still cannot read a page.

The normal scan is good news. It also does not mean nothing happened.

Knowing why the clean image belongs in the definition requires knowing what those machines do. A CT scanner is a fast X-ray map of density, which is why it finds blood and bone so well. An MRI scanner maps water in tissue. Both photograph what the brain is made of. Neither photographs what the brain is doing.

In 2023 an American rehabilitation medicine task force spent three rounds of structured voting rebuilding the diagnostic criteria for mild traumatic brain injury. The goal was a definition that works the same way in a stadium, a car crash and a war zone. Seventeen working members drafted them.

Thirty-two outside clinician scientists reviewed them, along with public comment from sixty-eight individuals and twenty-three organizations. Ninety-one percent of the panel signed off on the final criteria, and ninety-four percent agreed that concussion and mild traumatic brain injury name the same thing. A normal CT or MRI does not exclude the diagnosis.

The sport side arrived at the same place. Jon Patricios, Kathryn Schneider and their co-authors published the Amsterdam consensus statement on concussion in sport, built on ten commissioned systematic reviews gathered across three and a half years. The imaging is normal because the imaging is looking at the wrong kind of thing.

Diffusion imaging finds a difference the clinical scans missed

Push the instruments harder and something does appear. An imaging group at the Mind Research Network in New Mexico asked whether anything measurable separated people whose clinical scans had been read as normal. Twenty-two patients were studied about twelve days after a mild head injury, against matched healthy controls, using diffusion tensor imaging. That method photographs no tissue at all.

It tracks how water molecules move along nerve fibers, and the freedom of that movement reports on how the fibers are arranged. Clinical imaging: no difference. Neuropsychological testing: no difference. Water diffusion in the corpus callosum and several left-sided tracts: different. The corpus callosum is the bridge of fibers joining the two halves of the brain. That measure separated patients from controls better than any standard test in the study.

Twenty-two people is a small study, and the direction of the effect is the opposite of the classic chronic white-matter pattern. The measure is fractional anisotropy, which is how strongly the water runs along a fiber instead of across it, and here it went up. It has not replicated cleanly.

Clinical imaging read as normal. Water diffusion in the corpus callosum separated patients from controls better than any standard test.

Blood tests find the bleeds, and concussion is defined by their absence

Blood tests have not closed the gap either. The emergency physician Jeffrey Bazarian led an international group that tested two proteins released by injured brain tissue, GFAP and UCH-L1, in nearly two thousand adults with suspected head injury.

The combined test picks up bleeding and bruising visible on CT with high sensitivity, which is exactly what it was designed for. Specificity is low, so most positive results are false alarms. And the people whose CT is clean are the people the word concussion actually describes. No blood test finds them.

The question that matters is the one the scanner cannot answer. If nothing is broken, what is wrong?

The head injuries that are not concussions

A minority of head injuries involve bleeding or swelling inside the skull, and those need a surgeon rather than an explanation. The warning signs are specific, and a 42,412-child study defined who can safely skip the scanner.

Get emergency care immediately if any of the following appear after a blow to the head. A headache that keeps getting worse. Repeated vomiting. A seizure. Pupils of unequal size. Increasing drowsiness or confusion, or a person who is hard to wake. Weakness, numbness or clumsiness in an arm or a leg. Slurred speech. Any of these can mean bleeding or swelling inside the skull.

Two rules follow, and neither is negotiable. Anyone suspected of a concussion comes out of the game and is assessed by a physician. Nobody returns to play the same day. The evidence behind the second rule is stronger than most people assume: young athletes kept in play after a concussion took twice as long to recover.

The warning signs have a 42,412-child evidence base

Those warning signs are not folklore. An American pediatric emergency research network enrolled 42,412 children who arrived within twenty-four hours of head trauma and were alert or nearly so. The question was who could safely skip a CT scan and its dose of radiation. Clinically important brain injury occurred in 0.9 percent of them. Sixty children needed neurosurgery.

For the youngest, six simple features together identified a group in which no important injury occurred. Normal mental status. No scalp bruise except at the forehead. No significant loss of consciousness. An injury mechanism that was not severe. No palpable skull fracture. And a parent reporting the child was behaving normally. The dangerous signs are specific, and they are known.

That rule governs the first day. Deterioration after it is a separate clinical question that needs a fresh assessment, not a recollection of what the first scan showed.

A model of regulation does not replace a surgeon. Structural, vascular and surgical causes are real, and they are treated as such. Every injury has a tonal expression, and the regulatory account offered here explains the course of a concussion rather than replacing the pathology of a bleed. Moderate and severe traumatic brain injury, where the imaging is abnormal and tissue really is destroyed, is a different problem with a different literature, and it belongs to the brain injury page.

What remains after the dangerous minority is removed is the large majority: people with a genuinely injured brain, a genuinely normal scan, and a set of symptoms that nobody has yet explained to them.

Firing is cheap, and the reset is what costs the neuron

A neuron spends almost nothing to send a signal and pays heavily to reset afterward, because the sodium-potassium pump that restores its gradients burns ATP with every stroke. A concussion is an injury to that budget.

A neuron is a living wire. It holds a voltage across its outer membrane by keeping sodium ions on the outside and potassium ions on the inside, a separation it maintains every second of your life. Sending a signal is the cheap part. Channels open, sodium rushes in, the voltage collapses in a wave that runs down the wire, and the message arrives.

The reset is what costs. A molecular machine sitting in the membrane, the sodium-potassium pump, drags the ions back where they belong against their own tendency to spread out. Every stroke of that pump burns ATP, the cell's energy currency. ATP is made from glucose and oxygen, and glucose and oxygen arrive by blood.

Two consequences follow, and the physiology of concussion hangs on them.

The first is that electrical activity and fuel are one account. A brain that has fired hard has spent, and it has to buy back its gradients before it can fire well again.

The second is that a working brain needs fuel exactly where and when it is working. So blood flow has to follow neural demand from moment to moment and region to region. It does, and the matching has a name: neurovascular coupling. Functional MRI exists only because of it, since the scanner measures blood and reads it as a proxy for activity. That proxy holds because in a healthy brain the two move together.

The central integrative state is tone at the scale of one cell

Everything just described, the charge across a membrane and the running cost of holding it, already has a name at the scale of one cell. The central integrative state is the sum of every excitatory and inhibitory influence arriving at a neuron, and it sets how that neuron answers the next signal.

That much is established. What this model adds is an identification. The central integrative state is tone read at the smallest scale. Blood flow matching, autoregulation, and heart rate control are the same state read across a whole person.

Why a rotation injures without tearing anything

Now the injury. The brain is soft, it floats in fluid, and the skull it floats in is hard. A rapid rotation of the head shears and stretches tissue that has nowhere to go. Membranes deform, channels open that should not, and the careful separation of ions is lost across a great many cells at once. Nothing has to tear for this to happen, which is why the scan can be clean.

This library has a word for what comes next. Load is what holding a state costs. Concussion is the fastest and cleanest example of load anywhere in the body, because demand rises steeply in the same minutes that supply begins to fall.

A glutamate flood, a glucose burst, then an energy debt

Three experiments across seven years measured the concussion cascade: potassium outside the cells rising to four to six times baseline, glucose burning as much as ninety percent above normal, then metabolism depressed for days. One laboratory asked the next question each time.

1990: the potassium surge and the glutamate release

In 1990 a head-injury laboratory whose members included the brain-injury researcher David Hovda set out to measure the first seconds. Their method was fluid percussion, a controlled pulse of fluid delivered to the exposed brain of an anesthetized rat, with a fine probe sampling the fluid in the spaces between cells.

After the harder impacts, potassium outside the cells rose to between four and six times its baseline and stayed up. A drug that silences ordinary nerve firing blunted the small early rise and left the large one untouched, so something other than firing was driving it. Sampling alongside, they found glutamate, the brain's main excitatory messenger, pouring out indiscriminately. Blocking glutamate receptors blunted the potassium surge.

A concussion begins as a mass release of transmitter that no signal ordered.

1991: the burst of burning, then the crash

The next year the same laboratory asked what the release costs. Ninety-eight rats received the same controlled fluid pulse, and glucose consumption was mapped at eight time points from immediately after injury out to ten days.

Straight after injury the cortex and the hippocampus were burning sugar far above normal, the hippocampus by as much as ninety percent. The hippocampus is the seahorse-shaped structure that lays down new memories. Then, from roughly six hours onward, those same regions dropped into a depressed metabolic state that lasted up to five days.

Demand spikes. Then supply falls. The injury lives in the gap between them, which is why the scan stays clean.

1997: the human demonstration, at a higher dose

Both of those were rats. In 1997 a group in California carried the question into people, scanning twenty-eight severely head-injured patients with FDG-PET. In that method a faintly radioactive form of sugar is injected and the scanner maps where the brain consumes it.

In six of the patients, oxygen use was measured at the same time. All six were burning glucose well outside the expected range, by at least two standard deviations, while oxygen consumption failed to keep pace. Within the first week, fifty-six percent of the patients studied showed the pattern.

Those were severe injuries and not concussions. The study establishes that a human brain does this, and says nothing about the dose required. That distinction matters, and it is one the field has sometimes let slide.

Giza and Hovda ordered the cascade in four stages

The neurometabolic cascade of concussion, published in 2001, gathered more than a hundred animal and human studies into one ordered sequence, and it appeared in a journal read by athletic trainers rather than by neurosurgeons.

By 2001 the pieces existed and nobody had assembled them. Christopher Giza and David Hovda, brain-injury researchers at UCLA, gathered more than a hundred animal and human studies into a single ordered sequence. The choice of journal tells you who they thought needed it.

The sequence runs in four stages.

Read the fourth stage twice. Reduced metabolism and reduced blood flow, arriving together and outlasting the headache, is the whole clinical problem in one line.

The 2014 revision tied stages to symptoms

In 2014 the same two authors rewrote the cascade, because imaging had finally made it possible to watch the process in living people. This time they tied stages to symptoms. The ionic and glutamate surge maps onto the migraine-like features of the first hours. The metabolic depression maps onto the window in which a second injury does damage out of all proportion to its force. Impaired connections between regions map onto the fog.

Both papers are reviews, and the links between individual stages and individual symptoms are proposed correspondences rather than demonstrated chains in people. The sequence is quoted so often that it has begun to sound like a measurement, and it is a synthesis.

The cascade describes a bill rather than a lesion. Energy is spent in seconds, then repaid on a shrunken budget.

The chemistry has not finished when the symptoms have

Brain chemistry after concussion takes thirty days to normalize in athletes whose symptoms resolve within days, and repeat injuries cluster inside ten days of the first. A spectroscopy series and a 2,905-player cohort mark out the same window.

If the debt is real, it should be visible in a living athlete after the symptoms have gone. An Italian group went and looked.

They used magnetic resonance spectroscopy, which uses an MRI scanner to measure chemistry instead of building a picture. It reports how much of a given molecule sits in a given piece of tissue. The molecule they tracked was N-acetylaspartate, a marker of how well mitochondria are working. Thirteen concussed athletes were scanned at three, fifteen and thirty days, against five controls.

At three days the marker had fallen by about eighteen percent. The athletes at that point reported that their symptoms were already gone. It took the full thirty days to normalize.

Three of them kept training and were concussed again inside that window. Their chemistry dropped further and took forty-five days to return.

Thirteen athletes, five controls, and a subgroup of three: a finding this important is carried by a very small study. What gives it weight is that the same window shows up independently, in a stadium, with thousands of players in it.

The ten-day cluster in 2,905 football players

The sport-science researcher Kevin Guskiewicz, Michael McCrea and their collaborators followed 2,905 college football players across 4,251 player-seasons. There were one hundred and eighty-four concussions. Twelve players were concussed a second time inside the same season, and most of those repeats fell within ten days of the first injury. Players who reported a history of three or more previous concussions were three times more likely to sustain a new one.

Concussion history in that study was self-reported and retrospective, and twelve repeat injuries is a thin base for a clustering claim. Even so, the ten-day cluster and the thirty-day chemistry are describing the same thing from opposite ends of the field: a period during which the system is spending more than it earns.

That period is the beginning of concussion's time course, and every same-season return decision is a bet about where in the window the athlete stands.

Blood flow stops tracking demand after concussion

Cerebral blood flow keeps changing after concussion symptoms resolve. In eighteen high-school players it was lower at day eight than at day one, and flow to the insula was still recovering at one month in college athletes cleared well before.

Neurovascular coupling, the matching of blood flow to neural work from moment to moment, is one of the tightest partnerships in the body, and it can be measured directly.

A Milwaukee group led by the radiologist Yang Wang scanned eighteen concussed high-school football players within twenty-four hours of injury and again at eight days, with nineteen non-concussed teammates as controls. The method was arterial spin labeling, a form of MRI that measures blood flow by magnetically tagging the blood already in the neck and watching where it goes.

Nothing is injected. Symptoms and cognitive scores were clearly abnormal at twenty-four hours and back to baseline by day eight. Blood flow moved the other way. It was significantly lower at day eight than at day one, while the controls did not change at all.

The player felt recovered while the supply line was still adjusting. That is the first crack in the idea that feeling better and being better are the same event.

The insula is still refilling at one month

A group in Tulsa, working with the neuroscientist Timothy Meier, followed forty-four college football players with serial scans at roughly one day, one week and one month. Reaction time normalized within a week and mood symptoms within a month. Blood flow in the insula was still recovering at one month.

The insula is the strip of cortex that reads the body's internal state, the region that turns a heartbeat and a gut and a breath into a feeling of how you are. Its perfusion was lowest at one month in exactly the athletes who had been slowest to be cleared. Only seventeen of the forty-four had all three scans, and the finding needs replication.

Symptoms recovered in a week. Flow to the region that reads the body was still climbing back at a month, and it was lowest in the players who took longest.

The tissue is turned down rather than starved

Falling flow on its own could mean starvation. Start with what is wrong with the study that tested it. Fifteen concussed athletes is a small sample for a claim this structural, and the group difference was confounded by resting carbon dioxide levels, which the authors report openly.

Now the work. A neuroscience group at Queen's University, with Allen Champagne as lead author, used a calibrated MRI method with breathing challenges of high carbon dioxide and high oxygen. That lets a scanner separate three quantities which usually travel together.

How much blood arrives, how much oxygen is pulled out of it, and how much oxygen the tissue actually consumes. Those fifteen athletes, about eight days after concussion, were compared with twenty-seven healthy ones. Perfusion was down across the gray matter. Perfusion means the blood actually reaching tissue, and gray matter is the layer where the nerve cell bodies sit. Oxygen consumption was down in the same places. The fraction of oxygen extracted from the blood did not change.

Supply had fallen, and demand had fallen with it. The tissue was not being starved. It had been turned down.

Coupling is coherence between systems, and a concussed brain has measurably lost some of it.

Concussion reads normal at rest and abnormal under challenge

Three instruments agree on the shape of concussion physiology: vessel reactivity, cerebral autoregulation, and the match between activity and supply all read normal at rest and abnormal once the system is asked to work.

The vessels answer carbon dioxide late

Start with the vessels themselves. Brain arteries are not passive pipes. They widen when carbon dioxide rises in the blood and narrow when it falls, which is how the brain protects its own supply during exertion and during sleep. A group at the University of Regina led by Trevor Len tested whether that responsiveness survives a concussion.

Twenty recently concussed people held their breath for twenty seconds and then hyperventilated. They did it on days two, four and eight after injury, and again after returning to play. Ultrasound tracked blood velocity in a major artery feeding the brain throughout.

At rest there was no difference on any day. Under the breathing challenge, day two was clearly abnormal, and it had settled by day four. Twenty participants, no control group, and a narrower window than any other physiological measure here. The shape of the result is what matters.

That literature does not run all one way, and the study in the last section is the reason. The same calibrated MRI scans that found perfusion and oxygen use reduced also found reactivity higher in the concussed athletes, in exactly the voxels where flow differed. Take the two results together. What a concussion changes is the relationship between demand, supply and the vessel's answer, and a relationship can be wrong in either direction.

Autoregulation loses its timing for a month

Next, the buffer. When you stand up, blood pressure drops for a beat or two, and the brain holds its own flow steady anyway by adjusting vessel tone faster than the pressure can fall. That trick is called cerebral autoregulation.

A group at the University of British Columbia whose authors included Alexander Wright, Jonathan Smirl and Paul van Donkelaar tested one hundred and seventy-nine junior elite hockey and football players before the season. They then followed the eighteen who were concussed during it. Repeated squat-to-stand maneuvers swung blood pressure up and down on a fixed rhythm while ultrasound tracked brain blood velocity, so the buffering could be watched in action.

The timing offset between pressure and flow is the signature of a buffer that is working. It was reduced by twenty-three percent at seventy-two hours and by eighteen percent at two weeks, recovering by one month. Those athletes were cleared to play at a median of fourteen days.

At rest the concussed and the healthy looked identical. Ask the system to do something and they came apart. The injury lives in the answer to demand.

In persistent symptoms, the coupling itself is off

And in people whose symptoms never settle, the coupling itself is still off. An imaging group compared thirty-eight people carrying a diagnosis of post-concussion syndrome with thirty-eight controls, matching local neural activity against local blood supply point by point. Almost every sense passes through the thalamus on its way to the cortex, and the loop between the thalamus and the front of the brain is part of how attention gets set.

Excitatory drive from thalamus to frontal cortex fell on both sides, and the inhibitory return from the right frontal cortex to the right thalamus fell with it. One deep structure beneath the cortex, the putamen, pushed harder on the frontal cortex, which the authors read as compensation. Thirty-eight per group, cross-sectional, recent and unreplicated. It is support rather than foundation.

Three instruments: a vessel answering carbon dioxide, a buffer timing itself against a swing in blood pressure, and the match between activity and supply. One shape.

Heart rate control degrades after concussion, mostly under exertion

Thirty-six studies of autonomic function after concussion qualified for a 2018 systematic review, and only three failed to find abnormalities. The clearest single study found the disturbance during exercise and none at rest.

The nervous system runs the body's automatic settings through two opposing arms. The sympathetic arm is the accelerator: it raises heart rate, tightens vessels, mobilizes fuel. The vagal arm is the brake: it slows the heart beat by beat and lets the system stand down.

You can read the balance between them at the wrist. A healthy heart never keeps perfect time. The interval between beats shifts continuously, lengthening as you breathe out and shortening as you breathe in, and the size and structure of that shifting is heart rate variability. High variability means the brake is present and responsive. A metronome heart is a heart with no brake on it.

The systematic review, graded to neurology standards

The neuropsychologist Jon Pertab led the first systematic review of autonomic function after concussion built to formal neurology standards, with two independent teams extracting and grading every study. Thirty-six studies qualified, and only three failed to find abnormalities. The review concluded that concussion likely causes them. Likely is the authors' own careful word, chosen after grading every included study for bias against American Academy of Neurology criteria, in a literature made of small samples.

The study that shows why the caution is warranted, and why the finding is real anyway, was run in 2004 by a kinesiology group at Simon Fraser University with Brent Gall as lead author.

Fourteen recently concussed athletes and fourteen matched controls were compared first at rest, then during ten minutes of steady cycling at moderate intensity, twice, five days apart. At rest there was no difference at all. During exercise the concussed athletes showed shorter intervals between beats, and both the quick beat-to-beat corrections and the slower swings shrank.

That is the second instrument giving the same answer as the breathing challenge. Rest hides it.

The disturbance outlasts the clearance decision

A Toronto group, with the researcher Arrani Senthinathan as lead author, then asked whether the disturbance outlasts the clearance decision. Eleven concussed university athletes were tracked at three moments: while symptomatic, once symptom-free, and one week after being cleared to play, each against a matched teammate.

Sitting and standing measures were used, so the shift between postures could be read as well as the resting value. A sympathetic shift while sitting and a blunted change between postures were still present after return to play. Eleven athletes, described by the authors themselves as preliminary, and prior concussion history was tangled up with the findings.

The contest is not settled. A group of physical therapists led by Brent Harper asked whether heart rate variability is reliable enough to be a concussion biomarker. They found only three relevant studies, with conflicting systematic reviews behind them. One found disturbance during exertion that persisted past symptom resolution. One found differences by concussion history but also by age and sex. The third found no age or sex effect and concluded that resting measurements are much less informative than measurements taken under exertion.

Three studies agreeing that the signal lives under load is not three studies agreeing on the signal. What they agree on is where to look.

Graded exercise below the symptom threshold speeds recovery

Aerobic exercise dosed just below an individually measured symptom threshold raised peak tolerated heart rate from 147 to 179 in the first Buffalo series. The multicenter trial that followed cut the risk of symptoms persisting past 28 days by 48 percent.

For decades the advice after a concussion was rest until symptom-free. It is the intuitive treatment for damage, and every instinct a person has supports it. A group in Buffalo tested it anyway, from the other direction.

First, a threshold nobody had measured

John Leddy, Barry Willer, Karl Kozlowski and their colleagues had twelve people whose post-concussion symptoms had refused to settle, six athletes and six non-athletes. They walked each person on a treadmill until symptoms began. That gave a number that had not existed before: an individual heart rate at which this particular brain stopped tolerating exertion. Then they trained each person five or six days a week at eighty percent of that heart rate, staying deliberately below the threshold.

Peak tolerated heart rate rose from 147 to 179 beats per minute without symptom flare. The rate at which symptoms fell changed in every subject, and mean symptom count fell in eight of the eleven for whom that analysis was possible. Everyone returned to work or sport. Twelve people, no control group, no blinding, and a sample that had selected itself by failing everything else. The authors said outright that a randomized trial was needed.

Then the randomized trials

Nine years later they ran it. One hundred and three adolescents seen within ten days of a sport-related concussion were treadmill-tested to find their own symptom threshold. Each was then randomly assigned to twenty minutes a day of aerobic exercise just below it, or to a stretching program built not to raise heart rate.

A physician blinded to assignment judged recovery. The exercise group recovered in a median of thirteen days against seventeen. The reduction in delayed recovery, four percent against fourteen, did not reach statistical significance, and participants obviously knew which activity they were doing.

One trial is not a finding. So the same group repeated it across three centers with heart rate monitors verifying that people actually trained where they were told to. One hundred and eighteen adolescents, individualized sub-threshold aerobic exercise against placebo stretching, for up to four weeks. The exercise group recovered faster and had a forty-eight percent lower risk of symptoms persisting past twenty-eight days. No adverse events occurred.

Rest is the intuitive treatment for damage. If the injury is to regulation, graded demand is the input the system uses to find its settings again.

Adults two years out

Then the harder case. A Calgary group with the researcher Leah Mercier as lead author asked whether this reaches adults who are already years out. Fifty adults averaging twenty-five months post-injury, all with exercise intolerance, did twelve weeks of sub-threshold aerobic exercise, some starting immediately and some after six weeks of stretching.

Symptom burden fell significantly in both groups after each had done its twelve weeks of exercise. Waiting bought nothing: the group that stretched first improved less, not more, once it finally trained. That was a cohort nested inside a trial and not a randomized comparison against a control, seventy-four percent female, with a self-reported primary outcome.

Notice what every one of these studies had to do first. Before treating anyone, they measured that person's individual threshold. The dose was never the same twice.

Strict rest was tested, and it slowed recovery

Five days of strict rest after concussion produced more symptoms across ten days than usual care, 187.9 against 131.9, in the randomized trial that finally tested the advice.

The rest-until-symptom-free instruction had never been tested against an alternative. A Milwaukee emergency department group led by the pediatrician Danny Thomas tested it.

Ninety-nine adolescents were randomized within twenty-four hours of concussion to five days of strict rest, or to usual care, which meant one to two days of rest and then a stepwise return to activity. Cognitive and balance outcomes came out the same. The strict-rest group reported more symptoms overall across ten days, 187.9 against 131.9, and slower symptom resolution.

Symptom reporting was not blinded, and the authors note that the recommendation itself may have shaped what teenagers reported. That is the interesting part rather than a flaw in the trial. An instruction to do nothing is an input, and it arrives at a nervous system already asking whether it is damaged.

3,063 children, observed rather than instructed

The population data point the same way. A Canadian network enrolled 3,063 children and adolescents across nine pediatric emergency departments and simply recorded what they did in the first week.

Among the 2,413 with complete data, persistent symptoms at twenty-eight days occurred in 24.6 percent of those who had done some physical activity within seven days, against 43.5 percent of those who had done none. The gap held after the authors reweighted the sample so that the active and the inactive children matched on everything they had recorded.

The authors stated the objection themselves. This is observational, and children who feel better move more. No statistical correction fully removes that. The authors called explicitly for a randomized trial, and the trials in Buffalo are the closest thing the field has to one.

The whole treatment literature: thirteen studies from 6,533

The treatment evidence as a whole is thin, and it points one way. As part of the Amsterdam process, Kathryn Schneider and colleagues screened 6,533 studies of treatment after sport-related concussion and found thirteen that met inclusion. One was rated high quality. Seven were acceptable. Five were at high risk of bias.

The interventions and outcomes varied so much that pooling was impossible. The strongest single signal was for treating the neck and the balance system in people whose dizziness, neck pain or headache had lasted beyond ten days. Those patients returned to sport at close to four times the rate of those who did not receive it.

Thirteen studies is a thin evidence base for a condition this common. It is also better than it was ten years ago, and it points in a consistent direction: what helps is graded, individualized demand delivered to a system that has lost its settings.

Nobody strikes a head without moving a neck

The head, the neck, and the eyes are wired into one loop that keeps gaze steady while the body moves, and a blow that accelerates the head loads all of it. After concussion, neck strength and position sense worsen, and seventy percent of examined adolescents carry at least one oculomotor diagnosis.

Concussion without any blow to the head

Benjamin Elkin, James Elliott and Gunter Siegmund study how bodies are hurt in collisions. They asked why so many people walk away from a rear-end car crash with concussion symptoms when nothing struck their head. They took head-motion data from real rear-end crash tests and from instrumented football helmet impacts. Both went into a finite element model of the human brain, a computer reconstruction that divides the tissue into thousands of small blocks and works out how far each block is stretched and sheared.

Brain strain tracked best with how fast the head rotation changed and not with how hard anything was hit. One crash in which the head wrapped over the head restraint produced strains comparable to a helmet impact known to cause concussion. That is a modeling study and not a patient study, and brain models carry large uncertainties in their tissue parameters.

The neck is a sense organ

The neck is also a sense organ. Muscle spindles are stretch receptors buried inside a muscle, reporting its length moment by moment, which is how you know where your head is without looking. Anatomists in Melbourne counted them at autopsy in sixteen people aged four to seventy-seven, in tissue taken from the fifth to seventh cervical segments.

The deep flexor at the front of the spine carried a high density and multifidus behind it carried a low one, and the gap between the two was the finding. So not every small deep neck muscle is built to report position. The ones that are sit exactly where a collision loads them, and post-mortem counts depend on how the tissue was prepared.

What measurably changes, and the null beside it

So what actually changes after a concussion? Kathryn Schneider, Carolyn Emery and colleagues tested elite youth ice hockey players before the season and again about four days after injury, so each player was his own control. Neck flexor endurance, anterolateral neck strength and the response to a sudden head perturbation were all significantly worse after concussion. Two measures were not. The vestibulo-ocular reflex, the brainstem loop that holds your gaze on this line while you shake your head, was unchanged, and so was dynamic balance.

The null carries weight. Sixty-nine of ninety-seven players completed paired testing, no single test had more than fifty-five pairs, and the bedside reflex tests may be too blunt to catch what a laboratory would. Or the reflex may genuinely be intact in most concussed adolescents while the neck is not. Either way the null sits beside the oculomotor findings, and both stand. A claim that these instruments read one organization has to answer for the times they disagree.

Position sense predicts who gets concussed

The neck may also help decide who gets injured in the first place. A group working inside professional rugby tested one hundred and sixty-five players three times across a season on a joint position error test. A blindfolded player rotates the head, then tries to return it to a marked starting point. Forty-five concussions occurred that season.

Each ten percent increase in repositioning error for gross right rotation was associated with a five percent higher concussion rate. One sport, male professionals, one direction of rotation out of several tested, and a small effect. It is an association and not a demonstrated cause. It is also the cleanest available evidence that the quality of a body signal helps set what an event does to the body carrying it.

Convergence, accommodation, and the vestibulo-ocular reflex

Now the eyes. Convergence is the inward turn of both eyes to hold a near target as a single image, and a vergence problem is a failure of that turning. Accommodation is the lens changing shape to focus.

The vestibulo-ocular reflex, again, is the brainstem loop that holds your gaze steady while your head moves, and vestibular simply means the balance organs of the inner ear. Each of these is a prediction the brain makes about its own body, and each needs a clean position signal to work.

A Pittsburgh group including Anne Mucha, Michael Collins and Anthony Kontos noticed that sideline tests were missing the eye-movement problems they saw in clinic. So they built a five-part provocative screen: smooth pursuit, saccades, near point of convergence, the vestibulo-ocular reflex and visual motion sensitivity.

Each item is designed to provoke, and the tester asks whether it makes symptoms worse. Among sixty-four patients and seventy-eight controls, sixty-one percent of patients had symptoms provoked by at least one item. It is a new tool validated by its own developers, in one cross-sectional sample, which is where every screening tool starts and not where it should end.

Screening is not examining. The optometrist Mitchell Scheiman and a sports medicine group examined one hundred and thirteen adolescents properly, four to twelve weeks after concussion. There was no uninjured control group, and the sample was referred, which selects for people still symptomatic at a month. Inside those limits the numbers are striking.

Seventy percent had at least one oculomotor diagnosis, and forty-seven percent had more than one. Sixty percent had a vergence problem, the two eyes failing to turn together. Fifty-seven percent had an accommodative problem, the lens failing to change focus. Physician screening had caught these in only forty-three to sixty-three percent of the cases the full examination found.

Treating the neck and the balance system

One small trial asked whether treating that loop shortens the course. Thirty-one people aged twelve to thirty with persistent dizziness, neck pain or headache were randomized to eight weeks of usual care, or usual care plus cervical spine physiotherapy and vestibular rehabilitation. Eleven of fifteen in the treatment group were cleared within eight weeks, against one of fourteen. Thirty-one people at a single center, and a control clearance rate of one in fourteen that inflates the contrast considerably.

The cervical spine has its own page at neck pain, and vestibular assessment belongs to balance and coordination. What matters here is narrower. Head, neck and eyes form one coupled loop, and a concussion is an injury to the loop and not to the skull.

Most people recover on a schedule, and children keep their own

In 1,631 baseline-tested college football players, balance recovered within three to five days of concussion, cognition within five to seven, and symptoms by about a week. Children run slower curves now plotted as centiles, and the physiology finishes after all of it.

The familiar figure is seven to ten days. It comes from one study, and it is worth knowing exactly what that study measured.

Michael McCrea, Kevin Guskiewicz and a collaboration across fifteen colleges baseline-tested 1,631 football players, then followed the ninety-four who were concussed and fifty-six uninjured teammates at eight time points out to ninety days. Balance recovered within three to five days, cognition within five to seven, and symptoms largely by about a week. Those were college men playing one sport, and recovery was defined by the instruments in use at the time.

A population curve is not a promise to an individual. A Canadian network made that explicit by drawing the pediatric curve properly, plotting centile trajectories by age band and sex the way growth charts are plotted. A clinician can then say where a particular child sits rather than whether the child is late. The result is a family of trajectories rather than a deadline.

The tests go quiet before the physiology does

Now the harder question. Does the biology finish when the tests do?

Heather Belanger and Rodney Vanderploeg pooled twenty-one studies covering 790 concussions against 2,014 controls. Within twenty-four hours the cognitive effects were large, especially for delayed memory. Beyond seven days the pooled effects were null on neuropsychological testing, with one exception they name. In the studies that used a separate control group instead of each athlete's own baseline, delayed memory was still impaired at seven days.

Set that against a review convened to ask the same question of the physiology. Joshua Kamins and colleagues screened 5,834 articles and extracted eighty, sorted by what each measured: functional MRI, diffusion imaging, spectroscopy, blood flow, electrophysiology, heart rate, exercise tolerance, fluid markers and magnetic stimulation. The theme that recurred across the modalities was that physiological dysfunction can outlast clinical recovery. They could not put a single number on it, and they noted how few studies had followed both clocks to the end.

The tests go quiet at seven days. Blood flow, autoregulation and heart rate control are still moving at a month. Recovery measured by output is not recovery measured by regulation.

Children keep their own curve

Children are not small adults here. A Pittsburgh group including Melvin Field, Michael Collins and Mark Lovell assumed, as everyone did, that a younger brain would bounce back faster. They had baselines on 371 college and 183 high-school athletes, then followed fifty-four concussed players serially against thirty-eight controls. High-school athletes were still significantly worse than matched controls at seven days. College athletes matched their controls by day three, despite having sustained the more severe injuries. Self-reported symptoms did not predict who was still impaired.

Removal from play, and what actually justifies it

Which brings the removal-from-play rule to its evidence. A Pittsburgh group compared thirty-five young athletes pulled from the field immediately with thirty-four who kept playing after a concussion. Recovery took 44.4 days against 22.0, and continuing to play made a recovery longer than three weeks 8.8 times more likely.

Removal status predicted protracted recovery better than any other variable measured, including sex. That study was not randomized, and athletes who keep playing may have had milder presentations or stronger reasons to hide symptoms. The weakness is real, and the rule survives it.

Keep it for the right reason. The rule is usually justified by second impact syndrome, and a Minnesota group led by the sports physician Steven Stovitz systematically reviewed ninety-one sources describing that condition. Eighty-seven percent defined it as catastrophic swelling or death after a second concussion sustained before the first had resolved.

Thirteen percent described nothing more than two concussions in a row. Of the sources quoting a mortality rate, thirty-three of thirty-five called it fifty to one hundred percent. Because nobody knows how often a second concussion is sustained before recovery, the true rate is unknown.

So the removal rule stands on the metabolic window, on the ten-day clustering of repeat injuries, and on the removal study itself. It does not need a mortality figure that the evidence cannot support.

What decides who takes a week and who takes a year

Nine variables predict persistent symptoms in children after concussion, and none of them measures how hard the head was hit. Across 101 studies, the strongest predictor of a slow recovery is the severity of the earliest symptoms.

Most people follow the curve. A minority does not, and the question of what separates them is where this condition stops being about the blow.

Roger Zemek and a Canadian network enrolled 3,063 children within forty-eight hours across nine pediatric emergency departments, with a separate validation cohort, to build a tool that could tell a family what to expect. Thirty-one percent still had symptoms at twenty-eight days. Nine variables carried the prediction. Female sex. Age thirteen or older. A physician-diagnosed history of migraine. A prior concussion with symptoms lasting over a week. Headache, noise sensitivity, fatigue, answering questions slowly, and four or more balance errors.

Not one of those nine measures how hard the head was hit.

The authors call the score's ability to separate the children who will and will not have persisting symptoms modest, and say plainly that it should not be adopted in practice before external validation. A risk tool quoted without its authors' caution becomes something they did not build.

Early symptom burden is the signal that survives

Grant Iverson, Andrew Gardner, Douglas Terry and colleagues screened 7,617 articles and included one hundred and one in a review of what predicts recovery. The most consistent predictor in the whole literature was the severity of early symptoms. Subacute headache or depression raised the risk of symptoms lasting past a month, and a pre-injury history of mental health problems raised it further. Attention deficit disorder and learning disability did not.

A Boston clinic pushed harder on the same question. William Meehan, Rebekah Mannix and colleagues used a statistical method that builds its own decision tree, and offered it every candidate predictor they had in 531 patients aged seven to twenty-six.

Age, sex, loss of consciousness, amnesia, prior concussions, prior headache treatment, migraine history, family history and computerized cognitive scores all went in. Only the total symptom score at the first visit survived, at about four percent higher odds of symptoms past twenty-eight days for every point.

The migraine phenotype lengthens the course

Headache type may matter as much as headache. Joshua Kamins and a multi-center youth registry classified a post-traumatic headache as migraine-phenotype when it was new or clearly worse, moderate to severe, and came with nausea or with sensitivity to both light and sound. Then they tested the classification against recovery time in 281 patients with 286 concussions, aged five to eighteen.

Recovery took a median of ninety-five days with a migraine phenotype against seventy without one, and forty-four days in the children who had no post-traumatic headache at all. Registry data from specialty clinics over-represents people who did not recover quickly, and the phenotype was assigned from symptom inventories instead of by a headache specialist. The move itself is the interesting part: the concussion literature reaching for a distinction the migraine literature made decades ago.

And here the field complicates itself, usefully. Douglas Terry and colleagues used a surveillance system across eleven college athletics programs to test whether pre-injury migraine really slows recovery. Of 1,409 concussed athletes, 111 reported migraines. They took longer to return to the classroom without accommodations, a median of seven days against five, and that gap held in a model adjusted for age, sex and symptom severity.

Return to play was messier. Median days did not differ, and migraine did not survive the adjusted model. Yet by twenty-eight days only seventy-two percent of the migraine group had returned to play, against eighty-three percent of the rest. By fifty-six days the gap was eighty-four against ninety-three.

Migraine history was self-reported and only 7.9 percent of the sample had it. Read the split for what it is. The signal was strongest and most consistent for getting back to class, and mixed for getting back to sport, which suggests a specific vulnerability meeting a specific demand rather than a general slowness. The model did not forecast that split.

It is being read after the result, and a reading built after a result is worth very little on its own. The prospective version can be stated now: in a new cohort, specify in advance that a history of migraine delays the cognitive return more reliably than the physical one. If that does not reproduce, the reading was a story.

The symptom list that uninjured people can produce

Uninjured people asked to imagine a concussion reproduce the post-concussion symptom cluster almost exactly, and depressed people report more of it than healthy controls. The cluster is real, and it belongs to a strained nervous system rather than to concussion alone.

In 1992 a group of psychologists asked people who had never had a head injury to do something odd. Imagine you were concussed in a car crash six months ago. Now tick the symptoms you would expect to be experiencing today.

The imagined cluster came out almost identical to the syndrome real patients report. Meanwhile the actual patients systematically underestimated how many of these symptoms they had before their injury, compared with the base rate in uninjured people. Expectation accounted for about as much of the variance as the injury did.

That was an analogue study, and the imagining group was not injured. It shows expectation is sufficient to generate the cluster. It does not show injury is unnecessary.

Athletes, then the comparison groups the theory had not faced

The team then repeated it in contact-sport athletes, who might have been expected to resist. Head-injured athletes reported symptom rates no different from uninjured athletes. What differed was memory of the before. Injured athletes underestimated how many of these symptoms they had experienced previously, and athletes with no head-trauma history overestimated how much a concussion would change them.

Then two researchers added the comparison groups the theory had not faced: one hundred and forty-one people including healthy controls, healthy athletes, depressed individuals, head-injured athletes and chronic headache sufferers. Depressed people with no head injury reported more of these symptoms than controls, which shows the cluster is not specific to concussion at all.

But both the injured athletes and the headache sufferers recalled their earlier selves as better than their present selves and not as better than the baseline of uninjured controls. A pure memory bias needs that recalled past to sit below the normal baseline, and it did not. Both mechanisms are doing some work, and the cells in that study are small.

Headache, fatigue, poor sleep, irritability and fog are what a strained nervous system produces. They are not a fingerprint of concussion, which is why concussion has no fingerprint.

Expectation is an input, and prediction is the machinery

None of this says persistent symptoms are imagined. It says something more interesting. A nervous system does not wait for evidence before acting; it runs on a model of its own condition and allocates resources against that model, which this library calls prediction. Expectation, meaning and context are high-order inputs to that model, and they change lower-order regulation through the same machinery as a night of lost sleep.

A brain that has concluded it is injured spends differently. It guards more, samples less, and reads ordinary signals as evidence. That is a mechanism, not a dismissal. And it is one of the few concussion mechanisms that responds to being explained to the person carrying it.

What the CTE brain-bank studies do and do not show

Chronic traumatic encephalopathy is a real tau pathology that can currently be diagnosed only after death, and the widely quoted 87 percent came from brains donated because families were already worried. In an unselected bank, the pathology appeared in 27 of 300 former athletes.

Four facts, in ascending order of what they license.

The pathology is diagnosed after death

The neuropathologist Ann McKee, working with Robert Stern, Robert Cantu and colleagues, examined brains from eighty-five people with histories of repeated head impacts. What they were looking for was tau. Tau is a protein that normally stabilizes the internal scaffolding of a nerve cell and, in several diseases, clumps instead. It is found by treating a slice of brain with a dye that sticks only to the clumped form.

They found the characteristic tau pathology in sixty-eight of them, at ages of death from seventeen to ninety-eight. They sorted it into four stages, from small clusters of tangles around blood vessels in the frontal cortex through to widespread disease. That paper put the condition on a formal footing. It is a brain-bank case series with no denominator, and its authors do not claim otherwise.

The 87 percent describes a donated sample

The famous number is not a prevalence. The same bank, with Jesse Mez as lead author, reported two hundred and two former football players, of whom one hundred and seventy-seven met criteria. That is eighty-seven percent, including one hundred and ten of one hundred and eleven who had played professionally. Families donated those brains because the person had shown symptoms. The authors state that selection openly in the paper. The figure describes the sample, and it was never a measurement of risk.

Unselected banks change the number

Look in a bank not assembled by worried families and the number changes. Kevin Bieniek, Dennis Dickson and colleagues reviewed records for 1,721 men in a bank of neurodegenerative disease. Twenty-one of sixty-six former contact-sport athletes had consistent pathology. None of one hundred and ninety-eight people without contact-sport exposure did, including thirty-three with a documented single head injury from a fall, crash or assault.

The same group then built a closer approximation to a denominator, searching obituaries and yearbooks against 2,566 autopsies to identify three hundred athletes and four hundred and fifty non-athletes, screening the tissue blind to exposure.

Forty-two cases in total, twenty-seven athletes and fifteen non-athletes, American football highest at fifteen percent of cases, with an odds ratio of 2.62. An odds ratio of 2.62 means the odds were about two and a half times higher, which is not the same as the risk being two and a half times higher.

The association is real. Repetitive exposure, rather than any single blow, is what tracks with it. Fifteen non-athletes with the pathology complicates a clean exposure story, and that group reported it rather than explaining it away.

Later-life risk, measured by designs built to measure risk

As part of the Amsterdam process, Grant Iverson, Rudolph Castellani and colleagues reviewed cohort and case-control studies only: ten of former amateurs, eighteen of professionals. In former amateur athletes, five studies found no increased risk of depression, and none of nine found an association with suicide.

Some studies comparing professionals against the general population did report associations with dementia and with motor neuron disease as a cause of death. Most of those failed to control for confounders and were graded at high risk of bias. No post-mortem or imaging study met inclusion, which is a limit of scope and not a verdict on that literature.

If you are thinking about harming yourself, contact emergency services or the 988 Suicide and Crisis Lifeline.

The accurate summary is short. Repetitive head impacts are associated with a real neuropathology, and the diagnosis can currently be made only after death. The widely quoted figures come from brains selected because someone was already worried. The incidence among athletes generally is not established.

One boundary is worth stating plainly. A tauopathy is a real neurodegenerative process and it does not reduce to dysregulated tone. Structural, genetic, infectious, toxic and malignant causes are real, and they are treated on their own terms.

What this model claims is narrower and testable: every disease has a tonal expression, and many are initiated, maintained or amplified by failures of tonal regulation. Here that buys an account of how a nervous system receives repeated impacts, what it spends compensating for them, and why the same exposure history runs a different course in two people. It does not buy the pathology.

One injury, read as tone

Ionic flux and hyperglycolysis. Metabolic depression. Reduced cerebral blood flow. Impaired vessel reactivity under challenge. Degraded autoregulation. Reduced heart rate variability under exertion. Cervical position sense. Vergence and accommodation. A recovery curve, and a minority who fall off it. Nine different specialties own those findings, and each specialty reads its own instrument.

The Unified Model of Tone reads them as one thing.

Tone is the integrated organization of the body's interacting state transitions, the way its mechanical, electrical, chemical, fluid and neural processes are related at a given instant, taken as one bound state. Not any single rhythm. The heartbeat, the breath and the pressure wave carry tone without being it. The chord rather than the notes. Health is the width of the range that organization can move through, and disease is its narrowing.

Concussion, in those terms, is three things at once.

Load: the bill

It is load, in the strict sense: the cost of holding a state. The cascade is a bill. Gradients are spent in seconds, and the pumps that restore them run far beyond baseline. The burst of glucose burning is followed by reduced metabolism, with reduced blood flow arriving at the same time. That is a system meeting demand it cannot fund, and then operating on a smaller budget while it repays.

Coupling: the relationships that come apart

It is coupling failure, and this is what makes it a tonal injury rather than a fuel shortage. Blood flow stops tracking neural work. Vessel reactivity fails under a breathing challenge while looking normal at rest. Autoregulation loses its timing. Autonomic balance shifts and stops answering posture. What the neck reports, what the eyes do and what the balance organs say fall out of register with each other. Each of those is a link between two systems that used to stay in step.

Nothing in a concussion is broken, because a relationship is not a part. What is injured is the relationship between systems, and a relationship does not show on a scan.

Time course: the schedule, and the minority who fall off it

And it is a time course. Most systems re-register on a schedule, which is why the population curve is real and useful. The physiology finishes later than the symptoms, which is why the ten-day window exists and why a second injury inside it costs so much. A minority does not re-register at all, and the pattern stabilizes into something that maintains itself.

What the reading dissolves

The reading dissolves several puzzles at once. The scan is clean because organization is not a structure. The symptoms are diverse and non-specific because a change in the whole organization surfaces wherever a given body has least capacity to absorb it.

Rest at first seemed obvious and then proved counterproductive, because a system re-registers through graded demand rather than through silence. The predictors are prior migraine, prior mental health, early symptom burden and neck position sense rather than impact force, because an input meets a tone and the outcome is the product of the encounter.

It also explains why the studied inputs look nothing alike. A graded ride below a treadmill-measured threshold enters through cardiovascular demand. Neck and balance work enters through the mechanoreceptors and the proprioceptive field, changing what the body reports about where it is.

An instruction to lie in a dark room enters through meaning and prediction, and the strict-rest trial is the evidence that an instruction is an input at all. A pill enters through receptor coupling. These are not the same act and they are not interchangeable.

They are inputs delivered at different points of one continuous loop, and a continuous loop carries a change introduced anywhere in it through the rest of itself. The treatment literature has that shape. Thirteen studies could not be pooled, because the interventions and the outcomes varied too much to add together. Inputs entering one loop at different points get measured by whatever instrument sits nearest each entrance. What the model claims here is the loop, and never what any of these inputs achieves.

The wager, and the result that would settle it

That reading is a wager, and one result on record bears directly on it. After concussion the neck measures moved and the vestibulo-ocular reflex did not. Measures of one organization are supposed to move together, and those two did not. The study that decides it is a plain one. Record several of these measures in the same people, and test whether they share a common factor. Loading on a common factor establishes the organization as one thing and confirms this section.

Two people take the same hit. One is back in nine days and the other is still unwell at nine months. They did not receive the same event, because they were not the same system when it arrived.

Quieting an output against restoring a range

Relief and restoration are different targets in a concussed nervous system. A pill that quiets a headache changes the output for the duration of the dose. Sub-threshold exercise widens what the system tolerates, which is a change in the range rather than in a symptom.

A person three weeks out from a concussion has a headache that will not let them sleep. A pill quiets the headache and they sleep. That is worth having, and the model says so plainly.

Someone kept awake by pain is carrying a regulatory burden on top of the injury, and lifting it can return the reserve the system needs to reorganize at all. Analgesia, sleep support, migraine-directed care and mental health care all have an honorable place, and decisions about any of them belong with the person's own physician.

What relief does not do is move the setting. An agent that damps a signal pushes in one direction for as long as the dose continues, and when it stops, the organization is where it was. The model calls that masking, and the word is descriptive rather than dismissive.

The model draws that distinction across the entrances rather than between them. A pill, a graded ride and a program built around the neck's position sense can each be delivered with either aim. Every entrance holds both. The difference is the aim, and never the entrance.

Restoring is a different target. It means the system regains its own ability to match supply to demand, hold a value against a challenge, and change state and come back. The measurable signature would be a widening range rather than a lowered number.

The intervention that behaves like restoration

The concussion literature has one intervention that behaves like that, and it is worth seeing why. Sub-threshold aerobic exercise does not treat a symptom. It delivers a graded, individually calibrated demand to a system whose tolerance for demand has narrowed, and the tolerance widens.

The case series in Buffalo measured it directly: peak tolerated heart rate climbed from 147 to 179 without symptom flare. The randomized trials then showed faster recovery and a lower risk of persistence past four weeks, and the adult cohort showed symptom burden falling in people twenty-five months out.

That same adult cohort carries the null that belongs beside it. Heart rate, heart rate variability and blood pressure were measured before and after the twelve weeks, sitting and standing, and none of them moved. Symptom burden, quality of life, dizziness and exercise tolerance all improved while the resting autonomic numbers sat still.

The authors attribute that partly to the intensity and duration of the intervention. The other available reading, stated here as a reading and not as a result, is that a resting number was the wrong place to look for a change in what the system can tolerate.

A promising mechanism can still fail: melatonin

A genuinely promising mechanism can still fail. Melatonin is neuroprotective in animal models of brain injury, and a Calgary group led by the pediatric neurologist Karen Barlow tested it properly. Ninety-nine children with symptoms persisting four to six weeks were randomized to one of two doses or to placebo, double blind, for twenty-eight days. Everyone improved, by a median of twenty-one points. Melatonin did no better than placebo at either dose.

A promising mechanism is not yet a treatment. The trial may also have been underpowered against a strong natural-recovery effect, and the authors say so. Both readings are allowed, and the result stands either way.

Four tests of this reading of concussion, and what a recovered system looks like

Three results would confirm the tone reading of concussion: challenge that separates groups where resting measures do not, physiology that stays disturbed after symptoms clear, and prior state that out-predicts impact force. A fourth test, never yet run here, says whether an input restores regulation or masks a symptom.

One caution belongs in front of the list. None of these studies was designed to test this reading, and agreement noticed after the fact is worth less than a result specified before it. The evidence pointed at each test is a reason to run it, and never a test already passed.

Challenge against rest

First and sharpest. The claim is that challenge separates a concussed system from a healthy one better than rest does. Vessel reactivity was normal at rest and abnormal only under a breathing challenge. Heart rate variability was identical at rest and abnormal only during cycling. A formal appraisal of that literature concluded that resting measurements carry much less information than measurements taken under exertion.

Note what this does not say. Resting perfusion and resting oxygen consumption were genuinely reduced, so a concussed brain does not look normal at rest. Run both panels in the same people, and dynamic measures that separate the groups where the resting measures do not confirm the claim.

The two clocks

Second. Physiological measures still moving after the symptoms have cleared confirm the claim that the injury is to regulation and not to experience. The systematic review of physiological recovery found dysfunction that can outlast clinical recovery, and blood flow, autoregulation and autonomic measures are all still moving after clearance. Its authors also noted how few studies followed both clocks to the end, which is the current state of it. That is the study waiting to be done rather than the study that settles it.

Prior state against impact force

Third. Pre-injury state that carries information about outcome beyond what impact magnitude carries confirms this reading. The validated pediatric risk score for a head injury contains nine variables, and not one of them is a measure of the hit.

The strongest single predictor across a hundred-study review is early symptom burden. Cervical repositioning error measured across a season was associated with concussion rate. None of that was gathered to test this model. Build a predictive model in which prior state beats accelerometer-measured impact magnitude, and the argument is settled.

The bidirectional test, stated in advance

Fourth, and this one is a prediction rather than a finding. The model's signature claim is bidirectional restoration. A restored regulator moves a value toward the middle of its healthy range from either side. A directional agent pushes one way regardless of where the person started.

That two-tailed design has never been run in concussion. The exercise trials are one-tailed, moving exercise tolerance and symptom burden toward normal from a starting point of intolerance. The accurate statement is that the concussion literature is consistent with bidirectional restoration and has not tested it.

It could be tested. Take a measure with a defined healthy window, autonomic balance or resting cerebral blood flow, and assemble two groups of concussed people, one above the window and one below it. Specify the input in advance, deliver it to half of each group, and give the other half a sham matched for contact and attention.

The model predicts convergence toward the middle in the treated arms, exceeding whatever the sham arms produce. A uniform shift in one direction marks the input as one that pushes the output, helping whichever group it points at and carrying the other group further from the middle.

What a recovering system looks like

Recovery is already measurable on instruments that exist. Vessel reactivity under a carbon dioxide challenge. Autoregulation under squat-to-stand. Heart rate variability under exertion instead of at rest. Exercise tolerance expressed as a symptom threshold on a treadmill. Cervical joint position error. Near point of convergence. Perfusion in the insula. A restored system widens on those measures together, and its symptoms move in step with them instead of independently.

One fact matters more than any of it. A normal scan was never evidence that nothing happened. It was evidence about where the injury lives, and it lives in regulation, which is the one thing in a body that was always built to change.

How concussion relates to the rest of the library

Concussion sits at a junction in the library, because one rotational load disturbs systems that other pages treat one at a time. Each page below carries a specific piece of this condition.

• Load is the fastest complete demonstration in the library of what holding a state costs: gradients spent in seconds and repaid over weeks on a shrunken budget.
• Coupling owns the deeper claim, that what a concussion injures is the agreement between systems, and every challenge-test finding above is a coupling measurement.
• Time course is why the same injury is an acute event in most people and an entrenched pattern in a minority, and why the ten-day window governs the return to play.
• Prediction explains the strict-rest result, because an instruction is an input and a brain that has concluded it is injured spends differently.
• Input quality is the cervical position sense that tracked concussion risk across a rugby season before any injury occurred.

• Among conditions, brain injury takes over where the imaging turns abnormal and tissue is destroyed, which is a different problem with a different literature.
• Neck pain owns the cervical spine that receives the same rotational load as the brain in every collision.
• Balance and coordination carries the vestibular assessment behind the strongest treatment signal in the Amsterdam review.
• Migraine supplies the phenotype distinction that separated a 95-day recovery from a 70-day one.
• Heart rate variability is the instrument that reads a concussed brain at the wrist, and only under exertion.
• And the tone pillar holds the full definition behind the reading: tone as the integrated organization of the body's interacting state, whose health is the width of its range.

Frequently asked