For millions of people, COVID-19 didn’t end when the fever broke. Brain fog, crushing fatigue, anxiety, depression, and a host of other neurological and psychiatric symptoms have lingered for months or even years after the initial infection, a constellation of problems now widely recognized as long COVID (LC). One leading theory is that the brain stays inflamed long after SARS-CoV-2 has cleared the body, a state of smoldering neuroinflammation that drives ongoing symptoms.
A new study published in the Journal of Neurology in April 2026 put this hypothesis to the test using one of the most powerful tools available for measuring brain inflammation: PET scanning. What they found was more nuanced and arguably more interesting than a simple “yes, the brain is inflamed.”
What Is TSPO PET, and Why Does It Matter for Long COVID?
To understand what the researchers measured, it helps to understand how inflammation in the brain actually works. Unlike the rest of the body, the brain has its own specialized immune cells, called microglia and astrocytes, that become activated in response to infection, injury, or stress. This glial activation is a central feature of neuroinflammation, and it’s measurable in living people using PET (positron emission tomography) imaging.
The specific target the researchers tracked is a protein called TSPO (translocator protein, 18 kDa), which sits on the outer membrane of mitochondria, the energy-producing structures inside cells, and is upregulated when glial cells become activated. By injecting participants with a radioactive tracer that binds to TSPO and then imaging its accumulation in the brain, researchers can create a detailed map of glial activation.
This technique has been used extensively in conditions like multiple sclerosis (MS), where neuroinflammation is well-documented and severe. Using MS patients as a positive control group a group in which elevated TSPO signals are expected allowed the Finnish research team at Turku University Hospital to validate their method and compare the long COVID brain with both healthy and diseased brains on the same scale.
Who Was Studied?
The study included 14 people with long COVID, 11 healthy controls (HC), and 13 people with multiple sclerosis. The long COVID participants had confirmed SARS-CoV-2 infection (via PCR or antigen test) and had been experiencing neurological symptoms, fatigue, brain fog, or cognitive difficulties severe enough to interfere with work or study for at least two months, with symptoms present at three months post-infection. Their long COVID had been ongoing for a median of 15 months at the time of scanning.
All participants underwent brain PET imaging with [¹¹C]PK11195 a well-established TSPO radiotracer alongside MRI to assess brain structure and volume. Blood samples measured two key biomarkers: neurofilament light chain (NfL), which rises when nerve fibers are damaged, and glial fibrillary acidic protein (GFAP), which increases when astrocytes are stressed or injured.
Long COVID participants completed a battery of validated mental health questionnaires covering quality of life, fatigue, anxiety, depression, insomnia, and disability.
The Main Finding: Brain Inflammation Levels Were Not Elevated Overall
The headline result may surprise those who assumed long COVID involves persistent, widespread brain inflammation. When the researchers compared TSPO availability across the entire brain — and across specific regions, including the hippocampus, amygdala, thalamus, white matter, and cortex they found no significant differences between long COVID patients and healthy controls in any area.
By contrast, the MS group showed meaningfully higher TSPO availability in several brain regions, particularly the thalamus, which confirmed the imaging technique was working correctly and that the long COVID-versus-healthy comparison wasn’t simply a case of the method being too insensitive.
Blood biomarkers told a similar story: NfL and GFAP concentrations didn’t differ between long COVID patients and healthy controls, suggesting no measurable ongoing nerve fiber damage or astrocyte injury at the time of imaging.
This doesn’t mean long COVID has no biological basis in the brain far from it. But it does challenge the idea that persistent, whole-brain neuroinflammation comparable to what occurs in MS is the primary engine driving symptoms in people with established, chronic long COVID.
The Timing Clue: Earlier = More Inflammation
One of the most intriguing secondary findings involved the timing of imaging relative to infection. Long COVID participants who were scanned within 16 months of their SARS-CoV-2 diagnosis had higher white matter TSPO availability than those scanned more than 16 months after infection a statistically significant difference (p = 0.04).
In plain terms: the closer to the original infection, the more measurable brain inflammation. By the time participants had been living with long COVID for over a year, those inflammatory signals had mostly faded even though their symptoms had not.
This is a meaningful finding. It suggests that COVID-19 may indeed trigger a period of elevated neuroinflammation in the weeks to months following infection, but that this inflammation resolves over time even as symptoms persist. The implication is uncomfortable but important: by the time most long COVID patients are showing up in specialist clinics, the inflammation that may have kicked off their symptoms may no longer be the dominant driver of what they’re experiencing.
The Most Striking Finding: The Limbic System and Mental Health
Here’s where the study becomes genuinely fascinating. Even though overall brain inflammation wasn’t elevated, the researchers found a striking, highly specific pattern when they examined the relationship between TSPO levels in individual brain regions and how sick long COVID participants actually felt.
In the hippocampus, amygdala, and thalamus three structures that together form the core of the brain’s limbic system, the network most involved in emotion, memory, and stress responses higher TSPO availability was strongly and consistently correlated with worse mental health outcomes:
- Lower quality of life correlated with higher TSPO in the hippocampus, amygdala, and thalamus (correlation coefficients ρ = −0.83 to −0.70)
- Higher anxiety (measured by GAD-7) correlated with higher TSPO in the hippocampus and amygdala (ρ = 0.88 and 0.83)
- Higher depression (PHQ-9) showed similar correlations (ρ = 0.87 and 0.75)
- More severe mental fatigue also tracked with higher hippocampal and thalamic TSPO
The researchers ran a permutation test to check whether this concentration of significant findings in the limbic system could be a statistical coincidence — i.e., just the result of testing many brain regions at once. It wasn’t: the enrichment of significant correlations in limbic areas represented more than a threefold excess over what would be expected by chance (p = 0.006).
This means the pattern is real: among long COVID patients, those with more glial activation in their limbic system tended to suffer more depression, anxiety, and reduced quality of life.
What Might This Mean?
The researchers are careful to note that TSPO isn’t exclusively a marker of inflammation. It is also associated with mitochondrial activity and other non-inflammatory cellular functions, meaning elevated TSPO in the limbic system of long COVID patients could reflect heightened metabolic activity in these regions, not necessarily destructive inflammation.
This distinction matters. The limbic system is central to mood regulation, stress responses, and the processing of physical sensations like fatigue. Chronic psychiatric and neurological symptoms, whether from long COVID, depression, or ME/CFS (myalgic encephalomyelitis/chronic fatigue syndrome), may alter how glial cells function in these specific areas, not by inflaming them in the classical sense, but by changing their activity levels, metabolic demands, or signaling patterns.
Indeed, similar patterns have been reported in ME/CFS and in major depressive disorder, two conditions that share considerable symptom overlap with long COVID, including fatigue, cognitive difficulty, sleep disruption, and low mood. The authors suggest this might reflect a shared neurobiological mechanism underlying the chronic psychiatric symptoms across these conditions, one involving glial-mediated changes in limbic system function.
What This Research Does and Doesn’t Tell Us
This is a small study; 14 long COVID participants is a modest number for drawing firm conclusions, and it uses a first-generation TSPO radiotracer that has lower sensitivity to subtle inflammatory changes than newer tracers used in some other long COVID PET studies. This may partly explain why the overall group-level differences in inflammation weren’t detected here, while some other studies using newer tracers in patients imaged closer to their initial infection have found elevated signals.
What this study does add, thoughtfully and rigorously, is a picture of long COVID that evolves over time. Early post-infection: possible neuroinflammation. Later, in established long COVID, that inflammation appears to subside, but glial activity in the limbic system correlates tightly with how miserable and impaired patients feel.
Whether that limbic signal reflects residual inflammation, cellular adaptation, or an entirely different process remains to be determined. What seems clear is that long COVID is not, at the chronic stage, simply a condition of runaway brain inflammation — and finding its real biological drivers will require looking beyond simple inflammation markers into the subtler metabolic and functional changes in specific brain circuits.
If you are experiencing persistent symptoms following a COVID-19 infection, speak with a healthcare provider. Long COVID is a recognized medical condition with a growing evidence base, and specialist clinics exist in many countries to support those affected.
References
Tuomaala J, Saraste M, Smith E, Kuusi M, Westerberg E, Honkonen E, Kargar R, Laaksonen S, Lehto J, Luoma A, Matilainen M, Misin O, Atosuo J, Kanerva M, Liira H, Laakso S, Posharina T, Saunavaara V, Wahlroos S, Rajander J, Airas L. Association between post-COVID-19 neuropsychiatric symptoms and persistent glial activation in the limbic system: a TSPO PET study. Journal of Neurology. 2026; 273:298. https://doi.org/10.1007/s00415-026-13842-w