Hacker News Reader

Comment by elric | Hacker News Reader

It's taken 12 years to get from "this stuff exists in mice" to "we now know it actually exists in humans and is not vestigial". That seems like a long time. People get brain MRIs with contrast all the time, is there any reason why this never showed up? Because no one was looking? Or because it's a slow mechanism?

View on HN

Looking up “brain MRI with contrast”, the biggest difference seems to be that in a regular MRI the contrast goes to the blood, but here it goes straight to the brain cerebrospinal fluid. You need open brain for that, so indeed doesn’t seem trivial.

Cerebrospinal fluid is produced in the brain, I would imagine that the contrast would end up in newly produced fluid as well, but maybe that assumption is wrong.

Even assuming that the contrast could show up in newly excreted CSF (maybe, maybe not), MRI contrast elimination half lives are very short (mostly all under ~2 hrs, excepting cases like renal dysfunction), and cerebrospinal fluid doesn’t replenish particularly quickly.

Unfortunately no, that is not how the contrast agent gets around. Most Gd agents are huge and don’t leak out of the vasculature unless there is a disruption (tumour etc).

Two things:

1. Blood brain barrier and CSF should be separate for all but tiny molecules. It's why CT angiograms are able to visualize distinct vessels. So it is pretty hard to directly interact with this sort of thing in vivo

2. A good chunk of the neuro community have been operating under the assumption that some of those mouse model findings are mechanisms in humans too. Since we couldn't easily prove it, people used a bunch of next best tools with fancy imaging that demonstrated it was very likely. On top of furter proof, this sort of study allows us to begin pinpointing exactly how close our next best tools are at estimating in vivo processes without opening up the head.

They injected a dye into the cerebrospinal fluid to confirm the pattern of diffusion into the brain.

I imagine this procedure would have to be confirmed safe in humans first. Also they needed subjects who already happened to be undergoing a specific type of brain surgery.

Its because we don’t normally inject Gad in CSF. Its kind of wild west to even do so, and these researchers were able to identify this unique dataset and ask some interesting questions.

The main reason is because 12 years ago we didn't have ChatGPT to tell the researchers the answer.

Reply on news.ycombinator.com