Newly Discovered 'Rose Hip' Neurons May Be Unique To Humans | Hacker News Reader

Newly Discovered 'Rose Hip' Neurons May Be Unique To Humans

npr.org

103 pointssnapspans44 comments

While I am usually sceptical of anything that has even a whiff of human exceptionalism, this is a fascinating result. Ed is extremely interested in and well versed in comparative neurobiology, and despite the deathgrip that Allen's PR team has on all external communication, the fact that he seems to have approved of the spin is a good sign. Second, the primary outstanding question that I have is whether we have actually sampled the population in L1 sufficiently in other species. Another (maybe unpublished?) result that has come out of Ed's group (and collaborators) is that human neurons are extremely metabolically robust. They are almost like turtle neurons, they can be patched for extended periods of time and live for days to weeks. When compared to rodents this means that this cell type might simply be living long enough to see it, and dying too quickly/being too unhealthy to patch in rodents. The transcriptomic data suggests that this might not be the case, but the results from the sequencing of marmoset cell types are not in yet, and good luck getting this data for a great ape.

Once an aspect of human exceptionalism is noted, I would imagine one would look at chimpanzees as the next analogue, and not mice.

Nearest neighbor, top down, based on genetic similarity. If you’ve got something, and it seems that it’s only detected in humans, aren’t chimpanzees (or may other great apes) the gold standard for human comparative biology?

Then follow the chain, back through old world monkeys, before reaching rodentia.

I had some old pictures of stained rodent brains and saw something looked qualitatively like what they described in the first (and only) image I checked. Since they dont give any definite criteria for me to compare against, what can I do?

Well, they give about the most definite criteria they can given current techniques, which is the transcriptomic profile. In fig 2e [0] they also give a bouton density profile. If you still have the pictures and have neurogliaform and basket cells stained with the same technique, you should be able to determine whether you are in the ballpark.

https://www.nature.com/articles/s41593-018-0205-2/figures/2

I looked at the paper. I mean there needs to be a table that tells us what they consider the acceptable range for spine density, bouton density, soma size, soma ellipsity or whatever shape stats, branching stats, etc.

I dont really know what to do with that figure 2e... it looks like they cherry picked stats where they saw a "significant difference" from the other types of cells they looked at.

I'd imagine there are slides, and slides, and slides of ape neurons in circulation, possibly imaged as photomicrographs and stored in digital libraries, and heaping data sets of genomic primate data.

Are data rights costly? I doubt that sort of information has a shelf life. Consider that HELA cells are still around.

If they release the specific results, and how they arrived at such conclusions, others should be able to take that in hand, and move quickly to look at primate corollaries.

It's a lot cheaper and ethically less problematic to use lab rats instead of apes, so that's what we usually do.

Well, yeah, but ape brains from deceased apes in captivity are probably short-listed for science experiments and not cremation or natural burial.

So it’s probably less difficult than finding a match for an organ donor, no?

I think they have some currency to start a search for samples, with this first stage finding. But really, as we’re finding out, replication will be important too.

Let's say you have rats and human brain samples and are ready to present a finding. Do you hold off until you find an ape brain, find expertise in dealing with ape brains (if reqiured), deal with IRB issues related to using those materials, run the tests, then publish? Obviously not. You use this preliminary work to run a more definitive series of experiments. This is how science works..as a series of studies, not a single definitive study.

If money and logistics were not an issue, you might be right about it being the most logical step, but primate research is not something that is largely accessible to most researchers. In any case, a null in rodents tends to put a limit on how out from humans such a neuron goes, and they are cheap to work with.

>"While I am usually sceptical of anything that has even a whiff of human exceptionalism, this is a fascinating result...the primary outstanding question that I have is whether we have actually sampled the population in L1 sufficiently in other species."

What do you mean by "sufficiently"? Did they check in any animals? Did they provide a list of criteria that, if met, allow definite identification of these cells in an animal? Can you quote these parts of the paper?

This paper did not check. This was a purely human study. I'm talking about known results in the field.

> I am usually sceptical of anything that has even a whiff of human exceptionalism

Why are you skeptical? There is clearly a huge leap in rational intelligence when you go from other animals to humans. I would expect we would eventually find a reason for it.

Empiricism. There have been many claims to human exceptionally that have been spurious or found elsewhere. And there are exceptionally human things that are probably just chance because every species likely has something odd going on (rats! Have no gall bladder and have two insulins)

>There is clearly a huge leap in rational intelligence when you go from other animals to humans

Citation needed. Monkeys/apes even fight wars between tribes. All differences between humans and other animals seem to be purely quantitative. I'm sure the difference between your average cat and the smartest species of non-human monkeys/apes are greater than the difference between the smartest species of non-human monkeys/apes and your average human.

I wonder how dolphins and elephants would compare to human intelligence if they had vocals pathways capable of speaking and limbs capable of grasping.

"If a lion could speak we wouldn't understand him" -

https://en.wikipedia.org/wiki/Philosophical_Investigations

I'm pretty sure I've heard dolphins speaking and seen elephants grasping.

Elephants have a grasping limb, not grasping limbs. That may seem pedantic, but having at least two hands is probably a prerequisite for a lot of simple technological development.

”It is still possible that these newly identified neurons will also be found the brains of primates like monkeys or chimps, Lein says.”

Possible and, in my universe, quite a bit more likely than that these cells are “unique to humans” (cursorily reading the paper, these scientists didn’t even look for this type of cell in anything but humans, but only claim that, if rodents had them, they would have been found there, given the amount of time spent looking at rodent brains)

Disclaimer: I’m not a neuroscientist (but that’s an advantage, too. I don’t have to keep hunting for funding. That removes incentives to sensationalize results)

By that logic, if humans had them, they would have been found before now, given the amount of time spent looking at human brains.

It seems you have to be looking in the right place, in the right way, for the right thing.

But it's wrong to knock a new discovery - even if not unique to humans, it's still a discovery. Let's celebrate the feature for itself, not benefits that might follow.

There's probably been orders of magnitude more time spent analyzing rodent brains, simply because there is an effectively unlimited quantity available and an extremely limited quantity of human brains available. Essentially all invasive human brain studies are preceded by animal model studies (usually starting with rodents before proceeding to "higher" animals if that's called for), but conversely not all animal model studies result in a human follow up. Meaning a hell of a lot more rodent brains get put under the microscope. So in general you would expect to find them in rodents first, if the probability of serendipitous discovery is purely a function of time spent conducting unrelated analyses.

I wouldn't expect the number of human autopsies to be more than a few million over the entire history of the human species. Nobody looks at human brains. It's a huge problem.

US alone does over 3 million autopsies a decade so it’s easily 10’s of millions over human history. The issue in terms of science is the vast majority of these have nothing to do with scientific research.

Right, no one is doing mouse autopsies to determine the murder weapon but we don't wholesale kill humans in the quest for scientific research either. Don't discount a medical examiner's job as not being science. They aren't discovering new cells in the body but they are more akin to a GP doctor (versus a clinical research doctor). Just like your GP doctor, they can read the latest medical research, try new techniques and practices, and report experiences at conferences. Or they can work out of a 1950's handbook. But just like an engineer that can't be bothered to adapt to new technology, they are quickly going to fall behind their peers.

We don't remove the brain from someone murdered because they likely didn't give researchers permission to do so. Moreover, to study cells requires some skills not found at the coroner's.

Uniqueness as an artifact caused by limits on current knowledge is a salient point. To give an idea of the state of the art in the understanding of living brains, this page presents a pie chart of known neuron types by species:

http://neuromorpho.org/byspecies.jsp

If an alien stumbled across that data they might surmise that rodents "rule" this planet, followed by fruit flies as their symbiotes/slaves/pets, followed by everyone else.

However, if they had a moral code like ours they might instead surmise that rats, mice, and flies are viewed as expendable to the actual "rulers" of the planet and that these have an aversion to slicing themselves up for science. So they might conclude the rulers to be one of the primates given the existence of three close species: human/chimp/baboon in one family (monkey).

I read the article, and it seems like it could be an important difference for some types of biology research, if they actually verify it isn't found in other species that they haven't checked yet.

But my initial reaction to the headline was about AI research. My personal belief is that if you are looking at biology for help creating more powerful general purpose AI (AGI or "strong AI"), the idea that humans are special is actually mostly wrong and is counter-productive. For example there were many earlier efforts that emphasized reproducing the more rational aspects of human cognition (things animals seemed unable to do), but those turned out to be dead ends for AGI. Computers can do automated mathematical problem solving or proofs or play chess much better than humans, but it turns out that much of what people do doesn't have cut-and-dried rules that we can just enter in.

Most, if not all of the capabilities that have been more difficult to achieve in AI are shared between humans and animals. If anyone has had a pet they might have noticed some abilities that may or may not have seemed particularly smart at that time. But when you start to look at it from the perspective of reproducing those abilities in robots, no one has come close. And those capabilities are shared with humans and can be mapped to higher-level abilities.

For example, start with the advanced vision and motor systems that work together in the average house cat. Add in the ability to communicate and coerce humans into providing fresh food and water at will (ok, my cat was spoiled). The desire and ability of the cat to play in a slightly creative way -- my cat would look at me, arch her back, puff up her tail, tilt her head, as a clear communication that she wanted me to play chase with her. Then she would run and find a new place to hide. Or she would run up and smack me with both paws on the ankle at the same time, check that I had noticed and then run away -- another invitation to play. (Coincidentally, tag was also by far my nephew's favorite game when he was little.)

I guess I don't really have the motivation today to make a really in-depth and convincing argument here. But at least consider making artificial cats before you make artificial humans. Even if you don't agree that it is a big step towards reproducing human abilities. Because cats _can_ be pretty great, but litter boxes are disgusting.

But we’re totally going to be able to simulate the human brain and upload consciousness into a computer, right?

What if singularity already happened and we are its result?

Then we could probably make it happen again, for us.

eventually

I'm quite curious as to what provides for the kinds of mental 'structures' that make for so much of our inner lives.

This is an old blog post, and seems to have lost the formatting, but this is the concept I'm pointing to:

https://www.gapingvoid.com/blog/2004/07/31/pillar-management...

The conclusions we come to, wind up constraining the thought space of our future thoughts. I believe this cognitive process demands explanation.

I feel discoveries like this that can either be a promising lead that turns out to be nothing (e.g. mice don’t have these neurons but other animals do) or could wind up being the single biggest discovery in the study of sentience ever are the perfect example of what real science is like.

When has it ever been the latter?

Penicillin was pretty big. Amusing discovery too. As penicillin is fundamentally a derivative of the blue-green mold that can appear on decaying matter, the idea of injecting it into yourself to combat bacterial infections is something that few would believe. It only began to be regularly used some 14 years after its discovery, even the discover rapidly shared and published his findings.

On the other hand, maybe these discoveries are a bit different. It seems the big discoveries often come in the form of falsifying previously held beliefs rather than immediately introducing novel 'actionable' discoveries. For instance Mercury's orbit goes off by a pesky 0.01 degree change per century that could not be accounted for by Newtonian planetary mechanics. That falsification is perhaps at least part of the long chain of events that would culminate in the discovery of relativity.

> A scientist named Gábor Tamás and members of his lab were studying brain cells called inhibitory neurons, which act like the brakes in a car. They tell other brain cells when to slow down.

Is this "explain it like I'm five" passage really necessary? Surely any adult would understand what an inhibitory neuron does from the name alone. Would a writer feel the need to explain the meaning of the word inhibit? Inhibition? Uninhibited? And if they did, would they do so in such childish terms?

Not necessarily.

I self-assess as 'fairly smart', but my neurological understanding is pretty poor, and inhibitory neuron may mean a neuron that inhibits other neurons, or a neuron that inhibits something else entirely. If that's to 'slow down' the pace of neuron activity or to actually suppress it (ie. inhibit it) isn't 100% clear from the name, either.

Does an inhibitory neuron prevent other neurons from firing, negate the effect of their firing, or slow them down?

It can do any of those or the opposite. The effect of the signals from one neuron depend on the properties of the receiving neuron.

Ie, the same neurotransmitter released can be "excitatory" to neighboring neurons in one case and "inhibitory" in the next: https://www.sciencedirect.com/science/article/pii/S030645221...

"These types of cells [inhibitory neurons] are extremely important," he says. And when there's dysfunction in them, he says, that can "directly be linked to different types of neuropsychiatric disease, like schizophrenia."

That's a really bold claim with nothing in the article backing it up. If I put my thinking cap on, perhaps there is an established link between failure of inhibitory neurons and neuropsychiatric disease (this would not surprise me but it's not called out in the article), but even if that has been established it's a big leap to assign so much importance to this particular variant of inhibitory neuron that has just been discovered (in dead people) and is barely understood.

The claim isn't that bold, there is a well established link between inhibitory neuron disfunction and neuropsychiatric disorders.

I think you're mis-reading the comment. The passage says that inhibitory neurons are important, not the 'Rose Hip' neurons.

You are correct. I interpreted that as saying these particular cells are important, not the general class of inhibitory neurons. The article also didn't (unless I missed it) spell out the connection between inhibitory activity and neuropsychiatric disorders.

I had not heard of the relation between lack of inhibitory action and such disorders, but I know glycine is an inhibitory neurotransmitter and its deficiency can play a role in such disorders. Now I know of a more general framework to put that into. So this all make sense now, but it wasn't clear to me strictly from the article. Thanks.

Reply on news.ycombinator.com