Inflammation and the Nervous System

The four signs of inflammation are actions the body takes on purpose

Redness, swelling, heat and pain have defined inflammation since Celsus listed them around the year 30. Behind each sign is a program with a beginning and an ending, and the difference between health and disease is whether the ending runs.

A splinter goes into a thumb. Within hours the skin around it is red, swollen, hot and sore to the touch. Everyone knows that sequence, and almost everyone reads it the same way, as the damage showing itself.

The list is older than the microscope, and every medical student still learns it in Celsus's order. A fifth sign, loss of function, is routinely credited to Galen, the Greek physician whose writings governed Western medicine for fifteen centuries. In 1971 a medical historian went looking for that attribution in Galen's own writing and could not find it. The phrase entered medicine far later. The list of four is genuine. The story attached to the fifth is folklore.

Look again at what those four describe. Redness is blood vessels opening. Swelling is fluid crossing a wall that normally holds it back. Heat is blood arriving in quantity. Pain is a nerve that has lowered its threshold.

None of that is damage. Every item on the list is something the body is doing, deliberately, in a sequence it has run a thousand times.

Chronic inflammation runs the same four signs without the ending

The splinter comes out and the thumb is normal inside a week. That is the whole program, opened and closed on schedule. Now hold it against the other kind. A gut that has been inflamed for nine years. A joint that swells with no injury behind it. A blood result that keeps coming back raised while nobody can say what it is reporting. The same four signs, and no ending.

That gap separates a working response from a disease. Inflammation is a program. It has a beginning, a middle and an ending, and each of the three is separately regulated. The beginning is famous. The ending is the part almost nobody is taught, and it is where much chronic inflammatory disease lives.

The nervous system sits inside the program at every stage. It reports the inflammatory state upward, continuously, the way it reports temperature. It sets how large the response is allowed to get. And it can produce inflammation from a nerve ending, with no pathogen present anywhere.

Sepsis and the red flags that need a doctor today

Inflammation is also how people die of infection. In 2016 a task force of nineteen intensive care specialists defined sepsis as life-threatening organ dysfunction caused by a dysregulated host response to infection, and that response can kill within hours.

The task force rewrote the definition because the old criteria had been missing people who were septic and flagging people who were not. The danger sits in the response rather than in the organism that provoked it.

The red flags, listed

So the practical line comes first. Fever with severe illness needs a doctor today, not tomorrow. Get urgent care for fever with a stiff neck, for confusion, or for a rash that does not fade when you press it.

The same goes for severe or one-sided abdominal pain, a single hot swollen joint, a spreading patch of hot red skin, or new shortness of breath. A new headache in an adult over fifty needs same-day assessment, especially alongside scalp tenderness, jaw pain on chewing, or any change in vision.

One inflammatory condition presents in exactly that pattern and can take sight permanently. Unexplained weight loss, drenching night sweats, or a fever that keeps returning with no explanation should be investigated rather than watched. Anyone taking an immunosuppressive drug or a biologic, meaning a medicine built from a large protein rather than a small chemical, should be assessed the same day if they become seriously unwell. Those drugs quiet the very signals a doctor would otherwise read.

What regulation never replaces

The second boundary matters as much. Infection, autoimmunity and malignancy are real causes with real treatments, and reading them through regulation replaces none of that. The cancer biologists Lisa Coussens and Zena Werb assembled the case that inflammation is a component of tumor progression rather than a bystander, since many cancers arise at sites of chronic infection and chronic irritation.

A tumor is not a mis-set thermostat. None of it is a reason to decline, delay or alter immunosuppressive or biologic therapy, and that decision belongs with the doctor who prescribed it.

The formulation this library holds to is exact. Every disease has a regulatory expression, and many are initiated, maintained or amplified by failures of regulation. That is a claim about how a body receives a cause. It is never a claim that the cause was not there.

Inflammation also reaches mood directly. If you are thinking about harming yourself, contact emergency services or the 988 Suicide and Crisis Lifeline.

Macrophages run the inflammatory response, and cytokines carry its orders

Three signals organize inflammatory disease: tumor necrosis factor, interleukin-6 and interleukin-1 beta. The cell releasing most of them is the macrophage, whose discoverer was fought for decades by researchers certain that immunity lived in serum.

In the late nineteenth century a zoologist who had turned to pathology, Elie Metchnikoff, proposed that wandering cells inside an animal swallow invaders whole. Researchers certain that immunity lived in the blood serum fought him for decades. Serum is the clear liquid left when blood is spun and the cells are taken out.

The immunologist Jean-Marc Cavaillon has traced that whole dispute, including the later classification schemes that led the field down blind alleys. Both camps had hold of something real, and the Nobel committee eventually split one prize between the two accounts.

The macrophage and its neighbors

The cell Metchnikoff described is the macrophage. It lives in tissue, patrols it, and eats debris, dead cells and microbes. Two other white cells belong in the cast. The neutrophil arrives first and in enormous numbers, lives briefly, and dies at the site, which is why clearing dead neutrophils is a job the body has to schedule. The monocyte is the macrophage's traveling form, riding in the blood until tissue calls it in.

The macrophage also does the thing that matters more here. It signals.

Dinarello counted more than a hundred cytokine genes, thirty-three of them interleukins

The signal it releases is a cytokine. A cytokine is a small protein a cell releases to change the behavior of other cells, and its defining oddity is that the same molecule means different things in different tissues. Charles Dinarello, the immunologist who isolated interleukin-1, counted more than a hundred genes coding for such activities, thirty-three of them named interleukins. The cytokine does not carry the outcome. The cell that receives it does.

TNF, interleukin-6, interleukin-1 beta

Three of those words run inflammatory disease. The first is tumor necrosis factor, and its name is an accident of the experiment that found it. Cancer researchers in the 1970s were testing an old clinical observation, that tumors sometimes shrank in patients who had developed a severe bacterial infection. They primed mice, dosed them with bacterial endotoxin, and transferred the serum into tumor-bearing mice.

Endotoxin is a fragment of the outer coat of certain bacteria, and the immune system treats it as proof of invasion. The serum alone made the tumors blacken and die, and the substance responsible came from the host's own cells rather than from the bacteria. They named it for what it did that day. Its actual work is inflammatory signaling, and it is the most consequential molecule in inflammatory medicine.

The second is interleukin-6. It appears quickly after injury or infection and drives fever, along with the acute-phase response, which is the liver's switch to manufacturing a different set of proteins while the body is under attack. Production is normally shut down again by tight control.

When it does not stop, as the immunologists Toshio Tanaka and Tadamitsu Kishimoto set out, the same molecule maintains chronic inflammation and autoimmunity. Kishimoto's laboratory cloned interleukin-6, and their work on blocking the docking point it uses became a drug now approved for rheumatoid arthritis.

The third is interleukin-1 beta, the signal most directly responsible for fever. It returns in the experiment that revealed how the brain learns there is an infection, and in a trial of 10,061 people with heart disease.

Inflammation runs without an invader

Ruslan Medzhitov named the quiet version para-inflammation in 2008: the same response, run at low grade by resident macrophages, triggered by ordinary tissue stress with nothing to fight. Much modern chronic inflammation may be a maintenance response nobody switched off.

Ask what this program is for and the obvious answer is defense. The obvious answer is incomplete, and the incompleteness is where chronic disease hides. Medzhitov, an immunologist, put the question the textbooks skip. What is inflammation doing when nothing has invaded? His answer was that infection and injury sit at one end of a much larger set of triggers. Para-inflammation is a proposed category with no agreed measurement, argued for rather than demonstrated.

On that reading, the low-grade inflammation of modern chronic disease is a maintenance response nobody switched off. The distinction changes what you would go looking for. A false alarm needs silencing. An alarm correctly reporting an unresolved condition needs the condition resolved, and needs its ending to work.

The mast cell against the nerve ending

The nervous system is already inside this picture, and the mast cell shows it first. Mast cells sit in tissue loaded with granules, tiny packets of ready-made chemicals they can empty in seconds, and they cluster against nerve endings. The neuroimmunologist Paul Forsythe has reviewed what that proximity means. Neurotransmitters, the chemicals a nerve releases to speak to the next cell, activate mast cells directly.

So the nervous system can order an immune release with no pathogen anywhere in the story. The traffic runs the other way as well. Mast cells secrete mediators, among them neurotransmitters and nerve growth factors, that change how excitable the neighboring nerve is, sometimes long afterwards. Most of that evidence comes from rodents, and Forsythe says plainly that its weight in human health is unclear.

Hold that picture, because it is the shape of everything ahead. The nerve ending and the immune cell are two halves of one circuit, which is the whole of what the word neuroimmune means.

The molecules that end inflammation were found before anyone knew what they did

Charles Serhan isolated the lipoxins in 1984 while mapping what white cells make from arachidonic acid. Years passed before timed sampling of resolving tissue revealed their job: they are the chemistry inflammation uses to stop recruiting cells.

Prostaglandins open the response

When a cell membrane is damaged it releases a fatty acid called arachidonic acid. Enzymes inside the cell convert that raw material into short-lived local signals built from fat rather than from protein, and the best known of those are the prostaglandins.

A prostaglandin opens blood vessels, draws fluid into tissue and lowers the firing threshold of nearby pain nerves. Two of Celsus's four signs are prostaglandins doing their job. The enzyme that makes them in inflamed tissue is called COX-2, and it is what aspirin and the ordinary anti-inflammatory drugs act on.

Lipoxins, and the timing question

In the early 1980s chemists in Stockholm were mapping what else human white blood cells make out of arachidonic acid. Charles Serhan, working with Bengt Samuelsson, isolated a new class of compound formed when two separate enzyme pathways inside a single cell interact.

They named them lipoxins. In that first paper the new molecules looked like activators, and one of them stimulated white cells rather than quieting them. The off switch was sitting in the data for years before anyone read it as an off switch.

The reading came from a question about timing. If inflammation ends reliably, when is the ending decided? Bruce Levy and Serhan, working in the same Boston laboratory, sampled the fluid inside inflamed tissue over time in patients and in animals. They tracked which lipid signals appeared when. The molecules that recruit white cells came first. Then the same cells switched their chemistry and began producing lipoxins, which stopped further recruitment. The switch was triggered by the first-phase signals themselves.

The ending is written into the beginning. Inflammation issues its own shutdown order at the moment it starts. That conclusion is reconstructed from the timing of mediators in a limited set of samples, and it is worth naming as such. Nobody has watched the switch happen inside a living person.

Aspirin reveals the resolvins

Then aspirin produced a surprise. Aspirin blocks COX-2, which is why it brings a fever down and eases a swollen joint. It turns out that it also chemically modifies the enzyme, so treated tissue makes a mirror image of the usual product instead. Serhan's group gave mice aspirin and an omega-3 fatty acid, then sampled the fluid during the resolution phase rather than the onset phase, which almost nobody had thought to do.

They found a family of molecules nobody had described. At nanogram doses these cut the influx of white cells by 40 to 80 percent. They named them resolvins. This is mouse work with synthetic mediators, and it says nothing about whether eating omega-3 raises these compounds usefully in a person.

Resolution is an act the body performs with its own chemistry

Lipoxins, resolvins, protectins and maresins, grouped as the specialized pro-resolving mediators, are agonists: molecules that switch a program on. What they switch on is the clean-up, and its failure is where much chronic inflammatory disease lives.

Serhan states the difference in kind without hedging. The program these molecules start calls in the cells that clear debris and dead neutrophils, and it speeds the clearance of microbes rather than suppressing it. A drug that blocks a signal is doing the opposite kind of work.

Resolution arrives with its own chemistry, its own signals and its own timetable. It clears the site and rebuilds it.

Inflammation does not fade. It is ended, by a program as organized as the one that opened it. Much chronic inflammatory disease is that second program failing to run.

That is not a fringe position. Five researchers who study how inflammation ends, among them Christopher Buckley, Derek Gilroy and Serhan, were asked how you turn off inflammation that will not stop. Their shared answer was that resolution is as elaborate and as actively organized as onset.

The disease in many chronic inflammatory conditions is the failure of that second program rather than an excess of the first. It was a discussion piece rather than a systematic review, and they posed the therapeutic promise as an open question.

The consequence is a change of target. Suppressing onset and completing resolution are different goals, and a drug that achieves the first does not by itself achieve the second. That distinction arrives here from inside mainstream immunology, and it is the same one this library draws between managing a value and restoring the regulator that sets it.

It is also why time course carries part of inflammation's signature. An inflammatory disease is often an acute program stalled halfway, held open for months or years, at a site where the second crew never came.

The same enzyme opens the response and closes it

Derek Gilroy set out to confirm COX-2 as the inflammatory enzyme and found it running the resolution too. Blocking it early reduced rat pleurisy, and blocking it at forty-eight hours made the same inflammation significantly worse.

The demonstration that timing is a property of the system came from an experiment set up to show something else.

By the late 1990s COX-2 was regarded as the inflammatory enzyme and the obvious drug target. Derek Gilroy, working in a group that tested anti-inflammatory drugs, tracked its expression across two days in a rat model of pleurisy, which is inflammation of the lining around the lung.

It peaked at two hours alongside the classic inflammatory prostaglandin, exactly as expected. Then it did something nobody expected. It peaked a second time at forty-eight hours, 350 percent greater than the first peak, at the precise moment the inflammation was resolving. This time it was making different prostaglandins.

So they tested what the second peak was for. A COX-2 inhibitor given at two hours reduced the inflammation. The same inhibitor given at forty-eight hours made the inflammation significantly worse. Replacing the missing prostaglandins reversed that harm.

One enzyme runs the opening and the closing, and the clock decides which. Nothing about the molecule changed between hour two and hour forty-eight. What changed was the state of the tissue receiving its product. The input did not carry the outcome. The system that met it did, and that is the principle the whole neuroimmune story runs on.

Two things this result is not. It is one rat model of one kind of inflammation, and it does not show that ordinary human use of anti-inflammatory medicines prolongs anything. It is also not advice about anybody's medicine cabinet.

Questions about any anti-inflammatory, prescribed or bought off a shelf, belong with a pharmacist or a physician. What the experiment does show is worth carrying. A pathway can be load-bearing at one phase and harmful to block at another, and that difference is invisible to any measurement taken at a single moment.

Whether pro-resolving mediators can be measured in people is disputed

The analytical challenge is blunt: in blood and tissue, resolvins and lipoxins are often reported below the concentrations the instruments can reliably see. The dispute is about measurement, and resolution as an active program rests on other evidence.

In 2023 fifteen lipid chemists, led by Valerie O'Donnell and including Garret FitzGerald, published a challenge to the measurement underneath the pro-resolving mediator literature. The argument is analytical rather than biological. These molecules are identified by mass spectrometry, an instrument that names a compound by weighing it with great precision. In blood and tissue samples, they say, the compounds are frequently reported at concentrations below the standard limits of detection and quantitation for the instruments in use.

If a machine cannot reliably see a molecule at that concentration, reporting it as present is not justified by the data. Jesmond Dalli and Esteban Gomez, lipid chemists whose own measurements were among those challenged, replied, defending their analytical methods and their identification criteria. The argument has run in short correspondence rather than in a study built to settle it, which is why it is still open.

What that challenges, and what it leaves alone, is worth being precise about. It contests whether these specific molecules can currently be quantified in human samples. It does not contest that resolution is an active program, because that claim rests on other evidence.

The timing of the class switch. The second COX-2 peak and the harm done by blocking it at the wrong hour. The reframing of chronic inflammatory disease as failed resolution. None of those depends on measuring femtogram quantities in a tube of blood.

The settled position is narrow. Resolution is a program. Whether we can yet read that program's molecules in a living person is a question about instruments, and it has no agreed answer today.

The inflammatory reflex was found by accident

Kevin Tracey's laboratory injected an experimental anti-inflammatory into a rat's brain at doses a million times below the effective intravenous dose, and swelling in a paw still fell. The messenger was a nerve, and the discovery reorganized neuroimmunology.

The vagus nerve page owns the anatomy of this nerve and the measurement of vagal tone. The immune half of the story begins with the accident, and the accident is worth telling properly.

Kevin Tracey ran a laboratory working on shock, the state in which the body's own inflammatory signaling floods the system and organs begin to fail. In the late 1990s the group was testing an experimental compound built to blunt that flood.

They injected it into the fluid spaces of a rat's brain, then measured inflammation in a paw at the far end of the animal. Doses at least a million times lower than the dose needed by vein still suppressed the swelling. Something was carrying the message, and it was not the drug, because there was not enough drug in the animal to do the job.

Cut the nerve and the effect vanishes

The tests came fast. The vagus is the long wandering nerve that runs from the brainstem down to the organs of the chest and abdomen, and it carries the body's main parasympathetic brake. At its endings it releases acetylcholine, the chemical the parasympathetic system speaks with. Cutting both vagus nerves in the neck abolished the effect. So did atropine, a drug that blocks one of the two receptor families acetylcholine acts on.

A receptor is a lock on a cell's surface, shaped to fit one key. Given the same compound by vein instead, recording from that nerve showed its firing rate climb from 87 to 229 impulses per second within minutes. Then the decisive step. Electrically stimulating the cut peripheral vagus reproduced the anti-inflammatory effect with no drug at all.

The compound had not been treating the inflammation. It had been pressing a switch, and the switch was a nerve.

Acetylcholine turns down the alarm

The same year, in the paper that anchors this field, the group took the mechanism apart. Acetylcholine added to cultured human macrophages cut their release of TNF, interleukin-1 beta, interleukin-6 and interleukin-18, while leaving the anti-inflammatory interleukin-10 untouched. That selectivity is the finding to keep. The nerve's transmitter turns down the alarm signals without shutting the cell down. In living rats given a lethal dose of endotoxin, stimulating the vagus prevented shock.

Naming the reflex

Two years later Tracey named it the inflammatory reflex. The argument for the name is simple. Too little inflammation kills you and too much kills you, so the quantity has to be regulated continuously, and the nervous system regulates continuous quantities by reflex.

A reflex is a loop with three parts. Something senses, something computes, something is sent back out, and none of it reaches awareness. Blood pressure is held that way. Body temperature is held that way. On this account, so is the size of an immune response.

How the nerve's signal reaches the macrophage

A claim that a nerve quiets a macrophage owes three answers: which receptor hears the signal, where the meeting happens, and what delivers the transmitter. The answers were the alpha-7 receptor, the spleen, and a T cell that makes acetylcholine itself.

The answers arrived in that order, and each one created the next question.

The alpha-7 receptor

The lock came first. Hong Wang and Luis Ulloa, working in Tracey's group, found it. It is the alpha-7 subunit of the nicotinic acetylcholine receptor, the same receptor family nicotine acts on. Acetylcholine failed to suppress TNF in macrophages lacking alpha-7.

Vagus stimulation, which reliably cut TNF in normal mice, did nothing at all in mice bred without it. Those are knockout animals, and alpha-7 sits on many cell types besides macrophages, so the experiment establishes that the receptor is required without establishing whose receptor matters.

The spleen, and a noradrenergic surprise

The place came next. The spleen is a fist-sized organ in the upper left abdomen that filters blood and houses a large resident population of immune cells, and it generates much of the body's systemic cytokine response. Mauricio Rosas-Ballina and colleagues in the same laboratory worked in mice with endotoxemia, the controlled whole-body inflammation produced by injecting endotoxin. Vagus stimulation suppressed TNF specifically in macrophages of the spleen, in the red pulp and the marginal zone, rather than everywhere at once.

Nerve endings lay close against those macrophages. Those endings did not carry the machinery for making acetylcholine, and they were noradrenergic, meaning they release noradrenaline. That is the transmitter of the sympathetic nervous system, the accelerator branch that runs a body under demand. Cutting the splenic nerve abolished the effect, and so did emptying its terminals of their transmitter stores.

The T cell that speaks acetylcholine

Which leaves a hole in the chain. The macrophage needs acetylcholine, and the nerve inside the spleen cannot make any. The same group found the missing link among the spleen's T cells. A T cell is a white blood cell that learns, carrying a record of what the body has met before.

A small population of them, bearing that memory signature, turned out to synthesize acetylcholine themselves. Remove those cells and vagus stimulation no longer suppressed cytokine production. This was mouse work using adoptive transfer, which means the cells were taken from one animal and put into another, and whether a human equivalent carries the role is unestablished.

Pause on what that means. A synapse is the narrow gap where one nerve cell hands its signal to the next. The last handoff in this circuit is not to a nerve cell at all. It is to a lymphocyte, an immune cell doing the work of a nerve terminal. At that final gap the distinction between the two systems stops being useful, which is what coupling means at its most literal.

Which nerve carries the outgoing arm is disputed

The vagus does not enter the spleen. A Melbourne group labeled 883 spleen-projecting neurons, found no vagal synapse on any of them, and concluded that sympathetic splanchnic nerves carry the outgoing arm of the inflammatory reflex.

There is a connection in that account nobody had ever seen, and the Melbourne group went looking for it.

Tracing the missing synapse

The problem is anatomical. The vagus does not enter the spleen. For the reflex to run outward through the vagus, its motor neurons would have to hand off to the sympathetic neurons that do supply the spleen, and that handoff had been proposed rather than observed. The physiologists Robin McAllen, Davide Martelli and Michael McKinley, with colleagues, tested it with tracers. A tracer is a substance a nerve cell takes up and carries along itself.

Different tracers travel in different directions, so two of them meet at a connection. They put a forward-traveling tracer into the brainstem cluster of cells that sends the vagus out to the body in rats, and a backward-traveling tracer into the spleen. They waited seven to nine weeks and looked. They labeled 883 spleen-projecting neurons and found no synaptic contact from vagal terminals onto any of them. Stimulating the vagus produced no change in splenic nerve activity, or in fifteen individually recorded spleen-projecting cells.

If the vagus is not the outgoing wire, what is? The same group gave anesthetized rats endotoxin and measured TNF ninety minutes later. Cutting both vagus nerves changed nothing. Cutting the greater splanchnic nerves, which are sympathetic, raised TNF sharply, and recordings showed splanchnic and splenic traffic climbing once the endotoxin arrived. Their title states the conclusion. Reflex control of inflammation by sympathetic nerves, not the vagus.

Notice how much of the original account their critical review accepts. Stimulating either cut vagus suppresses inflammation. The spleen is where the systemic response is generated and where it is suppressed. The effect requires noradrenergic terminals in the spleen. The acetylcholine-making T cells are an essential non-neural link. Alpha-7 receptors are essential. They reject one thing. The direct vagus-to-splenic-nerve connection, and with it the claim that this particular pathway is the outgoing arm of the body's own reflex to endotoxin.

What the vagus does carry

Then they showed it directly. Selectively stimulating the sensory fibers of the abdominal vagus cut TNF by 88 percent. Cutting the vagus upstream blocked that effect. So did cutting the splanchnic sympathetic nerves, and vagal sensory stimulation was shown to raise sympathetic firing reflexly. The vagus is the incoming arm. The splanchnic sympathetic nerve is the outgoing arm. The two are one loop.

A narrow disagreement, and a warning about shams

The disagreement is narrower than it looks. Two laboratories differ about which wire carries the outgoing signal. They agree that a reflex exists, that it is powerful, that it runs in every healthy animal, and that it sets the size of an inflammatory response. Coupling and gain survive the argument untouched, because neither depends on the anatomy landing one way. The model does not need the contested version.

That experiment carried an incidental finding. Merely exposing the nerve surgically was itself anti-inflammatory. Anything done to this system is an input to it, which is why a sham arm is not optional in any study of a physical input. A sham arm is a control group given everything except the active ingredient of the procedure. It is matched for contact, time and attention, so the effect of being handled can be subtracted from the effect of the treatment.

The brain reads the immune system the way it reads temperature

Cut the vagus below the diaphragm and interleukin-1 beta no longer produces a fever. Immune status is a sense, reported to the brainstem continuously, and what the brain does with the report is reorganize behavior around the illness.

The incoming arm was mapped first, and it changes what the immune system is.

Cytokines are large proteins, and the brain is walled off from the blood by a barrier that keeps most large molecules out. So how does the brain find out there is an infection? For a long time the assumed answer was that enough cytokine gets through by blood. Linda Watkins and Steven Maier, who spent decades on how immune signals reach the brain, tested it differently.

They cut the vagus nerve below the diaphragm, let the animals recover, then gave interleukin-1 beta and waited for the fever. The fever was blocked. The immune signal had been traveling up a nerve. Later work found several parallel routes, so the vagus is one channel rather than the only one.

State the consequence plainly. Immune status is a sense. The body reads its own inflammatory state the way it reads temperature or blood pressure, continuously, and reports it to the brainstem below awareness.

Two reports, not one alarm

How finely it reads was shown recently by a laboratory better known for its work on taste. Charles Zuker's group asked whether the brain tracks an immune response as it unfolds. Pro-inflammatory and anti-inflammatory cytokines signal separate populations of vagal sensory neurons, so what reaches the brainstem is two distinct reports rather than one alarm. The brain then modulates the response it is hearing about.

Silencing the circuit produced unregulated, out-of-control inflammation. Activating the same circuit suppressed the pro-inflammatory arm and strengthened the anti-inflammatory one. This is mouse work, using genetic silencing and chemogenetic activation, which means the neurons were engineered so a drug could switch them on and off. Whether the circuit can be reached in a person is untested.

Sickness behavior is a strategy

Then ask what the brain does with the report. Everybody already knows the answer from the inside. Appetite gone, interest gone, the day spent lying still in a dim room.

Benjamin Hart, a veterinary behaviorist, asked in 1988 whether that state is the illness happening to an animal or something the animal is doing. Reviewing the fever literature of his period he argued the second. Lying still conserves the energy a fever costs, and a fever costs a great deal. Not eating withholds the iron invading bacteria need to multiply. Sickness behavior is an organized strategy at a moment when the outcome is in doubt.

The human version was tested with an unusually clean design. Twenty healthy men received either a very low dose of bacterial endotoxin or saline, double-blind and in random order, on two separate occasions. Double-blind means that neither the volunteer nor the investigator knew which had been given. There was no visible illness. Blood pressure and heart rate did not move. Body temperature rose half a degree.

Cytokines rose. So did anxiety and depressed mood, by a moderate amount, and both verbal and non-verbal memory got measurably worse. Effect sizes were around 0.55 to 0.66, which means the shift was a little over half the size of the ordinary spread between one person and the next. Twenty young men, and a laboratory model of infection rather than an infection.

The mind moved before the body showed anything.

From sickness to depression

Robert Dantzer and colleagues, who spent twenty years on how immune signals change behavior, followed the signal further. The brain answers a cytokine with the coordinated state described above, and that state ends when the signal ends.

Where peripheral immune activation continues without stopping, in systemic infection, cancer or autoimmune disease, the same signaling can carry the state past sickness into clinical depression in vulnerable people. That link rests on association plus a plausible mechanism, and anti-inflammatory drugs have not proved generally effective for depression. An anti-TNF depression trial, split by C-reactive protein, shows why.

Through the model the reframe is exact. The defensive state is not the pathology. Entering it is a competence. The pathology is the narrowing of the range, so that a system can enter the state and cannot leave it. Health is the width of what a body can do and then undo, and that width is a property of its tone. The depression page carries that argument further.

A reflex sets the gain on the inflammatory response

Removing a nerve rather than stimulating one produced the sharpest number in the neuroimmune literature: the same challenge coming out roughly fivefold larger. The inflammatory response of an intact animal runs under continuous restraint nobody feels, a setting rather than a quantity.

Cutting the greater splanchnic nerves in an anesthetized rat given endotoxin raised TNF roughly fivefold. Nothing was being stimulated. A nerve was removed, and the TNF response came out five times larger. Which means that in the intact animal, those nerves had been holding down about 80 percent of the TNF response to that challenge, silently, while nobody was doing anything to it.

That is gain. Gain is how loudly a system answers relative to what it was asked. It is neither the input nor the output. It is the ratio between them, and here the ratio is set by a reflex nobody feels.

A dial, read from both sides

The dish experiment says the same thing from the other side. Acetylcholine cut TNF, interleukin-1 beta, interleukin-6 and interleukin-18 from human macrophages while leaving interleukin-10 alone. An off switch would have quieted the cell. What happened instead was a change in proportion, which is what a dial does. The brainstem circuit in the last section behaved the same way. Removing it did not lower inflammation. It let inflammation run loose.

Two directions of failure

A dial can sit wrong in either direction, and both failures are recognizable. Let the restraint slip and an ordinary provocation produces a response out of all scale with it, which is the shape of a flare that seems to arrive from nowhere.

Hold the restraint too hard and the response is too small to clear what provoked it. The catastrophic version of the first already has a name. Sepsis is organ failure produced by a dysregulated response, the same setting failing fast enough to kill within hours.

The septic shock trial

One trial shows what follows from reading a setting as a quantity. Too much TNF kills people in septic shock, and animals given a TNF blocker survive it. So 141 patients received placebo or one of three doses of a soluble TNF receptor.

Mortality ran 30 percent on placebo, 30 percent on the low dose, 48 percent on the middle dose and 53 percent on the high dose. More drug meant more death, and the trend was statistically solid. How abundant a mediator is does not tell you what removing it will do inside a system already failing to regulate itself.

The setting also accounts for something clinics see constantly and cannot explain. Two people meet the same trigger, the same virus, the same injury, the same bad week, and one of them runs a response several times the size of the other's. Nothing about the input explains that. The tone that received it does.

A setting is a different kind of thing from a quantity. A quantity can be lowered by removing some of it. A setting has to be changed, which is a different job with different evidence behind it, and section 18 takes it up.

A nerve can produce inflammation, and it can hold it down

Stimulating sensory nerves in rats produced plasma leak and swelling with no injury and no pathogen anywhere, a result Hungarian pharmacologists published in 1967. The nerve was not reporting the inflammation. It was making it.

So far the nerve has been reading the immune system and turning it down. It does the opposite as well.

A group of Hungarian pharmacologists, among them Janos Szolcsanyi, asked whether a sensory nerve could produce inflammation rather than report it. Stimulating sensory nerves in rats produced the full local picture, plasma leaking out of the vessels and the tissue swelling, with no injury and no pathogen anywhere. Cutting the nerve and waiting for it to degenerate prevented it. So did pre-treating with capsaicin, the compound that makes chili peppers hot, which selectively destroys those particular endings.

Neuropeptides open the vessels

What the ending releases are neuropeptides, short chains of amino acids built in the nerve cell body and stored at the tip. Two matter most, substance P and calcitonin gene-related peptide. The second had just been discovered, and nobody knew what it did to blood vessels. Susan Brain, a pharmacologist, and the team she worked in put it into skin in femtomole amounts, quantities almost too small to weigh.

In a rabbit it opened the microcirculation and raised blood flow. In human skin it produced a lasting local redness. In a hamster's cheek pouch it visibly widened the arterioles, the small vessels that meter blood into a tissue. That is the first of Celsus's four signs, produced by a nerve firing backward into the tissue it was supposed to be monitoring.

Wiring density sets severity

How much this matters to a real disease showed up in rats with experimental arthritis, where some joints reliably get worse disease than others in the same animal. The pain researcher Jon Levine, working in the team that ran the study, asked why. Ankles, which developed severe arthritis, turned out to be far more densely supplied with substance P fibers than knees, which developed mild disease.

Infusing substance P into the knee made its arthritis worse. The severity of an inflammatory disease at a site tracked how heavily that site was wired. The blocker available at the time, a molecule that occupies a lock without turning it, did not reduce severity, so the causal chain was left incomplete in that paper.

The model reads that result as a distortion surfacing where the local wiring gives it least resistance. It is one reason the same process shows up as one disease in one body and another in the next.

Pain neurons restrain the immune response

Then the direction reverses again. Isaac Chiu and Clifford Woolf, neurobiologists asking whether a nerve does more in an infection than report it, infected mice with Staphylococcus aureus. Pain tracked the live bacterial load rather than the swelling or the immune response, because the bacteria were triggering the pain neurons directly. Deleting those neurons abolished the pain and increased local immune infiltration and lymph node swelling.

The pain nerve had been holding the immune response down. The autoimmune page tells that experiment in full. The two of them have also reviewed the wider picture. Pain-sensing neurons carry many of the same danger receptors immune cells carry, and they reach the scene first, because nerves move faster than cells.

Both results live in one system. A nerve that raises inflammation and a nerve that restrains it, in the same tissue, depending on what is happening. That is coupling with the sign reversed, and it is why treating inflammation as a substance the body has too much of cannot account for either result.

Inflammation has an amplifier inside the spinal cord

Microglia, the immune cells of the central nervous system, turn up pain from inside the cord. Blocking one microglial receptor reversed touch-evoked pain after nerve injury in rats, and injecting stimulated microglia produced the pain state from nothing.

Much of the nervous system is not made of neurons at all. It is glia, treated for a century as packing and housekeeping, and two kinds matter here. Microglia are the immune cells of the central nervous system. Astrocytes support and regulate the neurons around them. Both carry immune-type receptors.

Linda Watkins and Steven Maier's group argued that these cells amplify pain. Microglia and astrocytes respond to viruses and bacteria, release inflammatory cytokines directly into the cord, and those cytokines make pain worse. Certain sensory signals arriving from the body switch them on the same way an infection does. On that account persistent pain carries an immune component inside the nervous system itself, which is a different thing from inflammation in the tissue that hurts.

P2X4, the microglial mechanism

The crisp mechanism came from Makoto Tsuda and Kazuhide Inoue, cell-signaling researchers hunting the cell type responsible, working with the neuroscientist Michael Salter. After a nerve injury, ordinary light touch can become painful, a state called tactile allodynia. Blocking one spinal receptor reversed that hypersensitivity while leaving normal pain intact in uninjured animals.

The receptor is P2X4, and what it answers is ATP, the small molecule cells use to carry energy and spill when they are stressed. After injury P2X4 appeared only in microglia, not in neurons or astrocytes. Injecting stimulated microglia carrying that receptor into the spinal cord of a normal rat produced the pain state from nothing.

The minocycline trial, and what averaging hides

The human test has not gone well. Minocycline is an antibiotic that quiets microglia. The anesthesiologist Pascal Vanelderen and colleagues, at a multidisciplinary pain center, asked whether it relieves pain in people. They recruited sixty patients with subacute lumbar radicular pain, meaning pain running down the leg from an irritated nerve root in the low back.

The comparators were placebo and amitriptyline, an old antidepressant used at low dose for nerve pain, over fourteen days. Both active drugs beat placebo on the leg pain rating by about 1.5 points out of ten. Neither changed the neuropathic pain questionnaire at any time point. The authors judged the effect too small to be clinically meaningful. Twenty patients per arm, two weeks, one drug.

The model expects that result and says why. A trial that gives one fixed dose of one glia-directed drug to everyone carrying one diagnosis, for fourteen days, has averaged the people whose pain is currently glia-driven with the people whose pain is not. The mean it reports describes neither group.

The model states that expectation before the data arrive, and it is a claim about how the trial was assembled rather than a complaint about how it was run. Stratify the sample first, using a measure taken before the outcome is known, and the model predicts the average moves. The gain and pain pages carry that argument in full.

The immune system here is setting the volume in the pain system, a different measurement from the size of an inflammatory response in tissue. Both are gain. They are read with different instruments.

Inflammation is one face of a single state

Five literatures hold five discoveries: a scaling reflex, a sensory channel, nerve endings that start inflammation, glia that amplify it, and a resolution program. Read together they describe one organization, and the Unified Model of Tone calls that organization tone.

A reflex that sets the size of an immune response. A sensory channel reporting inflammation to the brainstem continuously. Nerve endings that produce inflammation on their own. Glia amplifying a signal inside the cord. A lipid program that ends the response on its own clock. Those are usually filed as five discoveries in five literatures. They are five readings of one thing.

This library calls that thing tone. A body runs countless processes at once, and tone is not any one of them. Tone is their organization, how all of them stand in relation to one another at a given moment, taken as one state rather than a list of parts. The body's oscillations carry that organization the way notes carry a chord. The two-day arc of an enzyme and the clock the second crew runs on are carriers.

Tone is the chord they compose, never any one note in it. Autonomic balance is one reading of it. Muscle tension is another. Cortical excitability is another. Inflammatory state is another, which is why it travels with sleep, with heart rhythm and with mood instead of sitting quietly in a compartment of its own. The tone page builds the variable from the beginning.

Coupling, gain and time course are that variable seen from three sides. Coupling is tone read between systems. Gain is tone read as how loudly the system answers. Time course is tone read across hours and years, which is whether a state can be entered and then left. One variable, three instruments.

One loop, many entrances

One organization has a consequence worth naming. A continuous loop can be entered at more than one point, and three entrances have already appeared. A drug reached the loop at an enzyme, and the tissue answered by building a different class of mediator. Stimulating a sensory nerve reached it through the body's sensory field, and plasma leaked into tissue with nothing to fight.

Cutting a nerve reached it through the traffic that nerve carried, and the same TNF response came out five times larger. Those are not the same act, and they are not interchangeable. They are inputs delivered at different places in one loop, and a loop carries a change introduced anywhere through the rest of itself.

Health is the range

Health, on this reading, is the range, because a body that cannot inflame dies of the first infection it meets. The capacity to mount a full response, hold it as long as the threat lasts, and then end it completely, including the ending's own chemistry. Disease is that range narrowing. The system keeps the response and loses the exit.

An inflammatory disease is a defense the body can enter and cannot leave.

A pattern that outlives its cause

One experiment makes the persistence concrete. Tamar Koren and Asya Rolls, neuroimmunologists asking whether the brain holds a record of a particular illness, tagged the neurons in the insular cortex of mice that fired during two different inflammatory illnesses. The insular cortex is the region that maps the body's internal state. Later, with the inflammation long resolved, they switched those tagged neurons back on, and the corresponding inflammation reappeared in the right tissue with no new trigger.

Mice, artificially tagged cells, untested in people, and the autoimmune page takes up what it might mean for a disease that attacks its own tissue. What it settles here is narrower. A pattern can persist after its cause is gone, because the pattern is held in how the system is organized rather than filed in a record the body has to look up.

The discipline that goes with a claim this size has to be stated in the model's own words. Every disease has a tonal expression, and many diseases are initiated, maintained or amplified by failures of tonal regulation. A tumor is not a bad tone. An infection is not a bad tone. What the model adds is an account of how an organism receives such a cause, what it does with it afterwards, and why the same insult runs a different course in two people.

Where the reading dominates, and where it does not

Where does this reading dominate? Where the trigger has gone and the response has not. Autoimmune disease, where the loop keeps running against the body's own tissue. Persistent pain, where the volume stayed up long after the injury healed. Gut inflammation, where the largest concentration of immune tissue in the body sits under a dense sensory supply. The depressive states that ride on continued immune activation. These are rarely one pattern.

The presenting inflammatory pattern commonly sits on a deeper driver, such as lost sleep, unresolved load, or a sensory field sensitized for years. Reading only the loudest layer is how a condition keeps returning unchanged. Where does this reading not dominate? A new infection needs the organism identified and treated. A tumor needs oncology. A single-gene disorder needs genetics. The reading earns its place after those questions are answered, never instead of them.

What a C-reactive protein result reports, and what it cannot

C-reactive protein can rise ten thousandfold within hours of almost any stimulus, and across 194,418 people, gene variants that raise it for life left heart disease risk unmoved. The marker reads the inflammatory state without producing it.

Somewhere in the last twenty years inflammation became a number people get told about, and the number is usually C-reactive protein.

Start with the claim underneath it. Twenty researchers across immunology, aging and psychology assembled the case for a low-grade inflammatory state running far below the level of any symptom. They argue it contributes upstream to most of the conditions responsible for disability and death. Cardiovascular disease, cancer, diabetes, kidney disease, fatty liver, autoimmune and neurodegenerative disease.

The drivers they list are ordinary life. Infection history, inactivity, diet, environmental exposures, disturbed sleep, sustained psychological demand. It is a perspective article with no agreed diagnostic threshold for the state it describes, and for several of those risk factors the direction of causation is unsettled.

A sensitive, non-specific gauge

C-reactive protein is made by the liver in response to interleukin-6. Mark Pepys, who spent a career on this one protein, described the two properties that matter to a patient. It can rise ten thousandfold, climbing within hours of a stimulus, peaking around two days and then clearing with a half-life of about nineteen hours.

That makes it exquisitely sensitive. It is also completely non-specific about where the inflammation is or what is driving it. A raised result says something is running. It does not say what.

It is also not a cause. Genetics supplies a natural experiment here, because the gene variants that set a person's lifelong CRP are assigned at conception and are not confounded by how anyone lives. Across 47 studies, 194,418 participants and 46,557 cases of coronary heart disease, those variants shifted CRP by up to 30 percent per copy.

They shifted heart disease risk essentially not at all, a risk ratio of 1.00 with confidence limits of 0.90 to 1.13. A risk ratio of 1.00 means the two groups had the same risk, and the limits are the range of values the data still allow. Measured CRP predicts events. Genetically raised CRP does not produce them.

Heart rate variability and inflammation

The other number in this argument is the heartbeat. Heart rate variability is the beat-to-beat variation in the interval between heartbeats. The vagal brake acts fast enough to change that interval within a single beat, so the size and structure of the variation index how much vagal restraint a person is running. The vagus nerve page and the page on the measurement itself carry that story in full.

The psychophysiologist Julian Thayer ran the obvious test. If a reflex restrains inflammation, people running more of that restraint should carry less inflammation. His group measured 611 apparently healthy employees of an aircraft plant in southern Germany, taking heart rate variability and overnight urinary noradrenaline together, so the vagal and the sympathetic contributions could be told apart.

The vagally mediated component was inversely related to CRP even after controlling for sympathetic activity, and more strongly in women than in men. It is cross-sectional. It cannot say which came first.

The literature as a whole is weaker than that one study suggests. A meta-analysis combines many studies into a single estimate. One that screened 2,283 studies and pooled 51 found the expected direction overall, higher vagally mediated variability with lower inflammatory markers.

Two indices carried it, the plain spread of the beat-to-beat intervals and the fastest component of that variation, which is the one that tracks the vagal brake most closely. The same analysis found a genuinely mixed literature, with a number of studies reporting the opposite association. The relationship is real, weak and inconsistent. Report it that way.

A single resting mean of one marker, set against a single resting index in a sample nobody stratified, compares two quantities. It does not read the coupling between them, and the coupling is where tone lives. The model's expectation is that dynamic measures carry the information resting means do not.

How far the system moves when challenged. How fast it comes back. Whether the two move together while it happens. That is a prediction with a way to fail, and this literature is the reason to state it.

Blocking a messenger and changing a setting are different achievements

In 1994 one anti-TNF infusion moved rheumatoid arthritis in 19 of 24 patients against 2 of 24 on placebo, and the era it opened is among the finest things medicine has done. Restoring the reflex that sets the inflammatory response is a different job, and it has barely begun.

An antibody is a protein the immune system makes to grip one specific target and mark it for removal. A monoclonal antibody is one such protein, chosen for its target and then manufactured in quantity. A biologic is a drug built out of a large protein of that kind rather than a small chemical, which is why these drugs are infused or injected rather than swallowed.

The anti-TNF era and its price

In 1994, 73 patients with active rheumatoid arthritis received a single infusion of a monoclonal antibody against TNF or placebo, double-blind, across four centers. The immunologist Marc Feldmann and the rheumatologist Ravinder Maini were among the authors, and they had spent years arguing an unfashionable case. In a joint full of cytokines shouting at once, they held that one of them sat far enough above the others that removing it alone would move the disease.

At four weeks, 2 of 24 placebo patients had responded. Nineteen of 24 on the high dose had. At their peak, mean improvements in tender and swollen joint counts and in CRP exceeded 60 percent on the high dose. One infusion, one molecule blocked, and a disease that crippled people was altered. That trial opened an era, and the drugs that followed are now standard care in rheumatoid arthritis and in several other diseases.

The cost was then found, measured and managed, which is what medicine working looks like. In animals TNF is the molecule that holds tuberculosis inside its granuloma, the walled-off structure that keeps a dormant infection dormant. The human evidence arrived as an unwelcome surprise.

A safety review at the United States regulator found seventy cases of tuberculosis after infliximab treatment, most within three infusions, forty of them outside the lungs and seventeen disseminated. Most came from countries where tuberculosis is rare. Screening for latent tuberculosis before starting these drugs is now routine, and that review is why.

Those are different achievements rather than a better one and a worse one. The line runs by aim and not by instrument. A molecule that switches on the body's own resolution program is doing restorative work, and a physical input delivered only to quiet a report is doing masking work.

Every profession contains both. For a person whose joints are being destroyed now, the messenger being blocked is exactly the right thing to be happening, and none of it is a reason to decline or delay it.

That the line can run across the professions says something about the system they share. They are not working on separate ones. What differs between them is the entrance. A biologic enters at a receptor. A current enters through the traffic a nerve already carries. Diet and sleep enter as metabolic substrate and as recovery, which is why they sit on the same list of drivers as an old infection. Different places to stand beside one loop, and only the aim sorts them.

Two heart trials, one lesson

Blocking also has to hit the pathway that is actually running, and two trials from the same group make the point precisely. In the first, 10,061 people who had already had a heart attack and had CRP at 2 mg per liter or higher received an antibody against interleukin-1 beta or placebo.

CRP fell 26 to 41 percent below placebo depending on dose, and lipids did not move. The 150 mg dose reduced further cardiovascular events. Inflammation was shown to be a target in its own right.

The trial was industry funded, the benefit was modest and dose-dependent, and fatal infection was more common on the drug, which is the price of removing a defensive signal. In the second, 4,786 high-risk patients received low-dose methotrexate, an old anti-inflammatory long used in rheumatoid arthritis.

It did not lower interleukin-1 beta, interleukin-6 or CRP, and it did not reduce events, a hazard ratio of 0.96. A hazard ratio compares how fast events accumulate in the two groups, and 0.96 is as close to no difference as the arithmetic gets.

The pair says one thing. Lowering inflammation helps when you lower the pathway actually running in that patient, and does nothing when you do not. There is more than one dial, and turning down the wrong one changes nothing. The diagnosis names the endpoint and never the tone that produced it, which is why the same strategy succeeds in one trial and fails in the next.

The reflex as therapy, so far

So how far has the reflex itself come as a treatment? Not far. The rheumatologists Frieda Koopman and Paul Tak, working with Tracey, implanted stimulators in seventeen rheumatoid arthritis patients, and TNF production fell as disease scores improved. Open label, no sham arm, and the vagus nerve page tells it in full. Open label means everyone knew who was being treated. A later first-in-human study tested a miniaturized device in fourteen patients whose arthritis had failed two or more advanced therapies.

It was powered for safety rather than efficacy, and it did include a small masked sham arm, four patients implanted with the device left switched off. In Crohn's disease, nine patients had an electrode placed on the left cervical vagus and stimulated for a year, with five in clinical remission at twelve months and no control group at all. Crohn's disease remits and relapses on its own. The gut page carries that axis in detail.

The one clean controlled test came out negative. Healthy volunteers were given endotoxin to produce a controlled inflammatory response. Ten had a stimulating catheter threaded into the jugular vein alongside the vagus and stimulated for thirty minutes. Ten had the catheter and no current, and everyone was masked. The stimulation demonstrably reached the nerve, because subjects felt their larynx vibrate. Fever, symptoms, hemodynamics and the entire cytokine response were the same in both groups.

The biology is established. The human therapeutic evidence is small, mostly unblinded, and it includes a well-designed null. That is what early looks like.

What a restoring input does in inflammation that a masking one cannot

A restored regulator moves a dysregulated value toward the middle from either side, a signature no directional drug can produce. The test has a design: stratify by tone in advance, fix the input in advance, and match the sham for contact.

A model that explains this much owes a result it can be held to.

A drug is directional by construction. It lowers a value in everyone who responds, including people in whom the value was already low, because that is what a molecule with one mechanism does. A restored regulator behaves differently. It moves a dysregulated value toward the middle of its healthy range from whichever side that person sits on, because what has been corrected is the system's ability to find the middle.

One result above carries that shape. Silencing the brainstem circuit let inflammation run loose, and activating it moved the two arms of the response in opposite directions at once. That is a regulator behaving like a regulator. It is two arms of one response moving oppositely rather than one variable converging from either side, so it is the shape of the signature and not the signature itself.

The same input, a different system, a different event

Two further results belong to a different claim, and collapsing them together would overstate the case. Their claim is the law this whole page runs on. The same input, meeting a differently organized system, becomes a different event.

The first is about site and strength. Shenbin Liu, Qiufu Ma and colleagues, neurobiologists asking why an identical stimulus produces different immune effects at different places on the body, used electroacupuncture in mice with endotoxin-driven inflammation. Electroacupuncture is a fine needle placed at a defined point on the body and driven with a small electrical current, so the site and the strength can both be specified exactly.

Low-intensity stimulation at the hindlimb drove a vagal-adrenal pathway and was anti-inflammatory. High-intensity stimulation at the abdomen drove a spinal-sympathetic pathway to the spleen, and that one came out anti-inflammatory or pro-inflammatory depending on the animal's disease state and its receptor profile at that moment. What selected the pathway was where the input landed and what state it met. Strength is a second axis and not the same one.

The second is closest to a person, and it is a drug rather than a restored regulator, which is exactly why it belongs here. An anti-TNF antibody was tested in sixty outpatients with treatment-resistant depression, three infusions across twelve weeks. Overall it did nothing, 50 percent responding in each arm. Baseline CRP interacted significantly with treatment. Above 5 mg per liter the drug outperformed placebo, 62 percent against 33 percent.

At or below 5 mg per liter the direction reversed, 41 percent against 57 percent. Be careful with both halves. Neither comparison on its own reached statistical significance, which means neither subgroup held enough patients to rule out chance. The split between the two subgroups did hold up, and the analysis was exploratory in sixty patients. One molecule at one dose helped one group and left the other no better, with the direction reversed. Averaged together, it described nobody.

Stress behaves the same way, which is why the sentence about stress raising inflammation cannot be settled as written. In mice, circuits that drive movement pushed neutrophils out of the bone marrow into the tissues within minutes. At the same time a small region of the hypothalamus, the paraventricular nucleus, pulled monocytes and lymphocytes the other way, back into the marrow.

The hypothalamus is the small area at the base of the brain that runs the body's housekeeping. It did this with cortisol-like hormones, the body's own steroid stress hormones, acting directly on those cells. The two shifts changed disease susceptibility differently. The hypothalamic shift protected against acquiring autoimmunity while impairing defense against influenza and SARS-CoV-2.

The test, stated in advance

Now the test itself. Take a cohort and stratify it in advance by a measure of tone recorded before any outcome is known. One group sits above its healthy window on a given inflammatory or autonomic variable, and one group sits below it. Fix the input and the site from that same measure, in advance, so that matching cannot be judged after the fact. Deliver it to half of each group, and give the other half a sham matched for contact, time and attention.

The model predicts convergence toward the middle in the treated arms, exceeding whatever the sham arms produce, and it predicts several separately regulated measures sharing that direction in the same person, at the pace compensation allows each one.

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. Convergence no greater than sham marks it the same way. Given that surgical exposure of the nerve alone suppressed inflammation, the sham arm is the whole experiment.

What to record is not exotic. Heart rate variability and its internal structure. The trajectory of an inflammatory challenge rather than a single value, meaning how high it goes, how fast, and above all how completely it resolves. Recovery time. Sleep. A system regaining its range shows it in how it returns, and returning is the measurement almost nobody takes.

One question remains open. The ending of inflammation is a program with its own chemistry and its own clock, and nobody has yet shown that program being started on purpose in a living person. The instrument that would settle it is not exotic. Take a person through a controlled inflammatory challenge, before and after a specified input, and measure the resolution phase instead of the peak. Nobody has run that study.

How inflammation relates to the rest of the library

Inflammation is where the immune system and the nervous system stop being separate subjects: one reflex scales it and one nerve reports it. The same failure to close appears as autoimmune disease, persistent pain, gut inflammation and depression.

The foundations that carry inflammation's signature each hold their own page. Coupling is why a splenic T cell can do the work of a nerve terminal. Gain is why cutting one nerve quintuples a cytokine response. Time course is why a chronic inflammatory disease can be an acute program stalled halfway.

Frequently asked