Neuroinflammation
Neuroinflammation is a mechanism, not a diagnosis.
It is a real and heavily studied biological process. It occurs in traumatic brain injury, stroke, central nervous system infection, multiple sclerosis, autoimmune encephalitis and the neurodegenerative diseases. It is also the single most oversold idea in wellness medicine, sold on symptoms it cannot be diagnosed from and measured by panels that do not measure it. This page is the honest version. What the cells actually do. What the process actually is. Where it is genuinely documented, how it is genuinely measured, and what the evidence supports doing about it.
Begin reading
Who is involved
Neurons
Microglia
Astrocytes
Oligodendrocytes
Capillaries
Circulating immune cells
The sequence
Where it becomes a problem
The trigger is still present
Tissue is still being injured
Activation keeps repeating
Resolution itself is impaired
Four different problems with four different answers. Suppressing the response is not automatically the right one.
Two things are true at once, and most writing on this subject picks one. Neuroinflammation is genuine, consequential neuroscience with thousands of papers behind it. And almost nobody being told they have it has had it demonstrated.
Holding both is not fence sitting. It is the only position that lets you take the biology seriously and still protect someone from a two year detour.
This page covers what the brain’s own immune cells actually do. Why the M1 and M2 labels fell apart. Why failure to resolve matters more than activation. How the body signals the brain, and what sickness behaviour is. Where neuroinflammation is genuinely established, and how researchers actually measure it. What the evidence supports doing, and what is promising but not yet proven.
The central idea
The question is not whether inflammation is present. It is why it has not finished.
Activation is how tissue is defended. Failure to resolve is where the damage lives.
Free guide
Get Brain on Fire
The guide explains how to work out which neurological function has changed. What belongs in a proper history. Which foundational tests find the reversible causes, and where inflammation genuinely fits rather than where it is usually claimed to.
Brain on Fire
How immune activity in the brain changes how you think and feel.
The neighbourhood
Neurons are outnumbered by the cells that keep them alive.
Microglia are the resident immune cells of the central nervous system, and they arrive early in development from a different lineage than the rest of the body’s immune system. Their processes are in constant motion even in a completely healthy brain, surveying the tissue, contacting synapses, clearing debris. Surveillance is not a state of alarm. It is the job.
Astrocytes regulate the chemical environment at the synapse, buffer potassium, recycle neurotransmitters, help control local blood flow, and form part of the barrier with their end feet. Oligodendrocytes build myelin, which is what makes fast conduction possible and is metabolically expensive to maintain. Capillaries deliver everything.
Microglia also prune synapses, which is one of the more striking findings of the last two decades. The same machinery that tags cellular debris for removal tags weak synapses. That process is essential in development and appears to be reactivated inappropriately in some disease states, which is a far more interesting story than inflammation as a fire.
The support network
Ordinary jobs
Ordinary function
Microglial surveillance is ordinary function, not an alarm state. Their processes are in constant motion in a completely healthy brain.
A model that did not survive contact
Why M1 and M2 fell apart.
You will see microglia sorted into M1, the inflammatory bad ones, and M2, the healing good ones, often with a supplement attached to the second column. That framework came from cultured macrophages stimulated in a dish. It was useful shorthand and it does not describe what happens in living tissue, which matters because an entire commercial edifice was built on it.
The old model
M1, inflammatory and harmful
M2, healing and protective
What single cell work actually shows
Many overlapping states rather than two
Signatures specific to each disease
States that differ by brain region
The same cell changing over hours
Protective and harmful features coexisting in one cell
None of this can be inferred from how a person feels, or from any blood panel currently sold. The binary was a laboratory convenience, not a description of living tissue.
This is not a small correction. If microglial states are a shifting, region specific, disease specific landscape, then the idea of pushing someone from M1 to M2 with a supplement is not an oversimplification of the science. It is a description of something that does not exist.
The part nobody sells
Resolution is an active programme, not an ending.
For most of the twentieth century, inflammation was assumed to stop when the stimulus ran out. It does not. Resolution is switched on deliberately, with its own signalling molecules, its own cellular choreography and its own failure modes.
The mediators involved are derived from omega-3 and omega-6 fatty acids and are collectively called specialised pro-resolving mediators: resolvins, protectins, maresins and lipoxins. They actively stop the recruitment of new inflammatory cells, promote clearance of the ones already there, and switch tissue toward repair. This is well established biology, and it reframes the whole question.
Because if resolution is a programme, then chronic inflammation is not simply too much activation. It can be a programme that failed to run. Those are different problems and suppressing the first does nothing for the second.
Chronic inflammation is not always excess activation. Sometimes it is a resolution programme that never ran.
Worth being precise about what follows from this. That resolution biology exists and is fatty acid derived is established. That taking fish oil reliably resolves neuroinflammation in humans is not, and the trial evidence is mixed. The mechanism is real. The consumer conclusion drawn from it usually overshoots.
Four stages
And four failure modes
A complete response
Four ways it becomes chronic
The trigger is still present
Tissue is still being injured
Activation keeps repeating
Resolution itself is impaired
Only the first of these is straightforwardly addressed by suppressing the response. The other three call for finding the trigger, stopping the injury, or supporting resolution.
Four routes
How the body tells the brain it is inflamed.
The brain is not sealed off from systemic immune activity. There are several well described routes by which inflammation elsewhere in the body changes how the brain works. This is why serious infection, major surgery and severe illness reliably produce confusion, low mood and cognitive slowing in people whose brains were fine the week before.
Body to brain
What this produces
Confusion during serious infection
Cognitive change after major surgery
Withdrawal and low mood in severe illness
The size of the systemic signal matters. These routes operate at the scale of significant illness, not at the scale of a mild food reaction or a minor digestive symptom.
An evolved programme
Sickness behaviour is the brain doing its job.
When you have influenza you withdraw, you sleep, you lose appetite, you lose interest in things you normally enjoy, and thinking becomes effortful. That cluster is not a side effect of being unwell. It is a coordinated programme the brain runs deliberately in response to immune signals, and it conserves energy and redirects it toward fighting infection.
This is one of the cleanest demonstrations that peripheral inflammation changes brain function, and it can be induced experimentally in healthy volunteers. Give someone a low dose inflammatory stimulus and you can measure lowered mood, social withdrawal and slowed processing within hours.
It also explains something clinically important. The overlap between sickness behaviour and depression is not a coincidence, and it is the strongest reason to take the inflammation and mood question seriously. It is also why the honest version of that story is a subset of people rather than a universal mechanism.
Withdrawal, fatigue, anhedonia and slowed thinking are what an immune signal to the brain feels like from the inside.
The programme
Adaptive, then costly
What the brain switches on
Withdrawal
Fatigue
Loss of appetite
Loss of interest
Increased sleep
Slowed thinking
Heightened pain sensitivity
Over days
Adaptive. Energy is conserved and redirected toward fighting infection
Over months
Costly, and difficult to distinguish from depression from the outside
This programme can be induced experimentally in healthy volunteers, which is among the cleanest evidence that peripheral immune signals change brain function.
Selective, not sealed
The blood brain barrier, and what leaky brain gets wrong.
The barrier is endothelial cells joined by tight junctions, wrapped by pericytes and astrocyte end feet, with dedicated transporters carrying glucose, amino acids and other essentials across. Its job is not exclusion. It is control. Permeability genuinely does change in stroke, traumatic brain injury, multiple sclerosis, central nervous system infection, seizures and severe systemic inflammation. That is established neurology, measured with contrast imaging and cerebrospinal fluid studies.
The structure
Endothelial cells
Tight junctions
Pericytes
Astrocyte end feet
Where permeability genuinely changes
Stroke
Traumatic brain injury
Multiple sclerosis
Central nervous system infection
Seizures
Severe systemic inflammation
How it is actually measured
Contrast enhanced imaging
Cerebrospinal fluid studies
Not by a serum antibody panel and not from a symptom questionnaire. Antibodies to brain proteins are found in healthy people, and assay results vary substantially between laboratories.
The phrase leaky brain borrows the credibility of the real phenomenon and applies it to a claim that cannot currently be tested outside research settings. Someone can hold both facts: the barrier is dynamic and clinically important, and no commercially available blood test tells you the state of yours.
The established territory
Where neuroinflammation is genuinely documented.
These are the conditions in which neuroinflammation has been demonstrated in human tissue, human imaging or human cerebrospinal fluid, rather than inferred. Some are emergencies. Several are treatable, and one of them is missed often enough to be worth a section of its own further down.
Acute, and sometimes prolonged
Microglial activation follows injury and has been demonstrated on imaging years afterward in some people. This is the strongest human evidence for inflammation persisting long after the event, and it is the basis for the priming hypothesis discussed below.
Injury, then a second wave
The initial damage is loss of blood supply. What follows is an inflammatory cascade that contributes to the final size of the injury, which is why the biology of that second wave is an active target in stroke research.
Meningitis and encephalitis
Direct infection of the nervous system or its coverings. Medical emergencies, diagnosed with imaging and lumbar puncture, and the clearest case of inflammation in the brain there is.
Immune attack on myelin
An autoimmune demyelinating disease with inflammation as its defining mechanism, visible on MRI and treated with immune modifying therapy that genuinely changes its course.
Treatable, and frequently missed
Antibodies against neuronal surface proteins producing rapid onset psychiatric symptoms, memory loss, seizures and movement abnormalities, often in young people. It responds to immunotherapy and is misdiagnosed as primary psychiatric illness with real regularity.
Present, and the causal direction is debated
Microglial activation is consistently found in Alzheimer’s and Parkinson’s tissue. Whether it drives the disease, responds to it, or does both at different stages is genuinely unsettled, and anti-inflammatory trials in these conditions have largely disappointed.
Sepsis and critical illness
Systemic inflammation of sufficient magnitude produces measurable brain effects, including delirium during the illness and cognitive change persisting afterward in some survivors.
Under investigation, not yet settled
Persistent cognitive symptoms after infection are real and measurable. Whether ongoing neuroinflammation is the mechanism is an open research question rather than an established fact, and it is being studied properly.
Notice what is absent. Fatigue with a normal examination, brain fog, food reactions, chemical sensitivity and non-specific mood symptoms are not on this list, because in those situations neuroinflammation has not been demonstrated. That does not make the symptoms unreal. It makes the explanation unproven.
The measurement problem
How researchers actually measure it.
This section matters more than any other on the page, because the gap between how neuroinflammation is measured in research and how it is sold in clinics is where most of the harm happens.
In research it is measured with TSPO PET imaging, which uses a tracer binding a protein upregulated in activated glia, or with markers in cerebrospinal fluid obtained by lumbar puncture. Both are specialised, neither is screening, and TSPO imaging has known interpretive limitations including a genetic polymorphism that changes tracer binding between individuals.
There is real progress in blood based markers. Neurofilament light chain reflects axonal injury and glial fibrillary acidic protein reflects astrocyte activation, and both are moving toward clinical use in specific settings. Neither tells you that you personally have neuroinflammation causing your symptoms, and neither is what consumer panels are selling.
What is sold as a neuroinflammation panel measures antibodies of uncertain meaning, not inflammation in your brain.
Research versus retail
The actual gap
Measured in research
TSPO PET imaging, with known interpretive limits
Cerebrospinal fluid markers, by lumbar puncture
Neurofilament light chain, reflecting axonal injury
Glial fibrillary acidic protein, reflecting astrocyte activation
Sold as a brain panel
Antibodies to brain proteins
Barrier permeability markers
Both found in healthy people
Results varying substantially between laboratories
Blood based markers are genuinely progressing toward clinical use in specific settings. None of them establishes that a given person’s symptoms are caused by neuroinflammation.
A hypothesis worth knowing and holding loosely
Microglial priming.
There is reasonable evidence, much of it from animal work and some from human imaging, that after a significant neurological insult the immune cells of the brain remain more reactive than before. A later challenge may then produce a response out of proportion to its size. This is plausible, it is useful, and it is one of the most overextended ideas in this field.
Disproportionate symptoms do not prove priming. Prior neurological injury simply changes the questions we ask.
There is no clinical test for priming, no threshold that confirms it and no treatment validated against it. What it earns is a place in the history. If someone had a concussion at nineteen and their recovery from illness has been different ever since, that is worth knowing, and it is not a diagnosis.
What to actually do
What the evidence supports.
This is the part most pages on this subject skip, because the honest answer is less marketable than a protocol. None of these is sold as a neuroinflammation treatment. All of them have human evidence for reducing inflammatory signalling, improving the systems that support neural function, or both, and several are more aggressive than what a routine visit will suggest.
The first question, not the last
Untreated sleep apnoea, active periodontal disease, an unrecognised autoimmune condition, persistent infection, ongoing exposure, uncontrolled metabolic disease. Chronic inflammation with a live trigger does not resolve because you took something. Find the driver.
The most consistent evidence there is
Regular aerobic activity lowers systemic inflammatory markers, improves cerebral blood flow, supports metabolic health and is associated with better cognitive outcomes across large populations. Nothing in a bottle has evidence approaching this.
Specifically, not just sleep generally
Obstructive sleep apnoea produces intermittent hypoxia, sympathetic activation and elevated inflammatory markers, and it is common and frequently undiagnosed. If someone is being worked up for neuroinflammation without ever having been screened for apnoea, the order is wrong.
Adipose tissue is an inflammatory organ
Visceral fat produces inflammatory signals continuously, and insulin resistance is associated with elevated markers. This is one of the more direct levers on systemic inflammatory tone available. Blood Sugar and Metabolic Health covers it properly.
Restriction raises markers measurably
Experimental sleep restriction increases inflammatory markers in healthy volunteers within days. Sleep is also when metabolic clearance in the brain is most active. This is foundational rather than optional.
Pattern, not a single food
Dietary patterns high in vegetables, legumes, fish, olive oil and whole grains are associated with lower inflammatory markers. The evidence is for the pattern. It is much weaker for any individual food credited with the effect.
A direct and reversible contributor
Alcohol raises inflammatory markers, disrupts sleep architecture, affects the gut barrier and at higher intake risks thiamine deficiency. Reducing it is one of the few interventions that acts on several of these mechanisms simultaneously.
Real mechanism, honest evidence
They are the substrate for the pro-resolving mediators described above, which is genuine biology. The human trial evidence for improving cognition or mood is mixed rather than compelling. Worth including in a dietary pattern, not worth presenting as a treatment.
What is conspicuously absent: broad anti-inflammatory supplement stacks, long term NSAIDs, which failed in Alzheimer’s prevention trials, and any protocol claiming to switch microglia from one state to another. The mechanism-based reasoning behind those is often elegant. The human outcome data is not there.
Being explicit about the frontier
Promising, and not yet proven.
These are areas where the mechanism is interesting, early data exists, and the human outcome evidence has not arrived. Saying so plainly is what separates following the science from selling it. Any of these could become standard. None of them should be sold as though it already is.
The pattern to notice across all five is the same: strong mechanistic and animal work, thin human outcome data. That is not a reason to dismiss any of them. It is a reason to keep them in the column marked promising rather than moving them into the column marked treatment, and to be suspicious of anyone who has already moved them.
Seek emergency assessment
Some brain inflammation is an emergency.
Meningitis, encephalitis and autoimmune encephalitis are treatable and time critical, and the last of those is regularly mistaken for primary psychiatric illness. The pattern that should trigger urgent assessment is rapid onset, meaning days to a few weeks rather than months. It matters most when psychiatric and neurological features appear together in someone with no such history.
Fever with severe headache and a stiff neck
Confusion with fever
New seizure
Rapid onset psychiatric symptoms in someone with no history
Hallucinations with memory loss developing over weeks
New abnormal movements or rigidity
Rapidly progressive memory loss
Personality change over days to weeks
Reduced consciousness
Speech disturbance with confusion
Severe headache unlike any before
Any of the above after a recent infection
Autoimmune encephalitis responds to immunotherapy and the response is better the earlier it starts. If a young person develops psychiatric symptoms alongside seizures, memory loss or movement abnormalities over a short period, that combination deserves neurological assessment rather than a psychiatric admission alone. If you are having thoughts of suicide or you are in immediate danger, call or text 988 in the United States, or call 911.
Free guide
Understand the signal before blaming the brain
Brain on Fire walks through which neurological function has changed, and what belongs in a proper history. Which foundational tests find the reversible causes. And how to tell a real inflammatory mechanism from a marketed one.
Own your biology
Take the biology seriously enough to be strict with it.
Neuroinflammation deserves better than both of the usual treatments. It is not a fashionable word for feeling unwell, and it is not a fringe idea to be waved away. It is a mechanism with real cells, real signalling, real measurement methods and real diseases attached.
Respecting it means insisting on the same standard you would want applied to anything else. What is the trigger. What is the evidence. What would change the decision, and what would tell us we were wrong.
Mechanism from diagnosis A process that occurs in disease is not itself a diagnosis someone can be given.
Activation from failed resolution Four different problems hide inside chronic inflammation, and only one calls for suppression.
Research measurement from retail panels Imaging and spinal fluid are how this is measured. Serum antibody panels are not.
Promising from proven Several of the most interesting ideas here are exactly that, and saying so is the point.
Find the trigger before treating the fire.
No pressure, and nothing to buy. Bring your history, including any head injuries, illnesses and surgeries. How you sleep, and whether apnoea has ever been assessed. Everything you take, and any results you already have. From there we can work out what is actually driving the inflammatory signalling rather than assuming it.
Common questions
Questions about neuroinflammation.
Short answers, including to the ones with uncomfortable answers.
What is neuroinflammation?
Can neuroinflammation be tested for?
Does brain fog mean I have neuroinflammation?
Are microglia M1 or M2, and can I switch them?
Does the blood brain barrier become leaky?
What actually reduces inflammation in the brain?
Is inflammation the cause of Alzheimer’s disease?
What is autoimmune encephalitis and why does it keep being mentioned?
