Thread by @ZdenekVrozina

New study out of Amsterdam UMC asks a question most Long COVID imaging papers don’t tackle at once - does inflammation in the brain actually track with how well different brain regions talk to each other? 45 people, roughly 27 months post-infection!🧵

TSPO PET is a scan that lights up wherever immune cells in the brain (microglia) are activated - basically a map of where inflammation is happening. This version is fully quantitative, with blood sampling during the scan, not a shortcut estimate.

The second scan, resting-state fMRI, measures which brain regions sync up while someone just lies there doing nothing.

They split people into two separate pairs of groups
high brain inflammation on PET vs not (10 vs 35),
ongoing cognitive complaints/fatigue vs not (31 vs 14).
Then they checked whether network connectivity looked different across each pair.

The authors propose
First - brain inflammation tracks with weaker connectivity in the visual and attention networks - and between the brainstem and a control network. Second - the complaints (fatigue, brain fog) track with weaker connectivity in the default mode network (the network active when your mind wanders, reflects, remembers) - and between the thalamus and motor cortex.

So inflammation drives one pattern of problems, the complaints themselves drive a different pattern. Two independent lines of evidence that happen to converge on the same brain regions.

But every single one of the 10 people with high brain inflammation also had ongoing complaints. Zero people had high inflammation without complaints. Maybe the inflammation group is a subset of the complaints group - not two separate observations.
The authors admit this in their discussion, calling it shared variance.

This is a group-level tendency, not something usable for an individual person.
There’s also no non COVID comparison group - all 45 people had COVID. The comparison is only COVID with complaints vs COVID without complaints.

A single snapshot in time. 27 months is a longer window than most studies in this space, which deserves credit - but it’s still different people scanned at different points, not the same brains followed over time.

In the discussion, the authors bring up Alzheimers and MS as precedent for chronic inflammation eventually damaging networks and leading to neurodegeneration.

Yes, this connects to something bigger. The brainstem disruption here lines up with other recent work - one find reduced brain energy use in a similar region (limbic system, smell-processing areas, brainstem), another raised doubts about whether the brainstem itself is even a clean, unaffected reference. Three different imaging techniques, all circling the same small piece of anatomy.

It’s exactly the region current models point to as the starting point for Parkinson’s disease and related synucleinopathies. Three different machines, three different things being measured, all converging on the region you’d expect to light up first in an early synucleinopathy - not generic neurodegeneration.

And it’s a pattern getting harder to wave off as coincidence. The closest viral parallel is HIV. Chronic immune activation in the brain rewires how networks function long before anything resembling classic neurodegeneration shows up.

This is another real signal in a condition the system still often frames as psychogenic. There’s too much signal now to keep operating without prevention. @szupraha @ZdravkoOnline @adamvojtech86

Visser at al., Decreased functional connectivity in post-COVID syndrome patients with high neuroinflammatory activity. sciencedirect.com/science/articl…