Discordant voxel clusters likely coincident with large vessels, and what you're seeing is big changes in blood gases with attention.

...defining voxels by their response (-ve) implies reduced activity but doesn't prove it, and I'd prefer a principled neuroanatomical basis, or EEG, etc., which leaves me assuming a lot of vascular physio artifacts are going to be interpreted. A case for subject-specific HRF, not canonical, perhaps!

So it's all a bit beyond my expertise, but things that I worry about: what's going on with the brain edges in Fig1c? They discuss partial vol. effects, but why the crescents? Next, since I assume the DMN is primarily vascular anyway, I can't interpret +ve or -ve meaningfully there! Finally,...

Yep, that’s what I was implying by arousal, as shown by Korpanay et al last year.

I’ll give it a look too. At first glance I worry about the separate PCASL and BOLD runs. Attention & arousal effects could differ systematically.

In their defense it was a demo of a custom system rather than a novel neuroscience piece!

That part can be neural, just like cerebellar diaschisis after stroke. I’m only pointing out the initial event as having a strong potential for being vascular.

PPS it might help to think of transient ischemia as a model. It can lead to spreading depression - enhanced firing for a brief period - followed by prolonged suppression. Not saying tFUS is creating transient ischemia, only that we do have related knowledge to draw on in physiology.

PS the "online" results might also have a contribution from a (weak?) Lorentz effect in the MRI magnetic field, e.g. pubmed.ncbi.nlm.nih.gov/37236172/. This would indeed be neural, although I suspect dominant vascular effects because we know vessels sense pressure normally.

Final note, and mirroring the real impetus behind my original post: Nowhere in that paper do the words blood, vessel, capillary or artery appear. BOLD is as good as it gets. In the Discussion they mention all sorts of explanations, but none of the plausible mechanisms is vascular, apparently!

...activity. That's even easier to explain as a persistent suppression of vasodilatory capacity. SMCs and pericytes all have mechanosensitive ion channels for autoregulatory purposes. Had they used the same paradigm online as offline we might have had greater mechanistic insights.

They do two radically different expts in that paper, making it even harder to determine what's actually happening. "Online" they demo enhanced V1 activity. That could arise, e.g. from retrograde signaling in LGN vasculature which is sensed by astrocytes. Then, "offline," they show reduced V1...

That could still be a vascular first mechanism.