I was sixteen when they showed me the X-ray.
Eighty-five degrees. Kyphoscoliosis. The orthopedic surgeon traced the curve with his finger like he was drawing a road on a map. He talked about Harrington rods, about fusion, about what the hardware would and would not do.
I looked at the film on the lightbox and thought: How did my body do that?
Nobody answered. Nobody could. Because the honest answer, the one printed on every chart and whispered in every referral, was this: We don’t know.
They had a word for that honesty. They called it idiopathic.
The word that replaced the answer
If you or someone you love has been diagnosed with scoliosis, you have almost certainly encountered this word. Idiopathic. It sounds clinical. It sounds precise. It sounds like a category of knowledge.
It is not.
Idiopathic comes from the Greek idios (one’s own) and pathos (suffering). It literally means “a disease of its own kind.” In clinical practice, it means: we measured the curve, and we do not know why it is there.
Eighty percent of scoliosis cases carry this label. Eight out of ten. That is not a rare gap in the research. That is the majority of every scoliosis diagnosis ever written. The most common spinal deformity in adolescents, and the field’s official position on its cause is a shrug dressed in Latin.
Why is most scoliosis called idiopathic? Because the standard diagnostic model measures the shape of the spine but has no framework for identifying what is generating that shape. “Idiopathic” is not a diagnosis. It is an admission that the model is incomplete.
This matters, because what you believe about the cause determines what you believe is possible.
If the cause is unknown, the options are narrow: watch it, brace it, or fuse it. Manage the shape. Accept the curve.
But what if the cause is not unknown? What if the model was just looking in the wrong place?
The spine is the output, not the input
Here is the foundational error in the standard scoliosis model: it treats the spine as the problem.
The curve shows up on the X-ray. The Cobb angle gets measured. The vertebrae are rotated, wedged, shifted. So the interventions target the spine. Braces compress it. Rods fuse it. Exercises try to straighten it.
But the spine does not move itself. It never has. The spine is moved by muscles, which are driven by nerves, which are organized by the brain’s model of where the body is in space. The vertebral column is the last thing in the chain to respond, not the first thing to go wrong.
The spine is the printout. The question is: what is running the printer?
This is not a metaphor. It is the literal biomechanical reality. Your nervous system generates a pattern of muscle activation. That pattern produces forces on the skeleton. Those forces, sustained over months and years of growth, shape the bone. The curve you see on the X-ray is the structural record of a neuromuscular instruction that has been running, without interruption, for as long as the curve has been developing.
The standard model describes the shape. It does not explain the generator.
The generator model
So what is generating the curve?
Over the past two decades, research from multiple independent groups has converged on an answer. It was not discovered in a single breakthrough. It was assembled, piece by piece, from vestibular neuroscience, animal models, sensory integration research, and clinical observation. And it points upstream from the spine to something the standard model almost never examines: the vestibular system.
> “The idiopathic generator model proposes that scoliosis labeled ’cause unknown’ is actually driven by a feedforward loop involving vestibular asymmetry, aberrant righting reflexes, and cerebellar miscalibration.”
Here is the chain.
Step 1: Vestibular asymmetry. Your vestibular system, housed in the inner ear, tells your brain which way is up. Specifically, the otolith organs (the utricle and saccule) detect the direction of gravity. They are your internal plumb line. If the otoliths on the left and right sides are not symmetrical, if one is structurally different from the other, the brain receives a biased signal about vertical. The internal model of “which way is up” is tilted.
This is not theoretical. CT and MRI studies have documented structural asymmetries in the semicircular canals of adolescents with idiopathic scoliosis. The left lateral semicircular canal tends to be more vertical and further from the midline. An intercanal angle of 170 degrees or less is 100% specific for scoliosis. And critically, these morphological differences are determined before birth, well before any spinal curve appears.
Step 2: Asymmetric righting reflexes. When the vestibular system says “you are tilted,” the brain does not just note the information. It acts on it. The vestibulospinal and reticulospinal pathways send descending commands to the trunk muscles to correct the perceived tilt. This is the righting reflex. It is automatic. It is continuous. And if the vestibular signal is biased, the correction is biased too.
The muscles on one side of the trunk receive more activation than the muscles on the other side. Not because those muscles are weak or tight. Because the brain is issuing an asymmetric instruction based on an asymmetric perception of gravity.
Step 3: Asymmetric loading during growth. During adolescence, the skeleton is still growing. The growth plates in the vertebrae are responsive to mechanical loading. This is governed by the Hueter-Volkmann Law: increased pressure on one side of a growth plate slows growth on that side, while reduced pressure on the other side allows faster growth. Asymmetric muscle activation produces asymmetric spinal loading. Asymmetric loading, sustained through the growth spurt, produces vertebral wedging. Vertebral wedging produces the curve.
Step 4: The feedforward loop. Now the curve exists. The proprioceptive system detects it. But the vestibular reference has not changed. The brain still perceives “up” as tilted. So the righting reflex continues to drive asymmetric activation. The curve deepens. The proprioceptive system adapts to the new shape. The body schema accepts the deformity as baseline. The Gamma Loop, which maintains muscle tension set points, recalibrates around the bracing pattern. The contraction becomes automatic, and then invisible.
The generator keeps generating. The output keeps printing.
The evidence nobody connected
This is not one rogue hypothesis. The evidence comes from multiple independent lines of research, across species, methodologies, and decades.
The Xenopus study (Lambert & Straka, 2009). This is the landmark. Researchers removed one vestibular organ from frog larvae. The frogs developed scoliosis as they matured, complete with lateral curvature, sagittal deformation, and vertebral rotation matching human scoliosis patterns. The critical insight: in aquatic vertebrates, there is no limb proprioception to compensate for the missing vestibular signal. The asymmetric descending drive persists permanently, progressively deforming the skeleton as it ossifies.
A follow-up study (Lambert et al., 2013) confirmed the specific mechanism: restricted neural plasticity in vestibulospinal pathways after unilateral vestibular loss. The spinal circuits cannot compensate. The asymmetric tone becomes permanent.
Pinealectomy in chickens (Machida et al., 1997-1998). Remove the pineal gland from a young chicken and approximately 75% develop scoliosis. The mechanism was initially attributed to melatonin disruption, but further investigation revealed the connection: melatonin modulates vestibular nucleus function. Disrupt melatonin, disrupt vestibular processing, produce a curve. And here is the telling detail: pinealectomy does NOT produce scoliosis in rats or hamsters. Only in bipedal or semi-bipedal species where axial postural control depends more heavily on gravity sensing. The more gravity-dependent the posture, the more vulnerable the spine is to vestibular asymmetry.
Subjective visual vertical testing (Cakrt et al., 2011; Simoneau et al., 2018; Bailly et al., 2021). When you ask someone with idiopathic scoliosis to align a line with “true vertical” in a dark room, they get it wrong. Consistently. Their internal model of vertical is tilted. This is not a postural problem; it is a perceptual problem. The body is curved because the brain’s reference frame for “straight” is biased.
Vestibular evoked myogenic potentials (Kucerova et al., 2025). A comprehensive assessment found that 38% of AIS patients had abnormal VEMP responses, indicating dysfunction in the otolith-cervical and otolith-ocular pathways. The vestibular hardware is measurably different.
Sensory reweighting (Simoneau group; Wang et al., 2025). Healthy postural control dynamically adjusts how much it relies on vision, proprioception, and vestibular input depending on conditions. AIS patients show a specific failure in this reweighting: they assign greater weight to vestibular information compared to controls, even though their vestibular system is the compromised channel. They are leaning harder on the biased signal. This creates a vicious cycle: the body relies more heavily on a tilted reference, which reinforces the asymmetric postural strategy, which maintains the curve.
99 out of 100. Across multiple species, unilateral vestibular stimulation or removal has produced scoliosis with a dose-response relationship. Ninety-nine out of one hundred scoliotic animal subjects showed associated equilibrium defects. The greater the vestibular disruption, the greater the spinal deformity.
This is not weak evidence. This is convergent evidence across species, methodologies, and research groups. The vestibular-scoliosis connection is one of the most consistently replicated findings in the field. It has simply never been integrated into the clinical model that most patients encounter.
For more on how this genetic thread connects to the nervous system, see The Scoliosis Gene That Proves It Is a Nervous System Problem.
Why the standard model missed it
If the evidence is this strong, why does the diagnosis still say “idiopathic”?
Because medicine is organized by organ systems, not by causal chains.
Scoliosis belongs to orthopedics. The vestibular system belongs to ENT or neurology. The righting reflexes belong to developmental neuroscience. The sensory reweighting research belongs to motor control physiology. No single specialty owns the full chain. Each department sees its piece. Nobody sees the loop.
The orthopedic surgeon looks at the X-ray and sees a curved spine. That is the problem they are trained to address. They are not trained to test vestibular function. They are not trained to assess sensory reweighting. They are not trained to trace the curve back to its generator.
This is not a criticism of individual clinicians. It is a structural observation about how the model is organized. The diagnosis describes the shape because the diagnostic framework was built to measure shapes. It was not built to identify generators.
And so the word “idiopathic” persists. Not because the cause is truly unknown, but because the model that would explain it spans too many departments to fit in a single referral.
The same pattern plays out in other areas of posture. Your diagnosis described the shape, not the generator. This is systemic, not specific to scoliosis.
The jaw, the eyes, and the rest of the upstream inputs
The vestibular system is not the only upstream input the standard model ignores. The jaw, the eyes, and the trauma history all feed into the same body schema that generates the postural pattern.
The TMJ (temporomandibular joint) shares neural real estate with the vestibular nuclei and the upper cervical spine through the trigeminal-vestibular complex. A jaw that closes asymmetrically sends an asymmetric signal into the same brainstem circuits that are trying to calibrate vertical. For many people with scoliosis, the jaw is a secondary generator that nobody examined because it “belongs to dentistry,” not orthopedics. The jaw-posture connection is one of the most underappreciated links in the entire postural chain.
Trauma operates through a different but parallel mechanism. A fall, a surgery, a sustained period of threat produces the Trauma Reflex: an asymmetric protective contraction (hiked hip, lateral trunk shift, rotated ribcage) that is mediated by the tectospinal tract. If the reflex is never resolved, it consolidates into the brainstem as an autopilot pattern. The brain stops perceiving the contraction. The muscles are not weak or tight by accident; they are following a protection order that was never rescinded. The trauma-posture connection is another thread in the generator web.
The curve on the X-ray is the output of all of these inputs running simultaneously. Vestibular asymmetry. Jaw asymmetry. Unresolved trauma reflexes. Proprioceptive degradation. Sensory reweighting failure. The spine is the last domino. It falls in the direction determined by everything upstream.
What changes when you see the generator
This reframe changes what is possible.
If scoliosis is a structural defect with no known cause, the options are limited: brace it, fuse it, or live with it. Monitor the shape. Manage the symptoms. Accept the curve as permanent.
But if scoliosis is generated, if there is an identifiable upstream process that is producing and maintaining the curve, then the generator can be addressed.
This does not mean the curve vanishes overnight. It does not mean bracing or surgery are never appropriate. It means the conversation changes from “we don’t know why this happened, so let’s manage the output” to “here is what is driving this, and here is how we address it upstream.”
Vestibular calibration. Sensory integration training. Proprioceptive restoration. Trauma reflex resolution. Body schema updating. These are not replacements for structural interventions when structural interventions are needed. They are the missing layer, the upstream work that the standard model does not include because it does not have a framework for it.
I know this because I lived it.
My 85 degrees
I carried that curve for seventeen years.
From sixteen, when they showed me the X-ray, to thirty-three, when I finally stopped trying the standard model and went looking for the actual cause. I tried physical therapy. I tried chiropractic. I tried exercise protocols. I tried “core strengthening.” Everything the system offered, I tried. And everything helped for a day, maybe two. Then the body returned to what it knew.
Because the interventions were targeting the output. The printer kept printing the same pattern because nobody addressed the code.
At thirty-three, I stopped working as normal. Eight months of complete focus. Not a gradual process. A concentrated, sustained search born from exhaustion with incremental approaches. And what I found, piece by piece, was the generator.
The vestibular miscalibration. The asymmetric pressure organization in my diaphragm. The Trauma Reflex that had been running uninterrupted since adolescence. The body schema that had accepted the deformity as baseline.
I did not straighten my spine with willpower. I updated the system that was generating the curve. And the body reorganized. Two inches of height. Not from surgery. Not from a brace. From changing the input so the output could change.
The curve did not vanish. The organizing forces that were maintaining it were renegotiated. The nervous system updated its model of where my body was in space, and the structure followed.
The diagnosis describes the shape. We address the generator.
Here is what I want you to take from this.
“Idiopathic” is not a final answer. It is a placeholder. It means the diagnostic model ran out of framework before it ran out of cause. The cause is not unknown. It is unexamined by the model that wrote the diagnosis.
The spine is never the starting point. It is always the endpoint. The curve is the structural record of an upstream process that has been running, unchecked, for years. Vestibular asymmetry feeding asymmetric righting reflexes feeding asymmetric muscle activation feeding asymmetric skeletal loading. A feedforward loop that produces a curve and then maintains it.
If you have been told your scoliosis has no explanation, you have not been told the truth. You have been told the limits of the model. The body always has a reason. The question is whether the framework you are using is capable of finding it.
Your body did not randomly curve. Something is generating the pattern. And if it is generated, it can be addressed.
—
The diagnosis describes the shape. We address the generator.
If you have scoliosis and you are ready to look upstream, our scoliosis-specific cohort works with the generator model directly. Not exercises for your spine. A systematic protocol for the system that is organizing your spine.
[Learn about the scoliosis cohort at Posture Dojo →]
—
Sam Miller is the founder of Posture Dojo and the author of UPRIGHT. He was diagnosed with 85-degree kyphoscoliosis at age 16, declined surgery, and spent 17 years searching for the answer the diagnosis could not provide. He found it.