Looking inside the brain: the long search for what HBOT changes

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.

Looking inside the brain: the long search for what HBOT changes

In 2017, a neurosurgeon walked into a hyperbaric chamber to find out what was happening inside his own head.

Joe Maroon was 81 years old. He had spent his career operating on brains, including those of professional athletes — he is the team neurosurgeon for the Pittsburgh Steelers. He had also spent decades pushing his own body well past most people's idea of reasonable, completing Ironman triathlons into his late seventies. He was, in other words, an unusually well-equipped subject for asking a question that most people who try hyperbaric oxygen never get to ask seriously: what does this actually change?

Maroon designed an experiment with himself as the only subject. Forty sessions in a hyperbaric chamber, breathing pure oxygen at 1.5 ATA. And before, during, and after, he measured himself with whatever tools he could get his hands on. Not just one. All of them.

He had cognitive testing across multiple domains — memory, attention, processing speed. He tracked athletic performance — VO₂ max, lactate threshold, time trials. He had blood drawn for proteomics, a panel that maps thousands of inflammatory and metabolic proteins circulating in the body. He had his telomeres measured — the protective caps on chromosomes that shorten as we age, and that biologists use as one marker of cellular age. And he sat through diffusion tensor imaging, or DTI — a form of MRI scan that maps the white matter of the brain, the bundled cabling that connects different regions to each other.

The results, when they came back, were the kind of thing scientists describe carefully and patients describe loudly. The DTI showed that several of the major cabling pathways in his brain had become more orderly and intact — the pattern you would expect to see in a younger brain. The change was not subtle. Some of these tracts measured between 9% and 22% better-organised after the protocol than before. His telomeres had roughly doubled in length, meaning that by this one biological marker, his cells looked considerably younger. Some of his thinking-skill scores improved. Some of the proteins in his blood shifted in directions associated with reduced inflammation. Others didn't move at all.

Reading the paper, what's striking isn't a single finding. It's the picture as a whole. Different windows showed different things. Some of the change was dramatic; some was modest; some was probably noise. To answer the question of what HBOT did to this particular 81-year-old neurosurgeon, no single tool was enough. He needed all of them, and even then, the answer he got was honest rather than clean.

Which is exactly the situation everyone in HBOT research faces. There is no single brain measurement that tells you what hyperbaric oxygen does. There is a long search, conducted across decades and across continents, with many different tools, each of which can see only part of the picture.

This is the story of that search.

Why measuring a brain is harder than it sounds

Here's the trouble with brain science: nobody has ever measured a thought.

This isn't a failure of equipment. It's a fundamental constraint. When researchers want to know what's happening in your head, they can't observe thinking, or memory, or attention directly. What they can observe are the things that correlate with those processes — blood flow shifting toward an active region, electrical rhythms organizing into specific patterns, glucose being consumed, water diffusing along the bundles of myelin in your white matter, oxygen being extracted from haemoglobin as cells work harder.

Every brain measurement, in other words, is a proxy. You are not seeing the thing itself. You are seeing the shadow it casts on something measurable.

This isn't unique to HBOT research. It's the situation across all of neuroscience. But it matters more than usual in HBOT contexts, for one specific reason: brain images are shown in HBOT communications more often than in almost any other wellness intervention. Walk through any HBOT clinic's website and you will likely see at least one before-and-after scan, often with bright colours suggesting dramatic change. These images are real. The change in them is real. But what they actually mean — what biological story they're telling — depends entirely on which proxy is being measured, how it is being interpreted, and what is being compared to what.

Researchers have built many windows over the decades. Some watch blood — where it is flowing, how fast, how oxygen-rich. Some watch the structure of brain tissue itself, or the integrity of the cabling between its regions. Some watch electrical rhythms. Some watch where the brain is actually burning fuel. Some watch the molecular signals cells release into the bloodstream. And some simply test what a person can think and do.

Each is a window. None is the room.

Asking what hyperbaric oxygen therapy does to the brain turns out to be a layered question. The engineering side — pressure, flow, dose — is the easier part. The biological side, what actually changes inside when more oxygen reaches the brain, requires every one of these windows, and more besides.

The skill of reading HBOT research, then, is the skill of knowing which window each tool opens. Knowing what it sees clearly, what it sees blurrily, and what it can't see at all. The story of how researchers came to see what they currently see — and what they still can't — runs through several decades and several continents.

It begins, as a lot of HBOT history does, with electricity.

Before there were scans, there was electricity

In the late 1950s, a Dutch surgeon named Ite Boerema was using hyperbaric oxygen for an entirely different reason than the brain. He was trying to make open-heart surgery survivable.

Boerema's insight was that if you could pressurize a patient with pure oxygen, you could buy minutes of survival when the heart was stopped — enough time for surgical work that would otherwise be impossible. He published his results in 1959. Within a few years, hyperbaric chambers were being used not just for surgery but for carbon monoxide poisoning, gas gangrene, decompression sickness, and a slowly expanding list of other indications.

What followed was decades of clinical observation that something was happening to brains in chambers — patients with stroke, with carbon monoxide injury, with traumatic brain injury, who seemed to recover function in ways that were difficult to explain. But observing recovery is not the same as measuring it. To know what was actually happening inside the skull, researchers needed a window.

The first window was electrical. EEG — electroencephalography — had existed since the 1920s, and it had one specific virtue that mattered for hyperbaric research: it could record from inside a chamber. The wires running through the chamber's bulkhead seal could carry signals out to recording equipment in the open air. Russian and Soviet researchers in particular built an extensive tradition of EEG-monitored hyperbaric work through the 1960s and 1970s, much of it conducted at lower pressures than would later become standard in the West. Some of this Russian-language literature is only now being mapped systematically by Western reviewers.

EEG was, and is, an extraordinary tool. It captures cortical electrical rhythms in real time, with millisecond precision. It can detect patterns that disappear when subjects fall asleep, or relax, or focus, or seize. But it has two persistent limits. The first is depth: EEG primarily reads from the cortex, the outer rind of the brain, and is largely blind to the deeper structures — the hippocampus where memories are formed, the thalamus that routes information, the brainstem that runs your most basic functions. The second is anatomical specificity: even within the cortex, EEG is fuzzy about exactly where the electrical activity is coming from. Two people's EEG signals during the same task can look quite different, even though their brains are doing approximately the same thing.

For decades, this was almost everything researchers had. EEG could tell you that something was different about the brain after a hyperbaric session. It couldn't tell you, in any spatially resolved way, what.

That changed in the 1980s, when a different kind of imaging started showing up in HBOT research — and with it, the possibility of seeing not just rhythms but blood.

How perfusion became a window

Single-photon emission computed tomography — SPECT — works by injecting a small amount of a radioactive tracer that flows along with blood. Wherever the brain is using more blood, more tracer accumulates. The scanner detects the gamma rays the tracer emits and reconstructs a three-dimensional map of perfusion — a snapshot of where blood is and isn't flowing in significant quantity.

Richard Neubauer, a Florida physician, was among the first to apply SPECT systematically to hyperbaric work, beginning in the 1980s. He noticed something that would shape the next thirty years of HBOT brain research: the SPECT scans of stroke patients didn't always match what their MRIs showed. The MRI would show a clearly dead area — a stroke region. But the SPECT around it would show a larger zone of low perfusion, a kind of penumbra of underactive tissue. And after hyperbaric sessions, that penumbra sometimes lit back up.

Paul Harch, working in New Orleans, took this observation and built much of his career around it. By the early 1990s, Harch was using SPECT to map what he called the "idling neuron" — brain tissue that wasn't dead but wasn't fully functioning, and that might be coaxed back into activity by pressurized oxygen. He published case series, then a Phase I trial in 2012 in the Journal of Neurotrauma on blast-injured veterans with post-concussion syndrome and PTSD, and a 2017 case-control study in Medical Gas Research on civilian traumatic brain injury. The SPECT patterns shifted; the patients reported functional improvement; the standard deviations of regional perfusion (Harch's "Neubauer pattern") narrowed.

Harch's other window, used more selectively, has been PET — positron emission tomography. PET works on a similar principle to SPECT, with an injected tracer and a scanner that picks up its emissions, but with a different question. Where SPECT shows where blood is flowing, PET shows where the brain is actually burning glucose for fuel — which is what neurons do when they are working. In a 2018 case report, Harch and Fogarty used FDG-PET on a 58-year-old woman with Alzheimer's dementia, before and after 40 hyperbaric sessions at very low pressure (1.15 ATA). Her scans showed metabolic increases of between 6% and 38% across different brain regions — and those increases held when she was scanned again nearly two years later. PET findings like this complement SPECT in an important way. Blood flow can change for many reasons, including ones that don't necessarily reflect more neural work being done. Glucose burning is harder to fake. If a brain region is actively metabolising fuel, neurons there are doing something.

The most-discussed case from Harch's body of work isn't an adult veteran. It's a two-year-old girl named Eden Carlson, whose 2017 case report in Medical Gas Research showed something doctors found difficult to assimilate. Carlson had drowned and been resuscitated, but only after fifteen minutes underwater. The standard outcome of that timeline is severe, permanent brain damage. Her early MRIs showed exactly what they were supposed to show: cortical atrophy, white matter loss, the imprint of anoxic injury.

Then she received normobaric oxygen therapy, and then hyperbaric oxygen at 1.5 ATA, over a period of months. A follow-up MRI 162 days after the drowning showed near-complete reversal of the visible atrophy. Functionally, she had recovered to a level her doctors had not expected to see.

It's the kind of case that has made HBOT a topic of conversation far beyond research circles. It is also, importantly, a case that the original paper itself was honest about: Carlson received both normobaric and hyperbaric oxygen, and the relative contribution of each cannot be cleanly determined from a single case. The paper acknowledges this. The image is real; the recovery is real; what the active ingredient was, methodologically speaking, remains partially open.

This is the recurring tension of SPECT and structural imaging in HBOT research. The pictures are vivid. The questions about what the pictures mean are harder.

A lit-up region on a post-treatment SPECT could reflect neurons resuming activity. It could also reflect simple vasodilation — wider blood vessels delivering more tracer through tissue that is no busier than before. SPECT alone cannot distinguish those possibilities. Harch's specific interpretation of his pattern shifts has not been independently replicated outside his own laboratory, which doesn't make him wrong, but does mean the field has wanted additional windows to corroborate what SPECT was showing.

And in Israel, beginning in the early 2010s, a different research program started supplying those additional windows.

When one window isn't enough

The Sagol Center for Hyperbaric Medicine and Research at Shamir Medical Center, near Tel Aviv, became — over roughly the past decade — the most prolific source of brain imaging data on HBOT in the world. Under the direction of Shai Efrati, with Amir Hadanny and a wider team, the Sagol group built a research program organized around a specific question: if you take patients with brain conditions that conventional medicine considers chronic and stable, and run them through 40-60 hyperbaric sessions at 2.0 ATA with intermittent air breaks, what changes — measured by everything you can measure?

What they could measure was a great deal. The Sagol protocol, replicated across many of their published trials, layers several MRI types together. Structural MRI maps the physical shape and volume of the brain. Diffusion tensor imaging — DTI, the same modality from Maroon's self-experiment — maps the integrity of the white-matter cabling that links one brain region to another. Perfusion MRI tracks blood flow without requiring a radioactive tracer. Functional MRI (fMRI), in its resting-state version, looks at how different brain regions communicate with each other while the patient is just lying quietly. To this MRI stack, the Sagol protocol adds SPECT in some studies, plus a computerized cognitive battery called NeuroTrax that scores memory, attention, executive function, processing speed, and other domains. Patients are measured before, immediately after, and in some studies many months after a treatment protocol.

What's emerged from this program is a growing body of randomized, sham-controlled or pharmacologically-controlled trials. Boussi-Gross and colleagues published results in 2013 on chronic post-stroke patients, showing perfusion-MRI improvements in regions that matched cognitive recovery. Tal and colleagues published on chronic post-concussion syndrome in 2015 and 2017, with DTI showing increased fractional anisotropy — a measure of white-matter integrity — in patients whose perfusion-MRI also showed improved blood flow. Catalogna and colleagues published in 2022 in NeuroImage: Clinical on patients with persistent post-COVID symptoms, finding changes in resting-state functional connectivity that correlated with reductions in psychiatric symptoms. Zilberman-Itskovich and colleagues, also in 2022, published a randomized sham-controlled trial in Scientific Reports showing similar effects in a separate post-COVID cohort. Ablin and colleagues published on fibromyalgia patients with traumatic histories in 2023.

Three things are worth saying honestly about this body of work.

The first is that, methodologically, these are among the most rigorous HBOT brain studies ever conducted — randomized, controlled, with multimodal imaging and quantitative cognitive endpoints. They represent a genuine step forward from the case series and uncontrolled cohorts that dominated the field's earlier decades.

The second is that several of the principal investigators have direct commercial relationships with Aviv Clinics, the chain of consumer-facing hyperbaric centers that uses essentially the Sagol protocol. Hadanny serves as Chief Medical Officer of Aviv. Efrati has been involved in its scientific direction. This is openly disclosed in the published literature, and it doesn't make the science wrong. But it does mean independent replication outside the Sagol/Aviv ecosystem matters more than usual, and the field has been slow to produce that replication.

The third is that, even within the Sagol body of work, the modalities don't always agree. Cognitive scores improve more reliably than imaging changes; some imaging modalities (DTI, perfusion-MRI) tend to show clearer effects than others (resting-state fMRI, qEEG, MR spectroscopy); and the same trial can produce one finding that survives correction for multiple comparisons and several others that don't.

This isn't a flaw of the research program. It is what brain science looks like when it is being honest. Different windows show different things. Sometimes they agree. When they do, that's evidence that something real is happening — because three independent measurements of three independent proxies all moved in the same direction. When they don't agree, that is information too. It tells you that whatever is happening, it isn't simple.

Which raises a different question. What if, instead of measuring before and measuring after, you could watch the brain in real time while the chamber was actually running?

What if you could watch in real time?

Most of what's been described so far requires the patient to be measured outside the hyperbaric chamber. SPECT, PET, MRI, fMRI, DTI — none of these will operate inside a pressurized environment with metal walls and (in the case of MRI) a high-strength magnetic field. The patient gets measured before. Goes into the hyperbaric chamber for an hour. Comes out. Gets measured again. Whatever happened during the session itself stays a black box.

A few tools, though, can come along for the ride.

Transcranial Doppler ultrasound — TCD — has been the workhorse of in-chamber neuro-monitoring for decades. A small ultrasound probe held against the temple can measure blood flow velocity through the middle cerebral artery in real time. Omae and colleagues, in a 1998 Stroke paper, used TCD inside the chamber to document something interesting and counterintuitive: under 2 ATA of pure oxygen, MCA blood flow velocity actually decreased in healthy volunteers. Vasoconstriction from hyperoxia, partially offsetting the increase in dissolved oxygen carrying capacity. It is a finding that complicates the simple "pressure plus oxygen equals more blood flow" intuition.

EEG also goes into the chamber, with the right wiring. Paul Harch, in his current clinical practice, has integrated continuous in-chamber qEEG — the quantitative analysis version of EEG, which pulls statistical patterns out of the raw electrical signal — into the first several sessions of every patient's protocol. He uses the readout to individualize pressure dosing, the idea being that brain activity itself can guide what pressure works best for a given person. He describes this approach in his 2020 Townsend Letter article and uses it in routine practice; it has not been formally trialed in a controlled study at the time of writing.

A growing body of Chinese research uses EEG microstate analysis during hyperbaric sessions for patients with chronic disorders of consciousness — the population that emerges from severe brain injury into states of reduced awareness. Xu and colleagues published in 2025 in CNS Neuroscience & Therapeutics on 32 such patients, finding that microstate D — a specific configuration of cortical electrical organization — increased its dwell time during sessions in patients who showed clinical recovery. Ye and colleagues had earlier reported similar findings in 2023.

These are the existing in-chamber tools. They have been around for decades. They are useful. They are also limited — TCD measures one artery, EEG measures cortical rhythms, neither sees what is happening at the cellular or molecular level.

The next tool on this list is more recent, and is the one most likely to change what in-chamber measurement can do over the coming decade.

That tool is functional near-infrared spectroscopy — fNIRS. It uses near-infrared light, shone through the scalp, to measure changes in the relative concentrations of oxygenated and deoxygenated haemoglobin in the cortex. Where blood flow goes, oxygenation changes; where neurons are working, blood flow goes. fNIRS is not as spatially precise as fMRI, and it cannot see deep into the brain. But it has two virtues that no other functional brain measurement has at the same time: it is portable enough to fit inside a hyperbaric chamber, and consumer-grade versions of it are starting to appear at prices three orders of magnitude lower than research MRI.

A handful of researchers have begun using fNIRS in HBOT contexts. Vatcheva-Dobrevska and colleagues, in 2010, demonstrated that NIRS measurements at peripheral sites tracked oxygenation changes inside the chamber as expected. Yang and colleagues published in Sensors in 2025 a study using fNIRS pre- and post-HBOT in stroke patients, observing patterns of cortical activation that tracked motor recovery. The body of HBOT-specific fNIRS literature remains small, but it's growing — and this particular tool merits a deeper exploration than it can get inside this article.

That deeper exploration is the next article in this series.

The brain talks through the bloodstream

Imaging, even at its most sophisticated, sees only certain dimensions of what's happening in a brain. It sees blood flow, structure, electrical patterns, oxygen extraction. What it largely doesn't see is chemistry — the molecular and cellular signaling that ultimately determines whether cells thrive or atrophy, whether networks remodel themselves or stay stuck in old patterns.

For that, researchers turn to blood.

Several biomarkers have shown up consistently in HBOT brain research. BDNF — brain-derived neurotrophic factor — is a protein that supports the survival, growth, and differentiation of neurons; many HBOT studies have measured it before and after protocols, with mixed but generally positive findings on its concentrations. GFAP and S100β are markers of astrocyte activity, often elevated when there is brain injury, and tracked in studies looking at HBOT for traumatic brain injury and similar conditions. Neurofilament light chain (NfL) is a marker of damage to neuronal axons — the long fibres that carry signals from one brain cell to the next — and has gained traction recently as a sensitive index of ongoing neurological injury. Inflammatory markers — IL-6, TNF-α, CRP — are tracked in studies looking at HBOT's effects on systemic inflammation. Oxidative stress markers like SOD, catalase, and glutathione give a window into the redox state of tissues.

The most ambitious approach in this category has come from Hadanny and colleagues at Sagol, who in 2021 published in Aging a study using RNA-sequencing on blood samples from elderly subjects before and after a 60-session HBOT protocol. They found 1,912 differentially expressed genes — meaning that nearly 10% of the protein-coding genome was being read out at meaningfully different levels after the protocol than before. The pathways most strongly affected included angiogenesis, oxidative stress response, immune regulation, and DNA damage repair. This is an enormous result, with all the caveats that come with single-arm studies — but the effect size is hard to dismiss, and the implication is striking. Whatever the protocol was producing in this population, it was not a single localized effect. The pattern of altered gene expression was broad and distributed across many cellular systems.

A separate Sagol study, Hachmo and colleagues in 2020 in Aging, reported that 60 HBOT sessions were associated with increased telomere length — the protective caps on chromosomes that shorten with aging — in immune cells, alongside decreased markers of immunosenescence. Maroon's self-experiment, mentioned at the start of this article, reported a similar telomere effect in his own blood. Jason Sonners — a chiropractor and HBOT clinician currently completing a PhD comparing 1.3 versus 2.0 ATA effects on epigenetic markers and gene expression — is among several researchers extending this kind of work into questions of dose and pressure.

What this whole biomarker-and-omics approach adds to the picture is a different kind of question. Imaging asks: where is the brain different? Cognitive testing asks: what can the person now do? Biomarkers ask: what processes are running differently in the body, at what scale, and through which pathways?

When all three families of measurement point in compatible directions — imaging shows changed perfusion or microstructure, cognitive scores improve, biomarkers show shifts in pathways that would predict that kind of recovery — that is the strongest evidence the field currently has that something biological is genuinely happening. When they don't all point the same way, it usually means the picture is more complex than any single window can show.

Which brings us back, finally, to the most direct measurement of all.

The closest we get to "did it work?"

Imaging shows you something about what changed. Biomarkers show you something about the chemistry behind that change. But neither tells you whether the person whose brain you are imaging can now do anything they couldn't do before.

For that, you need to actually test them.

Cognitive testing has been a part of HBOT research for as long as HBOT research has existed, but the tools have evolved. Early studies used pen-and-paper neuropsychological batteries — Trail Making, Stroop, Wechsler subtests, Rey-Osterrieth — administered by trained psychologists in long sessions. These remain the gold standard for clinical assessment in many contexts, but they have a problem when used as outcome measures in trials: they are expensive to administer, they don't always have alternate forms for repeat testing, and practice effects can be substantial.

The Sagol Center program, like many modern HBOT trials, uses a computerized battery called NeuroTrax (also known as Mindstreams). NeuroTrax tests memory, executive function, attention, information processing speed, visuospatial perception, verbal function, and motor skills, with norms calibrated by age and education. It can be administered in roughly an hour, has alternate forms to reduce practice effects, and produces both domain-specific scores and a global cognitive score. It has been validated against traditional batteries for mild cognitive impairment, multiple sclerosis, Parkinson's disease, and other contexts.

Other batteries in use across HBOT research include Cambridge Brain Sciences (used in some longevity research), Cogstate, and ANAM — the Automated Neuropsychological Assessment Metrics, used in U.S. Department of Defense studies including Harch's earlier military work. None of these tools is perfect; all of them have practice effects of some magnitude, all of them have floor and ceiling effects in particular populations, and all of them measure constructs that are themselves abstractions of the much messier reality of how a person actually functions in their life.

The cognitive batteries are the closest the field gets to answering the question patients actually ask. They are not, however, identical to felt experience.

A finding from one of the Sagol Center fibromyalgia studies, published in 2018, is illustrative. Patients in the trial reported that their symptoms transiently worsened during the first 20 sessions of the protocol, even as SPECT imaging was already showing reorganization of brain activity in the regions implicated in chronic pain. Eventually symptoms improved; but the early dissociation — felt symptoms going one way, imaging going another — is exactly the kind of complication that imaging-only and symptom-only measurement traditions both miss.

Felt experience, biomarkers, imaging, and cognitive performance don't always tell the same story in the same time frame. A complete picture requires looking at all of them, accepting that they will sometimes disagree, and treating that disagreement as data rather than as an embarrassment.

What no machine can see yet

Even with all the tools described so far — SPECT, PET, MRI in its many forms, DTI, fMRI, MR spectroscopy, fNIRS, EEG, qEEG, MEG, transcranial Doppler, OCT, blood biomarkers, gene expression panels, telomeres, cognitive batteries — there are entire categories of what HBOT is hypothesized to do that current measurement essentially cannot see.

Mitochondrial biogenesis — the creation of new mitochondria inside cells — is one of the proposed mechanisms by which HBOT might support brain function, particularly in aged or injured tissue. There is no current way to image this in vivo in humans. It is tracked, when it is tracked at all, through indirect biomarkers and post-mortem tissue analysis in animal studies.

Stem cell mobilization is another. HBOT has been shown to increase circulating CD34+ progenitor cells in the bloodstream, but tracking what those cells go on to do — whether they engraft in the brain, whether they contribute to neurogenesis, whether they remodel vasculature — remains the territory of animal models with histology, not human imaging.

The glymphatic system, which clears metabolic waste from the brain primarily during sleep, has been imaged in animals and increasingly in humans, but has not yet been systematically studied in HBOT contexts.

The deepest structures of the brain — the brainstem, the hypothalamus, the deep cerebellar nuclei — are difficult to image with most current functional tools at the spatial resolution needed to detect HBOT-related change. fMRI can sometimes reach them; fNIRS cannot.

And one of the most fundamental questions the field has not resolved is the dissociation of pressure from oxygen. HBOT is, by definition, both pressure and oxygen at the same time. Animal studies and a small number of human studies have begun to tease apart which effects depend on pressure (mechanotransduction in cells) and which depend on oxygen (mitochondrial respiration, oxidative signaling). But there is no imaging tool that directly visualizes mechanical effects of pressure on cellular signaling pathways.

These are not failures of HBOT research specifically. They are the current edges of what brain measurement, broadly, can do. The field is honest about most of them in the published literature. They tend to be quieter in commercial contexts.

The democratisation question

A SPECT scan, in 2025, costs somewhere between a thousand and three thousand dollars, depending on the country and the facility. A multi-modal MRI protocol — structural, DTI, perfusion, and fMRI — typically costs between two and five thousand. PET runs from two to six thousand per scan. These are not small numbers, especially for any kind of repeat measurement across a 40-session HBOT protocol.

A clinical qEEG, by comparison, costs a few hundred. A specialized blood panel for biomarkers like NfL or BDNF, costing per analyte, adds up to perhaps a hundred to a few hundred dollars. A telomere length test runs in the high hundreds. A cognitive testing battery, administered remotely through services like NeuroTrax or Cogstate, can be done at home for under two hundred dollars.

And at the bottom of the cost gradient, in 2025, sit consumer devices that didn't really exist a decade ago. Mendi, a Swedish-made headband, retails for around three hundred dollars. Sens.ai and Muse S Athena, hybrid devices that combine surface fNIRS with EEG sensors, run from five hundred to twelve hundred.

It is worth being honest about what these consumer devices actually do, because the marketing language around them can suggest more than they deliver. They are not brain scanners in any meaningful sense. They look at one small patch of tissue — typically the area just behind the forehead, which is the prefrontal cortex — and track changes in blood oxygenation there over time. They cannot see the hippocampus where memories form. They cannot see the brainstem, the cerebellum, or the white-matter tracts that DTI follows. What they offer is a trend-line for one small region of the brain's surface, not a portrait of the whole brain. For some uses — meditation tracking, focus monitoring, biofeedback — that limited window can still be informative. For anything resembling clinical assessment, it isn't enough.

The validation literature for these specific consumer-grade devices is also thinner than for the underlying fNIRS technology that powers them. The technology is real. Whether the specific device a person buys delivers research-grade reliability is a separate question.

But they exist. They didn't ten years ago. And what they represent is a particular kind of trajectory: the steady, sometimes slow, sometimes surprisingly rapid migration of brain measurement out of the research laboratory and into the home.

This trajectory is not uniform across modalities. SPECT, PET, MRI, and MEG are essentially never going to be home-accessible — they require massive equipment, regulated radioactive materials or strong magnetic fields, and specialized facilities. But the optical and electrical methods — fNIRS and EEG — have a different trajectory. They use safe, low-power signals. They can be miniaturized. They can run on batteries. They don't require shielded rooms.

The question this raises, for anyone using HBOT in a wellness context, is practical. If you are going to do 40 sessions in a hyperbaric chamber, what — if anything — should you measure? At what cost? With what expected reliability? And what should you do with the data you collect?

There aren't clean answers to those questions yet. The reliability of consumer-grade brain measurement, specifically in the context of tracking individual response to repeated HBOT sessions, is largely uncharacterized. The cognitive batteries that are accessible at home have practice-effect issues that limit their utility for short-interval repeated testing. The blood biomarkers that are most informative tend to require clinical-grade analysis, not finger-prick home tests. And the consumer fNIRS devices that are within reach financially are almost entirely limited to prefrontal cortex — which is not the brain region most plausibly mediating HBOT's broader effects.

What we may say honestly, today, is this: brain measurement in HBOT is partly democratized at the level of cognitive testing, heart rate variability, and prefrontal optical monitoring. It is not, and will not soon be, fully democratized at the level of deep-structure imaging. The picture you can build at home is real, but partial. The picture you can build in a research-grade clinic is more complete, but expensive. The reasonable position, for someone navigating this space, is to know which is which — and to know what each one is and isn't telling them.

Reading the next brain scan you see

Joe Maroon's experiment ended with a paper in Frontiers in Neurology. The findings were measurable across several modalities. They were also, importantly, the findings of one person who designed his own experiment, knew the result he was hoping for, and was unblinded throughout. The DTI numbers were real. The telomere numbers were real. So were the limits.

What Maroon did, and what HBOT brain research has been doing collectively for forty years, is not search for a single magic measurement. It is the much harder, slower, more interesting work of opening windows. Each window shows a partial view. Some windows are old and well-understood — SPECT, EEG, structural MRI. Some are newer and rapidly evolving — DTI, resting-state fMRI, fNIRS. Some haven't been built yet. The in-vivo mitochondrial imaging that would resolve a major mechanistic question in HBOT does not yet exist.

The next time someone shows you a before-and-after brain image and tells you what it means, the questions worth asking, in roughly this order, are these. Which proxy is this image actually measuring — perfusion, structure, activity, or something else? What is being compared to what — same person before and after, person versus matched control, before-and-after with appropriate adjustment for practice effects and scanner drift? How many subjects produced this kind of finding, and across how many independent laboratories? Do other measurements on the same patient or population — cognitive, biomarker, functional — point in the same direction?

These are not skeptical questions. They are literacy questions. Every published HBOT brain study that has stood up to scrutiny has answers to them, even if the answers are nuanced. Every researcher operating with intellectual honesty can explain which of these questions their measurement protocol addresses and which ones it does not.

What this article has tried to do is hand you the basic toolkit for asking. The field is still searching. New windows are opening. Some old assumptions are being revised. The science is genuinely advancing, and it is also genuinely incomplete. Both of those things are true at once.

The next article in this series goes deeper on one of those new windows — fNIRS — and the question of whether brain measurement is actually about to become accessible in a way it has never been before.

References

• Boussi-Gross R, Golan H, Fishlev G, et al. Hyperbaric oxygen therapy can improve post concussion syndrome years after mild traumatic brain injury — randomized prospective trial. PLoS One 2013;8(11):e79995.
• Efrati S, Fishlev G, Bechor Y, et al. Hyperbaric oxygen induces late neuroplasticity in post stroke patients — randomized, prospective trial. PLoS One 2013;8(1):e53716.
• Efrati S, Golan H, Bechor Y, et al. Hyperbaric oxygen therapy can diminish fibromyalgia syndrome — prospective clinical trial. PLoS One 2015;10(5):e0127012.
• Tal S, Hadanny A, Sasson E, Suzin G, Efrati S. Hyperbaric oxygen therapy can induce angiogenesis and regeneration of nerve fibers in traumatic brain injury patients. Front Hum Neurosci 2017;11:508.
• Hadanny A, Daniel-Kotovsky M, Suzin G, et al. Hyperbaric oxygen therapy can induce neuroplasticity and significant clinical improvement in patients suffering from fibromyalgia with a history of childhood sexual abuse — randomized controlled trial. Front Psychol 2018;9:2495.
• Hachmo Y, Hadanny A, Abu-Hamed R, et al. Hyperbaric oxygen therapy increases telomere length and decreases immunosenescence in isolated blood cells. Aging 2020;12(22):22445–22456.
• Hadanny A, Forer R, Volodarsky D, et al. Hyperbaric oxygen therapy induces transcriptome changes in elderly: a prospective trial. Aging 2021;13:24511–24523.
• Zilberman-Itskovich S, Catalogna M, Sasson E, et al. Hyperbaric oxygen therapy improves neurocognitive functions and symptoms of post-COVID condition: randomized controlled trial. Sci Rep 2022;12:11252.
• Catalogna M, Sasson E, Hadanny A, et al. Effects of hyperbaric oxygen therapy on functional and structural connectivity in post-COVID-19 condition patients. NeuroImage: Clinical 2022;36:103218.
• Ablin JN, Lang E, Catalogna M, et al. Hyperbaric oxygen therapy compared to pharmacological intervention in fibromyalgia patients following traumatic brain injury: a randomized, controlled trial. PLoS One 2023;18(3):e0282406.
• Harch PG, Andrews SR, Fogarty EF, et al. A phase I study of low-pressure hyperbaric oxygen therapy for blast-induced post-concussion syndrome and post-traumatic stress disorder. J Neurotrauma 2012;29(1):168–185.
• Harch PG, Andrews SR, Fogarty EF, Lucarini J, Van Meter KW. Case control study: hyperbaric oxygen treatment of mild traumatic brain injury persistent post-concussion syndrome and post-traumatic stress disorder. Med Gas Res 2017;7(3):156–174.
• Harch PG, Fogarty EF. Subacute normobaric oxygen and hyperbaric oxygen therapy in drowning, reversal of brain volume loss: a case report. Med Gas Res 2017;7(2):144–149.
• Maroon JC. The effect of hyperbaric oxygen therapy on cognition, performance, proteomics, and telomere length — the difference between zero and one: A case report. Front Neurol 2022;13:949536.
• Omae T, Ibayashi S, Kusuda K, et al. Effects of high atmospheric pressure and oxygen on middle cerebral blood flow velocity in humans measured by transcranial Doppler. Stroke 1998;29(1):94–97.
• Vatcheva-Dobrevska R, et al. Tissue oxygenation measured with near-infrared spectroscopy during normobaric and hyperbaric oxygen breathing in healthy subjects. Eur J Appl Physiol 2010.
• Yang S, et al. Effects of hyperbaric oxygen therapy on cerebral activity in stroke patients based on fNIRS. Sensors 2025;26(6):1794.
• Xu L, Wang J, Wang C, et al. Evaluating brain activity in patients with chronic disorders of consciousness after traumatic brain injury using EEG microstate analysis during hyperbaric oxygen therapy. CNS Neurosci Ther 2025.
• Ye Y, et al. EEG microstate changes during hyperbaric oxygen therapy in patients with chronic disorders of consciousness. Front Neurosci 2023.

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 clinical medical settings using medical-grade interventions. The inclusion of research summaries does not imply that similar outcomes can be achieved using non-medical wellness devices.