Anxiety and the Nervous System

What anxiety is before it is a disorder

Anxiety is a normal function of a healthy nervous system. It is the body leaning forward into a future that has not arrived yet, raising its readiness in case that future turns out to be dangerous.

Two words get used as if they were the same thing, and separating them makes everything that follows clearer. Fear answers a threat that is here. A car swerves toward you and your body responds before you have finished seeing it. Anxiety prepares for a threat that might come. The road looks icy, the light is failing, and your hands tighten on the wheel although nothing has happened yet.

Both are useful. A person who cannot feel fear walks into traffic. A person who cannot feel anxiety does not check the weather, does not prepare for the meeting, and does not slow down on the ice. The forward lean is what lets a body meet a future it cannot yet see. Evolution did not install this state as a flaw.

So the disorder is the loss of the ability to leave the state, never its presence. The alarm sounds and does not stand down. The readiness that should last for the length of a threat lasts for months. The body stays braced against a danger that has either passed or never arrived, and the bracing itself becomes the problem.

This distinction matters for how a person understands themselves. Anxiety disorders are among the most common conditions in general medicine. The internist Kurt Kroenke measured how often they appear in ordinary clinic visits and found them frequently present and frequently missed. In his sample of 965 primary care patients, 19.5 percent carried at least one anxiety disorder, the disorders impaired daily function, and 41 percent of those affected were receiving no treatment for it.

None of what follows describes a broken person. It describes a control system that has narrowed, and the narrowing has a physiology.

Two autonomic lines wire anxiety into the body

Every bodily symptom of anxiety arrives through two nerve supplies: a sympathetic line that mobilizes and a vagal line that restores. The racing heart, the dry mouth, the churning gut and the cold hands are these two lines at work.

One word first, because everything that follows rests on it. A nerve is a living wire. It is a bundle of fibers that carries messages through the body as tiny electrical pulses, the way a cable carries a signal, except that this cable is alive and can change what it does.

Your organs are wired by two of these lines, and the two pull in opposite directions. The first is the accelerator, called the sympathetic system. When it fires, the heart speeds up, the vessels tighten, the airways open, the pupils widen, digestion is set aside, and the adrenal glands release adrenaline into the blood. This is the body configured for effort and for danger. Everything a person notices during a frightening moment is this line doing its work.

The second is the brake, called the vagal or parasympathetic system. Its main cable is the vagus nerve, which runs from the brainstem down through the neck and chest into the abdomen. When it acts, the heart slows, the body settles, digestion resumes, and the physiology of rest returns. The physiologists Erica Wehrwein, Hakan Orer and Susan Barman set out the modern architecture and pharmacology of this whole system, and their account is the plain map most clinicians work from.

Hold on to these two words. Accelerator and brake. Sympathetic and vagal. They are the two hands on almost everything anxiety does to a body: the racing heart, the dry mouth, the churning gut, the cold hands, the shallow breath.

One correction now, because it shapes everything later. The two lines are not a seesaw, with the accelerator up meaning the brake down. Sympathetic and vagal drive move independently, and in an anxious body both can be high at once. The physiology behind that correction arrives with the tone reading.

The amygdala fires before the thinking brain finishes

The amygdala can trigger the entire body response to a threat before the cortex has identified what it saw. In anxiety disorders it answers more readily, and the imaging shows it.

Deep inside each temporal lobe, roughly behind and above the ear, sits an almond-sized cluster of cells called the amygdala. It works as an alarm. It receives sensory information early and fast, and it can set off the full sympathetic response while the slower parts of the brain are still deciding what happened.

The neuroscientist Joseph LeDoux, who has spent his career at New York University working out how the brain learns fear, wanted to know exactly which part of the amygdala receives the incoming signal. His laboratory used a simple arrangement. An animal hears a tone, and the tone is followed by a mild shock. Very quickly the tone alone produces the whole defensive response.

LeDoux and his colleagues then asked which piece of the amygdala the sound has to reach for that learning to happen. They found that the lateral nucleus is the sensory interface of the amygdala. Damage it, and the animal never learns to fear the tone. The alarm has an input wire, and the wire is short.

That shortness is the point. A signal can reach the amygdala through a fast, coarse route before the slower cortical route has finished making sense of it. This is why you are already off the path before you know whether the shape on the ground was a snake or a stick. The alarm is designed to be wrong sometimes, because a false alarm costs a jolt of adrenaline and a missed alarm costs everything.

The alarm answers more readily in anxiety disorders

In people with anxiety disorders, this alarm tends to answer more readily. The psychiatrists Amit Etkin and Tor Wager pooled the brain imaging studies of anxious patients and healthy volunteers looking at threatening material, to see whether a consistent pattern existed across different diagnoses.

Their meta-analysis found greater amygdala and insula activity in post-traumatic stress disorder, social anxiety disorder and specific phobia. The alarm is louder. That finding is real, and it is also incomplete, because it describes reactivity to a cue. Most of what troubles an anxious person has no cue at all.

Reacting and waiting run on different brain structures

Brief fear of something present and sustained anxiety about something possible separate anatomically. The amygdala drives the first. A neighboring structure, the bed nucleus of the stria terminalis, drives the second.

The neuroscientist Michael Davis, working at Emory University, measured fear a clever way. He used the startle reflex, the involuntary jump you make at a sudden loud noise, which grows larger when an animal or a person is already afraid.

That gives a number for a state that is otherwise hard to see. Together with the clinical neuroscientist Christian Grillon at the National Institute of Mental Health, Davis asked whether a brief fear of something present and a sustained dread of something possible run on the same structure.

They do not. Their review of the rodent and human work concluded that brief cued fear and sustained uncued anxiety separate anatomically. The amygdala drives the short, cued alarm, the response to a signal that predicts harm. The bed nucleus of the stria terminalis, part of what anatomists call the extended amygdala, drives the long, uncued state of vigilance. One address for reacting to something bad. A different address for waiting for something bad.

Human imaging supports the same division. The psychologist Leah Somerville and colleagues at Dartmouth scanned volunteers to see which structures track a person's ongoing watchfulness rather than their response to a specific frightening picture. They found that activity in the human bed nucleus indexed hypervigilant threat monitoring, tracking sustained scanning for danger rather than a reaction to a particular cue.

What fails in anxiety is the standing down. The alarm itself is usually the right size.

Panic patients cannot stand down from unpredictable threat

Grillon then asked a sharper question. Are people with panic disorder simply hyper-reactive to all threat, or specifically to threat they cannot predict? He measured startle under two conditions: a shock announced by a signal, and a shock that could come at any time. Patients showed a normal response to the predictable threat and an exaggerated one to the unpredictable. When the danger was scheduled, they were unremarkable. When it could not be scheduled, they could not stand down.

Read that finding carefully, because it reframes the whole condition. The alarm is intact. The release is what fails.

Extinction builds a brake rather than erasing fear

The prefrontal cortex holds the brake on the amygdala, and extinction is how that brake is built. The original fear memory is never deleted. A competing safety memory is laid on top of it, and it can be lost.

The alarm does not run unsupervised. The front of the brain, the prefrontal cortex, can inhibit the amygdala. It is the structure that lets a person notice that the shape on the ground was a stick after all, and stand the body down.

To see how that brake is built, you need one idea from the laboratory called extinction. Take the animal that learned to fear a tone. Now play the tone many times with no shock following. The fear response fades. The obvious reading is that the memory has been erased. That reading is wrong, and knowing why it is wrong explains a great deal about why anxiety comes back.

The cells that hold the safety learning

The neuroscientists Mohammed Milad and Gregory Quirk wanted to find the cells that hold the new safety learning. They recorded from neurons in the medial prefrontal cortex of rats across fear learning and extinction, looking for cells whose firing tracked whether the animal remembered that the tone was now safe. They found them. Certain prefrontal neurons signaled memory for fear extinction, and how strongly those cells fired predicted how well the animal held on to its safety learning the next day.

So extinction lays a new memory on top of an old one. The original fear memory remains. What has been added is a competing memory, held in the front of the brain, which suppresses it. This is why old fears return under stress, after time, or in a new setting, and why a person who has recovered can be ambushed by the same feeling in an unfamiliar place.

Exposure therapy trains the brake

This is also exactly what good exposure therapy is doing. The clinical psychologist Michelle Craske and colleagues rebuilt the therapy around this understanding, arguing that treatment works by strengthening new inhibitory learning rather than by removing the old fear. Practicing in many contexts, tolerating the uncertainty, and violating the expectation of catastrophe all build a stronger brake. The alarm is never deleted. The system that governs it is trained.

That is worth stating plainly for anyone who has done this work and found the fear still there. Its presence is not failure. The aim was never erasure.

Cortisol's job includes ending the stress response

The slow arm of the alarm is a three-station hormone chain ending in cortisol, and a major class of cortisol's actions is suppressive. The hormone released at the peak of the anxiety response is part of the machinery that closes it.

The alarm has two arms, and they run on different clocks. The fast arm is the one we have met: nerves firing, adrenaline released, everything changing within seconds. The slow arm is chemical, and it takes minutes to hours.

It works as a chain of three stations. A region at the base of the brain called the hypothalamus sends a chemical signal to the pituitary gland just beneath it. The pituitary sends its own signal into the bloodstream. That signal reaches the adrenal glands, which sit like caps on top of the kidneys, and the adrenals release cortisol.

Physiologists call this chain the HPA axis, after its three stations. The neuroscientists Yvonne Ulrich-Lai and James Herman assembled how the brain governs it, describing the neural regulation of endocrine and autonomic stress responses as one coordinated output rather than two separate systems.

Now the part that is usually left out, and that changes the whole picture. A large part of cortisol's job is to contain and end the very response it belongs to. The neuroendocrinologist Robert Sapolsky and colleagues examined decades of work on what glucocorticoids actually do. They concluded that a major class of these actions is suppressive, restraining the stress response so that it does not overshoot and damage the body it protects.

Read that again in plain terms. A stress system that works is one that turns itself off. The hormone released at the peak of the response is part of the mechanism that closes it. Health here is a stress response with a functioning ending, and an anxiety disorder is a response whose ending has stopped arriving.

Cortisol measurements do not sort anxious from calm

The cortisol evidence tests the simple version of this story. If anxiety were too much cortisol, cortisol measurements would sort anxious people from calm ones cleanly. They do not. The psychiatrists Hesham Elnazer and David Baldwin worked through the cortisol studies in panic disorder, generalized anxiety disorder, the specific phobias and social anxiety disorder.

The findings were interesting one at a time and would not line up. Across the diagnoses they found no consistent pattern of endocrine disturbance, which they put down in part to differences in method and in the size and nature of the samples.

A model built on a single overactive chemical has to call that result disappointing. A model built on regulation calls it expected. What is dysregulated in anxiety is a loop with a beginning and an ending, and a loop can fail at either end. A system that cannot mount the response and a system that cannot close it are both out of range. A single hormone level at a single moment cannot tell you which one you are looking at.

The locus coeruleus sets how loudly anxiety hears everything

A few thousand neurons in the brainstem deliver noradrenaline to almost the entire brain, and their firing rate works as a gain control. In an anxious nervous system that setting sits high, so ordinary input starts to feel urgent.

In the brainstem, the stalk where the brain meets the spinal cord, sits a tiny blue-black patch of cells called the locus coeruleus. The name means blue spot. It contains only a few thousand neurons on each side, and yet it sends fibers to almost the entire brain, delivering a chemical called noradrenaline, the brain's own version of adrenaline.

The neuroscientists Gary Aston-Jones and Jonathan Cohen wanted to explain what this nucleus contributes to behavior, since its output is too widespread to carry any specific message. Recording from these cells in animals performing attention tasks, they proposed that the locus coeruleus functions as an adaptive gain control. It does not tell the brain what to think. It sets how strongly everything else is heard.

Picture the volume dial on a radio. At a moderate setting, the signal is clear and the static stays in the background. Turn the dial up and the music gets louder, but so does the static, until noise and signal are hard to tell apart. That is what raised gain does to a nervous system. Ordinary input starts to feel urgent. A door closing becomes a slam. A stray thought becomes a warning. A minor bodily sensation becomes something worth watching.

Nothing has gone wrong with the input. The setting has changed. A system whose contents are normal and whose organization is not: anxiety keeps producing that pattern, station after station.

Interoception is a sense with its own wiring

You carry a sense that reports on your own interior, with receptors, a dedicated spinal pathway, and a cortical map in the insula. The body-first account of anxiety runs on that wiring.

Everything so far has been the familiar account, running from the brain down. The more interesting account runs the other way, and it begins with a sense most people have never been told they have.

You know the outward senses. Sight, hearing, touch, taste, smell. They report on the world. There is a separate, older sense that reports on you: the state of your heart, your lungs, your gut, your blood chemistry, your muscles, the temperature of your skin. It is called interoception, and until fairly recently it was treated as a vague background hum rather than a sense with proper wiring.

Craig traced the pathway to the insula

The neuroanatomist Arthur D. Craig changed that. He was tracing where signals from the body's own tissues actually go, following them upward through the spinal cord to see whether they had a dedicated destination. They did. Craig described a distinct pathway carrying the physiological condition of the body upward.

It runs from a thin layer at the back of the spinal cord called lamina I, through the thalamus, to a region of cortex called the insula, folded deep in the side of the brain. In the insula those signals build a map of the body's internal condition.

That is the sentence to hold. Feelings from the body are a sense. They have receptors, a pathway, a relay and a cortical map, exactly as vision does. The tightness in an anxious chest arrives from somewhere. It is a reading.

The vagus mostly reports upward

The other half of the wiring is just as surprising. The vagus nerve is usually described as the brake, as though it were a control cable running from brain to body. Anatomically it is mostly the reverse. The neurobiologists Hans-Rudolf Berthoud and Winfried Neuhuber worked through the structure of the afferent vagal system, meaning the half of the nerve that carries information upward.

Its reach is enormous. It reports from the esophagus, the airways, the heart and the aorta, and below the diaphragm from the whole gastrointestinal tract, the liver, the biliary system and the pancreas. Where the fibers of the nerve have been counted in the neck, the sensory ones far outnumber the motor ones.

Take that seriously and the usual picture inverts. The brain is being told about the body far more than it is telling the body what to do. Whatever else the nervous system is, it is a structure listening hard to its own interior. That is the physical basis for the body-first account of anxiety.

The brain predicts the body and feels the difference

The brain issues a prediction of what the body should feel and compares it to what arrives. What reaches awareness is largely the difference. In anxiety, the prediction is set for danger, so ordinary bodily signals are amplified into alarm.

The neuroscientists Lisa Feldman Barrett and Kyle Simmons set out to explain how the brain handles this flood of internal signal. They proposed that the predictive architecture used for the outside world is used for the inside as well. In their account of interoceptive predictions in the brain, regions in the insula and nearby cortex send predictions downward about the expected state of the body. What travels upward and reaches awareness is largely the difference between prediction and arrival, the prediction error.

The cognitive scientists Anil Seth and Karl Friston built the same idea into a formal frame, arguing that emotional feelings arise from active inference about the causes of bodily signals. The brain guesses what the body feels, then acts on the body to bring the readings into line with its guess.

Now apply that to anxiety. If the prediction is set for danger, then ordinary bodily noise arrives already interpreted. A normal heart flutter, the kind everyone has several times a day, does not present itself as neutral information. It arrives against an expectation of threat, and it is read as the beginning of something.

The psychiatrists Martin Paulus and Murray Stein made this concrete. They asked what a single structure could contribute across the whole family of anxiety disorders. Their answer, an insular view of anxiety, is that anxious individuals hold an altered prediction of their own bodily state. They expect worse than arrives, so ordinary interoceptive signals are amplified into alarm.

The measurement of this sense trails the mechanism. The neuroscientist Sahib Khalsa and a large group of colleagues surveyed the field and produced a roadmap for interoception and mental health that sets out the state of the instruments. The measures used to assess this sense do not agree with each other well, the constructs are still being sorted out, and clinical application is early. The mechanism is well founded. The measurement of it is not finished.

A racing heart becomes evidence of danger

Put prediction and threat reading together and a loop appears that needs nothing from the outside world to run. A bodily signal read as danger opens the accelerator, and the accelerator raises the signal that started it.

A bodily signal rises. Perhaps the heart quickens for a perfectly ordinary reason: caffeine, standing up, a warm room, a poor night's sleep, the last flight of stairs. That signal is read against a prediction set for danger, so it is registered as threat. The threat reading opens the accelerator, because that is what a threat reading does. The accelerator raises the very signal that started the sequence. The heart is now genuinely faster, which confirms the reading, which opens the accelerator further.

The clinical psychologist David Clark described this pattern in 1986 and it has anchored the field since. He was trying to explain why panic attacks appear without any external trigger. He proposed a catastrophic misinterpretation of bodily sensations as the engine. A benign internal event is read as evidence of imminent disaster, and the reading itself produces more of the event.

Two things must be said carefully here.

First, nothing in this loop is imaginary. The heart really is faster. The chest really is tight. The dizziness is real dizziness produced by real changes in blood chemistry. Nobody is inventing a sensation. The loop concerns how a real signal is read, and the reading then makes the signal larger.

Second, this is a loop and never a chain of blame. There is no point in it where a person chose wrongly. The reading of an interoceptive signal is no more a decision than the perception of red. The reading happens beneath awareness, arrives as feeling, and only afterwards gets a story attached to it.

The loop closes inside the body. Nothing outside has to happen for it to run, which is precisely why it feels as though nothing outside caused it.

This also explains a detail that puzzles many people. The attack often arrives in a safe place. On the sofa. In bed. In a supermarket line. If the trigger were the world, that would be strange. If the trigger is a reading of the body, it is entirely expected, because the body is always present.

Panic can be produced through the body alone

A full panic attack can be manufactured with sodium lactate in a vein, a breath of carbon dioxide, or an adrenaline-like drug. None of these is a frightening thought. Each enters through the body.

A panic attack is the alarm firing at full strength with nothing to fire at. Ten minutes of pounding heart, breathlessness, sweating, trembling, tingling, unreality, and an overwhelming conviction of imminent death or collapse. People who have had one describe it as the most frightening experience of their lives, and often go to an emergency room certain they are having a heart attack.

Three agents that provoke attacks

The most important evidence about panic comes from a strange class of experiments. Researchers found that they could produce a full attack in the laboratory using agents that act on the body, and the results reshape what the condition can be.

The first was accidental. In 1967 the psychiatrists Ferris Pitts and James McClure gave intravenous sodium lactate to patients with what was then called anxiety neurosis, testing an idea about lactate metabolism in anxious patients. Lactate is the substance muscles produce during hard exercise. Infusing it triggered attacks in patients far more often than in healthy volunteers. A chemical delivered into a vein reliably produced the psychological experience of terror.

The second is inhaled carbon dioxide. Have a person breathe a single deep breath of a mixture containing about 35 percent carbon dioxide and, in susceptible people, a full panic attack follows within seconds.

The psychologist Nicole Leibold and colleagues tested this as a human experimental model of panic, giving healthy volunteers inhalations of 0, 9, 17.5 and 35 percent carbon dioxide in randomized order. Fear, discomfort and panic symptoms all rose with the dose, and so did blood pressure. The provocation is now a standard research tool precisely because it is so dependable.

The third is a drug. The psychiatrist Richard Balon and colleagues infused isoproterenol into patients with panic disorder. The compound acts on the same receptors adrenaline does. They cataloged the somatic and psychological symptoms of the attacks it produced. The drug speeds the heart and opens the airways from the outside, in the body.

Notice what these three have in common. Each of them enters through the body, and each of them can produce the full experience, including the conviction of doom.

Panic without an amygdala

Then there is the case that settles the matter. The neuroscientist Justin Feinstein and colleagues studied three patients with a rare disorder that had destroyed both amygdalae. These are people who cannot be frightened by the world. They handled snakes without alarm, walked through a haunted house without a startle, and reported no fear at horror films. The researchers then gave them a single inhalation of 35 percent carbon dioxide. All three experienced fear and panic, to the surprise of everyone involved, including the patients.

Sit with that result. The structure everyone calls the fear center was gone, and panic still arrived. It arrived through the body. The psychiatrist Donald Klein had proposed something like this years earlier, arguing that panic reflects a false suffocation alarm, a physiological monitor for impending asphyxiation firing when no asphyxiation is occurring. Klein's specific mechanism remains debated. The shape of his claim survived the test that could have killed it.

For a patient, the practical translation is short. A panic attack is a false alarm in a monitoring system, and false alarms are a property of monitors rather than a failure of nerve.

Over-breathing manufactures its own symptoms

Breathing hard when the alarm sounds blows off carbon dioxide, and the falling level produces tingling, lightheadedness and unreality out of blood chemistry alone. The anxious system reads those new symptoms as more danger.

One of the cruelest features of the loop is that the most natural response to it feeds it. When the alarm sounds, people breathe hard. Breathing hard produces a new set of symptoms, and the system reads those symptoms as more danger.

The physiology is worth teaching properly, because understanding it takes some of its power away. Every breath out carries away carbon dioxide, the waste gas your cells produce. Its level in the blood sets the blood's acidity, and it governs the diameter of the blood vessels in your brain.

Breathe faster or deeper than your body needs and you blow off more carbon dioxide than you are making. The level in the blood falls. Two things follow. The blood becomes slightly more alkaline, which changes how calcium behaves at nerve endings and produces the tingling in the hands, feet and lips that so many people report.

And the brain's blood vessels constrict, reducing blood flow to the brain. The physiologists Philip Ainslie and James Duffin worked through how carbon dioxide governs both breathing and cerebral blood flow. They described a tight coupling between carbon dioxide and the cerebral circulation. It is among the most reliable relationships in human physiology.

Less blood to the brain produces lightheadedness, visual disturbance, and the strange sense of unreality that frightens people more than any other symptom. The chest also tightens and the breath feels unsatisfying, so the person breathes harder still.

The evidence is mixed in an instructive way

The psychologists Alicia Meuret and Thomas Ritz reviewed the evidence on hyperventilation in panic disorder and reported a mixed picture. Over-breathing is common during attacks and between them, and it accounts for a real share of the symptoms. It is absent in some attacks, and the classic account of panic as a purely respiratory event does not hold for everyone.

That mixed result is what a regulation model expects. Breathing is one voice in the system. In some people the disturbance is carried loudly on that channel. In others it is carried on the heart, or the gut, or the muscles, and quieting the breath alone will not find it.

Meuret's own treatment work makes the point from the other side. She and her colleagues taught panic patients to raise their carbon dioxide deliberately, using a device that shows them the level in their exhaled breath, and tracked what actually mediated their improvement. The change in carbon dioxide predicted the change in symptoms. The intervention did not argue with anyone's thoughts. It corrected a physiological variable and the experience followed.

Slow breathing drives a control loop at its resonance

Breathing is the one autonomic lever under voluntary command, and near six breaths a minute it drives the blood pressure reflex at its natural frequency. The calming effect on anxiety is mechanical.

Almost everything the autonomic system does is out of reach of the will. You cannot decide to slow your heart. You cannot instruct your gut. There is one exception, and it is the reason breathing appears in every calming practice ever devised, in every culture, without anyone needing to know the physiology.

The respiratory gate and the baroreflex

Put a hand on your pulse and breathe slowly. The heart speeds slightly on every breath in and slows on every breath out. This is called respiratory sinus arrhythmia, and it is a normal, healthy feature of a resting heart.

The physiologist Dwain Eckberg spent years working out why. He was studying how the vagal brake behaves during the breathing cycle, and found that it is gated rather than applied steadily. In his account of the human respiratory gate, vagal outflow to the heart is opened and closed within each breath, so the brake is released during inhalation and reapplied during exhalation. The breath is a handle on the brake, mechanically, within a single cycle.

Now add a second control loop. In the walls of the large arteries of your neck and chest sit stretch sensors that feel how hard the blood presses with each beat. When pressure rises, they report it, and the brainstem answers within a single heartbeat by slowing the heart and relaxing the vessels. When pressure falls, it does the reverse. This is the baroreflex, the fastest guardian of blood pressure in the body.

Breathing moves blood pressure. The baroreflex answers pressure by moving heart rate. Two loops sharing a variable will, at some frequency, fall into step and reinforce each other. The psychophysiologists Evgeny Vaschillo, Bronya Vaschillo and Paul Lehrer set out to characterize that frequency. They paced the breathing of 32 asthma patients and 24 healthy adults across a range of rates, measuring the size of the heart rate oscillation each rate produced.

Each person had a resonant frequency of the cardiovascular system, a rate at which the oscillation in heart rate becomes far larger than at any other. It tracked with height and differed between men and women. It did not shift with age, and it was the same in the asthma patients as in the healthy adults. It held steady across ten sessions of training.

This is the physical fact underneath a practice that is usually taught as a mood trick. In most adults that resonance sits near six breaths a minute, so breathing at roughly that rate drives a real control loop at its natural frequency. Paul Lehrer and Richard Gevirtz laid out how and why heart rate variability biofeedback works, arguing that repeatedly exercising the baroreflex at resonance strengthens the reflex itself, the way repeated loading strengthens a muscle.

The breath reaches the arousal machinery

The breath reaches further up than the heart. The neuroscientist Kevin Yackle and colleagues were tracing what a small cluster of breathing-pacemaker neurons in the brainstem actually connects to. They found a subgroup that projects directly to the locus coeruleus, the gain control met earlier, and showed that silencing those neurons in mice left the animals calm rather than breathless. Breathing rhythm is wired into arousal at the source.

And the rhythm organizes the rest of the brain. The neuroscientist Christina Zelano and colleagues recorded directly from inside the human brain in patients being evaluated for epilepsy, asking whether the breath leaves a trace in deeper structures. It does. Nasal breathing entrained oscillations in the amygdala and hippocampus, and the phase of the breath affected how quickly volunteers recognized a fearful face.

The breath is a rhythm the alarm circuitry keeps time to. That is why the oldest instruction in every calming tradition on earth turns out to be the correct one.

Heart rate variability is reduced across the anxiety disorders

The beat-to-beat timing of the heart carries a measurable trace of anxiety. Pooled studies find reduced heart rate variability across panic disorder, generalized anxiety, social anxiety and post-traumatic stress, consistent in direction and modest in size.

A healthy heart does not beat like a metronome. The interval between beats changes constantly, by tens of milliseconds, breath by breath. This beat-to-beat variation is called heart rate variability. The psychophysiologists Fred Shaffer and Jay Ginsberg set out the standard metrics and normal ranges used to quantify it.

Stated at its most careful, heart rate variability is a validated index of autonomic state, and the high-frequency component is mainly a measure of vagal control of the heart. That is the established claim, and it is worth keeping precisely that size.

What does it show in anxiety? The psychologist John Chalmers and colleagues pooled the studies comparing patients with anxiety disorders to healthy controls. They found reduced heart rate variability across the anxiety disorders. The reduction appeared in panic disorder, generalized anxiety disorder, social anxiety disorder and post-traumatic stress disorder. A later systematic review and meta-analysis led by the physician Ying-Chih Cheng reached the same conclusion with more recent data.

The finding is consistent in direction. It is also modest in size, it overlaps heavily between groups, and it diagnoses nothing in an individual. Anyone who tells you a wearable can detect an anxiety disorder from a nightly number is selling something.

The raw nerve traffic complicates the picture

Physiologists can push a fine electrode into a nerve in the leg and listen to sympathetic bursts directly, a technique called microneurography. A group at the Baker Institute in Melbourne measured this in patients with panic disorder at rest, under laboratory stress, and during attacks that happened to occur in the laboratory. At rest the patients looked like the controls.

Muscle sympathetic activity, plasma noradrenaline and the whole-body and cardiac spillover rates were all similar between the groups, and responses to laboratory stress were almost identical. During the attacks, adrenaline secretion rose sharply, and sympathetic traffic to muscle rose steeply in two patients but not in the others. The authors stated their conclusion plainly, against the popular picture: the sympathetic nervous system is not globally activated during panic attacks.

Read that as information rather than defeat. A panic attack is not the whole accelerator flooring itself. It is a specific and patterned discharge, adrenaline prominent, nerve traffic selective, arriving in a body that at rest measures much like anyone else's. A model of too much sympathetic drive struggles with that. A model of a control problem does not.

How much belongs to the illness, how much to the treatment

The psychologist Gail Alvares and colleagues asked a question the field had largely avoided: how much of the low variability seen in psychiatric patients is the illness, and how much is the treatment? Their systematic review and meta-analysis put most of it on the illness.

Variability was reduced across every patient group, and the reduction held in people taking no medication at all. On top of that, specific antidepressants and antipsychotics were associated with a smaller but significant further reduction, over and above the disorder itself.

Keep that second finding at its real size. Most of the lost variability belongs to the condition. Some of it belongs to the treatment, which means a drug that reliably improves how a person feels can nudge this particular measure further in the direction associated with worse regulation. Hold that thought. It is a small and clean illustration of the difference between managing an output and restoring a range.

One distinction closes the section. Heart rate variability is a validated index of autonomic state. The Unified Model of Tone reads it as one channel of the body's whole organization, which is a broader claim than the instrument alone makes. The measurement belongs to physiology. The reading belongs to the model.

Anxiety is a chord sounded by coupled systems

Everything above is established physiology, credited to the people who established it. The Unified Model of Tone reads it as one thing: the pounding heart, the cortisol, the nerve traffic and the dread are one organization measured on different channels.

Two independent axes, not one seesaw

Start by repairing the wiring picture from earlier. The accelerator and the brake are not two ends of one seesaw. The psychophysiologists Gary Berntson, John Cacioppo and Karen Quigley worked out that sympathetic and vagal control are two independent dimensions rather than a single axis, a framework they called autonomic space. Both can be high at once. Both can be low. One can rise while the other holds.

Anyone who has been anxious already knows this in their body. The heart is pounding and the gut is churning at the same moment the hands are cold and the throat is closed. That is accelerator and brake sounding together. A single number on a single channel cannot describe it, and any model that tries will be wrong in some patients and right in others for no visible reason.

One organization, many channels

This is what the Unified Model of Tone means by a chord. A measured variable is never a note that one channel plays alone. It is a chord sounded by many coupled voices at once. The electrical traffic in the nerves, the mechanical stretch of vessel walls, the chemistry of the blood, the hormonal signal in the circulation, the rhythm of the breath.

The same information rides several mediums simultaneously and the voices are held in step with one another. Heart rate variability, cortisol, sympathetic traffic, respiratory rate and the sense of dread are five readings of one state.

Tone, in this system, is the integrated organization those voices hold together, and the capacity to change it as a moment demands and change it back when the moment passes. The neurophysiologist Julian Thayer and the psychiatrist Richard Lane built the closest established relative of this idea. They described a neurovisceral integration network in which the prefrontal and autonomic systems form one regulatory unit.

Vagal control of the heart is its visible face. Their model is established science and this model owes it a debt. The step this model takes further is to treat that organization as one variable running through every scale of the body, and to define health by the width of its range.

An anxious system has lost the distance between alarm and rest. Arousal alone does not describe it.

Read the physiology through that definition and the pieces stop being a list. The unpredictable-threat finding is a system that cannot release. The extinction work is a brake that has to be built and can be lost. The cortisol result is a loop that can fail at its beginning or at its ending.

The interoceptive prediction is a range narrowed at the level of expectation, so that ordinary readings fall outside it. The gain setting is that same narrowing expressed as volume. Panic is the whole range collapsing to one end in ninety seconds.

None of those is a broken part. Every one of them is a regulation that has stopped moving through its full width. That is why no lesion is found, and why looking harder for one has not helped. Tone that can move to alarm when a moment demands it and return afterward is health. Tone held at the alarmed end past every demand is what manifests as the disorder. Nothing is broken. The range has closed.

Treatment can mask an output or restore a range

Cognitive behavioral therapy beats placebo. Several drug classes beat placebo. The Unified Model of Tone asks a different question of every treatment for anxiety: does it push a signal in one direction, or does it widen the range the system can move through?

What the standard treatments show

Psychological therapy for anxiety works. The psychologist Joseph Carpenter and colleagues pooled the randomized trials that compared cognitive behavioral therapy against a placebo condition rather than a waiting list, which is the harder comparison. They found a moderate advantage for the therapy across anxiety and related disorders. That is a genuine and substantial result.

Medication works too. The team led by April Slee performed a network meta-analysis of drug treatments for generalized anxiety disorder, comparing agents against each other and against placebo across the whole trial literature. Several classes showed clear efficacy over placebo. For many people these medicines are the difference between a functioning life and a lost one.

Approaches that work on the regulation itself also have trial support. The psychiatrist Elizabeth Hoge and colleagues ran a randomized trial comparing eight weeks of mindfulness-based stress reduction against escitalopram, a standard antidepressant, in adults with anxiety disorders. The mindfulness training proved non-inferior to the drug.

The psychiatrist Brendon Stubbs and colleagues pooled the trials of exercise in anxiety and stress-related disorders and found a significant anxiolytic effect compared with control conditions. The psychologist Vera Goessl and colleagues meta-analyzed heart rate variability biofeedback, the resonance breathing taught above. They reported a large pre-to-post reduction in self-reported stress and anxiety. Many of the included trials were small, and the field needs larger ones.

Two different aims

Now the distinction this model cares about. A drug that blunts one output is doing something precise and valuable. It reduces the size of a signal in one direction, in everyone who takes it, for as long as they take it. That is what a good drug is for, and it is the accurate description of a benzodiazepine, a beta blocker, or a serotonergic antidepressant.

Restoring a range is a different aim. It asks whether the system can move to alarm and come back, and it treats the width of that movement as the target rather than the height of any one reading. Recall the smaller half of the Alvares finding.

Medication can improve how a person feels while nudging a regulatory measure further in the unfavorable direction. The effect is modest, and most of the reduced variability belongs to the illness rather than to any drug. Even at that size it shows the two aims are separable and measurable apart from each other, without any argument about which one matters more.

Restoring regulation versus masking a symptom, and how to tell

One prediction tells them apart, and it is the model's sharpest claim.

A genuine restoration of tone moves a dysregulated value toward the healthy middle from either side. An overriding intervention pushes one direction for everyone.

State it concretely. Give a masking intervention to a hundred people and the measure moves the same way in all of them, whether they started high or low, because the intervention is applying force in one direction. Give a genuine tonal correction to the same hundred and the model predicts convergence. Those who started high should trend down. Those who started low should trend up. The population variance should shrink around the middle.

The prediction is specific. It is currently untested, and existing instruments can run it. An input that pushes everyone in one direction regardless of where they started is masking the output rather than restoring the regulation. It eases whichever group it happens to point at and carries the other group further from the middle. That is a different prediction from the one a drug-style account makes, and the split-group result reads it directly.

The law of initial values is not this prediction

Physiology has an older claim in this neighborhood, and being precise about it matters. It is called the law of initial values, and it says that a body's response to a stimulus depends on the level it started from. The traditional version had high starting values responding less to anything that raises a function and more to anything that lowers it.

That version did not survive inspection. The psychologist Ping Jin worked through its statistical problems and argued for a reconceptualization of the law of initial value. His conclusion ran the other way for the general case. The higher the starting value, the greater the reactivity that follows, with reversed responses appearing only near the extremes.

So the old law does not hand this model its prediction. What it establishes is only the premise, that starting state governs response. Convergence toward the middle is a separate and stronger claim, and it belongs to this model to demonstrate or to lose.

Findable conditions can produce the symptoms of anxiety

An overactive thyroid, a pheochromocytoma, a rhythm disturbance, stimulants, alcohol withdrawal and sleep apnea can each produce the sensations of anxiety, and each is findable.

An overactive thyroid produces racing heart, tremor, sweating, heat intolerance, insomnia and a sense of being wound tight. The psychiatrists Robertas Bunevicius and Arthur Prange reviewed the psychiatric manifestations of Graves' hyperthyroidism, describing how readily the presentation is taken for a primary anxiety disorder.

A rare adrenal tumor called a pheochromocytoma releases adrenaline in bursts and produces episodes almost indistinguishable from panic attacks. A recent review of its clinical presentation and screening describes the paroxysms of headache, sweating, palpitations and pressure that bring these patients to psychiatry before endocrinology. Heart rhythm disturbances, certain medications, stimulants, alcohol withdrawal and sleep apnea belong on the same list.

A findable cause is worth finding, because removing it can end the symptoms outright. The workup comes first. The regulation account addresses what remains when the workup is clean.

What to track instead of a level

With that floor in place, the useful question becomes what to track. If the target in anxiety is the width of a range rather than the height of a reading, then the endpoints follow, and most of them are already recorded in ordinary practice.

A person recovering from anxiety shows it in these terms. The variability rises. The recovery after a stressor shortens. The breath slows and the carbon dioxide settles. The day regains a shape. Those measurements are ordinary, and they are the ones that would turn this account from a frame into a tested claim.

The definitive question is open. How much of anxiety, in how many people, is a regulation that can be widened rather than an output that must be held down. What is already clear is where to look. A quieter nervous system is a smaller aim. A system that can go to alarm when a moment calls for it and come all the way back has its range back.

How anxiety relates to the rest of the library

Anxiety is a condition page, and the readings it rests on each have a page of their own. The claims below are the specific connections, one per page.

• Tone is the pillar this page reads through: the integrated organization the body holds across its systems at once, and the range that organization can move through.
• Prediction owns the machinery behind the forecast set for danger, the reason an ordinary heartbeat can arrive as evidence.
• Gain is the locus coeruleus setting taught here in volume terms, the dial that makes a closing door read as a slam.
• Load carries what months of bracing cost, including the cortisol arm that fails to close the response it belongs to.

• The autonomic nervous system is the anatomy the alarm runs through, and its page dismantles the seesaw picture properly.
• Heart rate variability is the instrument behind the reduced-variability findings pooled in the 2014 and 2022 meta-analyses, and its page explains what a personal number can and cannot certify.
• The vagus nerve is the mostly sensory cable that makes the body-first account of anxiety anatomically literal.
• Depression is the condition that most often travels with anxiety, and its biology runs flat where anxiety's runs high.
• Trauma is an alarm trained by an event that keeps predicting the event's return.
• Stress and physical symptoms reads the same threat physiology through the body's other channels, the gut, the muscles and the skin.
• Sleep is where the alarm is supposed to stand down nightly, and a readiness that will not release takes the night with it.
• Dysautonomia is where autonomic regulation itself is the presenting problem rather than the threat forecast.
• And mental health is the wider psychiatric account this regulation reading belongs to.

Frequently asked