The mind of a worm, the mind of a human, and mind uploading

In 1986, a key paper on the nervous system of the nematode worm Caenorhabditis elegans appeared. It was a fairly complete description of the entire nervous system of the animal: it counted all the neurons (302 then) and a good first pass at counting neural connections (5000 chemical synapses, 600 electrical synapses, and 2000 synapses to muscles).

The paper was subtitled, “The mind of a worm.” 

This subtitle was despite there being nothing about the behaviour or internal states of the species. It was all anatomy.

I think it was called that for a couple of reasons. 

First, co-author Sydney Brenner was a playful writer. He wrote a great column for Current Biology for many years that often showed his wit.

Second, it represented a hope in invertebrate neuroethology that understanding the neural connections of an organism would allow you an almost complete ability to predict the behaviour of that organism. A few papers since have used the phrase “mind of a worm” in homage to the 1986 paper.

So the subtitle was probably a joke, but one the represented a real goal of invertebrate neuroethology: to crack the neural circuitry of animals. Then we’d understand behaviour.

Fast forward to 2022. Our understanding of nervous systems is better. Technology has gotten batter. But a preprint about C. elegans released this year says, “We find that functional connectivity differs from predictions based on anatomy.” 

Wait, what? Wasn’t the promise that that once we had the neural connections, we would have near omniscience about the behaviour of thise worm?

In the world of invertebrate neuroethology. this is kind of yesterday’s news. It’s been decades since neuroethologists gave up on the idea that neural circuitry alone would give you a relatively complete understanding of behaviour. I suspect that idea was already wobbly in 1986 but was pretty clearly abandoned by the 1990s. It was killed by things like the study of neuromodulation in the stomatogastric nervous systems of decapod crustaceans.

But it seems that the news has yet to reach other fields in neuroscience.

After my last post on mind uploading, Ken Hayworth helpfully sent me a link to an article he wrote about mind uploading. I appreciated the writing, particularly this concise summary.

The core of the scientific argument: I am my connectome and (aldehyde stabilized cryopreservation) preserves the connectome

Most of what I said in my review of a book length treatment of the connectome argument is relevant here. In brief:

Invertebrate neuroethology has already tested the connectome hypothesis. The connectome hypothesis didn’t exactly fail, but it certainly under performed. Knowing the connectome of a worm has not revealed the mind of a worm.

The only reason to think, “Sure, the connectome didn’t give us everything we wanted to know in small neural circuits in invertebrates, but it will totally work in humans” is special pleading.

External links

Vitrifying the Connectomic Self: A case for developing Aldehyde Stabilized Cryopreservation into a medical procedure (PDF)

References

Randi F, Sharma AK, Dvali S, Leifer AM. 2022. A functional connectivity atlas of C. elegans measured by neural activation. https://doi.org/10.48550/arXiv.2208.04790

Related posts

Brainbrawl! The Connectome review

If you know how to do mind uploading, please do tell us

 

If you know how to do mind uploading, please do tell us

Here we go again.

Ken Hayworth is mad that more neuroscientists aren’t working on the problem of immortality.

(I)t is still really bugging me that we could ALL likely get to the future together (via quality brain preservation and far future mind uploading) if only the neuroscience community would start taking its own findings seriously and consider the implications.

He seems to think that neuroscientists are avoiding this topic because we are all secretly dualists?

Neuroscientists should not succumb to magical thinking about the mind/brain.

I don’t know any other working neuroscientists who think the brain in magic. Citation needed.

Neuroscientists should know that we already have the tools to preserve brains fantastically well at the ultrastructural and molecular levels using aldehyde fixation.

But Hayworth does not say how he can demonstrate that the current level of detail of preservation is “enough” detail. Sure, synaptic connections will be important. What about resting membrane potential? Gasotransmitter levels? Blood flow? Can any of those be captured in a static preserved brain?

Of course mind uploading will be possible in the future… just read the damn literature on models of brain function… they are all computational.

But Hayworth does not say how he can demonstrate how structural information can be converted into a computational model.

Why refuse to openly discuss/debate what should be considered the most important long-term application of your research?

Here I am, discussing it. I’ve answered a lot of questions about mind uploading on Quora. Let’s see... (Searches Quora) In no particular order:

Is mind uploading digital immortality a real possibility for the future?  

How is it possible to copy our brains into computers?

How realistic is mind uploading and what should I learn in order to research it?

Why is there no focus on uploading our brains to travel in efficient small spaceships? Should we fund mind uploading research for this end purpose?

Are we working on a way to upload brilliant minds in various scientific disciplines into a digital repository to hasten scientific advancement?

Does quantum computing make it possible to upload a neural net of your brain into a simulation so that when your body dies a copy of your brain itself could live forever in a computer network with other people?

Could thoughts and memories be digitalized at some point?

What is the current state of progress on mind uploading? What are the specific major obstacles and progress on each? 

How long will it be until I can upload my consciousness?

Why are people so against the idea of mind uploading and even question its possibility? 

Do any laws of physics prevent mind uploading?

What percent chance is there that whole brain emulation or mind uploading to a neural prosthetic will be feasible by 2048?

Will human consciousness ever be transferable to a new body or a machine e.g. a robot cyborg computer or avatar? Can-you transfer your own consciousness and memory and leave your biological body without creating two selves? 

Can the-mind be uploaded to a computer?

What do we need to know in order to be able to upload a human mind? 

It is like physicists refusing to discuss the possibility of fusion power plants. It is like NASA refusing to contemplate human settlements on Mars.

Hayworth thinks mind uploading is a mere engineering problem – like space travel in the 1920s. It was pretty clear that you could get into space with a big old rocket. (A movie about going to the moon, Frau im Mond, was made in 1929.) Conceptually, we knew how to solve the problem. Creating the rockets was a mere engineering problem. 

But I argue that mind uploading in the 2020s is not like space travel in the 1920s. “Mind uploading” is an unsolved conceptual problem, as far as I can see. That is, it’s more like NASA admitting they don’t know how to create artificial gravity or faster-than light travel than refusing to think about sending humans to Mars.

Liquid fueled rockets were flying in the 1920s. What’s the “mind uploading” equivalent of Goddard’s rocket? Can Hayworth point to a single working model of taking a preserved nervous system and successfully reconstituting it as a complete digital model that can faithfully reproduce behaviour?

For me, science is the art of the solvable. If there is anything I have learned in science, it’s that the only way to tackle big questions is by tackling small questions.

And I am selfish. I want to contribute to problems I can solve in my lifetime. Or, at least, see some significant progress towards solving. I don’t want to chase pipe dreams of the “far future.”

Many a career has been wrecked on the rock of “big ideas.” Immortality is a exactly such a rock. 

 Do you have scientific objections that will stand open scrutiny?

Wait. Burden of proof is on the claimant. If Hayworth thinks mind uploading is possible, it is up to him to specify why it is possible. In considerably more detail than a six tweet thread.

Human life not worth saving?

Hayworth may consider a biologically based computer model to be a life. But many will not. Nowhere in his description is how this “uploaded mind” will interact with the world. Hayworth probably considers Alex Murphy’s existence in RoboCop not just acceptable, but positively luxurious:

I don’t think a disembodied simulation of a dead brain warrants the name “life.”

Related posts

Overselling the connectome

Brainbrawl! The Connectome review

Brainbrawl round-up

“Mind uploading” company will kill you for a US$10,000 deposit, and it’s as crazy as it sounds

Belated Dalek Day

 Ooh, it’s been a while since I’ve used the “personality quizzes” label on here...

Yes, I am a day behind on Dalek Day, the 59th anniversary of the best television monsters. I should have posted my results for the “Which Dalek are you?” quiz yesterday, but better late than never.

Grad schools should not require students to publish to graduate

Times Higher Education and other news venues is reporting on changes to doctoral programs in India, one of which is not requiring doctoral students to publish work as a requirement to graduate. This change was recommended back in 2019, but only just took effect.

While this is described as maybe the most controversial of several reforms to India’s doctoral programs, only one person is quoted speaking against it.

Dr Mukherjee said that this was the wrong approach. “If we want good quality research, and are worried about the rise of predatory journals that cash in on the students’ need to publish, the concern is the predatory journals,” she said.

Most of the article is about what students have to do to enter doctoral programs, not about what you need to do to complete it.

The news story is a good kick in the pants to me to finish this post, which I have been thinking about and drafting for months. Because India is far from the only place that has been requiring its doctoral students to publish papers to get their degrees.

A while ago, I asked about what graduate programs required their students to publish before they were eligible to graduate. I was surprised that the practice was quite widespread, and by the number of publications some institutions require their students to put out to graduate.

The highest requirement reported was three first-author published papers with two more submitted (five total). Another reported authorship on four papers, of which two have to be first author. But even requiring a paper or two was much more widespread than I expected.

This replies to my Twitter thread were abundant and helpful, so you might want to look through them.

A little later, DrugMonkey ran a poll that indicated a little more than half of programs required a publication to graduate.

To be blunt, I think requiring publications before a degree is awarded is bad, and I think India made the right move.

Publication requirements are extremely unfriendly to students. Neither a student nor supervisor nor university has any control over the peer review and editorial process. So it is entirely possible for a student’s graduation to be delayed through no fault of the student. And there is a real financial cost in many cases for students have to pay tuition and expenses for extra semesters, when all the work is done.

Worse, there is a completely predictable outcome of publication requirements. A publication requirement creates an incentive for fast publication and high acceptance, both of which are most easily achieved by shoddy work and editorial shortcuts.

Researchers found 88% of “approved” journals were problematic. India Express reported  did other investigations that suggested India was a major source of articles for dubious publishers. More than one person has suggested that such requirements are partly responsible for the enormous growth of MDPI and similar publishers.

I am annoyed hearing researchers loudly tell others not to publish in MDPI or Frontiers or what have you, and then get upset that these publishers are growing all the time. But they don’t seem to look at things like graduate publication requirements that create a vast, worldwide market for the services of journal articles where decisions are fast and acceptance is likely.

I completely understand the counter arguments. Training a student to produce professional level academic work is the point of a doctorate. The best way to prove that work is publishable is to publish it. It is harder to argue, “That shouldn’t have been published” than “That won’t be published.” 

I also get that public institutions are always being asked to justify their expenses to prove they are not wasting taxpayer money. Journal articles are a “product” that is easily tracked and counted.

I am all in favour of getting graduate students to publish. This should absolutely be the goal and students should be encouraged – or even pushed (with kindness) – to write up work and get it in the hands of editors. 

But Goodhart’s law comes in action once again. When something becomes used as a measure, it stops being a good measure.

Reference

Patwardhan B, Nagarkar S, Gadre SR, Lakhotia SC, Katoch VM, Moher D. 2018. A critical analysis of the ‘UGC-approved list of journals’. Current Science 114(6): 1299-1303. https://www.jstor.org/stable/26797335

External links

India axes publication goal for PhDs to tackle predatory journals (Registration required)

Not mandatory to publish in journals before final PhD thesis: UGC

No paper, no PhD? India rethinks graduate student policy

New podcast interview on aquarium crayfish

Back in October, I recorded a podcast interview with three University of Ottawa students about my papers about crayfish in the aquarium trade. It was tons of fun to do, and I hope that comes across in the recording!

BEaTS Research Radio's Podcast: Come to their aid! - The crayfish pet trade - Special Episode

https://beatsresearchradio.buzzsprout.com/591520/11837184

Thanks to Julie Pan, Malithi Thanthridge, and Ishika Tripathi for their great questions, and editing our conversation down to a tight 15 minutes.