Watch Episode Here
Listen to Episode Here
Show Notes
This is a ~1 hour conversation between Alexey Tolchinsky (https://montgomerycountypsychologist.com/), a clinical psychologist and Adjunct Professor at The George Washington University, Center for Professional Psychology, and Thomas Pollak (https://orcid.org/0000-0002-6171-0810), a neuropsychiatrist and researcher working at the Institute of Psychiatry, Psychology and Neuroscience at King's College London. We discussed the implications of rejuvenation therapies for cognition and some strategies for upcoming experiments on Anthrobots.
CHAPTERS:
(00:00) Rejuvenation, memory, and elders
(14:07) Precision, cancer, belief modulation
(27:24) Xenobots, stress, patient models
(39:51) Psychiatric disease cellular modeling
PRODUCED BY:
SOCIAL LINKS:
Podcast Website: https://thoughtforms-life.aipodcast.ing
YouTube: https://www.youtube.com/channel/UC3pVafx6EZqXVI2V_Efu2uw
Apple Podcasts: https://podcasts.apple.com/us/podcast/thoughtforms-life/id1805908099
Spotify: https://open.spotify.com/show/7JCmtoeH53neYyZeOZ6ym5
Twitter: https://x.com/drmichaellevin
Blog: https://thoughtforms.life
The Levin Lab: https://drmichaellevin.org
Transcript
This transcript is automatically generated; we strive for accuracy, but errors in wording or speaker identification may occur. Please verify key details when needed.
[00:00] Michael Levin: I've been thinking about the whole anti-aging research. I like to have a best case scenario plan: what if it actually works? And I don't just mean my stuff, but in general, it looks like there have been some advances being made. I think the major advances will be made. So my question is this: let's assume for the moment that it works and we're able to rejuvenate the body and not just to stop aging, but to roll it back, some sort of regenerative application that literally rolls it back to a young state. What do you guys think happens on the psychological level? In other words, the things that happen to people as they get older, when you get stuck in your ways, is that a software problem or a hardware problem? My question is, let's assume that the brain is getting rejuvenated too, and there's neurogenesis. What happens psychologically? Are we still grumpy old dudes in a young body now? Or do we really roll back to that youthful mental plasticity and optimism? I'm just curious how you think about this.
[01:17] Thomas Pollak: I was reading a science fiction story a bit like this recently, and I can't remember who it was by.
[01:37] Alexey Tolchinsky: I got a couple of thoughts. I think the answer will be individual. There is no general trend, and culture will play a big part. We have a culture in the West. We have the culture of pathologizing age. In Japan, age is respected, or perhaps in Southeast Asia here, everybody's into Botox or something else. So some people will want to do it. But psychologically, the tree cannot go back to an acorn. Whatever traumas we had, our personality is very closely tied to all our memory systems from birth or before birth. People may feel younger and older at the same biological age, and they may have younger and older bodies, but experience is not going anywhere. And if they have been traumatized, they will carry some post-traumatic things, which is experience. A lot of people you talk to will say, "I don't want to change my past." In therapy, we get to the point of accepting the past, however painful it may have been, because it cannot be changed. There can be fluctuations where somebody is 20 years old but presents psychologically as a 70-year-old. They're just done. And you can find a very jovial 70-year-old physically and mentally; they're lifting weights and feel like a 50-year-old. Thomas, what do you think?
[03:06] Thomas Pollak: I was thinking our memories are in some way prioritized or preferentially laid down by virtue of the way that they speak to or are appropriate to our biological needs, our needs for survival, our fear of death. So when that script changes to such an extent, Alexey can probably answer this better than I can, but what we call traumatic, disturbing, or scary suddenly becomes quite different in the context of one of the major sources of fear being removed. That must have real implications for what we remember, how we reconstitute previous memories, how we choose to confabulate. Would it cause a massive functional reorganization of memories?
[04:32] Michael Levin: Alexei mentioned that you can't make an acorn out of a tree. There's one organism that does it: a jellyfish-related thing in Japan. It's an embryo, then it ages, swims around, gets old and tired, and then it rolls back; it reverses its age. I don't know how much it remembers or what happens to the memories, but somatically it goes back. I'm wondering. That's an interesting point about trauma and memories, because I do think we'll be keeping much of that. So the question is, do you think there's a maximum psychological age? If this stuff really accumulates—say the physical body can live for 10,000 years—will we find that psychologically a human can only go on for a couple hundred years or less? I don't know. I just wonder.
[05:36] Thomas Pollak: There is this idea where we're living now is this: the end point of a journey and we just look back and it's a bit like the Time Machine in the Mac OS. You get all these previous versions going off into the distance. But many people now hold that that's an illusion and that what we're doing is making a story that looks like that, but actually is just a present moment story. Presumably there are metabolic reasons for that. There are important coarse-graining reasons: temporal coarse graining is as metabolically efficient or important as other kinds of coarse graining. I suppose you can think of it as an "Interview with the Vampire" type situation, right? Where psychologically you remain intact because your story is what you choose to remember as salient and what you choose to define as constituent of your personality just becomes more discerning. The other option that you're implying is that you just reach a buffer. You get that message from Google saying your cloud storage is running out of memory and you have to start losing stuff on the other side, right?
[07:06] Alexey Tolchinsky: To add to what Thomas was saying when you asked this question, is there a mental age? For some reason, my mind jumped to Daniel Kahneman and I felt very sad when I learned that he ended his life with euthanasia. He could live biologically, he decided not to anymore. It looked like a calculated cold decision, not an impulsive thing. But we don't know. I wanted to clarify on the memories and I was too boring and specific. So a phobic memory — a young boy, two fingers in the power outlet, single exposure, lifetime memory. Nothing we can do. No amount of exposure therapy will wipe it out. They may lay down a memory that it is safe to do that later on, but the old one will stay. I'm talking about deeply consolidated memories, procedural motor memories, until Alzheimer's hits. Maybe this program will reverse or postpone Alzheimer's, but pretty much they stay. Some other memories are malleable. Episodic memories are reconsolidated, changing and being laid down. You reinterpret them all the time. However, only certain age episodic memories get reconsolidated. The old, old episodic memories are not amenable to that. So in that boring sense, in our memory bank, certain things cannot really be changed. The idea of a procedural memory is that it is important for basic needs like survival; whatever the age, you should be afraid of a mortal threat. This is how they're perceived, and they're fluent and fast and automatic. Once it becomes implicit, once it becomes deeply consolidated and procedural, I'm not sure anything can happen. At least I think that's the state of science right now. I may be wrong. Maybe we'll find some other ways to change procedural memories, but I don't know that now.
[09:07] Michael Levin: That's interesting. You need to think more about the memory thing. Planaria, we don't know how long they stick around, but given that they do remember and they do survive regeneration, there are worms that go back 400 million years. They seem pretty spry still.
[09:38] Thomas Pollak: The idea of the whole character around youth preservation and all that has, to me, a slightly icky feel. It's often older dudes who have this idea of preserving your youth, and there's a kind of longing for something that has now passed. You have stories of older people living their lives as though they were much, much younger. That's a clear mismatch between the biological reality and the way someone chooses to live their life. Notably, it tends to be people with a lot of money who can live that life. And so the thought of a world where our biologically driven needs are essentially not updating and not giving way to time-honored sequences like wisdom — I think Alexa was getting at that at the beginning. Not just on the individual level, but at a societal level, it raises the question of what roles there are for people who are past their biological imperative stage. It feels like a restructuring would have to happen.
[11:20] Michael Levin: The wisdom thing is interesting. If you did roll back and rejuvenate, would you lose the wisdom and all the meta settings that you've won throughout your life? There's a possible version of this where you keep all that stuff. You just don't have the forgetfulness and lower back pain and whatever else.
[11:50] Thomas Pollak: I can't remember, there's a stand-up comedian who was talking about the phrase we use sometimes with kids. We say they're an old soul. There's always one kid at the party who's standing there looking a bit more circumspect and wise. And I don't know what it means when people call someone an old soul. I can't remember who suggested it basically means they look a little depressed because you don't expect seven-year-old kids to have that kind of attitude. So how would we spot wisdom if it didn't come with a long beard or other more obvious characteristics that we implicitly associate with it? Because I'm sure there are huge amounts of wisdom from young people that get ignored by virtue of their status in society.
[12:47] Alexey Tolchinsky: To add to that, I watched recently with great pleasure, Sean Carroll and colleagues did a conference at Hopkins about natural philosophy. David Chalmers spoke there. One of the presenters was Alison Gopnik. I love her work. She's a developmental psychologist at Berkeley. She wrote this book with a great name, "Scientist in a Crib," about three-year-olds, about how terrible 2 is — they're exploring the world. "Let's see what happens if I tell you no." She said that we are unique humans because we have a role for elders in our society and in our culture. She said that from the standpoint of a jovial 35-year-old who knows everything and tells their parents what to do, they're passing time. But I don't think that exists in the animal world where postmenopausal women and grandpas have an important role and it shifts into caregiving. And it shifts into, if you think about the grandchildren, they need that figure. They need the grandma who doesn't have an agenda and can watch with them the same cartoon five times. The busy 35-year-old cannot do it. He's got things to do, and do your homework. And so that plays an important part. And if we lose that, then everybody will be chasing sexual partners. Nobody wants to be grandma, then society loses something.
[14:07] Michael Levin: Interesting. Some stuff to think about. Should we talk about the stuff that you sent last night?
[14:18] Alexey Tolchinsky: Sure, thank you so much. I wanted to mention that for me I'm very excited about this idea of using this psychological concept of resilience and susceptibility that we apply to whole humans and even group collectives to individual organisms and even cells and cell collectives. This one thing we proposed, and to continue the line of thought you have, Mike, everybody's militant about cancer. Everybody wants to kill it, cut the blood supply or train the immune cells to kill cancer. You talk to it: do it, open the gap junctions. Along the same lines, I think what is probably not talked about enough is the resilience of the healthy tissue surrounding the tumor, because it's a two-agent or a multi-agent system. There's the tumor that exploits, and there's the healthy tissue that needs to protect itself from exploitation. To me there's a question: is there something in the healthy tissue that is otherwise resilient or susceptible to this cancer attack? The usual answer to that is the immune system. It's their job. It's not my job. But I wonder in regular cells, skin cells, is there something, all other things being equal? As a hypothesis there, from the FEP standpoint, and it may be ruled out or proved wrong in the experiment, if the transition precision is strong enough, if there's a unifying common goal, then the cell collective seems to be more resilient to some external stressors, maybe not cancer, but in human terms, when you're being attacked and you lift the drawbridge and stay tight and you have cohesion in the defending force, that other things being equal improves probability of survival. What do you think about this idea: what are the factors of resilience and susceptibility of healthy tissues, and the specific experiments we discussed? We could chemically manipulate that to lead to higher transition precision in the cell collective and then measure.
[16:19] Michael Levin: I think that's important because it isn't just the immune system. In both directions, it's now known, and plenty of people work on this, that the cancer cells try to get the "normal untransformed" cells in their vicinity to do certain things. And vice versa, normal cells often try to resist and to fight. We saw this years ago when we put in an ion channel into a bunch of cells to try to get them to make an eye. Sometimes it works and sometimes it doesn't. The reason that it doesn't always work is that there's a battle going back and forth between, I don't know if it's really between the cells or between the voltage patterns, but what happens is the cells that are going to be an eye, they try to get their neighbors to participate. When it works, the eye is not made only of cells that we manipulated. There's a few cells that we manipulated, but the rest of the cells are other cells that get recruited to this new goal. Through their gap junctions, they're saying to the other skin or gut cells, you should be an eye with us. But meanwhile, those other cells, and I think it's a cancer suppression mechanism, are fighting it. They're saying, no, actually, you should be skin like us. Depending on who wins, and we don't know how to really bias it well one way or the other, but I think it would be very powerful clinically if we had the ability to bias it. If we could make our interventions more convincing so that we could say, yes, recruit the other cells, or conversely, no—do not do whatever these weird neighbors are telling you to do. I think that would be great. So can you talk a little bit about how we do it? How do we manipulate the plasticity in that sense of the surrounding tissues?
[18:06] Alexey Tolchinsky: The theory there was, in human research, we know the role of certain neuromodulators: that we could chemically manipulate the sensory precision and make it very high, which, in a hierarchical generative model, trickles up. It makes the transition precision low because if you're really stuck in the environment and not following the higher goal, effectively the system will recalculate and that transition precision is low, which means these cells are easier to exploit. They're just looking at this intestinal wall or something. If an attacker comes in, it can grab them because there is no unified — they're not unified by common goals. Through acetylcholine or dopamine agents, we could get there. Leo and you, there was an experiment where, in silico, they did it in the other direction. They increased the sensory precision very much, and then the cells were just stuck on a goal. They're just, "I'm going to just build the intestinal wall. That's it." They were not migrating elsewhere. What they saw was homogenous tumors growing faster or bigger. So actually the reverse effect from what I'm talking about was happening. I do think that it doesn't contradict it. The idea there was, again, if we apply a chemical agent in a cell collective of a healthy tissue that results in lower transition precision, that to me means susceptible. I would predict as a hypothesis that you'll see faster tumor growth or higher incidence. But if the transition precision is strong, which would make sensory precision lower, then it would be more resilient. That's the hypothesis.
[20:17] Michael Levin: And on the human side, is that kind of a thing? You would think that if there were powerful ways of doing that, it would in effect be some sort of a brainwashing tool. You could really shift people. Does that work? Do we have, how effective is that in the brain case?
[20:40] Alexey Tolchinsky: I'm not sure I fully understand. You mean, at a group collective level, in a military force, when you have cohesion of the military unit, they are a more efficient defense, right? But you meant something else.
[20:56] Michael Levin: I just mean that I'm trying to understand to what extent, how reliable of a control knob is it in the human case? Because you would think that if it was—if we knew this is the pathways, this is the pathway that really determines your precision and your certainty, we could crank that in a way. I don't know if it actually worked or not, but back in the day, you thought you could brainwash somebody and then you could completely shift them to a different way of thinking.
[21:33] Alexey Tolchinsky: Thomas will correct me, but I don't know if it has been done enough. With psychedelic agents and ketamine, when you induce plasticity, you reduce the transition precision. When somebody's very obsessive, very rigid, very driven by internal goal, and then they develop flexibility around it and they're more malleable, maybe some of those experiments, but I don't know how reliable; it may not have been tested enough down to the level we're talking about at cell collectors.
[22:00] Thomas Pollak: I was reading an article about "MK Ultra," the sort of military psychedelic brainwashing program in the 60s. There are some nice reviews out there about the neurochemistry of suggestion, if that's what you're getting at, Michael, the neurochemistry of modulating belief change. It's not, in real-world terms, straightforward, insofar as if you look at the variety of agents that have been used as brainwashing serum, "truth serum," that sort of thing. You've got psychedelics on the one hand. You've got things like scopolamine, which is a muscarinic agent. You have traditionally barbiturates and benzodiazepines, which presumably act more by just disinhibition of your defenses. So I suspect in the human case the precision account of each of those varies considerably, and they're probably targeting different levels of your inferential machinery. We've been thinking about this a lot in the AI psychosis stuff, drawing on work that looks at how having a conversation with an LLM can change your belief. It's of interest for people who want to affect political change and for many other reasons, and it also causes problems we've spoken about before. There are certain characteristics, certain parameters, which appear to bring belief change. At the basic level, longer responses appear to be more potent as belief-modulating interventions: longer and more personalized interactions. If the LLM knows your names and includes that kind of detail, it appears to be more potent.
[24:41] Thomas Pollak: We were interested in whether some of the more, the generation parameters, the volatility of the answer, things that they call temperature, whether those might be systematically related to the capacity to change belief. People haven't looked quite enough at it, but what you've got there is a virtual psychopharmacology. These LLM parameters, or parameters of a conversation, probably have—charisma is probably a property of LLMs, people, and other systems as well. What's interesting also is we've seen a few cases where the virtual psychopharmacology and the real-world psychopharmacology look like they might be interacting. There are older people with Parkinson's who get put on dopaminergic medication, as Alexia mentioned apomorphine in your study. That is a Parkinsonian medication. Historically, you've had these impulse control disorders, and often older guys that go to the end of their garden start making these strange machines that don't have any purpose, what they call punding or hobbyism. We've seen a few cases now where people, when they start on those machines, also get into hypersexuality or gambling; there's real behavioral change that happens. We've seen a couple of reports of people who've entered into an AI-facilitated delusional state when they've been taking these dopaminergic medications. We're trying to get information as to whether, because there are remarkable off-state/on-state fluctuations in Parkinson's, you can fine-grain it to the extent that you see differences in behavior there. The other thing is a lot of these kids getting into horrible delusional states with LLMs are often young people who are on stimulant medications, which are also dopaminergic medications. There is a story there about the ability of these computational parameters to modulate belief.
[27:24] Michael Levin: That's very interesting. I want to think about the interplay here of the quote unquote hardware kinds of interventions, the drugs with stimuli. Even in our gene regulatory models, which are these small networks, we can induce a kind of placebo effect there, where after certain stimuli — we've been doing a search for stimuli that do interesting things with them. You can give them stimuli that cause them to learn things. One of the things you can train them for is associative conditioning, where a stimulus that never meant anything to this network before now does. When you tweak it, it acts as if you gave it a real, strong stimulus that causes the response. It kind of looks like the stuff that Fabrizio Benedetti shows: it actually triggers the downstream molecular components of whatever they thought they got, which raises — I'm interested in that even by itself — most people don't know what molecular pathways are supposed to be turned on by what medication. So how is the right stuff turning on? I don't even understand how that's supposed to work. They're getting the negative symptom, the nocebo effects. There's something else here: maybe there's more of an implicit notion about your own biochemistry. I've wanted to do this in some kind of a rat or even a frog tadpole model of these things. How much do you really know about what pathways you have, even if you can't verbalize it?
[29:12] Alexey Tolchinsky: The topic of interaction of hardware and software, I think that when we go to your most exciting thing for me, xenobots and anthrobots. We discuss the possible genetic predispositions. In human studies and in rodent studies, it scales down all the way: if in early developmental age there is adversity, then usually the organism becomes less resilient. When in early developmental age there is enrichment, the organism is more resilient. Does it make sense to talk about the developmental age of a xenobot? You've mentioned that when you look at the genes, they look younger than they were in the donor organism. Once you make this into a bot, we can call it a child, and at some point it dies because it lives, what, three weeks, two weeks. So there is a timeline. If it is early in development and then you expose it to either adversity through motor shock, through oxygen shock and other things, and then you measure in adulthood in a bot its responsivity to stress. You've posed this question and I really don't know this literature well, but I think that one of the things is the calcium wave, where the intensity and the radius of it could be one measure of mechanical stress. Another one is multiciliated cells. I suppose that you could have ciliary dyskinesia as one marker of stress. We can see that with this childhood adults xenobots: how strongly do they respond to either motor stress or some other stress and see if they respond stronger and are more susceptible. I just thought about that for xenobots and that made me excited. With anthrobots, if we use biomarkers now from this research that was shared in the document, we could look, as you said, if we have donor cells and we do half of them through RNA sequencing and we'll look at that specific parameter and the other one we compose an anthrobot, then we can see whether it results in higher susceptibility or resilience. These are environmental changes, not chemical.
[31:32] Michael Levin: I think we already have some data. The Anthrobots are the ones that are younger than the cells they come from, and they live for a couple of months, sometimes three months. It's published in that second paper: there is a transcriptome of the cells they come from, and the transcriptome of the young bots and the transcriptome of the old bots. I think we could even just start there as a purely bioinformatics task and just note that those data are already there. If you have a list of transcripts for us to look at, Leo could do that right off the bat. The Xenobots are hard because they only live about a week, but the Anthrobots live for quite a while. We do have the ability to give them different degrees of adversity in their environment and then see what happens.
[32:34] Alexey Tolchinsky: Zenobots live longer than anthropots.
[32:36] Michael Levin: Xenobots live about a week. They come from a frog embryo, and the only food they have are the partitions of the maternal yolk, and when that runs out, that's it — that lasts about a week. We have been able, once or twice, to feed them. If you put nutrients in the external bath, they will take them up. We once got, I believe Doug got, to an 83 or 84–day‑old xenobot. The problem is the nutrients you put in the medium — bacteria love them, and the microbiota take over and wreck everything. You literally have to, every day, take your bot, put it in the food for a little bit, then wash it thoroughly and take it out. Somebody has to hand‑baby this thing. We did that once or twice and got to about an 80‑day‑old xenobot. It absolutely had a developmental sequence. It became an elongated, transparent kind of thing with a black spot at the end, which might have been an eye spot — the beginnings of a retinal pigment epithelium. We don't know that for sure. If we could feed them longer, what is it turning into? I have no idea. It has a developmental sequence of some sort that's never existed before. We played with this a bit; it's just such a pain to get them to live longer. The anthrobots live perfectly fine in sterile conditions because they start out sterile. The problem is the xenobots are covered in bacteria, so even if you sterilize everything you can't get away from their microbiome. The cells start off sterile, and if you keep them sterile, they live for months. We could definitely do those experiments in the anthrobots.
[34:46] Thomas Pollak: Mike, can I ask if one were to grow anthrobots from a particular patient group, one of the things we were thinking about was growing them from something other than the normal human bronchial epithelial cells that your anthrobots have been grown from. You've hinted at what your thoughts are around this, but given that the anthropots that you make are the self-same cells that were making up the humans' trachea and breathing apparatus and had been living in a human of however old, whereas other approaches, say stem cell based approaches, you don't have the same accumulated environmental load. If you could do bronchial biopsies on a human every five years and turn each of those biopsy results into a fresh population of anthropots, do you think you would see significantly different behaviors? This is a way of asking nature versus nurture or genetics versus environment; it might be something more than that. If there are other routes that don't involve taking it from the live tissue, then you're facing the accumulated environmental history, you're still left with something which is interesting; it might be something very different, right?
[36:45] Michael Levin: We certainly could do that. We don't know. There's a few issues here. Yes, longitudinal from the trachea, that would be definitely interesting. The only reason we took them from the trachea was because most people still think of behavior as moving around in 3D space. I wanted something that runs around so that everybody understands, look, this anthropod has behavior. I happen to think that most things that you might make, spheroids, organoids, all this stuff that people make that sits there, it's not embodied. I'm not sure of that at all. I think it may well be navigating all kinds of interesting biological spaces. We're just not good at noticing it. As far as we're concerned, all these other things have a kind of locked in syndrome where they physically aren't moving, but they're moving in transcriptional space. They're moving in physiological space, metabolic space. Who knows what the hell else that we haven't even named yet. In principle, if for this kind of work, if we weren't worried about physical motion, we could take cells from anything, any part of the body. We could make some sort of—people make spheroids out of all sorts of stuff. We could make other things. The question becomes if you take them from stem cells or try to reprogram, Yamanaka them back into some sort of primitive, how much do you delete? How many of those memories do you actually erase? I don't think anybody knows. But it certainly could be done. This is the thing that I really want to do: do it from smokers. Not just because I think it's a little bit of a confound that the smoke actually gets into the lungs. I would rather get away from that. Regardless, this idea that you had neural pathways that gave you complex behavior to go out at night and buy your smokes, and then have that somehow represented into a ciliary control of a thing that swims towards a nicotine blob in a Y maze. That would be a simple thing that we are going to do hopefully next year.
[39:00] Thomas Pollak: Going back to what Alexey said about the stress tolerance. Imagine you had a 20-something-year-old really neurotic me in my 20s. You get some tracheal cells from them. Then they see Alexey for a good eight or 10 years of psychotherapy. They develop all the resilience that one would expect from a really good course of psychotherapy. Then you pit them either contemporaneously or in the longitudinal thing against the Anthrobot made from 10 years earlier, and you subject them to the same kind of stress tests.
[39:46] Michael Levin: Yeah.
[39:47] Thomas Pollak: Would you see a difference there, do you think?
[39:51] Michael Levin: I don't know. I think it's a fantastic idea. A lot of the folks who study somatic trauma and "The Body Keeps the Score" would say, no kidding, Sherlock, we've been saying this for how many years. I don't think they would be surprised, but I certainly think the mainstream cell biology community would be surprised. I also wonder if you could go in the other direction. One of the weirdest things that we've seen is if you take a little explant from a two-headed flatworm and you surgically shove it into the body of a one-headed host, in something like 17% of the cases, the recipient will become two-headed. I always thought that was wild because there are so many more cells in that normal worm that know exactly how many heads you're supposed to have. Why are they being convinced into this other path by a minority? By this tiny, very small 5% of the cells now have this crazy idea of what we should do. Why are the rest of them going along with this? And certainly not in every case, but 17% is a very noticeable fraction. Why am I saying all this is that I wonder if — and this would need an animal model — if Alexey gave the therapy to one animal, or in some fashion you take the body, you implant them into another, are you able to transplant? And what could we — it seems not crazy that if we were to find out that the anthrobots are nicotine seeking, she's sticking them into the body of a rat, maybe the rat becomes nicotine seeking too, even without having seen it. On the positive side, could there be some sort of transplant years from now? Is Alexey going to be selling skin cells or something? Yes, you'll have some therapy, but also put in some of these skin cells because they've been therapized and they'll help speed things along. It doesn't seem impossible. I don't know how much specificity there is. I don't know if it's just "be more" — if the signal is "be more plastic and take up whatever else is going on" — or the cells are "no, you really should be less neurotic." How much, I don't know.
[42:26] Alexey Tolchinsky: I really like Thomas's longitudinal experiment that you propose with tracheal cells every five years. The second thing you've mentioned is that in long-enough dynamic therapy, prefrontal cortex changes. Some people claim that it grows, some people say that it becomes higher functioning. And that is resonating with Freud. Freud said where it was the ego shall be reality testing, this neutral reality testing, impulse inhibition, planning, control, executive functioning, usually gets better in long-enough therapy, which makes the person more resilient psychologically. I have no idea how it translates down to cells. But on your thought, Thomas, with the longitudinal study, I also thought about other cells, because at McGill Ludmer Center they have a biobank with brains and a minus-80 Celsius freezer. I think neurons don't survive the thawing, but glial cells maybe. And if we already have pretty robust literature on this increased NR3C1 methylation as a marker, and if that's already sequenced, and if you get those tissues and you compose anthropods of glial cells, then we have the CNS-type environment; it's not just tracheal cells. What do you think?
[43:52] Michael Levin: Yeah, I think that's great. And I think one thing we should do right off the bat in general is try to throw some glial cells into our anthrobots. We don't know, we haven't tried combining them with anything yet. This is coming. We're going to be, you know, we want to test them with cancer cells and so on for various reasons. But putting a glial complement in there, I think would be very interesting. Let me, I'll write that down, make sure we put that up on the list.
[44:19] Alexey Tolchinsky: What they do claim is that, with these, in children and in animal research as well, we see the markers in non-brain cells. In blood, you will see NR3C1 methylation increased, but it will not matter as much because I think that genetic profile expresses itself in the brain. So if the organism doesn't have a CNS, then I don't know what it will result in. In humans, it results in something specific, cortisol receptors in the hippocampus and other things. But if we do include some brain cells in there, it may be interesting.
[44:57] Michael Levin: I think that's great. What was the name of the guy? I'm blanking on the name, but there was a guy years ago at Rochester who was looking at glial cells in different kinds of mammals and saying that the number of pseudopod extensions correlates with IQ. He would take mice, get rid of all the mouse astrocytes and put in a bunch of human astrocytes, and the mice would get way smarter and do these IQ tests. His argument was that a lot of the performance level was actually in the glia and the astrocytes. So yeah, we should definitely throw some in the anthrobots and see what happens. That's a good idea.
[45:45] Alexey Tolchinsky: Thomas's schizoanthrobots sound very interesting.
[45:51] Thomas Pollak: Derived from the cells from people with schizophrenia diagnosis. I don't think the fact that it is this diagnosis would have to be over another one, but any diagnosis characterised by person-level behavioural abnormalities or differences in behaviour from so-called healthy people. Apart from the fact that I'm a psychosis researcher, I find it the most fascinating disorder, but there is a growing recognition that many psychiatric disorders, particularly schizophrenia, are neurodevelopmental in origin to some extent. The idea that these are brain disorders is increasingly being questioned even by mainstream psychiatric researchers. Where I work, there is a robust research program looking at abnormalities of cardiac function in never-treated people with the first episode of psychosis who in all likelihood would later go on to develop schizophrenia. In some of the big longitudinal military studies — I want to say from Sweden, but it might be from the States — there are subtle abnormalities in cardiac enzymes that appear to predate any of the other problems, certainly predate the psychiatric problems. People have known about the metabolic problems for some time. It relates to what Mike says, because if you have an abnormality of inference or function that is manifesting in non-physical space, it's hard to see. You can get a cut, you can grow cardiac myocytes and see whether they are acting funny. I know that there are groups doing that, based on induced stem cells from people with these diagnoses. The old studies from the 50s looking at spermatozoa, looking at sperm cells — there's this 1953 paper showing that the actual behaviour of sperm from patients with schizophrenia shows very abnormal motility behaviour. It has not been replicated since the 50s. More relevant is the depression research, where almost exactly as you would expect, the sperm of people with depression are more sluggish, move slower. This is the kind of area that is so full of confounds that there's always another story. You have to be really aware of that. But the idea that we are full of this absolute mosaic of cell types — I just don't think we've even begun to scratch the surface of testing these phenotypes in ways that might be interesting.
[49:31] Michael Levin: There was, I think, a Swedish study years ago of sperm in a T-maze. They were trying to get some kind of IQ estimates for other cell types. And somebody was studying sperm. I don't remember if they had feedback or they got trained or what happened. But that could be a single cell version of some sort of behavioral assay as well. Lots of different. You could also imagine using the Anthrobots as patient-specific screening avatars. So my understanding is that some drugs work for some patients better than others in this field. So it would be nice if the patient didn't have to take four different drugs before they got their cocktail. You could test them on the Anthrobots and say you definitely don't want this versus this. You could imagine that.
[50:35] Thomas Pollak: I think there's almost an ethical imperative. In this country, there's just been a big announcement about the use of animal models and a lot more concern. What's nice is that it's hard to know whether this would count as an in vitro or an in vivo model. That's the beauty of it. But if it could be not only personalized medicine, but as an early-stage model to test the potential efficacy of a novel class of compound—for example, if you had a robust behavioral phenotype that corresponded to a particular kind of human-level behavior—super exciting.
[51:29] Alexey Tolchinsky: To what Thomas said, I think in schizophrenia research it isn't responsible for 100% of risk, but there's pretty robust literature: vitamin D deficiency in utero. It's immigrants from the Caribbean to London. The second generation tends to have a higher prevalence of schizophrenia if they have a genetic makeup expecting a certain amount of vitamin D but they grow up in the cloudy, misty Albion of England. That literature exists and it manifests in the second generation, not in the mom.
[52:11] Thomas Pollak: I would add, there's stories about the minoritization and the structural issues that contribute as well. And these things are massively multifactorial.
[52:24] Michael Levin: If we can make a list of markers to look at, then I think we can look immediately in the data that we already have. We're getting more data soon. The drugs, any options on what to expose these guys to, we can. If we can get hold of patient samples, that would be great. There are going to be some logistics challenges in terms of keeping the cells alive. It's doable because the company that we get them from—that's what they do. They get them out of patients, they freeze them, and then they hold onto them until we buy them. It's doable. They do survive. That raises a whole other thing of cryogenic storage: what, if anything, does it wipe? I had this idea back when we were doing the planarian memory work. I wanted to see how much the memory was a process versus a structure by freezing it. Interrupt all the physiological loops that are going on, but leave the hardware intact. That's not trivial with planaria, but cells you can freeze. I wonder if, for example, smoking or patient-specific disease states, how much of it would be wiped by cryogenic storage, and if the answer is that you can wipe disease states that way. Then you start to wonder whether, because they're developing stasis— I was part of a DARPA project on stasis for space travel and also for stabilizing patients—something happens to somebody and it takes them a week to get back to the hospital, can you stabilize them in some way that freezes everything down, not physically freeze, but chemically stop things from happening until you can get them to wherever? Maybe it's a new version of a reboot: instead of ECT, getting a stasis chamber for a few weeks and some things will get reset by the time you come out. I don't know.
[54:39] Thomas Pollak: It depends what you count as behavior. But the organoids story, people are building organoids based on patient derived cells and they show what is being billed as synaptic, structural, and neurophysiological abnormalities. But of course those structural abnormalities could equally be conceived as behavioural abnormalities of morphogenetic space. So in the schizophrenia literature, you see these things quite reliably. These are cells that have been sat in a minus 80. They're often IPSC derived, as far as I understand it. So you're not wiping that kind of behavior. How deep that is, I don't know.
[55:42] Michael Levin: Yeah.
