MIT scientists discover neurons that light up whenever we see images of food | Hacker News Reader

MIT scientists discover neurons that light up whenever we see images of food

news.mit.edu

75 pointsrntn27 comments

It's been awhile since I read through this literature, so I'm not an expert and things might have changed. However, with face neurons at least there was always an argument about the extent to which the neurons were specific to faces versus extremely motivationally significant stimuli that required individual recognition. I vaguely remember there being studies showing that if you showed people something that was very emotionally significant to them, and that they were expert in discriminating between exemplars (for example, many similar but different models of vintage cars to a vintage car enthusiast), similar neural populations would be recruited. The argument was that faces just happened to be something that tend to be really emotionally salient to most people, and that people have to differentiate between large numbers of exemplars of.

Maybe that interpretation has been disproven now, or maybe it doesn't quite apply, but I wonder if it's really food or something else. Or if maybe at some level the neurons are responding to something so similar to "food" as a category the distinction is moot.

This reminds me of something I read about babies (I have one). When they cry for hunger, is it that they are trying to say "I want food"? Probably not - it's more like, my stomach feels emptyish and that feels dangerous and I want something to come help me feel different.

I read recently that even the smell of food kept fruit flies from getting life extending gains of calorie restricted diet:

https://medicalxpress.com/news/2022-07-serotonin-dopamine-mo...

Would be interesting to see the link between these neurons and the hormones that control hunger in our bodies. Perhaps there is a way to deactivate neural pathways when we see food, v. just targeting hormonal pathways with obesity-related medications.

The paper: https://www.cell.com/current-biology/fulltext/S0960-9822(22)...

One of the highlights: a novel food-selective component is discovered in the ventral visual cortex

Ah yes, totally not circular. When I see me dog the neuron says no food. When I see a cow it says food–Oh wait, that's actually Bessy my pet cow, so it doesn't light up.

There's no "Halle Berry neuron", it's just correlation. Not really groundbreaking that there's so-called neurons for "food" whatever that may be.

do discoveries need to be ground breaking in order to be interesting?

Are they really a "discovery" if they aren't? Seems more like a plain observation.

Yes they are. (Often, but not always)

Depends on context. In fact, just finding a new context that fits existing discoveries is a discovery.

And then there is whether we have fully considered a given observation.

Say we make a discovery. It will have an obvious scope. Usually. And then a broader scope will be identified. These discoveries are high value and may even eclipse the original!

And then over a very long period of time the more complete implications and ramifications of a discovery should be known. Ever learn something significant that ripples through your mind? Like it has high impact, but it takes processing. And that takes triggers too.

For an example, tell me all you know about cars. We can't really do that.

Now, say I gave you a profound car fact or dynamic, rule... And then we start talking about cars. You will no doubt realize some discoveries, insight as various things about cars comes to mind, and that fact, or insight ripples through... And of course, you will muse about it on your own, insight spreading rapidly that way too.

All of this can make seemingly mundane or obvious things seem boring and ordinary to one of us and powerful and new to others among us. We really are limited in some basic ways we so not often think about.

And...

There are a lot of us people, a lot of discoveries and so damn much to communicate! We are forever behind in many ways, not just the basic latency that lies between us and reality.

I could go on, but the hard fact is we really do ourselves a disservice weighting things like this. Very few of us are qualified enough for it to make sense and for tons of us the data adds next to no value.

The better thing is to communicate whether it is new to us and whether a conversation might help to make the most sense.

All IMHO of course!

I think of this the same general way I do answers to the who is smarter question. Actually answering that is time consuming and we find there are many axis. The answer may not mean much!

But there is a related observation I find very useful, and that is we are all likely smart enough! The difference boils down to the amount of work needed to proceed.

I will stop there. Hope you all get something valuable from this perspective. I did.

Nope, see comment below.

Not being dismissive - did we not already know this? What about this finding was not already known?

I think it was not known that it are always the same neurons lighting up, even between humans. So not learned behavior, but baked in through evolution.

But I didn't read the article, so I could be wrong.

In what sense can it be the "same" neuron across different people? And if the answer is that both light up under the same conditions, isn't that circular?

It can be the "same" neurons if they are found in the same area of the brain. E.g. in some layer behind the retina.

"To do that, the researchers applied a mathematical method that allows them to discover neural populations that can’t be identified from traditional fMRI data. An fMRI image is made up of many voxels — three-dimensional units that represent a cube of brain tissue. Each voxel contains hundreds of thousands of neurons, and if some of those neurons belong to smaller populations that respond to one type of visual input, their responses may be drowned out by other populations within the same voxel.

The new analytical method, which Kanwisher’s lab has previously used on fMRI data from the auditory cortex, can tease out responses of neural populations within each voxel of fMRI data.

Using this approach, the researchers found four populations that corresponded to previously identified clusters that respond to faces, places, bodies, and words. “That tells us that this method works, and it tells us that the things that we found before are not just obscure properties of that pathway, but major, dominant properties,” Kanwisher says.

Intriguingly, a fifth population also emerged, and this one appeared to be selective for images of food."

I think what the research is showing is that our visual cortex is primed to see 5 different categories: faces, places, bodies, words, and food.

I want to know what happens when they read text describing food.

I want to know what happens when they read text.

Actually the opposite. The story of a single neuron on/off state for everything we experience is just that, a story. If you tried to build a skull large enough to hold all that, it would burst. The reality is that paradigm of hardwired neurons works well for the initial layers of processing (Eg recognizing lines and circles in visual stimuli), but that same model can't be repeated all the way up to the point of having a "grandmother neuron" for recognizing Grandmas face. There's just too many possible states of grandmas. There has been some evidence (like this study) that maybe sometimes there are "sparsely coding neurons" (Eg, some monkeys apparently have a brain cell that reliably responds to seeing Jenifer Aniston and nothing else). But such sparse coding (if it exists) is the exception not the rule.

https://www.youtube.com/watch?v=_njf8jwEGRo

Weird/random antidote - back in my late 20's I was going through fairly extreme dieting/working out (no carbs, calorie cutting, hours in gym, 6-pack abs yada yada) and everyday I would do a 45min incline walk (I did this for over a year straight) - the key to my success here was watching youtube videos of the people who binge eat all sorts of foods and try to have insane calorie days eating delicious take-out etc.

I always looked forward to the 45min treadmil/food-video combo & it felt liked it helped combat any diet depression (regardless of how masochistit may've seemed).

Similarly, when I was doing 500 calorie a day cuts, watching people binge eat food kept me sane and satisfied any food cravings I had.

Luckily, ever since eliminating sugar from my diet and doing monthly one day fasts I get no way as hungry as I used to

So you basically discovered mukbang.

nahhh, none of the gross stuff. More like fit people eating 50K calories a day - all delicious food.

My wife says my face lights up when I see food. Good to know my neurons do too.

So, how certain are we that these neurons and no others will also activate when we see food, rather than images of food? Or when we both see and smell food?

I know it may seem like a small jump to make, but our understanding of the brain is fairly limited.

So... seeing images of food fires certain neurons.

Do the same neurons fire when viewing actual food irl? Or do the neurons know what an image is? What about imagining food, will that fire the same neurons? How about imagining images of food?

The advertisement industry will eat this up.

Mmm.. picture of burger

Reply on news.ycombinator.com