This article is part of the HBOT Education series, where we explore what research tells us about hyperbaric oxygen therapy and specific health topics.
The Evidence, Clearly Explained — Brought to you by Brain Spa Hyperbaric
Disclaimer: This article is intended for educational and informational purposes only. It synthesizes published medical research and does not evaluate Brain Spa Hyperbaric products. The hyperbaric chambers offered on this website are non-medical wellness devices and are not intended to diagnose, treat, cure, or prevent any disease. Do not make medical decisions based on this article — consult a qualified healthcare professional.
In 2013, a team at University College London put an oxygen probe into the spinal cord of a rat with an experimental form of multiple sclerosis and watched the reading fall. The tissue was starved of oxygen. Not because the animal was breathing badly, and not because its blood was short of oxygen. The blood was fine. The inflamed tissue itself had become a place oxygen could not reach in the quantities it needed.
Then they gave the rat oxygen to breathe. Within an hour, its lost neurological function began to come back.
That result sits underneath a question that has followed the hyperbaric chamber for forty years. The hyperbaric oxygen therapy trials of the 1980s tested pure oxygen at double atmospheric pressure in people with MS and mostly found nothing. The patient-run centres that kept using it insisted otherwise. We told that story in a previous article. But there is a deeper question the trials never asked, because in the 1980s nobody could have asked it: what does oxygen actually do to myelin?
The answer turns out to be strange. Myelin, the insulation around your nerve fibres, is not just kept alive by oxygen. It is built from it.
Your brain’s wiring is wrapped in fat
Every fast nerve fibre in your brain and spinal cord is wrapped in myelin, a sheath that works the way plastic insulation works on an electrical cable. It lets signals travel up to a hundred times faster and, just as importantly, it lets them travel cheaply. A myelinated fibre spends a fraction of the energy an unmyelinated one needs to carry the same message.
Myelin is unusual stuff. By dry weight it is roughly 70 to 80 percent fat, which makes it the fattiest tissue in the body. About a quarter of that fat is cholesterol. Another large share is a family of lipids called plasmalogens, which are more concentrated in myelin than anywhere else you have them. The rest is a mix that includes very-long-chain fatty acids, the sort your cells need to trim and manage carefully so that membranes stay flexible rather than stiff.
When MS strips myelin away, what is lost is fat. And when the body tries to rebuild it, which it does, often surprisingly well in the early years of the disease, what it has to make is fat. So the question of whether a damaged patch of brain can repair itself comes down, in large part, to whether the cells responsible can manufacture these particular lipids on site.
Here is the part most people never hear, and the part no hyperbaric oxygen therapy trial ever measured. Several of the enzymes that make them cannot work without oxygen.
Oxygen as an ingredient, not just a fuel
We usually think of oxygen as fuel. Your mitochondria burn it to make energy, and the brain burns more of it, per gram, than almost any other organ. That is true, and it matters. But there is a second role that gets far less attention: oxygen as a raw material that ends up inside the molecule being built.
Cholesterol is the clearest example. The chemical route from its starting material to the finished molecule needs eleven separate oxygen molecules, consumed at specific steps by enzymes that take an oxygen atom from the air you breathed and bolt it onto the growing structure. Without oxygen, the assembly line stops at the point where the first oxygen was needed. No cholesterol, no myelin.
Plasmalogens are the same, and more so. The step that turns an ordinary membrane lipid into a plasmalogen was a mystery for decades; the enzyme was only identified in 2019, in an unexpected piece of detective work that started with a light-sensing bacterium. It is an oxygen-consuming desaturase. It needs oxygen to make the one chemical bond that defines a plasmalogen.
The enzymes that keep membrane fats flexible use oxygen too. So do the ones inside peroxisomes, the small cellular compartments that break down the very-long-chain fatty acids which would otherwise clog the membrane. The first step of that breakdown consumes oxygen directly.
None of this is controversial. It is textbook biochemistry. What was not obvious until recently is why it should matter for a disease that was, for most of a century, understood as an immune attack.
The lesion that behaves as if it were suffocating
In 2003, Hans Lassmann’s laboratory in Vienna reported something that did not fit the immune story. A subset of MS lesions carried the same molecular fingerprint as brain tissue damaged by a stroke, the pattern you see when a region has been cut off from oxygen. In these lesions, the first myelin protein to disappear was the one that sits deepest inside the sheath, furthest from any blood supply, exactly as in oxygen-starved white matter.
The catch was that the blood flowing past these lesions was fully oxygenated. Whatever was happening, it was not the ordinary kind of oxygen shortage.
Six years later, Bruce Trapp and Peter Stys gave the phenomenon a name: virtual hypoxia. Their reasoning went like this. A nerve fibre that has lost its myelin needs several times more energy to conduct a signal, because it has lost the insulation that made conduction cheap. Its mitochondria, damaged by inflammation, produce less. Swelling and scarring push the fibre further from the nearest capillary, so oxygen has further to diffuse. The demand goes up, the supply goes down, and the cell ends up in the same metabolic position as a cell that has been deprived of oxygen, in a brain that has plenty of it.
It was an elegant argument. But it was an argument. The next step was to measure it.
A probe in the spinal cord, and a surprising recovery
That is what Kenneth Smith’s group at UCL did. In the rat model of MS known as EAE, Andrew Davies and colleagues measured spinal cord oxygen two independent ways: with a chemical probe that binds only to oxygen-starved tissue, and with a physical oxygen sensor inserted into the cord itself. Both said the same thing. The inflamed tissue was hypoxic, and the degree of hypoxia tracked the animals’ neurological deficits in time and in place. Where the cord was short of oxygen, function was lost. When the hypoxia eased, function returned.
Then came the intervention. The rats were given oxygen to breathe, at ordinary atmospheric pressure. Function improved significantly within an hour. When oxygen was continued for a week, the improvement held for at least that long.
Three years later, the same group built a different model: not the whole-body immune disease of EAE, but a single lesion, induced in the spinal cord by a small dose of a bacterial molecule that provokes local inflammation. What they saw looked remarkably like the Vienna lesions. A week after the injection, myelin began to disappear, but not at the injection site. It disappeared at a vascular watershed nearby, the zone between two blood supplies where tissue is inherently most vulnerable to a drop in oxygen. Before the myelin went, there was a brief window of hypoxia. Oligodendrocytes, the cells that make and maintain myelin, died first. Demyelination followed.
Then the experiment that matters. When the researchers raised the animals’ inspired oxygen to 80 percent during that hypoxic window, for two days, the demyelination was significantly reduced or, in some animals, did not happen at all.
Read that again, because it is easy to slide past. In an animal model of the early MS lesion, giving oxygen at the right moment prevented the loss of myelin.
In 2020 the group closed the loop. The hypoxia, they showed, came from a collapse in local blood flow through the inflamed cord. A drug that restores that blood flow, nimodipine, restored oxygenation, restored function, and reduced demyelination in both models. Hypoxia was not a side effect of the disease process. In these models, it was part of the mechanism.
These were animal studies. That limitation is real and we will come back to it. But there is one human measurement worth knowing. In 2015, Runze Yang and Jeff Dunn at the University of Calgary used near-infrared light to measure oxygen saturation in the cortex of 72 people with MS and 12 healthy controls. In the MS group, 42 percent of readings fell more than two standard deviations below the healthy mean, and lower oxygenation went with greater disability. The authors called the finding supportive but not conclusive, which is the right description. It is one method, in one cohort, and it cannot see inside a lesion. But it is the first time anyone measured, in living patients, what Lassmann had inferred from tissue and Smith had measured in rats.
What oxygen shortage does to the myelin repair crew
So a lesion in multiple sclerosis can be short of oxygen. Why should that stop myelin from coming back, rather than just slowing things down?
Two findings answer that, and both concern oligodendrocytes, the repair crew.
The first is about peroxisomes. When a cell senses that oxygen is scarce, it responds by dismantling its own peroxisomes; a 2014 study traced this to the same oxygen-sensing switch that controls the rest of the cell’s low-oxygen response. That is a reasonable economy for a cell trying to survive a temporary shortage. It is a disaster for a cell whose job is to build and maintain myelin, because peroxisomes are where the long fatty acids get trimmed and where much of plasmalogen synthesis begins. Kassmann and colleagues showed in 2007 what happens when oligodendrocytes lose their peroxisomes: mice engineered that way develop demyelination, axon loss and inflammation that look strikingly like MS, with no immune trigger at all. The immune cells arrive afterwards, drawn to the damage.
The second finding is about maturation. Remyelination, the rebuilding of the sheath, is done by young precursor cells that must first mature into full oligodendrocytes. In 2014, Yuen and colleagues found that the low-oxygen signal inside these precursors acts as a brake: while it is switched on, they stay immature. They can only finish the job once the signal switches off, which is to say, once there is enough oxygen where the lesion is. Remyelination, in other words, needs the hypoxia signal off.
Put the pieces together and the chain is short:
- A lesion runs short of oxygen.
- The cells that should repair it dismantle the machinery they need and stall in an immature state.
- The lipids that myelin is made of cannot be built, because building them needs oxygen as an ingredient.
- The fibre, still uninsulated, keeps burning energy it cannot afford, which deepens the shortage.
This is the argument for why oxygen might matter in MS. It is coherent, it is built from work in good journals, and every link in it has been demonstrated. What has not been demonstrated is the chain as a whole, in a person, in a hyperbaric chamber. We will come to that. First, the complication.
The same oxygen that builds can burn
Before going further, it is worth meeting the person who has argued this chain most forcefully, because he also arrives at its most awkward implication. The Canadian biochemist Dayan Goodenowe has spent two decades on plasmalogens, and his model of MS runs the causal arrow the opposite way from the textbook. In his account the disease begins with cells whose mitochondria cannot keep up with demand. The energy shortfall produces oxidative stress; oxidative stress damages the fats in cell membranes, producing oxidised phospholipids; and those damaged lipids are what draw the immune system in. The injury comes first, and the injury attracts the immune cells. It is an unconventional model, and its ordering is unproven. But it makes a prediction that the rest of this article has been circling: if MS is at root a disease of cellular energy and membrane fats, then oxygen sits at the centre of both, and it sits there as a double-edged thing.
Oxygen is reactive. That is why it is useful and why it is dangerous. The unsaturated fats that make myelin flexible are exactly the fats most easily damaged by oxidation, and myelin is also rich in iron, which catalyses that damage. Oxidised phospholipids are not a side issue in MS. Lassmann’s group found them concentrated at the edges of fresh demyelination in 2014; Dong and colleagues showed in 2021 that they are directly toxic to nerve cells and to oligodendrocytes; and van den Bosch and colleagues found in 2024 that the small clusters of immune cells scattered through apparently normal MS white matter are busy engulfing these very molecules.
Plasmalogens, interestingly, are part of the defence. The oxygen-built bond that defines them is also a sacrificial target: it reacts with oxidising species before the more valuable parts of the membrane do. So a tissue rich in plasmalogens is protected, and a tissue that has used them up is exposed. A cell short of oxygen cannot make plasmalogens. A cell flooded with oxygen may consume them faster than it can replace them.
Whether a given dose of oxygen builds or burns, then, depends on the state of the tissue’s defences. And those defences can be trained. In a well-known series of experiments, Günter Speit and colleagues found that a first session of hyperbaric oxygen therapy caused measurable oxidative damage in volunteers’ blood cells, and that the damage did not recur with later sessions, because the first one had switched on the cells’ own protective enzymes. Used at a pace the body can match, the exposure strengthens the very defences that decide whether oxygen builds or burns. Pushed faster than those defences can grow, it does the opposite.
Clinicians who work with the injured brain have seen the same shape from the other side. Paul Harch, who has used hyperbaric oxygen therapy in chronic neurological injury for over thirty years, has argued that the injured brain responds within a narrow dose window and that pushing past it produces worse outcomes, not better ones; we went through his reasoning in an earlier article. In a 2024 review of his own records, he and a colleague documented 52 cases of chronic central nervous system oxygen toxicity at doses below 2.0 ATA, most of them reversible once the dose came down. His observation is clinical rather than biochemical, but it is the same curve: a range where oxygen helps, an edge past which it hurts, and an edge that sits lower for brain tissue than the textbooks assumed.
That is the biological reason the dose question in hyperbaric oxygen therapy is not a footnote. It is also the most plausible reading of the one adverse signal the 1980s trials did detect: the Newcastle team found reversible oxygen damage in the retina, concentrated in the patients whose disease was most active, the patients with the least defensive capacity to spare.
Where the evidence for oxygen and myelin is strong, and where it stops
It helps to lay this out plainly, because the argument is seductive and seductive arguments deserve extra scrutiny. What the research shows about oxygen, hypoxia and myelin repair in multiple sclerosis:
Established. Myelin is mostly lipid, and several of the enzymes that make those lipids consume oxygen as a substrate. A subset of MS lesions carries the molecular signature of oxygen starvation. In two rat models, inflamed nervous tissue is measurably hypoxic, breathing oxygen restores function within an hour, and breathing oxygen during the hypoxic window prevents demyelination. Low oxygen makes oligodendrocytes dismantle their peroxisomes and stall in an immature state. Oxidised lipids are toxic to the cells that make myelin.
Suggestive. Cortical oxygenation is reduced in a substantial minority of people with multiple sclerosis, by one measurement method, in one cohort.
Not yet tested. Whether any of this happens in a person with MS who breathes oxygen under pressure. Nobody has measured plasmalogens, oxidised lipids or remyelination markers before and after a course of hyperbaric oxygen therapy in MS. Not one of the nine randomised trials from the 1980s measured anything of the kind; they measured disability scores, which is what trials measured then. The animal work used oxygen at ordinary pressure for hours or days, a very different exposure from a one-hour session in a hyperbaric chamber. And EAE, however useful, is an immune disease induced in a rodent, not multiple sclerosis.
There is one more gap worth naming. The animal studies gave oxygen during a specific window, the days when a lesion was forming. Nobody knows whether the same intervention does anything for a lesion that formed years ago, or for the slow, lesion-independent decline that characterises progressive MS. The biology above gives reasons to think oxygen availability might matter for repair at any stage. Reasons are not results.
This research was conducted in laboratory and clinical settings under scientific supervision. Results from animal and clinical research cannot be directly applied to other settings or devices — always consult a healthcare professional.
Why this changes how the old hyperbaric chamber trials read
None of this existed when Fischer, Barnes and the others designed their hyperbaric chamber trials. The plasmalogen enzyme was unknown. Virtual hypoxia had not been named. Nobody had put an oxygen probe into an inflamed spinal cord. The hyperbaric oxygen therapy trials asked a reasonable question for their time: give a lot of oxygen for four weeks and see whether the disability score moves. Mostly it did not.
The biology described here suggests a different question, and a harder one. Not “does oxygen help MS?” but: is the tissue short of oxygen, at what dose does supply outrun the tissue’s ability to handle it, and what happens to the lipid chemistry of repair when the balance is right? Those are measurable questions. They have measurable answers. They have simply never been measured in a hyperbaric chamber study of multiple sclerosis.
The hypothesis Philip James floated in 1982 was that MS lesions are, at root, a problem of oxygen delivery. He was almost certainly wrong about the mechanism he proposed. It is starting to look as though he may have been asking a better question than the trials that set out to answer him.
Educational disclaimer
This content synthesizes findings from published medical research for educational purposes only.
The hyperbaric chambers sold on this website are non-medical wellness devices and are not intended to diagnose, treat, cure, or prevent any disease.
The studies discussed here were conducted in laboratory and clinical settings using medical-grade interventions. The inclusion of research summaries does not imply that similar outcomes can be achieved using non-medical wellness devices.
Sources
Hypoxia in MS and its models
Aboul-Enein F, Rauschka H, Kornek B, et al. “Preferential loss of myelin-associated glycoprotein reflects hypoxia-like white matter damage in stroke and inflammatory brain diseases.” Journal of Neuropathology and Experimental Neurology, 2003. PubMed
Trapp BD, Stys PK. “Virtual hypoxia and chronic necrosis of demyelinated axons in multiple sclerosis.” The Lancet Neurology, 2009. PubMed
Davies AL, Desai RA, Bloomfield PS, et al. “Neurological deficits caused by tissue hypoxia in neuroinflammatory disease.” Annals of Neurology, 2013. DOI
Desai RA, Davies AL, Tachrount M, et al. “Cause and prevention of demyelination in a model multiple sclerosis lesion.” Annals of Neurology, 2016. PMC
Desai RA, Davies AL, Del Rossi N, et al. “Nimodipine reduces dysfunction and demyelination in models of multiple sclerosis.” Annals of Neurology, 2020. DOI
Yang R, Dunn JF. “Reduced cortical microvascular oxygenation in multiple sclerosis: a blinded, case-controlled study using a novel quantitative near-infrared spectroscopy method.” Scientific Reports, 2015. PMC
Myelin lipids and oxygen
Braverman NE, Moser AB. “Functions of plasmalogen lipids in health and disease.” Biochimica et Biophysica Acta, 2012. DOI
Gallego-García A, Monera-Girona AJ, Pajares-Martínez E, et al. “A bacterial light response reveals an orphan desaturase for human plasmalogen synthesis.” Science, 2019. DOI
Summons RE, Bradley AS, Jahnke LL, Waldbauer JR. “Steroids, triterpenoids and molecular oxygen.” Philosophical Transactions of the Royal Society B, 2006. DOI
Walter KM, Schönenberger MJ, Trötzmüller M, et al. “Hif-2α promotes degradation of mammalian peroxisomes by selective autophagy.” Cell Metabolism, 2014. DOI
Kassmann CM, Lappe-Siefke C, Baes M, et al. “Axonal loss and neuroinflammation caused by peroxisome-deficient oligodendrocytes.” Nature Genetics, 2007. PubMed
Yuen TJ, Silbereis JC, Griveau A, et al. “Oligodendrocyte-encoded HIF function couples postnatal myelination and white matter angiogenesis.” Cell, 2014. PMC
Oxidised lipids and oxygen tolerance
Haider L, Fischer MT, Frischer JM, et al. “Oxidative damage in multiple sclerosis lesions.” Acta Neuropathologica, 2014. DOI
Dong Y, D’Mello C, Pinsky W, et al. “Oxidized phosphatidylcholines found in multiple sclerosis lesions mediate neurodegeneration and are neutralized by microglia.” Nature Neuroscience, 2021. PubMed
van den Bosch AMR, Hümmert S, Steyer A, et al. “Profiling of microglia nodules in multiple sclerosis reveals propensity for lesion formation.” Nature Communications, 2024. DOI
Speit G, Dennog C, Eichhorn U, Rothfuss A, Kaina B. “Induction of heme oxygenase-1 and adaptive protection against the induction of DNA damage after hyperbaric oxygen treatment.” Carcinogenesis, 2000. PubMed
Harch PG, Rhodes S. “Acute and chronic central nervous system oxidative stress/toxicity during hyperbaric oxygen treatment of subacute and chronic neurological conditions.” Frontiers in Neurology, 2024. PMC
Barnes MP, Bates D, Cartlidge NEF, French JM, Shaw DA. “Hyperbaric oxygen and multiple sclerosis: final results of a placebo-controlled, double-blind trial.” Journal of Neurology, Neurosurgery and Psychiatry, 1987. PMC
