Brief response to kalla724 on preserving personal identity with vitrification
About a month ago kalla724 posted a number of comments on this post, many of which were highly upvoted.
Synopsis: I) I think that kalla724 is too pessimistic about the practice of cryopreservation to preserve personal identity, because we don’t know what level of synapse/active zone/protein structure is preserved in human brains, and we also don’t know what level is required for personal identity. II) I think kalla724 is wrong about the required detail necessary to simulate a C. elegans. This is testable in the relatively near-term, and the results of that test might yield insight into whose argument in point I is stronger.
I
kalla724′s main argument: it is not possible (p = 10^-22) that cryonics will preserve personal identity, because replacing water with cryoprotectant will cause too much damage to proteins and lipids in the brain.
My view is that kalla724 is too pessimistic. To find a specific example to expand upon this intuition, I searched for “c elegans memory”. I chose one of the first reviews in the results: http://learnmem.cshlp.org/content/17/4/191.full, published in 2010 by Ardiel and Rankin. Here’s their first example:
Rankin et al. (1990) were the first to characterize learning and memory in C. elegans. They studied plasticity of the “tap-withdrawal response” (TWR), a behavior whereby worms swim backward in response to a nonlocalized mechanical stimulus generated by tapping the Petri plate containing the worm. The magnitude of this reversal response is around 1 mm (roughly the length of the worm), but this can change with experience. Repeated administration of the tap results in a decrement of both the amplitude and the frequency of the response
The specific neurons mediating this are known:
Using the circuits described by Chalfie et al. (1985) in conjunction with the neural wiring diagram (White et al. 1986), Wicks and Rankin (1995) identified the mechanosensory cells (ALM, AVM, PLM, and PVD) and interneurons (AVD, AVA, AVB, PVC, and DVA) mediating the TWR.
Through more science, they found that:
the locus of mechanosensory habituation is in a part of the circuit unique to the TWR, i.e., the touch cells and/or the synapses between the touch cells and the interneurons. Now the hunt for the underlying molecular mechanism could begin.
There is some evidence for how the short-term component of the tap-withdrawal response plasticity. This is it:
repeated activation of the touch cells results in autophosphorylation of the SHW-3–MPS-1 complex, thus diminishing K+ flux and prolonging the duration of mechanoreceptor potentials. This would slow the recovery from inactivation of EGL-19 (the L-type calcium channel mediating touch-evoked calcium currents) (Suzuki et al. 2003) and dampen cell excitability
This means that the complexes of proteins, working together, add phosphate groups to themselves as a post-translational modification. Each individual complex functions as a potassium ion channel, so changing its structure can alter the excitability of the cell.
Whether vitrification will preserve this specific post-translational modification is, as far as I know, an open question. The current cryoprotectant solution, M22, is pretty physiologic, which means that it functions similarly to water. But, we don’t have this data.
It’s likely that when the protein complex undergoes autophosphorylation, other changes occur in the cell as well. If this led to changes in the cell’s epigenome, which is very common, and the structure of the epigenome is retained by the cryopreservation, then the cell’s epigenome could allow reverse inference of the state of its ion channels. We also don’t have this data.
The authors also discuss evidence for the long-term component of the tap-withdrawal response plasticity:
the AMPA-type glutamate receptor subunit, GLR-1, was required for long-term habituation—glr-1 loss-of-function mutants habituated but did not retain decremented responses … long-term habituation was associated with a significant reduction in the size, but not the number, of the GLR-1::GFP clusters in the posterior ventral nerve cord
This means that the number and distribution of a well characterized protein at the synapses of cells is highly correlated with the strength of the memory. This is consistent with current paradigms of long-term memory.
Under ideal cryopreservation conditions, synaptic vesicles and receptor distributions are likely retained, even if some of the proteins may be a bit denatured. The data is far from perfect here, either.
It’s also important to stress that this only occurs under ideal conditions. Given the current practice of cryonics, cryoprotectant will not reach many or most areas of the brain. In these cases, there is a large amount of ice damage and the information is much more likely to be irretrievable.
II
kalla724 says:
Uploading a particular C. elegans, so that the simulation reflects learning and experiences of that particular animal? Orders of magnitude more difficult. Might be possible, if we have really good technology and are looking at the living animal.
kalla724′s requirement is that we look at live C. elegans to simulate them. But, the evidence above indicates a good correlation between AMPA receptor distribution and tap-withdrawal reflex. And there is good reason to believe that these features are retained by vitrification under ideal conditions.
So, it seems to me that if you were to emulate a particular C elegans, you could add more receptors (or just up the strength parameter) at those synapses, and thus mimic the plasticity of the tap-withdrawal reflex. Looking at live animals would not be required.
One more note:
Extrapolating results on personal identity from C. elegans to humans is not ideal. If the results are biased in one direction, we should expect more redundancy in mammalian neural systems than there are in nematode ones, because mammals have so many more brain cells.
Edit 6/15: fixed format of quotes.
Edit 6/16: added synopsis to clarify main points.
- 's comment on Brief response to kalla724 on preserving personal identity with vitrification by (17 Jun 2012 21:46 UTC; 12 points)
- 's comment on More Cryonics Probability Estimates by (18 Dec 2012 21:20 UTC; 5 points)
- 's comment on [LINK] blog on cryonics by someone who freezes things in a cell bio lab by (22 Oct 2012 23:45 UTC; 3 points)
- 's comment on Welcome to Less Wrong! (July 2012) by (20 Mar 2013 13:52 UTC; 3 points)
- 's comment on Brief response to kalla724 on preserving personal identity with vitrification by (17 Jun 2012 20:12 UTC; 1 point)
Nematodes have evolved a survival strategy under which they regularly undergo freezing with their own natural cryoprotectant. We can expect that unlike humans, every protein and system has undergone long term selection for cryogenic stability.
This is a good point. However, the post only discusses the substrates of personal identity in C. elegans.
Keep in mind that to freeze and unfreeze C. elegans you need much lower concentrations of the cryoprotectants as the freezing rate is much faster. Can’t generalize this onto freezing your whole head slowly (due to it’s size) with much larger concentrations of the cryoprotectants.
Yes, my takeaway from this is that it might be better to chop up my brain (provided it could be done with minimal damage) and freeze the pieces. The person recovered isn’t going to be exactly me. Will he be as close to me as myself waking the next day? Could we conceive of an empirical measure of such distance, even if we can’t implement one yet?
I think the chopped up brain would be as much you as you after a concussion assuming no bulk damage within slices. The neurons along the cut lines are destroyed, but it is still small percentile of all the neurons, and it’s not localized to one critical region. The neurons express distinct versions of a protein for self avoidance http://biologie.univ-mrs.fr/upload/p243/Dscam.pdf , so it is plausible that you can make sense of which wire goes where after scanning. Whereas the brain where all the strengths were screwed up by cryoprotectants, may likely be completely insane or entirely lack anything that you’d care to preserve.
This is assuming there’s no gotcha like the memory storage being dependent on a dynamic process.
I’ve argued elsewhere that this is the proper way to think about preserving identity. Being me is a property that various things possess to varying degrees, and to the extent to which I value continuing to exist I should value the extent to which things in the future are me.
AIUI, C. elegans has actually been frozen and revived. So if they learn, as you describe, then the experiment to do is obvious.
I am pleased to see that you agree that these ideas are testable.
Why would I not?
So why, in twenty years, has no cryonicist apparently done the experiment? How has this not happened yet?
Two experiments:
a) Teach a bunch of C. elegans the tap-withdrawal reflex, freeze them, thaw them, and see if they still know it. This is what I’m assuming you were referring to. I actually don’t think this is all that useful for testing the preservation abilities of cryonics. C. elegans have vastly different life cycles from humans and the ability to freeze them isn’t that generalizable. See Casio’s comment above: http://lesswrong.com/r/discussion/lw/d4a/brief_response_to_kalla724_on_preserving_personal/6u7e
b) Teach a bunch of C. elegans the tap-withdrawal reflex, freeze them, and do cryo-electron microscopy on them to determine the distribution and density of vesicles and receptors at the key synapses involved. Feed that data as parameters to a simulation and see whether the simulation reflects the tap-withdrawal learning experience of that particular C. elegans, as compared to controls. This will help shed light on whether personal identity is inferable from the kind of structural data that might be preserved by vitrification in human brains.
The experiment I prefer is the second. It will take time.
There is next to no funding for cryonics research. And massive regulations against doing animal-based experiments without government approval.
On the contrary, I think that given Casio’s comment, such an experiment constitutes powerful evidence if it finds that nematodes don’t remember after freezing—evidence for falsifying cryonics.
If it finds that nematodes do remember, then by conservation of expectations of course that’s only a little bit of evidence that cryonics is preserving the necessary information, but it’s still worth doing. (Cryonics costs a lot, so the VoI is high.)
There is an experiment testing something similar to this in rats. They retain their ability to navigate a maze following hypothermia. Andjus, 1956:
Yup. Subsequently for scientists there is a lot of inclination to test this (cost-benefit formula has positive term for disproof because taking money to freeze brains without memories would be unethical) while for pseudoscientists there is a lot of inclination not to test it (cost-benefit formula has negative term for disproof). The agent’s approximate utility can be deduced from the actions taken, especially from those concerning collection of information or generation of hypotheses. For the most part, liars do not even possess evidence they are lying, because they never looked for such evidence; that makes them seem more honest to naive people.
Yes, that first one is the experiment I thought was obvious (I was about to come back and edit my comment to detail this, but you responded first).
I think it would teach us whether freezing and reviving with learning preserved was actually possible or not. This strikes me as important and useful information. That C.elegans has some inbuilt ability to survive freezing would confound it slightly, but I still think it’s a necessary thing to at least look at.
That little? (I can believe it, though.)
Has this experiment, or something like it, even been postulated anywhere in the past 20 years, or is it not as obvious to everyone else as it is to you and me?
I really doubt the scientific exploitation of C.elegans is as hard as that would imply, compared to the numbers of mice and rats killed daily for science.
I agree it would be useful. My wording was less charitable than it should have been. Still, the second test seems more definitive.
True, C. elegans experiments wouldn’t be hard to do.
There are lots of worm people and I don’t know that much about the field. For all I know the experiment has already been done.
You should be a scientist!
As far as I know there are currently three labs in the world researching cryonics.
1) The de Wolf’s and Ben Best, researching at the lab they made, Advanced Neural Sciences. Their budget is tiny, $20,000/yr (pdf). And this seems to be almost all private. But this is the best out there.
2) Joao Pedro de Magalhaes. His lab was just funded by a public fundraiser ($12,000) to do an RNA-sequencing experiment to learn about mechanisms of cryoprotectant toxicity.
3) Brian Wowk and Greg Fahy at 21CM. They invented M22 and have done most of the useful work over the past 10 years. And even their website says,
So, compared to most other fields there is nobody researching this. Which annoys people like Mike Darwin and Ken Hayworth so much. This could work, but we don’t know, and we as a society are hardly trying to find out.
Also, a side note: IMO until someone freezes and unfreezes a mouse (or a vat grown neural network complete with it’s training) and until that can be done on whole brain level, the cryonics should not be approached like medicine but like tissue sample preparation. The goal is to preserve all the information; all the proteins in the cell membranes need to stay where they are; the proteins stuck to micro-tubules ideally should stay where they are, etc. Picture the children toy: magnetic letters on metal board. The letters should largely stay in place. You can’t have any chemicals that would detach anything critical and let it float around. Solvents are highly suspect.
Even then you can still have a big gotcha if some dynamic process is involved in memory.
To quote a discussion of long term memory and the specific synaptic changes that take place from Molecular Repair of the Brain:
What, exactly, might these changes be? Very strong statements are possible in simple “model systems”. Bailey and Chen, for example, identified several specific changes in synaptic structure that encoded learned memories from sea slugs (Aplysia californica) by direct examination of the changed synapse with an electron microscope[36].
“Using horseradish peroxidase (HRP) to label the presynaptic terminals (varicosities) of sensory neurons and serial reconstruction to analyze synaptic contacts, we compared the fine structure of identified sensory neuron synapses in control and behaviorally modified animals. Our results indicate that learning can modulate long-term synaptic effectiveness by altering the number, size, and vesical complement of synaptic active zones.”
“Morphological basis of long-term habituation and sensitization in Aplysia” by Craig H. Bailey and Mary Chen, Science 220, April 1, 1983, pages 91-93
Given plenty of time and some effort on my part, a future generation with the technology and gumption could re-animate me.
Given an infinite amount of time and no effort on my part, a future generation with the technology will re-animate me.
I call this passive cryonics [1]. In an infinitely large and timeless universe, plenty of combinations of things happen. One of them is me waking up again, somewhere, someday, no matter what I do. Even if I try to prevent it.
If I’m mistaken (in my preference) that the universe goes on infinitely beyond my perception, then I guess this is it. Y’all who make it, say nice things about me when I’m gone.
[1] http://ovo127.com/2010/01/27/trevor-blake-passive-cryonics/
You don’t find your subjective self go somewhere else when you get drunk or something of this kind, do you? Why should death be different? Why if you are to be anaesthetised, then suffer some minor brain damage, you don’t expect to be saved by vastness of the universe? If you end up in ‘closest match’, clearly the enormous spatial distance (where all the information within your head is encoded into spatial coordinates) should make those more distant from you. Are you sure they are closest to you when you die, rather than someone alive and well nearby?
I believe that such notions are easily falsifiable and have been already falsified in just about anyone’s subjective history. If I had a subjective history of crazy rare coincidences happening to prevent me from ever bumping my head on anything, or indeed prevent me from any other changes in the brain such as maturation, then yes i’d be a believer in this. Alas, none of that was observed, so no.
This is also known as quantum suicide and/or quantum immortality.
Note that it is dependent on an underlying reality with splitting branches, such as in Many-worlds. If the universe is highly deterministic and only has one time-path, then you may very well die for real. In either case, cryonics is a good hedge bet.
Many worlds don’t make quantum suicide a reasonable position if you care about measure, and you very likely should. See Preference For (Many) Future Worlds.
Should I care about Everett branches in which I do not exist? This seems like a tragedy of the commons scenario—I have a strong interest in having others care about such scenarios, but I see no reason why I myself should.
Note: I don’t think I could ever think like this is day-to-day life, but I’m trying to follow the logic here. Also, is the downvote to my parent in regards to an incorrect statement, or as a social signal against moral relativism?
Right now you should certainly care about branches descended from this one but in which you no longer exist. Allow the assumption “you care at least epsilon about anything except your own physical form”. Whether, at a time where there is an everett branch in which you already no longer exist, you should still care about that branch is more arbitrary. However, what is not arbitrary is that the you now should prefer to self modify to be the kind of person who does care about Everett branches descended from this one in which you no longer exist, since that gives better expected results (from your current perspective).
Makes sense, thanks for articulating it. This is the reason I don’t consider quantum suicide (via cryonics) a reasonable position until a normal end of life situation.
And without using cryonics, quantum suicide has the potential to be a very bad experience. Cryonics is a rather all-or-nothing approach, which is good in that it reduces the very large middle ground where I survive but with 40% brain function. (As noted in Valdimir_Nesov’s link)
I’m not sure where the cryonics/quantum suicide link is coming from. Do you just mean ‘cryonics’ in the sense “use something to make the patient clearly unconscious and then either wake them up or destroy them”? (As opposed to any long-term-storage connotations.)
I’m a fan of that link for some reason.
Yes. Cryonics being preferred over other methods because it’s a viable method for survival in the linear time case as well.
You at least likely should care about how many branches you exist in. Not existing in more branches means that you exist in fewer branches.
It’s interesting this post is being upvoted. It reads like jabber to me. I have little idea what it is trying to argue. Stuff like:
“It’s likely that when the protein complex undergoes autophosphorylation, other changes occur in the cell as well. If this led to changes in the cell’s epigenome, which is very common, and the structure of the epigenome is retained by the cryopreservation, then the cell’s epigenome could allow reverse inference of the state of its ion channels. ”
Is either meaningless or flawed, probably both. The whold post reminds me of the idea on LessWrong that one might as well just assume Omega will reconstruct you based on trace evidence in the physical world.
Post-translational modification of proteins is involved in some types of memory. Sustained post-translational modifications are likely to involve changes in gene expression. Otherwise, the system would not be very robust.
Changes in gene expression are likely to involve changes in the cell’s epigenome.
Even if the post-translational modifications are gone after you replace water with cryoprotectant, the epigenome might be still stable. This would allow you to see what the changes in gene expression were. So, you might be able to tell what the memory was.
This is not a new idea. Ben Best:
You also say:
It could definitely be flawed. We don’t know. That’s why I used two conditionals in the sentence you quoted.
What part of
did you not get?
My main point is that kalla724 is too pessimistic about cryonics. We don’t have a lot of answers to important questions and therefore can’t say with so much confidence (p = 10^-22) either way.
Saw a recent youtube with Aubrey De Grey where he expressed confidence in cryonics, and said a lot of recent progress had been made.
As for p=10^-22, that’s an unserious number. That can only come about analytically, premised on your assumptions. Ask yourself which assumptions, proven false, would overturn your conclusions. Are you confident in them with p>1-10^-22? I say no.
Jaynes had a nice practice in this regard. Always include a “something else I don’t know about” as an hypothesis in hypothesis testing, and assign some reasonable value for your ignorance of how the universe runs. That will keep your calculated confidence in your other hypotheses from reaching absurd levels.
Creationists will claim that it is impossible for evolution to have produced such and such feature in living creatures. Really? You’ve enumerated all the possible things that could happen in the universe, and understand the functioning of the universe so well, that you can rule it out and call it impossible? They don’t just believe in God, they believe in their own godhood.
Do you have a link to the video?
ETA:
I agree. It’s not my number. It’s kalla724′s. It would be difficult for me to assign a precise numerical probability.
I understood it was the other guy’s number.
And right you are about a link for the De Brey video. Lazy of me. Starts at 13:40:
http://www.youtube.com/watch?v=6eet44YacRg&feature=plcp