Comment by foobarbecue | Hacker News Reader

Comment by foobarbecue | Hacker News Reader

Another one of these, eh? I have a really hard time believing that FMRI studies like this tell us anything useful.

There are repeatibility concerns, e.g. see this article from the same journal as the original link posted here: https://doi.org/10.1177/0956797620916786 , but that's not what bothers me the most.

What bothers me is that FMRI only measures blood flow, and people get tricked into thinking it's reading minds or revealing personalities. It just measures blood flow -- not electrical signals, not neural architecture, not thought. Sure, if you map that in high res 4D you're going to get lots of data and when you have lots of data you can find lots of patterns. When you can find lots of patterns you can publish lots of papers and fund yourself.

If there is a significant effect here, the paper should at best be titled "Lonely individuals have blood going to slightly different parts of their brains than non-lonely individuals." Ok, maybe... so what? Is this going to lead to some breakthrough therapy? No, it's going to be used to build further castles in the air. It'll be cited in other FMRI studies.

I don't have medical training, but my experience publishing and reviewing papers in science gives me a strong gut feeling that this stuff is worthless other than as a way to amplify bias and mint papers.

It seems that the statement "blood flow doesn't tell you anything about neural activity" is definitely not true: https://www.nature.com/scitable/blog/brain-metrics/what_does...

Also, it seems dangerous to suggest rejecting studying a specific area because the scientific tools and our understanding around it are currently limited - this is what science is for after all. If we thought this way, we wouldn't have nutrition research.

Nutrition research is famously inconsistent though?

Terribly so. So we need more high-quality research, not to stop researching it. The scientific method can only get better by putting it in practice.

Or we need to recognize some fundamental limits on what is measureable

Feels tautological to me almost. And of course reinforces that “those others that don’t think like the hive mind are BAD!!!” Which I find problematic.

Brains are incredibly more complex than simple blood flows indicate. I’ve read that the DNA of individual neurons morphs in the storage and recall of memory. Feels a stretch to use fMI and really anything singular to grasp at what’s going on. And feels grounds for abuse taken too seriously.

Brains are different one day to the next. They’re clearly going to be different one person to another.

Exactly. Blunt instrument, tautological reasoning. "The different people's heads are different!" History will group these studies with phrenology.

>What bothers me is that FMRI only measures blood flow, and people get tricked into thinking it's reading minds or revealing personalities. It just measures blood flow

So? It can still be a great proxy for thought processing in different areas of the brain, which is what is used for.

For example one would expect a brain to show very small blood flow while creating a shallow HN dismissal.

> It can still be a great proxy for thought processing in different areas of the brain

It's a great proxy for energy being used in different areas of the brain.

The concept of 'thought' is not crisply defined, and lots of brain activity doesn't correspond to what we usually mean by the folk idea of 'thought'.

Well, it would be some amount of physical proof for this maybe?

"The models also revealed new evidence of specific and non-specific associations, including the first evidence of a potential benefit (creativity) associated with unsociability."

https://www.sciencedirect.com/science/article/abs/pii/S01918...

Too late to edit my comment, but I want to note that I was misinformed that FMRI just measures blood flow. BOLD-contrast imaging is sensitive to oxygenation of haemoglobin, and it measures haemodynamic response, I just learned. So it's a little bit closer to information about actual neural activity. Thanks to thfuran below for pointing out this error.

>FMRI only measures blood flow

Isn't it more like a measure of metabolic activity?

Indirectly. The idea is that when cell metabolic rate increases, the brain's circulatory system compensates by dilating vessels to increase flow to that area. (EDIT: this is wrong, see below)

But the fmri is measuring change in decay time caused by change in blood oxygenation rather than change in flow rate, right? And that's a relatively direct measure of local metabolic activity (at least for some metabolic pathways).

You're right, and I didn't know that. Thanks for educating me. Apparently it's called BOLD-contrast imaging. Time for me to go read about haemodynamic response since I didn't know that was a thing.

Reply on news.ycombinator.com