Hacker News new | past | comments | ask | show | jobs | submit login
Single-Celled Organism Appears to Make Decisions (2019) (the-scientist.com)
113 points by jwmhjwmh on Nov 1, 2020 | hide | past | favorite | 61 comments



I think they buried the lede, which is that it has implications for the evolution of brains

“This paper nicely settles a debate between those [researchers] willing to accept that non-neuronal organisms are also capable of processing information and acting on that information, and those that stick to the idea that only neuronal organisms are capable of complex decision making,” Madeleine Beekman, an evolutionary ecologist at the University of Sydney who was not involved in the study, writes in an email to The Scientist. “Clearly there is a fundamental difference between brained and brainless organisms,” she continues, “[b]ut the point is that a brain did not come out of nothing. The brain is the result of selection pressure placed on organisms with the most basic form of information processing. Gunawardena and colleagues show that this basic ground work is already present in unicellular organisms.”


> This paper nicely settles a debate between those [researchers] willing to accept that non-neuronal organisms are also capable of processing information and acting on that information, and those that stick to the idea that only neuronal organisms are capable of complex decision making,” Madeleine Beekman ... writes ....

I don't know who Dr. Beekman has been debating with, but, at best, this has only been a debate of semantics for at least 50 years.

Here is an excerpt from a review titled Bacterial Microprocessing (Cold Spring Harb Symp Quant Biol. 1990;55:539-45. doi: 10.1101/sqb.1990.055.01.052.). The review is from 1990, but the work goes back at least to the late 1960s.

To someone trying to understand the brain, the bacterium Escherichia coli must be an awesome beast. Its talents are legion, but its size is miniscule. E. coli is a cylindrical organism less than 1 μm in diameter by 2/μm long--20 would fit end-to-end in a single rod cell of the human retina or some 3000 in that of a frog. Yet, it is adept at counting molecules of specific sugars, amino acids, or dipeptides; at integration of similar or dissimilar sensory inputs over space and time; at comparing counts taken over the recent and the not so recent past; at triggering an all-or-nothing response; at swimming in a viscous medium; and as we shall see, even at pattern formation.

At the 53rd Symposium, I presented an overview of sensory transduction in bacterial chemotaxis (Berg 1988), describing the kinds of measurements that E. coli makes on its surroundings and noting how this strategy matches the physical constraints imposed by small size in an aqueous environment (i.e., by life at low Reynolds number; see Purcell 1977). I will restate that case briefly, comment on progress made in the field during the past 2 years, and mention some of the work going on in my own laboratory.


Thank you for writing out this comment!

Even in the Age of the Internet, being able to put your hands on the right review paper is not given at all

My girlfriend and I both have mental health issues and that is most of my saw in wanting to understand Neuro 101 or comparative animal neuro or Evolutionary Neuro or anything that will incrementally add to an understanding of human brains that will improve the ability to advocate for ourselves in the US health care system.

Serendipitously, I was independent studying on fluid mechanics and dynamics last year[0] and had been through Purcell 1977, so I'm excited to check out your work on that account as well!

[0] There are some nice YouTube videos out there to ramp you up to understanding Purcell 1977 better as a layman with undergrad maths

Numberphile -- Reynolds Number

https://youtu.be/wtIhVwPruwY


It's not noted in the article, but it may be important that Stentor is a ciliate (same group as paramecium). Ciliates are very elaborate single-celled organisms, with the unusual feature of having a "macronucleus" in addition to a "micronucleus" that is similar to the single nucleus of other eukaryote cells. The macronucleus contains numerous copies of DNA from the micronucleus, which could provide a lot of scope for states that are represented in terms of gene activations, allowing behaviours such as described in the article.


This is new to me.

This specializes the regular nucleus for reproduction, while the macronucleus becomes a sort of brain, using DNA as an information processing and storage engine?

It calls attention to the brain-like activity of the regular nucleus in cells that don't have a macronucleus, and of the raw DNA in a bacterium.


'Makes decisions' is a terrible headline for this story. A thermostat makes decisions.

'Makes a suprisingly sophisticated series of decisions' is more supportable. But whether you are surprised depends on your intuition about how complex a tiny machine with feedback from sensors and quite a lot of internal state can be. Since almost every biological system I've ever learned anything about is mind-bogglingly rich with complex dynamics, my default is to be unsurprised by rich externally-observable behavior.


> 'Makes decisions' is a terrible headline for this story. A thermostat makes decisions.

Well yes, that's the point. They found that prospective control goes faaaaaaaaaar back in evolutionary history.


"Cognition all the way down: Biology’s next great horizon is to understand cells, tissues and organisms as agents with agendas (even if unthinking ones)" https://aeon.co/essays/how-to-understand-cells-tissues-and-o...


From 2016: "How Slime Molds make decisions" https://www.scientificamerican.com/video/how-slime-molds-mak...


I don't understand the issue? Decisions are just a state machine getting different inputs. It's the same for trees and humans.


The article says the organism tries several approaches in succession to escape the unpleasant carmine. This still isn't super advanced behavior, but it's interesting considering the organism is single-celled.


> the organism tries several approaches in succession to escape the unpleasant carmine

This implies memory. Venus fly traps use calcium gradients for short-term memory [1]. It would be interesting to learn how this organism remembers.

[1] https://www.sciencenews.org/article/how-venus-flytraps-store...


> This implies memory.

Shouldn't be too surprising.

"In genetics, a promoter is a sequence of DNA to which proteins bind that initiate transcription of a single RNA from the DNA downstream of it. This RNA may encode a protein, or can have a function in and of itself, such as tRNA, mRNA, or rRNA. Promoters are located near the transcription start sites of genes, upstream on the DNA (towards the 5' region of the sense strand)."

In programming terminology, the promoter region listens for events, and is followed by a handler body. So protein concentrations can be used for state.

Regarding the calcium gradients you mentioned, I wonder... I think they are more about inter-cellular communication. Cells can use calcium to make electricity. Calcium is also used by neurons to send electric pulses.

I actually find it fascinating that neurons are not unique in using electricity to communicate - that an analog exists in the plant world.


What does it remember? The approach tried previously?


Probably stress level, and the approach that follows, follows from that.


Yes, but it’s a richer state machine than previously thought, and we aren’t aware of the mechanisms it’s using for internal state


I think this is the key. "Just a state-machine" is a gross over-simplification. Imagine a tiny fruit-fly brain, vastly smaller than a human brain. Remarkably we can slice that brain into over 4000 pieces and try and rebuilt the connectome, which is a huge challenge in and of itself. Describing what's going on there as "state-machine" miss-represents a computer-science definition of "state-machine", and doesn't remotely reflect the biological complexity going on.


Whether it's more complex than a state machine can model (doubtful[0]) and whether it "doesn't remotely reflect the biological complexity" may be true but the guy's point still stands, it's a mechanism that can be modelled.

[0] you'd have to inject non-determinism as well.


non-determinism can also be modeled as a state-machine (and in fact, you can prove non-deterministic state machines can always be represented by deterministic state machines).


Ugh, you're right and I used completely the wrong terminology. I meant some degree of randomness (I think the posh term is stochastic-something), noise, jitter etc. Not NFA/DFA stuff. Good catch and thanks.


They get to do more than a state machine, because they can do analog computation.

You can simulate that with a state machine, but only with one that is much more complicated. Nature doesn't exactly prefer simplicity, but it's very bad at discovering solutions that need a big jump in complexity compared to a previous generation.


Maybe. We don't know that. You take it on faith, apparently, but that doesn't prove anything.


If you believe in a soul, and if you believe that a single celled organism (SCO) has a soul, then you can indeed delegate decision making to a non-corporeal entity. Not saying that's wrong but I struggle to accept that.


Or you might believe that a single celled organism can reasonably be understood as a state machine responding to inputs, but that a human -- or a tree, or a cat -- is not merely a scaled-up version of this same model.

We are far from knowing whether this is the case. So you can believe one extreme or the other (or something in between), but either way it's an article of faith, not an established fact.


Please clarify because I don't know - are you invoking the existence of a soul in the decision making process?

If so, you're bringing in something that's purely a matter of faith as no evidence has ever been found of a soul (AFAIK).

If no soul then be it single-celled or human, we're back in the mechanistic world and one is, somehow, just a more elaborate version of the other.


We have no more evidence that what we experience as consciousness and decision-making arises from the operation of a purely mechanistic state machine than we have for the existence of a soul.

Some people believe that when the machine becomes sufficiently complex, consciousness will emerge. Others don't. Either way, it's a matter of faith.


So the answer is yes, you're invoking the existence of a soul. It is not honest to duck a simple but pivotal and explicit question.

> We have no more evidence...arises from the operation of a purely mechanistic state machine than we have for the existence of a soul.

Evidence of a soul: 0%. Evidence of atoms: 99.999999999999999999% or higher.


Appreciate the attempt at engagement, but precision is not seen as helpful for "that kind" of discussion.

Ultimately, "you do you" is as much resolution as can reasonably be expected, down such a rabbithole.


I'm not invoking the existence (or non-existence) of anything. I'm just noting that we are far from understanding what consciousness is, let alone how it works.

(Who said anything about atoms? I didn't.)


> I'm just noting that we are far from understanding what consciousness is, let alone how it works.

The fact that we don't know what consciousness is doesn't mean we can't make reasonable guesses. Now, given that we understand both macroscopic and microscopic physics pretty well these days, there is only so much room for guessing and some guesses are much more reasonable than others.


> we are far from understanding what consciousness is, let alone how it works

Couldn't agree more, but that wasn't your original assertion, to wit first para of https://news.ycombinator.com/item?id=24962220

> Who said anything about atoms? I didn't.

You: "...from the operation of a purely mechanistic state machine" So what's a 'purely mechanistic state machine' going to be made of if not atoms.


A link to the HN guidelines:

https://news.ycombinator.com/newsguidelines.html

Note this one in particular:

  Please respond to the strongest plausible interpretation 
  of what someone says, not a weaker one that's easier 
  to criticize. Assume good faith. 
Also, don't forget: when you assume, you make an ass of u and me.


I appreciate the intervention, and I really must learn to let things drop, that said I've been trying to work out what the heck he's saying. I haven't knowingly responded to a weaker argument.


Energy


Perhaps the least relevant, but I really enjoyed the moral outrage portion of the renewed study.


This seems an appropriate topic in connection with the recent posting[0] about Charles Turner. I hope his work appears in the reference list, alongside the Harvard eugenicist. I know which neglected body of work I think is the more significant.

[0] https://news.ycombinator.com/item?id=24949335


So an individual person is extremely complex -- so a simple storefront business could easily be simpler than a person. Similarly, a single cell is incredibly complex -- so a simple multicellular organism (like a worm) could easily be simpler than a cell.

We all began as single cells (zygotes) and somewhere along the way we suddenly became conscious, whereas a moment before, we were not conscious. What was that point in time?


>We all began as single cells (zygotes) and somewhere along the way we suddenly became conscious, whereas a moment before, we were not conscious.

You may not have realized it, but many would debate the assumptions in that sentence. ^_^

Sidestepping a decent into the definition of 'conscious' itself for now, one could propose that consciousness is an inherent nature of matter itself, and it just emerges further/grows in complexity as the zygote develops.


A more justified explanation would be that consciousness is the capacity to learn and adapt to the environment, so we don't have to say 'it just emerges'. It's there, even at cell level, but it's not inherent in matter, that makes no sense. It's probably inherent in self replicators that live in a competitive environment, but not just any matter. Rocks don't learn and adapt and have no reason to do so.


Since "conscious" is an aggressively undefined term, you can say anything about it that feels good. I say "aggressively" because any attempt at a precise operational definition brings random testy objections out of the woodwork, typically invoking varying degrees of woo.

Ultimately, nobody has determined a practical need for a definition, or really for the word, outside of discussion of people in comas, so reasonable people let it go.


Is there evidence of humans that skip the zygote phase of life?


lol, no -- but one could imagine splitting up stem cells to grow new humans


> We all began as single cells (zygotes) and somewhere along the way we suddenly became conscious, whereas a moment before, we were not conscious. What was that point in time?

You never 'become' conscious, you adapt and interact with the environment from the first cell.


I don’t get what’s surprising. DNA transcription is computation. Decision making only needs the most basic of conditional information processing. Pretty much any regulatory process leading to homeostasis is a form of decision making.


Ok so cells make decisions. But what about organelles with independent dna - what about mitochondria and plastids?


Yes, though the details vary quite a bit.

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


This reminds me of Donald Hoffman's theory of conscious agents. He posits the simplest conscious agent to be a binary one : it can make only one kind of decision, with only two possibilities.


Here is a discussion between Donald Hoffman, Sam Harris and Annaka Harris about Hoffman's ideas on consciousness:

https://samharris.org/podcasts/178-reality-illusion/


Biochemical signal transduction networks, consisting of receptors and enzymes, are very similar to artificial neural networks, in terms of computational structure.


It's all about oscillatory patterns. These occur in all living things, it's just that brains are especially good at creating and maintaining them.


Great! Maybe we (Britain) can appoint it as Prime Minister in the next election. Gotta be better than what we have now.


Serious question: if I reframe the carnot heat engine as a decision making process that heat undergoes, can I publish a biology paper on it?


Appropriate for the election.


So, a scientist tried to replicate an experiment and failed, then kept trying until he succeeded by changing one factor of the experiment - the stimulus.

To Fix the Reproducibility Crisis, Rethink How We Do Experiments

https://blogs.scientificamerican.com/observations/to-fix-the...


“So an inventor tried to make a lightbulb and failed, then kept trying until he succeeded by changing one factor - the filament.”

The paper also includes dozens of trials replicating the results across organisms.


Well, I think it is an important result. If single celled organisms make decisions, what does that tell us about volition?

Evidence of volition and decision making in this form simplifies the challenge of understanding "consciousness".


> If single celled organisms make decisions, what does that tell us about volition?

Not much. The organism is expressing a preference for staying alive. That is a default preference for life.

The question of volition only emerges with higher preferences or ones at odds with survival.


> If single celled organisms make decisions, what does that tell us about volition?

The cell is a mechanism tested and refined by evolution. It's purpose is to sustain itself, adapt to the environmental aggressions and self replicate. Its volition is the best strategy nature could find to keep it alive.


Yeah, I'm wondering if or why we should consider volition in multi-cellular organisms any differently?

Are not single celled organisms as "conscious" as multi-celled organisms? Or, alternatively, are multi-celled organisms not as non-"conscious" as single celled organisms?


Even if a cell is so different in complexity to an organism, they both solve the same problem - what action should I take now, given the state of the environment and previous experience.


What is your understanding of the term "consciousness"?

What organism/s experience it?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: