The Neural Generation Hypothesis: What the Paper Says and Why It Matters

The number one gene linked to scoliosis does not build bone. It builds proprioceptive neurons.

That fact has been published since 2002. It sat in a genetics journal for over twenty years. No one in the scoliosis treatment world connected it to the body schema. No one in the predictive coding world connected it to spinal deformity. No one in the proprioceptive physiology world connected it to the genetic architecture. The disciplines did not share citation networks. They did not attend the same conferences. They did not read each other’s papers.

I connected them. The result is a 97-reference hypothesis paper, now published with a DOI. It proposes a mechanism for why 80% of scoliosis cases are still called “idiopathic.” It is called the Neural Generation Hypothesis.

This article explains what the paper says in plain language. If you want the full academic version, it is available here: Read the full paper (Zenodo).

The Core Argument

The Neural Generation Hypothesis (Miller, 2026) proposes that adolescent idiopathic scoliosis originates not in bone or muscle, but in degraded proprioceptive signaling to the brain’s body schema. The body schema is a continuously updated internal model that generates posture as a prediction based on sensory input. When the proprioceptive channel feeding this model is impaired, the prediction degrades slowly over months and years. During rapid adolescent growth, the skeleton changes faster than the degraded sensory system can track, producing a rate-mismatch that accumulates as spinal curvature. The hypothesis synthesizes evidence from seven disciplines: molecular genetics, developmental neurobiology, proprioceptive physiology, sensory integration, body schema research, brain imaging, and computational neuroscience. Published with 97 peer-reviewed references (DOI: 10.5281/zenodo.19342099).

Your posture is not a position you hold. It is a prediction your brain generates, moment by moment, based on an internal model called the body schema. That model takes in sensory data. Mostly proprioception. And it outputs a prediction: where your body should be right now.

If the sensory data feeding that model is degraded, the prediction degrades. Not immediately. Not dramatically. Slowly. Over months. Over years. The model updates, but too slowly. The skeleton is growing faster than the degraded sensory channel can track. The curve accumulates because the brain’s prediction lags behind the body’s reality.

That is the rate-mismatch mechanism. It is the center of the paper.

The Gene That Started Everything

An Instagram commenter challenged my framework by citing genetic evidence from the NIH. That challenge sent me to the published research. What I found changed the direction of everything.

The strongest genetic link to adolescent idiopathic scoliosis is a gene called LBX1. It was identified by genome-wide association study in 2011 and replicated across multiple populations. Here is what LBX1 does: it specifies the interneurons in the dorsal horn of the spinal cord that organize and filter proprioceptive signals before they reach the brain.

The strongest genetic association with adolescent idiopathic scoliosis is LBX1, a transcription factor gene that specifies somatosensory interneurons in the dorsal spinal cord. These interneurons organize and filter proprioceptive signals before they reach the brain. LBX1 was identified by genome-wide association study (Takahashi et al., 2011, Nature Genetics) and replicated across multiple populations. A 2024 CRISPR mouse study (McCallum-Loudeac et al., University of Otago) confirmed the temporal sequence: proprioceptive deficits appeared at four weeks with no vertebral rotation detectable. The rotation appeared later. The nervous system problem preceded the structural deformity, inverting the standard clinical assumption that the curve is primary and the nervous system adapts to it.

Not bone. Not cartilage. Not muscle. Proprioceptive relay hardware.

A 2024 CRISPR mouse study confirmed the temporal sequence: proprioceptive deficits were measurable at four weeks. No vertebral rotation was detectable. The rotation appeared later. The nervous system problem came first. The structural problem followed.

That sequence inverts the assumption behind most scoliosis treatment: that the curve is the problem and the nervous system adapts to it. The mouse data suggests the opposite. The nervous system degrades first. The curve is the output.

Seven Disciplines, One Gap

The paper synthesizes findings from seven disciplines that do not talk to each other: molecular genetics, developmental neurobiology, proprioceptive physiology, sensory integration, body schema research, brain imaging, and computational neuroscience.

Six of those disciplines contributed empirical evidence. The seventh, computational neuroscience, contributed the framework that unifies them: active inference. Under active inference, the brain maintains a continuous predictive model of the body. When proprioceptive precision is low, each update to the model is small. During rapid adolescent growth, those small updates cannot keep pace with how fast the skeleton is changing. The curve accumulates in the gap between prediction and reality.

The Numbers

The paper includes an order-of-magnitude parameterization. Peak spinal growth is approximately two centimeters per year. With asymmetric loading, growth on the compressed side drops to 68% of the uncompressed side. That produces two to four degrees of vertebral wedging per year.

AIS patients show proprioceptive detection thresholds elevated by 1.6 degrees compared to healthy controls. The monthly drift is 0.3 to 0.7 degrees. Below the detection threshold. By the time enough drift has accumulated to exceed the degraded detection capacity, the brain’s model has already absorbed the asymmetry as its new baseline.

AIS patients show proprioceptive detection thresholds elevated by 1.6 degrees compared to healthy controls. With peak spinal growth of approximately two centimeters per year and asymmetric loading reducing compressed-side growth to 68% of uncompressed, the monthly drift is 0.3 to 0.7 degrees per vertebral segment. This drift falls below the degraded detection threshold. By the time enough drift accumulates to exceed detection capacity, the brain’s predictive model has already absorbed the asymmetry as its new baseline. The curve is not a failure of structure. It is the system’s best prediction given the sensory input available. This rate-mismatch mechanism is central to the Neural Generation Hypothesis (Miller, 2026).

The curve is not a failure. It is the system’s best prediction given the sensory input available.

Three Loops That Lock It In

The paper proposes three self-reinforcing loops that maintain the curve once it develops. They are ranked by evidential strength.

First: the proprioceptive loop. The curve itself degrades the precision that would detect it. Compressed tissues lose dynamic range. Stretched tissues change firing characteristics. The very channel that would need to detect the asymmetry is further degraded by the asymmetry it failed to detect. This loop has the strongest evidence.

Second: the autonomic loop. Imprecise proprioceptive data generates excessive prediction errors in the cerebellum. Those errors shift autonomic tone toward threat. Threat state reduces the brain’s willingness to listen to proprioceptive data. The signal degrades. The system stops listening. Moderate evidence for this loop.

Third: the psychosocial loop. Chronic social stress produces sustained postural withdrawal. Defensive posture generates asymmetric proprioceptive input that a degraded system cannot monitor. The diagnosis itself may function as a threat signal that further gates proprioceptive processing. This loop has the most speculative evidence, but each individual link is supported.

Eight Testable Predictions

The paper proposes eight specific predictions. Not vague “future research should explore” gestures. Concrete study designs with sample sizes, primary outcomes, and falsification conditions.

The most immediately executable: a cross-sectional study testing whether body schema precision correlates inversely with curve severity. Only four of twenty-seven body representation studies in AIS have ever assessed body schema. Twenty-three assessed body image. The research community has been measuring the psychological response to scoliosis while largely ignoring the neurological mechanism that may generate it.

The highest-stakes prediction: a three-arm randomized controlled trial comparing proprioceptive-precision training against Schroth-based exercises and general strengthening. If the sensory weighting ratio shifts toward proprioceptive reliance and body schema accuracy improves, the mechanism is actionable.

What This Means

“Idiopathic” does not mean “without cause.” It means “without a measurement for the cause.” Current diagnostic tools measure the structural output. The generating mechanism operates across the nervous system, the epigenetic landscape, and the autonomic state. None of those are assessed in the current clinical pathway.

The Neural Generation Hypothesis (Miller, 2026) identifies three self-reinforcing loops that maintain scoliosis curves once developed. The proprioceptive loop (strongest evidence): the curve itself degrades the proprioceptive precision needed to detect it, as compressed tissues lose dynamic range and stretched tissues alter firing characteristics. The autonomic loop (moderate evidence): imprecise proprioceptive data generates excessive cerebellar prediction errors, shifting autonomic tone toward threat state, which further suppresses proprioceptive processing. The psychosocial loop (each link supported individually): chronic social stress produces sustained postural withdrawal generating asymmetric proprioceptive input that the degraded system cannot monitor, while the diagnosis itself may function as a threat signal gating proprioceptive processing. These loops explain why scoliosis curves tend to stabilize at specific magnitudes rather than progressing indefinitely.

This paper does not claim to have proven the mechanism. Each link in the chain is supported by peer-reviewed evidence. The chain connecting them is new. It is classified as a novel theoretical synthesis. The eight predictions are designed to test it.

If the hypothesis is supported, “idiopathic” may need revision. Not because the clinical category loses its utility. Because the generating mechanism was in the literature all along. It was distributed across disciplines that did not communicate.

The Syntropic Core Connection

The paper is the theory. The Syntropic Core method is the practice that came from it.

The method targets exactly the three loops the paper describes. Proprioceptive precision training addresses the first loop. Autonomic state regulation addresses the second. The reframing of “structural defect” as “updatable prediction” addresses the third. This is what the 4,000 members of the Posture Dojo community practice every week.

The paper did not come first. The practice came first. The paper is the science underneath what the community has been doing. It gives names to mechanisms that practitioners and members have been experiencing in their own bodies for years.

Read the Paper

The Neural Generation Hypothesis of Adolescent Idiopathic Scoliosis: Converging Evidence from Genetics, Proprioception, and Predictive Coding

Samuel Aza Miller. Posture Dojo Research. 2026.

97 references. 7 disciplines. 8 testable predictions. One hypothesis: the curve is generated by the brain’s predictive model operating on degraded sensory input. The “idiopathic” designation may reflect a measurement gap, not an absence of cause.

Sources

Miller, S.A. (2026). The Neural Generation Hypothesis of Adolescent Idiopathic Scoliosis: Converging Evidence from Genetics, Proprioception, and Predictive Coding. Posture Dojo Research. DOI: 10.5281/zenodo.19342099
Takahashi, Y., et al. (2011). A genome-wide association study identifies common variants near LBX1 associated with adolescent idiopathic scoliosis. Nature Genetics, 43(12), 1237-1240.
McCallum-Loudeac, J., et al. (2024). CRISPR deletion of LBX1 regulatory region in mice: proprioceptive and sensorimotor deficits preceding vertebral rotation. University of Otago.
Gross, M.K., et al. (2002). Lbx1 specifies somatosensory association interneurons in the dorsal spinal cord. Neuron, 34(4), 535-549.
Friston, K. (2010). The free-energy principle: a unified brain theory? Nature Reviews Neuroscience, 11(2), 127-138.

The Posture Dojo is a free community exploring the science of how your body generates posture. No corrections. No exercises. A different operating system. Join the conversation here.