Scoliosis and the Nervous System

What a Cobb angle measures, and how much it wobbles

The Cobb angle is the number every scoliosis decision hangs on, and remeasuring the same film moves it by about five degrees for one observer and seven degrees for two.

The child stands, a radiograph is taken from behind, and the bones of the spine appear as a stacked column. A radiologist finds the vertebra at the top of the curve that tilts most, and the vertebra at the bottom that tilts most. A line is drawn along the top edge of the first and along the bottom edge of the second. Where perpendiculars to those two lines cross, an angle opens up.

Ten degrees is the threshold. Ten degrees or more, with the vertebrae also rotated around the long axis of the body, meets the definition of scoliosis. Below ten degrees the finding is called spinal asymmetry and does not meet the definition.

The measurement is less precise than the decimal points suggest. Orthopedic researchers led by Raymond Morrissy asked how much a Cobb angle wobbles when the same film is measured twice, and reported an intrinsic error of several degrees. One observer measuring the same radiograph twice varied by about five degrees. Two different observers varied by about seven. A curve has to move by roughly that much before anyone can be confident it moved at all.

The number describes a shape at one instant, from one angle, in one posture. It says nothing about the system that produced the shape. A photograph of a leaning tower tells you the tower leans, and tells you nothing about the ground.

Position sense is how the nervous system knows where the spine is

Muscle spindles and tendon organs report length and force at every joint hundreds of times a second, and buzzing one of them makes a motionless arm feel like it is moving.

Start with a nerve. A nerve is a living wire, a long thin cell that carries a small electrical pulse from one end of itself to the other. Some wires carry a command outward to a muscle. Others carry a report inward, from the tissues back to the spinal cord and brain. Much of the nervous system is listening.

Some of that listening is aimed at the outside world. A separate class of sensors is aimed inward, at the body itself. In 1907 the British physiologist Charles Scott Sherrington, working at Liverpool on how spinal reflexes coordinate opposing muscles, needed a name for that class.

He was mapping the machinery that lets one muscle relax exactly as its opponent contracts, and that machinery needs a source of information about the body's own state. In the journal Brain he named it the proprio-ceptive system, the sense of self. Sherrington later shared a Nobel Prize for his work on the function of neurons.

Two sensors do most of that work in muscle. Muscle spindles are stretch gauges, woven lengthwise inside the belly of a muscle, that fire faster as the muscle is pulled longer. Golgi tendon organs are tension gauges, sitting where muscle meets tendon, that fire according to how hard the muscle pulls. Between them the nervous system gets a continuous readout of length and force at every joint, without any conscious attention.

Vibrating a tendon moves a limb that never moved

Two French neurophysiologists, Jean-Pierre Roll and Jean-Pierre Vedel, went after the question directly in the early 1980s. They wanted to know whether spindle signals build the conscious experience of limb position. They sat blindfolded volunteers down and pressed a small vibrator against a tendon at the elbow. Vibration is a strong artificial stimulus for a spindle.

The volunteers reported an unmistakable illusion of the arm moving, at a smooth constant speed, in the direction that would stretch the buzzed muscle. The arm never moved. The same paper carries the direct evidence for what the buzz does to a sensor. Fine electrodes pushed into the nerves of the lower leg recorded spindle endings firing in time with the vibration.

Fool the gauge and the body moves, in the only place where the body's position exists at all.

Scoliosis begins in that inference. Where your spine is, as far as your nervous system is concerned, is worked out continuously from evidence.

Piezo2 is the molecule that turns a pull into a signal

Piezo2, identified in 2015, is the channel that converts a stretch into an electrical current inside position-sense endings, and mice lacking it in those neurons move as though they cannot find their own limbs.

Tissue being pulled is a mechanical event and a nerve impulse is an electrical one, so something has to translate. For most of the twentieth century nobody knew what did that job in mammals. Molecular neurobiologists at Scripps Research in California filled the gap. Seung-Hyun Woo and Ardem Patapoutian were hunting the specific channel that converts mechanical stretch into electrical current in proprioceptive nerve endings.

They found a channel called Piezo2, sitting in the sensory endings inside muscle spindles and Golgi tendon organs. It works like a trapdoor set into the wall of the nerve ending. Pull the tissue and the membrane deforms. The trapdoor opens, charged particles rush through, and an electrical signal is born. No stretch, no opening, no signal.

They then bred two independent lines of mice missing Piezo2 in their proprioceptive neurons. Those mice moved with severely uncoordinated bodies and held their limbs in abnormal positions. Recordings from the muscle nerves showed the proprioceptors barely firing when stretched. The gauge was still physically there.

It had gone deaf. Patapoutian shared the 2021 Nobel Prize in Physiology or Medicine for this line of work on how cells sense force. The same channel sits at the hinge of the scoliosis evidence, because losing it in position-sense neurons alone bends a spine that was built correctly.

Turn off the sensing, and a normal skeleton bends

Mice born with normal vertebrae develop scoliosis around puberty once the neurons carrying position sense are disabled, and two patients born without a working PIEZO2 channel followed the same course.

Two researchers at the Weizmann Institute of Science in Israel asked a plain question. One was an orthopedic surgeon named Ronen Blecher. The other was a developmental biologist named Elazar Zelzer. What keeps a growing spine aligned in the first place?

A stack of vertebrae has no built-in knowledge of straight. Something has to be watching, and correcting. To find out what, the team disabled the watching.

They bred mice missing Runx3, the gene that builds the TrkC nerve cells connecting position-sense receptors in muscle to the spinal cord. Those mice were born with normal vertebrae and symmetrical muscles. Nothing about the skeleton was malformed. Then, around puberty, they developed scoliosis. Deleting the same gene only in the peripheral nervous system did it. Deleting it only in peripheral sensory neurons did it. Mice missing Egr3, which lack muscle spindles while keeping their tendon organs, developed a milder version of the same thing.

The order of events is the whole finding. The sensing failed first. The bone bent afterward.

Deleting Piezo2 in bone and cartilage does nothing to the spine

Three years later the same group closed the obvious loophole. If PIEZO2 mutations in people cause skeletal problems, perhaps the channel does something structural inside bone and cartilage cells. Eran Assaraf, Blecher and Zelzer tested it by deleting Piezo2 in different tissues one at a time. Deletion in cartilage lineages did not reproduce the human deformity. Deletion in bone lineages did not either. Deletion in proprioceptive neurons alone produced spine malalignment and hip dysplasia. The skeleton was fine. The sense reading the skeleton was not.

Two patients born without working PIEZO2

Then the human case arrived. Alexander Chesler is a sensory neuroscientist and Carsten Bonnemann a neurologist, both at the United States National Institutes of Health. They were working with two young patients whose neuromuscular and skeletal syndrome fit no existing diagnosis. They sequenced the whole exome of each patient.

Both carried compound-inactivating variants in PIEZO2. Both had lost discriminative touch. Both had profoundly decreased proprioception, with a stumbling, misaimed quality to movement that worsened dramatically the moment they closed their eyes and lost vision as a substitute. Both had progressive scoliosis.

That is the mouse experiment, run by nature, in a human being.

What the word idiopathic is admitting about scoliosis

Adolescent idiopathic scoliosis affects roughly one to three percent of children in the at-risk band of ten to sixteen years, and the word idiopathic reports on the search rather than on the child.

In a review for The Lancet, the orthopedic surgeon Stuart Weinstein and colleagues set out those figures and wrote plainly that the origin of the disorder remains unknown. It is by a wide margin the most common form of the condition. Girls carry the risk of severe progression several times more often than boys.

The timing is the loudest clue in the whole condition. Curves appear and accelerate during the adolescent growth spurt, and mostly settle once growth finishes. That is why orthopedics tracks skeletal maturity so carefully. The Risser sign reads the progress of a bony rim across the pelvis as a proxy for growth remaining. James Sanders and colleagues later showed that a simplified maturity scale read from the hand predicts which curves will progress better than chronological age does.

Growth remaining governs how fast a curve can move. Notice what that sentence does not say. Growth is the speed of the process. It is not an account of why the process points sideways.

The dominant paradigm looks for a lesion: a malformed vertebra, a short muscle, a broken gene, a piece of tissue that can be pointed at. In idiopathic scoliosis there is no lesion to point at. Every part is normal. The paradigm treats that as a search that has not finished yet.

The Unified Model of Tone treats it as a category error. The search keeps failing because the disorder does not live in a part. It lives in the organization across parts.

Midline is computed from disagreeing reports, and scoliosis defends the wrong one

The spindles report muscle length, the tendon organs report pull, the inner ear reports gravity and the eyes report the horizon. Muscle is driven to defend whatever answer the weighing of those reports returns.

Midline is a computed value rather than a place in the body. Roll and Vedel showed how thin the computation is when they buzzed a tendon and their blindfolded volunteers felt a motionless arm moving at a steady speed. Change the evidence and the answer changes, in a body that has not moved. Posture, read this way, is a claim. It is the body's current best answer to the question of where it is.

Tone is the name for the coupled organization that generates that answer. Muscle expresses it and bone records it, and it belongs to neither. It is the way every channel is held in relation to every other, and the range across which the whole arrangement can move and still return.

Health is not a straight spine. Health is a spine whose owner can drift and come back.

A healthy system has a wide, self-correcting range around its computed midline. It sways, leans, twists, loads one side, and returns. The return is the health, and the width of the range is the health. Tone that keeps that width is what health means in a growing spine, because the system can meet a demand and recover from it.

Scoliosis is that range collapsing. The system stops moving freely around its midline and settles into one asymmetric configuration, which it then defends as correct. A growing skeleton, obeying the loads it is given, builds that configuration into bone. The curve is the printed record of a decision that stopped being revisable.

Nothing in that account requires anything to be broken, which is exactly why nothing broken is ever found.

The competing theories of scoliosis are channels of one variable

Proprioceptive genetics, vestibular asymmetry, spinal fluid flow, melatonin signaling and growth-plate loading are presented in the literature as five candidate causes, and each fails to account for the other four.

Every one of those findings holds up. Every one fails in the same way, by leaving the other four unexplained. Weinstein's 2008 review in The Lancet records the outcome of that search in a single line: the origin of the disorder remains unknown.

Alignment is sounded by many coupled channels at once, the way a chord is sounded by many notes. The same postural information rides several mediums simultaneously. Muscle spindles and tendon organs carry it mechanically. Piezo2 channels carry it as ion flow. The semicircular canals carry it as fluid inertia.

Vision carries it optically. The spinal fluid column and the fiber suspended in it carry it hydrodynamically and chemically. Back muscle drive carries it as sustained contraction. Growth plates under load carry it as changed growth rate. The autonomic background sets how strongly the whole arrangement answers.

These channels tune each other rather than reporting separately to a committee. A shift in one shows up as a shift in the others.

Read that way, the search for the cause of idiopathic scoliosis was always going to fail. Asking which voice sings the chord is not a question with an answer. A curve is a chord that has gone out of tune, and every voice in it is intact.

Cerebrospinal fluid flow is required to build a straight axis

Zebrafish with normal vertebrae twist into scoliosis when the cilia that drive their spinal fluid stop beating, and restoring the beat after the curve begins stops it progressing.

The brain and spinal cord float in a clear liquid called cerebrospinal fluid. It fills chambers inside the brain, called ventricles, and runs down a narrow channel through the middle of the spinal cord. The fluid is not still. Lining those chambers are cells carrying motile cilia, microscopic hairs that beat in a coordinated rhythm like rows of oars, driving the fluid along in a current.

Daniel Grimes and Brian Ciruna, developmental geneticists at Princeton University and the Hospital for Sick Children in Toronto, were studying a strain of zebrafish called ptk7 mutants. These fish grow normally and then twist into a curve during late growth, with no malformed vertebrae to blame. It is a close animal match for the human idiopathic pattern, and the team went looking for the bone defect behind it.

The defect sat in the motile cilia and the fluid flow they generate. Reintroducing the missing Ptk7 protein only in motile ciliated cells prevented the scoliosis. Mutating other, unrelated genes needed for cilia motility reproduced it. Restoring cilia motility after the curve had already started blocked further progression.

The Reissner fiber, and what a century-old curiosity turned out to do

Hanging in that current is one of the strangest objects in anatomy. In the nineteenth century the Baltic German anatomist Ernst Reissner described a fine thread running through the fluid from the brain down the spinal canal. It is built from a protein called SCO-spondin. For more than a century it sat in the textbooks as a curiosity with no known job.

Genetics gave it one. Yasmine Cantaut-Belarif and Pierre-Luc Bardet, working with Claire Wyart at the Paris Brain Institute, mutated the scospondin gene in zebrafish. Fish without the fiber had intact cilia and intact fluid flow, yet failed to build a straight body axis. The team also showed that cilia mutants systematically fail to assemble the fiber, which explained a puzzle that had sat unresolved for years.

Benjamin Troutwine and Ryan Gray at the University of Texas at Austin engineered weaker versions of the same mutation, to see what happens if the fiber assembles normally and only later falls apart. Those fish grew a straight body, then progressively disassembled the fiber during larval growth, and curvature emerged as it went. Watching a labeled version of the protein in living fish, they saw the fiber moving continuously from head to tail, constantly secreted and renewed.

From fluid flow to muscle drive, by way of a neuropeptide

A thread in a fluid reaches muscle through a class of nerve cells that dip a sensory ending directly into the spinal fluid, called cerebrospinal-fluid-contacting neurons. Xiaoli Zhang and Chengtian Zhao at Ocean University of China, with Sudipto Roy in Singapore, traced the chain.

Cilia-driven flow transports adrenergic signals that switch on urotensin neuropeptides in those fluid-contacting neurons. The urotensins act on receptors on slow-twitch muscle fibers along the back, and the sustained pull of those fibers holds the axis straight. Knocking out the urotensin receptor gave adult zebrafish severe scoliosis closely resembling the human disorder.

Hao Lu and Sudipto Roy joined the two halves. Fish lacking the Reissner fiber survived into adulthood with severe curvature and markedly reduced urotensin expression. Treating them with epinephrine rescued the gene expression and the axial defects. Supplying Urp2 directly in the fluid-contacting neurons rescued body curvature in larvae, and in adults.

Look at what that last result is. A curve already present, in an animal already grown, moved back toward straight when a missing chemical voice was restored. Nobody straightened a bone. They restored a signal, and the axis followed.

Removing one balance organ curves a growing spine

Tadpoles that lost a balance organ before any curve existed grew into frogs with lateral curvature, vertebral rotation and wedged vertebrae, which settles the order of events in the vestibular findings.

Behind each eardrum, buried in bone, sits a sensor for motion and gravity. Three fluid-filled loops, the semicircular canals, are set at roughly right angles to each other, one for each plane of rotation. Turn your head and the fluid in a loop lags behind, bending a small flap of tissue with hair cells attached, and the bending becomes a nerve signal. Two further organs, weighted with tiny crystals, report the steady pull of gravity and straight-line acceleration.

Signals from these organs run down dedicated pathways into the spinal cord, where they set the background drive to postural muscles on each side. An asymmetry up there becomes an asymmetry of muscular pull down here.

The obvious objection is direction of causation, since a twisted spine could produce vestibular oddities on its own. Francois Lambert and Pierre-Paul Vidal, neuroscientists working in Paris, with Hans Straka in Munich, designed a study that settles the order.

They took tadpoles of the frog Xenopus laevis and removed the balance organs on one side while the animals were still larvae, long before any curve existed. After the tadpoles matured, radiographs and three-dimensional micro-CT scans showed spinal curvature in both the frontal and side planes, with rotation along the body axis and deformed vertebrae.

The team also explained why the lesion stuck. On land, an animal that loses one balance organ rebalances over time, using weight-bearing information from the limbs to recalibrate. Floating in water, that limb loading is absent. The lopsided drive down the spinal pathways never gets corrected, and the soft growing skeleton slowly takes its shape.

Human vestibular findings that predate the curve

Martin Hitier and colleagues in Caen, France went looking for a vestibular difference that could not be a consequence of a curve. The bony shape of the inner ear is fixed before birth and does not change afterward. They imaged the vestibular organs of eighteen adolescents with idiopathic scoliosis and nine controls.

The left lateral canal was more vertical and further from the midline in the scoliosis group, and the two measurements correlated strongly with each other. Warm and cool water testing of canal function pointed the same way, with lower excitability and more canal weakness. Those functional differences did not reach statistical significance against controls.

The drive can also be measured directly. Galvanic vestibular stimulation passes a small current between two electrodes placed behind the ears, which biases the balance organ toward one side and makes a standing person lean. Jean-Philippe Pialasse and Martin Simoneau at Universite Laval in Quebec used it on adolescents with mild and severe curves, standing with eyes closed. Both scoliosis groups showed greater sideways body displacement and greater net sideways force than healthy controls, during the stimulation and after it stopped.

The persistence after the stimulus matters more than the size of the response. It falls in the window where a healthy system is busy re-weighting its senses and settling back down.

The balance deficit in scoliosis sits in the re-weighting, and it does not scale with the curve

Adolescents with idiopathic scoliosis sway more when ankle position sense is corrupted, and grow more variable rather than less when the senses are handed back. Postural control shows no correlation with the size of the curve.

Stand as still as you can. You are not still. You are falling in tiny amounts and catching yourself, continuously, in every direction. A force plate is a rigid platform with load sensors under it. It computes the center of pressure, the point on the floor where your weight is currently balanced, many times a second. Watching that point wander gives a readout of the balance controller at work.

A healthy nervous system re-weights its senses. In good light on solid ground it leans on vision and ankle position sense. Take the light away, and it dials vision down and the other channels up. Put the person on a soft surface, and it dials ankle information down. The weighting shifts to fit conditions, and shifts back.

Martin Simoneau and colleagues at Universite Laval built an experiment to test that machinery in scoliosis. They removed vision by asking participants to close their eyes. They corrupted ankle position sense using the same tendon vibration trick Roll and Vedel had used, which floods the spindles with false information.

Adolescents with idiopathic scoliosis showed a larger center of pressure range and more variable sway velocity whenever ankle proprioception was disturbed, with or without vision. They were leaning on ankle position sense more heavily than controls were.

What happens when the senses come back

The companion study went after the sharper question. The team perturbed vision and ankle information and then abruptly restored them. Healthy adolescents settled, and their sway variability dropped as the returning information was folded back in. The scoliosis group got more variable instead, most clearly with the eyes closed. Sway analysis suggested they were failing to scale their balance commands during the switch. The deficit sits in the re-weighting itself, which is a property of the coupling between channels.

The sensory finding does not track the size of the curve

A team led by Dan Wang tested thirty-five young adults with curves between ten and thirty-nine degrees against twenty-nine healthy controls. Postural control was measured under four standard sensory conditions plus a harder one involving rhythmic head movements to load the vestibular system. Spinal shape was measured separately using surface topography. The scoliosis group swayed significantly faster and further sideways under the hardest vestibular condition. Then came the result that matters most: no significant correlation appeared between postural control and spinal shape in either group.

The sensory abnormality does not scale with the size of the curve. That is the wrong shape for a consequence.

If a bent spine were producing the balance findings, worse bends would produce worse balance. They do not. The two things coexist and vary independently, which is what a shared upstream origin looks like.

The pinealectomy results split by species, and the split is the finding

Removing the pineal gland curved chickens, curved rats only when they were made to stand on two legs, and curved none of eighteen monkeys. One removal, delivered into differently organized bodies, produced different spines.

Deep in the middle of the brain sits a gland the size of a grain of rice. The pineal gland makes melatonin, the hormone of darkness, released at night and suppressed by light.

In 1993 Masafumi Machida and colleagues, including the French spine surgeon Jean Dubousset, removed the pineal gland from newborn chickens. They wanted to know whether a purely neurohormonal manipulation, with no mechanical insult to the spine, could produce a curve. It could. Every pinealectomized chicken developed scoliosis within two weeks, progressing over the following month or two into a three-dimensional deformity with lateral curvature and vertebral rotation.

Chickens given a pineal gland transplanted into their trunk muscle developed it only ten percent of the time. The brains showed no structural damage in either group. The birds that curved did show delayed cortical evoked potentials, which the authors read as a conduction disturbance above the brainstem.

It looked like a solved problem. The next twenty years took it apart.

Posture decides whether the same gland matters

Machida's own group tested whether the result transferred to mammals. They pinealectomized rats, some kept in the normal four-legged posture and some made to stand and move on two legs. Scoliosis developed only in the two-legged rats. The four-legged rats, missing exactly the same gland and exactly the same hormone, stayed straight. Melatonin pellets prevented the curve in all but one of ten two-legged animals.

Then the same experiment was run in a primate. Kenneth Cheung and colleagues in Hong Kong removed the pineal gland from eighteen young rhesus monkeys, with urine testing confirming complete loss of melatonin production in ten. Seven died prematurely and eleven survived to follow-up, at a mean of twenty-eight months. Not one monkey developed scoliosis. A null carried by eleven surviving animals is weaker than one carried by eighteen.

Even the chicken result loosened. Keith Bagnall and colleagues in Alberta ran a dose-controlled study and found that only fifty-six percent of eighty-nine pinealectomized chickens developed a curve at all. A physiological daily dose of melatonin, which restored the normal circadian rhythm on assay, changed neither the rate nor the pattern of scoliosis.

In human patients the receiving apparatus is what differs

Alain Moreau and colleagues in Montreal took bone-forming cells from forty-one adolescents with idiopathic scoliosis during surgery and tested how those cells respond to melatonin in a dish. Signaling was impaired in every scoliosis patient tested, and the patients sorted into three distinct response groups. What was measured there is the response to the signal rather than the size of it.

The standard reading of this literature is that the melatonin theory failed and animal models do not transfer. That reading throws away the most informative dataset in the field.

Read it through tone instead. The same input was removed every time, and what differed was the system it was removed from. In a chicken, upright on two legs with a fast-growing axial skeleton, losing the night signal collapsed the range and every animal curved.

In a rat kept on four legs the same loss changed nothing, because that animal's postural arrangement does not lean on that signal. Stand the same rat on two legs and the curve appears. In a monkey, with a mature and differently weighted postural arrangement, the loss was absorbed entirely.

Input meets tone. The same removal, delivered into different organizations, produces different bodies.

Whether the circulating hormone itself is low in these children is still argued over, with published studies on both sides. The response data do not depend on settling that. A system can be handed an adequate signal and still read it differently, which is a statement about organization rather than supply. That variability across species is the finding.

Autonomic measures in scoliosis shift modestly, in several places at once

Girls with adolescent idiopathic scoliosis carry a higher low-frequency heart rate variability band, at 631 against 540 milliseconds squared, together with elevated catecholamines.

Underneath everything the body does runs an automatic control system with two halves. The sympathetic half is the accelerator. It speeds the heart, tightens vessels, raises alertness and readies the body for effort. Its chemistry runs on catecholamines, the family that includes adrenaline and noradrenaline. The parasympathetic half, carried mostly by the vagus nerve, is the brake. It slows the heart, settles digestion and lets the body recover.

Neither half is good or bad. What matters is the balance between them and how freely it moves.

That balance is measurable in an unlikely place. The gaps between consecutive heartbeats are never identical, and the pattern of variation carries information about the two halves. A joint European and North American task force standardized the measurement and interpretation of heart rate variability in 1996, and it has been a validated index of autonomic state ever since.

What the numbers look like in adolescent curves

Zongshan Hu and colleagues in Hong Kong and Nanjing recorded heart rate variability and catecholamines in fifty-eight girls with adolescent idiopathic scoliosis and twenty-one age-matched healthy girls. The low-frequency band was higher in the scoliosis group, at 631 against 540 milliseconds squared, with a p value of 0.045. Catecholamine levels were elevated. The high-frequency band showed no group difference. The ratio between the two bands came in at 1.11 against 0.95, with a p value of 0.058, which does not clear the conventional threshold.

That is a modest result with a null inside it, and the low-frequency band is a debated marker. It is often used as a rough proxy for sympathetic activity, and it is contaminated by vagal and blood-pressure-reflex contributions, so it is not a clean sympathetic readout.

The model does not need it clean. One band shifted, a related ratio trending the same way without reaching significance, and elevated stress chemistry in the blood together make the signature of a distributed change in background gain. A louder accelerator in one isolated place would move one measure hard.

In scoliosis it moved a little, in several places, which is what a distributed retuning looks like. Heart rate variability is one of the few faces of tone a machine reads directly, and reading it that way is the model's claim rather than a finding the autonomic literature makes.

Growth plates convert a held posture into permanent geometry

Compressed vertebrae grow at 68 percent of the normal rate and distracted vertebrae at 114 percent, which is the machinery that turns a tilt into a wedge and a wedge into a curve.

A growing bone lengthens at a disc of cartilage near each end, called a growth plate. Cells there divide, stack into columns, swell, and then harden into bone. The rate responds to load.

Ian Stokes and colleagues at the University of Vermont set out to measure that response precisely. They were testing an old orthopedic rule, the Hueter-Volkmann principle, which says compression slows growth and tension speeds it, and which had conflicting evidence behind it.

They fitted small external rings to rat tail vertebrae and applied steady compression or steady distraction for several weeks, measuring vertebral length on serial radiographs. Compressed vertebrae grew at sixty-eight percent of the normal rate. Distracted vertebrae grew at one hundred and fourteen percent.

Why a tilt deepens itself

The consequence assembles in three steps. A spine held slightly tilted loads one side of each vertebra more than the other. The loaded side grows slower. The vertebra becomes a wedge instead of a block. Stack several wedges and the tilt increases, which increases the load asymmetry, which increases the wedging. Stokes later built a computational model of that feedback and showed that the loop is capable of driving progressive deformity during adolescent growth.

That vicious cycle is real, and it explains a great deal about the behavior of curves. It explains why they accelerate during the growth spurt. It explains why they mostly stop at skeletal maturity. It explains why the same number of degrees means something different in a twelve-year-old and a nineteen-year-old.

What it does not explain is the first tilt. A feedback loop amplifies whatever is fed into it. Growth-plate mechanics obey a direction that something upstream has already chosen.

Tone bends the body. Growth writes it down. The Cobb angle is the transcript.

This is also why time is the decisive variable in scoliosis. Every month of growth converts a little more of a postural decision into permanent geometry, and a wedged vertebra removes room the spine had to come back with. Restoring the regulation upstream has the most to work with when the least has been printed. The tension network that carries this loading is the subject of the constraint page, which sets out how a body holds its shape through balanced pull.

What must be ruled out in scoliosis, and what happens when it is fixed

Whole-spine imaging finds a structural abnormality of the nervous system in seventeen to twenty percent of children ten and under with curves over twenty degrees. In adolescents with a typical pattern and a normal examination the yield falls below two percent.

Some curves are driven by something specific, visible and treatable, and those must be found. The list is short and well known. A Chiari I malformation is a downward displacement of part of the cerebellum through the opening at the base of the skull, which obstructs the flow of spinal fluid.

A syringomyelia is a fluid-filled cavity inside the spinal cord itself. A tethered cord is a spinal cord anchored abnormally at its lower end, so that growth puts it under traction. A tumor inside or beside the cord distorts it as it grows.

Purnendu Gupta with Lawrence Lenke and Keith Bridwell imaged the whole spine in children ten years old and younger who had idiopathic-pattern curves over twenty degrees and a normal neurological examination. They found neural axis abnormalities in seventeen to twenty percent of them, and in half of the infants.

In adolescents the yield is far lower. Tu Do and colleagues at the Hospital for Special Surgery imaged 327 consecutive adolescents who had a typical idiopathic curve pattern and a normal physical and neurological examination before surgery. They found a syrinx in 0.6 percent and a Chiari malformation in 1.2 percent, and no patient required neurosurgical intervention.

Young age, an unusual curve pattern, a left-sided thoracic convexity, pain, or any abnormality on neurological examination all raise the priority of imaging. That judgment belongs with the treating physician.

Operating on the skull changes curves in the spine

Douglas Brockmeyer and colleagues in Salt Lake City reviewed twenty-two children who came in with scoliosis and turned out to have a Chiari I malformation with syringomyelia, twenty-one of whom formed the analyzed group. Each had a suboccipital decompression, an operation at the base of the skull that relieves the obstruction to fluid flow. Nobody operated on the spine. Sixty-two percent of the curves improved or stabilized. In children under ten years old, ten of eleven improved or held steady.

A systematic review and meta-analysis by Davide Palombi and colleagues in Rome pooled eleven studies covering 380 children with Chiari-associated scoliosis treated with posterior fossa decompression as the first operation. Fifty-five percent of curves improved or stabilized and forty-five percent progressed. Younger age at surgery and a smaller starting curve predicted the better outcomes. Set against that, 37.8 percent of these children still went on to need spinal fusion.

Surgeons operated at the base of the skull, on fluid dynamics and neural crowding, and about half of the spinal curves in a growing child changed course. The vertebrae were never touched.

No stronger evidence exists that a curve can be downstream of the nervous system. The failures are as instructive as the successes. The predictors of failure were older age and a bigger curve, which is to say less growth left and more of the pattern already printed into bone.

What bracing and fusion do to a curve, and what they leave unchanged

Bracing brought seventy-two percent of adolescents to skeletal maturity below the surgical threshold against forty-eight percent under observation, by constraining the geometry while the remaining growth runs out.

Whether bracing worked was genuinely uncertain until a well-built multicenter trial settled it. Stuart Weinstein, Lori Dolan and colleagues enrolled 242 adolescents with typical bracing indications into a study with both a randomized arm and a preference arm. Success meant reaching skeletal maturity without the curve progressing to fifty degrees, the usual threshold for considering surgery.

Across both cohorts, seventy-two percent succeeded with bracing against forty-eight percent with observation. In the randomized comparison the gap was seventy-five percent against forty-two. Success rose steadily with hours worn per day, and the trial was stopped early because the benefit was clear.

That result should change what a family does. Understand what produced it. A brace applies an external force that opposes the growth asymmetry described above. It constrains the geometry mechanically while the remaining growth runs out. It manages the output without touching the regulation that produced it. Remove the brace and the regulator is unchanged. A plaster cast manages a fracture the same way, and casts are worth having.

Surgery goes further. Instrumented fusion corrects deformity that nothing else corrects, and for large curves in a growing child it is sometimes the right call. It works by removing the motion of the affected segment and locking the geometry permanently. It is the most complete possible constraint on the output, and it exchanges range for stability by design.

The fifty-year natural history sets the stakes

Weinstein and colleagues followed 117 adults with untreated late-onset idiopathic scoliosis for fifty years, against sixty-two matched controls. Survival was comparable to the general population. Chronic back pain was more common, at sixty-one percent against thirty-five, though most of it was rated little or moderate. Breathlessness in daily activity rose mainly in those with curves beyond eighty degrees at a thoracic apex. Their conclusion was that untreated adults with this condition are productive and functional at a high level.

Hold both facts at once. Progression is worth preventing, and most curves are not a catastrophe.

Restoring tone aims at the width of the range around midline rather than at the degrees on the film. How freely can the system be perturbed and still return? How well does it re-weight its senses when conditions change? How much of its regulation is available rather than committed? The degrees are downstream of that. The model claims the range is where the disorder lives, and therefore where a correction has to land.

What a restoring input does to a scoliosis curve that a push cannot

Give the same tonal input to a child with a right-convex curve and a child with a left-convex curve. A mechanical push helps one and worsens the other. The model predicts both move toward midline.

A mechanical push has a direction. Press a spine from the right and it moves left, whoever owns the spine. A drug has a direction too. Lower a number and it goes down, in the person who needed it lowered and in the person who did not.

A restoration of tone should behave differently, because it acts on the regulator instead of the output. The prediction is bidirectional restoration. A genuine tonal input should move a dysregulated system toward the healthy middle from either side. If the input is mechanical, one child improves and the other worsens, and the effect averages out to nearly nothing across a mixed group.

The same test runs on a force plate. The force-plate studies above found scoliosis groups swaying more than controls under sensory challenge. If a subgroup sits on the other side of the healthy range instead, holding itself unusually rigid, the prediction has a second edge.

A one-directional intervention pushes everyone the same way, and helps one group while harming the other. The model predicts that excessive sway comes down and deficient sway comes up, both toward the middle. Measure the subgroups first, then apply the input.

An input that reliably improves right-convex curves while worsening left-convex ones is pushing the spine rather than restoring the regulator.

That contrast is stated in advance, in a form a properly designed study could deliver inside a single season.

The sensing changes first, and the skeleton follows

The most common objection to any nervous-system account of scoliosis is that a twisted spine could produce the sensory findings. The order of events in the strongest studies answers it. Blecher and Zelzer's 2017 mice had normal vertebrae and symmetric muscles when the sensory neurons were disabled. Assaraf's 2020 deletions spared bone and cartilage entirely.

Lambert's 2009 tadpoles lost a balance organ before any curve existed. Grimes and Ciruna prevented the curve by fixing cilia, and stopped its progression by restoring cilia motility after it had begun. In every case the sensing was altered first and the skeleton followed. Add Wang's finding that the sensory abnormality does not scale with curve size, and the consequence story runs out of room.

Five rival theories, one variable

Renaming five findings would advance nothing. The claim is that these are not five findings. Proprioceptive genetics, vestibular asymmetry, spinal fluid flow, autonomic gain and growth-plate loading appear in the literature as five rival theories. Each competes to be the cause and none explains the other four. The model reads them as five channels of one variable, which is why each produces real effects and none produces a complete account. The unification generates the bidirectional prediction above, and a label generates no predictions.

What this reading forbids

A model that accounts for every result risks forbidding nothing. This one forbids several things. It forbids a purely local skeletal defect turning out to explain idiopathic curves. It forbids any single-channel intervention working reliably across an unselected population, which is what happened to the melatonin literature. And it forbids the one-directional result described above.

Where bracing and surgery stand in this account

Bracing reduces progression and the trial that showed it was well designed. Surgery corrects deformity nothing else can correct. Findable causes must be found, and imaging must be ordered when the picture calls for it. The disagreement is about the level at which this disorder lives.

Idiopathic does not mean causeless. It means nothing was found in the place they looked, and they were looking inside the bone. A spine is aligned because a nervous system keeps deciding, thousands of times a day, where aligned is. When that decision loses its range and settles into one lopsided answer, a growing skeleton does what growing skeletons do. It believes the answer, and it builds it.

How scoliosis relates to the rest of the library

Scoliosis is the library's clearest case of a regulated value being written into anatomy, and each neighboring page carries one strand of the argument.

Three foundations of tone do the heaviest work here.

• Constraint holds the mechanics: how a body carries shape through balanced pull, what slack is for, and why a wedged vertebra is room the spine no longer has.
• Set point is the general physics of a defended value, taught there in pressure and temperature and here in a midline the muscles defend off center.
• Time course is why the same curve carries different urgency at twelve and at nineteen, and why an entrenched pattern behaves unlike a fresh one.

The remaining foundations each hold a strand.

• Gain is the volume control that turns a small vestibular current into an outsized lean.
• Oscillation covers the rhythms that carry regulation, including the ciliary beat that drives the spinal fluid current.
• Prediction is the system acting on its estimate rather than on the world, which is what the tendon-vibration illusion exposes.
• Load is the running cost of holding a posture, and in a growing spine the bill is paid in bone.
• Input quality is the fidelity of the body's reports about itself, degraded in the Piezo2 mutants and in the two PIEZO2 patients.
• Coupling is the re-weighting between channels that fails on the force plate.
• The autonomic nervous system is the anatomy behind the catecholamine and heart rate variability findings.

Six more pages complete the picture.

• Balance and coordination teaches the postural control system in full, the machinery this condition tests to failure.
• Senses covers position sense and gravity sense as senses in their own right, which is what makes an idiopathic curve legible.
• Pediatrics is where growth and regulation meet, and the window scoliosis lives inside.
• Neurophysiology supplies the wiring, from spindle to spinal cord to cortex.
• Low back pain holds the adult sequel, the sixty-one percent figure from the fifty-year cohort.
• And dysautonomia is where the autonomic strand of this page becomes the presenting problem in its own right.

Frequently asked