Most cases of adolescent idiopathic scoliosis have no single proven cause. Genetics plays a clear part, through many genes that each add a little. Growth, the nervous system and hormones are all studied, and none is settled. Once a curve exists, growth can make it progress. Our Neural Generation Hypothesis is a proposed explanation, not an established one.
This page is for parents, teens and adults who want to know what the research says about why adolescent idiopathic scoliosis (AIS) happens, and who want it separated into what is established, what is still argued over, and what is hypothesis. We cite the sources that clinicians use, including the Scoliosis Research Society (SRS), the US National Library of Medicine through MedlinePlus, and published reviews indexed in PubMed. Our own work comes last and is labelled as ours.
Conflict of interest: Posture Dojo Research is funded by Posture Dojo memberships, and I teach Syntropic Core, a practice that sits alongside this research. That is a reason to want our hypothesis to look good. So it is kept apart from the established findings below, labelled as synthesis [T3], and the established findings are cited from independent sources.
Quick answers
| Question | Short answer | Evidence tier |
|---|---|---|
| Is there a proven single cause? | No. Reviews describe a condition with no recognizable single cause and a likely mix of factors | Established [T1] |
| Does it run in families? | Yes. A close relative with AIS raises risk, though the size of the effect is uncertain | Established [T1]; degree of heritability [T4] |
| Are specific genes involved? | Yes, many, each with a small effect. LBX1 and GPR126 are among the replicated signals | Established [T1]; what they do in the curve [T2] |
| Does growth matter? | Onset clusters at the growth spurt, and growth loading can make an existing curve progress | Established [T1] for loading and growth; its full role in AIS [T2] |
| Is it the nervous system? | Sensory and balance differences are reported, but cause and effect are unsettled | Differences [T2]; as a cause [T4] |
| Is it hormones? | Melatonin and other hormones are studied; no direct cause-and-effect link is established | Mixed [T4] |
| Backpacks, posture habits, sleeping position? | No, according to the SRS | Clinical society guidance [T2] |
| Is the Neural Generation Hypothesis proven? | No. It is our hypothesis and a preprint | Synthesis [T3] |
How to read the evidence tiers: T1 independently replicated research, stated directly. T2 frontier research or clinical synthesis, sound but not yet unified, so one degree of hedging. T3 theoretical synthesis or our own contribution, always framed as synthesis. T4 unknown or mixed evidence. T5 an anti-pattern we recommend against. A compound claim gets a tier per part.
What is established
What the word idiopathic means
The SRS defines scoliosis as a side-to-side curve of the spine that measures more than ten degrees on an X-ray, and says idiopathic means the cause is not known. For most people with a curve, no specific cause is found. Conditions that are known to cause scoliosis include abnormally formed vertebrae present at birth, neurological disorders, muscular diseases and some genetic syndromes (SRS) [T2]. An international expert review describes adolescent idiopathic scoliosis as occurring in otherwise healthy people and says it currently has no recognizable cause, so that treating a direct underlying cause is not yet possible (Cheng 2015) [T1].
It helps to be precise about what that means. It does not mean the condition is mysterious in every respect. A lot is known about when it starts, who is more likely to get it, and how curves behave. What is missing is a single proven reason that explains why a particular child develops one.
It appears at the growth spurt, and severe curves are more common in girls
AIS appears during the adolescent growth spurt. Mild curvature is equally common in girls and boys, while severe and progressive curves are more common in girls, for reasons that are unknown (MedlinePlus Genetics) [T2]. A major review gives the same picture: it affects adolescents in the early stages of puberty and is more common in young women than young men (Cheng 2015) [T1].
It runs in families, and many genes are involved
Twin and family studies point to a hereditary component (Grauers 2012; Cheng 2022) [T1]. The size of that component is uncertain, because estimates differ widely between studies, and the known genetic variants explain only a minor part of it (Cheng 2022) [T4]. MedlinePlus Genetics puts it plainly: researchers suspect many genes, and a close relative with AIS increases a child’s risk (MedlinePlus Genetics) [T2].
Genome-wide studies have found specific signals. A locus near LBX1 was identified in a Japanese study (Takahashi 2011) and has since been found across several ethnic groups (Londono 2014) [T1]. A locus in GPR126 was identified in Japanese people and replicated in Han Chinese and European-ancestry groups (Kou 2013) [T1]. Each variant has a small effect. A variant is a liability, not a destiny: carrying one does not mean a curve will appear.
Some things do not cause it
Many families blame themselves. The SRS states that scoliosis does not come from carrying a heavy book bag or other heavy things, athletic involvement, poor sleeping or standing postures, lack of calcium, or a minor leg length difference (SRS) [T2]. If you are a parent reading this, nothing you did caused your child’s curve.
Once a curve exists, growth loading can make it progress
Bone grows more slowly under more pressure and faster under less. This is the Hueter-Volkmann principle, and animal experiments confirm that loading changes vertebral growth (Stokes 1996) [T1]. A modelling study predicted that a substantial part of progression during growth is mechanically mediated, and suggested that differences in how a person loads the spine may help explain why some curves progress and others do not (Stokes 2007) [T2]. This is about progression, which is a different question from why a curve starts. A clinical trial found that in growing adolescents at high risk of progression, bracing lowered the rate of progression to the point where surgery is considered (Weinstein 2013) [T1].
What is debated
This is where the research is active, and where honest summaries say “unsettled”.
- The nervous system and the senses. A systematic review and meta-analysis found proprioceptive differences in adolescents with AIS compared with controls, and called for work on the causal relationship (Lau 2022) [T2]. Whether those differences come before the curve, come from it, or both, is not established [T4].
- The balance organs. One systematic review could not confirm a link between an isolated vestibular disorder and scoliosis, and suggested a broader problem in how the brain combines vestibular, visual and body signals (Catanzariti 2014) [T4]. A later review of small studies reported asymmetries in the shape of the vestibular apparatus that seem to correlate with where the curve sits (Cortés-Pérez 2022) [T2]. The evidence is mixed.
- What the genes actually do. LBX1 has a documented role in specifying sensory relay neurons in the developing spinal cord (Gross 2002) [T1]. Whether LBX1 is the causal gene at that locus is not confirmed, so a sensing route is plausible, not proven [T2]. MedlinePlus Genetics notes that although many genes have been studied, few clear and consistent associations have been identified.
- Hormones. Alterations in melatonin are reported in people with AIS, but a review found the many variables do not allow a direct cause-and-effect relationship to be established (Gargano 2022) [T4].
- A mix of factors. A clinical review describes AIS as a multifactorial condition with a complex interplay of genetic, internal and environmental factors, and says the exact aetiology remains unknown (Marya 2022) [T2].
The honest summary is that several lines of evidence point to biology that is polygenic, tied to growth, and possibly involving how the nervous system senses the body. No one line has been shown to be the cause.
What is hypothesis: where the Neural Generation Hypothesis fits
Hypothesis, not established fact. Everything in this section is our hypothesis. It is published as an open-access preprint on Zenodo, which means it has not been through journal peer review. It makes no claim that any intervention reverses an established curve, and it is not a reason to skip monitoring, bracing or surgery.
The Neural Generation Hypothesis (NGH, Miller 2026) is our attempt to join the findings above into one account. Medicine defines scoliosis by shape: a curve over ten degrees. The hypothesis asks a different question, and proposes that scoliosis is “a curve your body keeps generating from its own map of itself, not just a bent bone.” The idea in short:
- Predisposition is polygenic. Small-effect variants add up to a liability. One of the strongest replicated signals, LBX1, sits at a gene that builds spinal relay cells for touch and position rather than bone. That makes a sensing route plausible, not proven, because the causal gene at that locus is not confirmed [T2]. PAX1 and GPR126 shape the spine’s own cartilage, disc and growth plate.
- Onset clusters at the growth spurt. The hypothesis proposes that when body sense is blurry, the small drift of growth is too small to register, and the brain’s map updates toward the curve [T3].
- Several routes could converge. A degraded sensory prediction channel, an unregulated expression of the left-right asymmetry that nearly all humans have, or, in secondary scoliosis, a primary structural load [T3].
- One neural lock. The brain’s forward model may absorb the curved geometry as its own normal and stop flagging it as an error. The preprint calls this predictive entrenchment [T3].
- Progression is a separate problem. Once a curve exists, its growth is driven by asymmetric loading on growing bone, as described above. The neural lock is not proposed to govern that [T1].
- Loops keep it going. Proprioceptive, autonomic and psychosocial loops could turn an ordinary mismatch into entrenchment [T3].
Most of the individual findings the model draws on are established or well studied (our reading: [T1] to [T2]). The unifying model that joins them is our synthesis [T3]. A hypothesis earns its place by being testable and open to being wrong. Studies that follow children’s sensory function over time, before and after a curve appears, are the kind of evidence that would bear on it. Until such evidence exists, treat the NGH as one proposed explanation among others.
The hypothesis also suggests a reading of the word idiopathic: it may describe a measurement gap. Current tools measure the structural output, the shape on the X-ray, while the process that generates it may run through the nervous system. That is a proposal, not a finding.
Read the full argument in the Neural Generation Hypothesis article, or the preprints themselves: The Neural Generation Hypothesis of Adolescent Idiopathic Scoliosis: Converging Evidence from Genetics, Proprioception, and Predictive Coding (doi:10.5281/zenodo.19342098) and the broader Generative Posture Framework: A Unified Account of Postural Generation, Persistence, and Change (doi:10.5281/zenodo.19580170). The wider idea, that posture is generated as a prediction, is explained in Posture Is a Prediction [T3].
What this does and does not mean for people with scoliosis
Knowing that the cause is unsettled does not change what to do next. If you or your child has a curve, follow your clinical team. Monitoring, bracing and surgery are decisions for a specialist who has examined the person and seen the X-rays. According to the SRS, options include routine monitoring while a curve remains small, a brace to try to stop a curve from progressing, and surgery for major curves that cause problems. Nothing in this article is a reason to delay, change or decline any of them or the check-ups that track a curve.
- If your child’s curve is being monitored, keep the appointments. Progression is a growth question, and your specialist is tracking it.
- If a brace has been prescribed, the best trial evidence supports it for growing adolescents at high risk (Weinstein 2013) [T1]. Follow the plan your team set.
- If you want to find a specialist, use the Navigator to find a specialist on this site.
- If you came here because you feel to blame, the SRS is clear that posture habits, backpacks and sleeping positions do not cause scoliosis.
Where Syntropic Core fits: it is somatic education, not treatment, and it makes no claim to change a curve. It teaches attention, breathing and pressure so a person can learn to notice, read and organize their own body, and it is built to sit alongside medical care, never instead of it. It is a place to start learning about your own body, not a place to take decisions that belong to your clinical team.
Where the research may go
A clearer answer will probably come from several directions at once: larger genetic studies, long-term studies that follow children before curves appear, and better ways to tell which curves will progress. Until then, an honest answer to “what causes AIS?” is the one at the top of this page. If you want to see how we weigh claims, the evidence tiers are explained above, and the sources below are the ones this page relies on.
Frequently asked questions
What causes adolescent idiopathic scoliosis?
No single cause has been proven. Medical reviews describe it as multifactorial: many genes that each add a little, together with growth and other factors that are still being studied. The word idiopathic means of undetermined cause.
Is adolescent idiopathic scoliosis genetic?
It runs in families, and a close relative with it raises a child’s risk. Genome-wide studies have found many variants, including near LBX1 and GPR126, each with a small effect. Known variants explain only a minor part of the family pattern, and having the genes does not mean a curve is certain.
Can a heavy backpack, poor posture or sleeping position cause scoliosis?
According to the Scoliosis Research Society, no. It states that scoliosis does not come from carrying a heavy book bag, athletic involvement, poor sleeping or standing postures, lack of calcium, or a minor leg length difference. Nothing a child did or did not do caused the curve.
Why does it usually appear during the growth spurt?
It is well established that onset clusters in the adolescent growth spurt, and that severe and progressive curves are more common in girls. Why is not established. The Neural Generation Hypothesis proposes that, during growth, blurry body sense lets a small drift go unregistered while the brain’s map updates toward the curve. That is a hypothesis, and it is untested.
Is the Neural Generation Hypothesis proven?
No. It is our hypothesis, published as an open-access preprint on Zenodo, which means it has not been through journal peer review. Most of the individual findings it draws on are established or well studied, and the way it joins them into one model is a synthesis that still has to be tested by others.
What should I do if my child has a curve?
Follow your clinical team. Monitoring, bracing and surgery are decisions for a specialist who has seen your child, and nothing on this page is a reason to delay or decline any of them. If you are looking for a specialist, the Navigator on this site can help you find one.
This article is education and research, not medical advice. Sam Miller is not a doctor, physical therapist or licensed clinician, and Syntropic Core is somatic education, not treatment. If you have a diagnosed spinal condition, have had spinal surgery, are pregnant, or are the parent of a teen whose curve is being monitored, talk to your clinical team first. Nothing here is a reason to delay or decline a brace, surgery or the check-ups that track a curve.
Sources
- Scoliosis Research Society. Scoliosis (patient information). srs.org/Patients/Conditions/Scoliosis. Accessed 24 October 2026. [T2]
Clinical society guidance: the definition of scoliosis, the meaning of idiopathic, what does not cause it, and the usual care options. - MedlinePlus Genetics, National Library of Medicine. Adolescent idiopathic scoliosis. medlineplus.gov/genetics/condition/adolescent-idiopathic-scoliosis. Accessed 24 October 2026. [T2]
US government patient resource: appears during the growth spurt, probably genetic and environmental factors together, many genes involved, a close relative raises risk. - Cheng JC, Castelein RM, Chu WC, et al. (2015). Adolescent idiopathic scoliosis. Nature Reviews Disease Primers, 1, 15030. doi:10.1038/nrdp.2015.30 [T1]
International expert review: the condition occurs in otherwise healthy people and currently has no recognizable cause; the definite aetiology is still unclear. - Weinstein SL, Dolan LA, Cheng JC, Danielsson A, Morcuende JA. (2008). Adolescent idiopathic scoliosis. The Lancet, 371(9623), 1527-1537. doi:10.1016/S0140-6736(08)60658-3 [T1]
Review stating that the aetiopathogenesis remains unknown and that bracing and physiotherapy had not been rigorously assessed at the time. - Marya S, Tambe AD, Millner PA, Tsirikos AI. (2022). Adolescent idiopathic scoliosis: a review of aetiological theories of a multifactorial disease. Bone & Joint Journal, 104-B(8), 915-921. doi:10.1302/0301-620X.104B8.BJJ-2021-1638.R1 [T2]
Clinical review: exact aetiology unknown, likely a complex interplay of genetic, internal and environmental factors. - Grauers A, Rahman I, Gerdhem P. (2012). Heritability of scoliosis. European Spine Journal, 21(6), 1069-1074. doi:10.1007/s00586-011-2074-1 [T1]
Large twin registry analysis: a genetic contribution, with unique environmental effects also important. - Cheng T, Einarsdottir E, Kere J, Gerdhem P. (2022). Idiopathic scoliosis: a systematic review and meta-analysis of heritability. EFORT Open Reviews, 7(6), 414-421. doi:10.1530/EOR-22-0026 [T1]
Twin and family studies indicate a hereditary component, but heterogeneity is large, the degree is uncertain, and known variants explain only a minor part of it. - Takahashi Y, Kou I, Takahashi A, et al. (2011). A genome-wide association study identifies common variants near LBX1 associated with adolescent idiopathic scoliosis. Nature Genetics, 43(12), 1237-1240. doi:10.1038/ng.974 [T1]
First genome-wide signal near LBX1. - Londono D, Kou I, Johnson TA, et al. (2014). A meta-analysis identifies adolescent idiopathic scoliosis association with LBX1 locus in multiple ethnic groups. Journal of Medical Genetics. doi:10.1136/jmedgenet-2013-102067 [T1]
The LBX1 association holds across several ethnic groups. - Kou I, Takahashi Y, Johnson TA, et al. (2013). Genetic variants in GPR126 are associated with adolescent idiopathic scoliosis. Nature Genetics, 45(6), 676-679. doi:10.1038/ng.2639 [T1]
A second locus, GPR126, replicated in Japanese, Han Chinese and European-ancestry groups. - Gross MK, et al. (2002). Lbx1 specifies somatosensory association interneurons in the dorsal spinal cord. Neuron, 34(4), 535-549. doi:10.1016/s0896-6273(02)00690-6 [T1]
Developmental biology: Lbx1 has a role in specifying sensory relay neurons in the spinal cord. Whether that explains the scoliosis association is not established. - Stokes IA, Spence H, Aronsson DD, Kilmer N. (1996). Mechanical modulation of vertebral body growth: implications for scoliosis progression. Spine, 21(10), 1162-1167. doi:10.1097/00007632-199605150-00007 [T1]
Animal experiment: compression slowed vertebral growth and distraction sped it, consistent with the Hueter-Volkmann principle. - Stokes IA, et al. (2007). Analysis and simulation of progressive adolescent scoliosis by biomechanical growth modulation. European Spine Journal, 16(10). doi:10.1007/s00586-007-0442-7 [T2]
Modelling study: a substantial component of progression during growth is predicted to be biomechanically mediated. A model, not a trial. - Lau KKL, Law KKP, Kwan KYH, Cheung JPY, Cheung KMC, Wong AYL. (2022). Timely revisit of proprioceptive deficits in adolescent idiopathic scoliosis: a systematic review and meta-analysis. Global Spine Journal, 12(8), 1852-1861. doi:10.1177/21925682211066824 [T2]
Proprioceptive differences were found in AIS, with the causal relationship still to be worked out. - Catanzariti JF, Agnani O, Guyot MA, Wlodyka-Demaille S, Khenioui H, Donze C. (2014). Does adolescent idiopathic scoliosis relate to vestibular disorders? A systematic review. Annals of Physical and Rehabilitation Medicine, 57(6-7), 465-479. doi:10.1016/j.rehab.2014.04.003 [T4]
Discordant results; the review could not confirm a link between an isolated vestibular disorder and scoliosis, and proposed a broader multisensory account. - Cortés-Pérez I, Salamanca-Montilla L, Gámiz-Bermúdez F, et al. (2022). Vestibular morphological alterations in adolescent idiopathic scoliosis: a systematic review of observational studies. Children, 10(1), 35. doi:10.3390/children10010035 [T2]
Small observational studies point to asymmetries in the vestibular apparatus; the reviewers report these seem to correlate with where the curve sits. - Gargano G, Oliva F, Migliorini F, Maffulli N. (2022). Melatonin and adolescent idiopathic scoliosis: the present evidence. The Surgeon, 20(6), e315-e321. doi:10.1016/j.surge.2021.07.008 [T4]
Melatonin alterations are reported, but the many variables do not allow a direct cause-and-effect relationship to be established. - Weinstein SL, Dolan LA, Wright JG, Dobbs MB. (2013). Effects of bracing in adolescents with idiopathic scoliosis. New England Journal of Medicine, 369(16), 1512-1521. doi:10.1056/NEJMoa1307337 [T1]
In growing adolescents at high risk, bracing reduced progression to the surgical threshold compared with observation. It did not test exercise. - Miller S. (2026). The Neural Generation Hypothesis of Adolescent Idiopathic Scoliosis: Converging Evidence from Genetics, Proprioception, and Predictive Coding. Open-access preprint on Zenodo, not journal published. doi:10.5281/zenodo.19342098 [T3]
A hypothesis about how idiopathic scoliosis starts and keeps going. It makes no claim that any intervention changes an established curve, and it does not replace monitoring, bracing or surgery. - Miller S. (2026). The Generative Posture Framework: A Unified Account of Postural Generation, Persistence, and Change. Open-access preprint on Zenodo, not journal published. doi:10.5281/zenodo.19580170 [T3]
Proposes that posture is generated by the body schema as a prediction, and names conditions under which that prediction updates.