BI 172 David Glanzman: Memory All The Way Down

David    00:00:03    Cells were around for billions of years before the nervous system came along. And the question is whether non neural organisms have memory. Then I thought, well, what does this mean? You know, it sounds like the synapses aren’t where the memory is. The memory is back in the cell body somewhere, and the synapses are simply a reflection of the memory back in the cell body. We do not have a satisfactory neurobiological explanation for any form of memory in any organism.  

Paul    00:00:51    Good day, you Awesome brain inspired people. I’m Paul. This is brain inspired, and, uh, you’re even more awesome. If you support this podcast on Patreon, where you can access full length episodes, join our Discord community, et cetera. Check that out if you’re feeling generous@brandinspired.co today. David Glanzman, uh, joins me on the podcast. So, David runs his lab at, uh, ucla, where he’s also a distinguished professor. David used to believe what is currently the mainstream view that our memories are stored in our synapses, those connections between our neurons. So the idea is, as we learn, the synaptic connections strengthen and weaken until they’re just right, and that serves to preserve the memory. So that’s been the dominant view in neuroscience for decades, and, uh, is the fundamental principle that underlies basically all of deep learning in ai. But, uh, because of his own and others’ experiments, uh, which David describes in this episode, um, he has come to the conclusion that memory must be stored not at the synapse, but in the nucleus of neurons, likely by some epigenetic mechanism mediated by R n a molecules.  

Paul    00:02:06    Um, if this sounds familiar, I had Randy Galles still on the podcast on episode 126 to discuss similar ideas, and David discusses where he and Randy differ in their thoughts. Um, this episode starts out pretty technical, so, um, because David describes the series of experiments that led him to change his mind, uh, and, and led him to his current view. But after that, we broaden our discussion to a lot of the surrounding issues regarding weather, and if his story about memory is true, and we discuss meta issues like, um, how old discarded ideas in science often find their way back eventually. Well, what it’s like studying non-mainstream topics like David is studying, um, including challenges, trying to get funding for it, <laugh> and so on. So I link to, uh, a bunch of what we talk about in the show notes at brand inspired.co/podcast/ 172. All right. I hope you enjoy our discussion. Here’s David. One of the things that I find most interesting about you is that you are a rare specimen in that you have changed your mind about something scientifically. Um, do you agree with me that that is rare in the, the world of science and  

David    00:03:21    Well, to the extent that I’ve changed my mind? Yeah. I think that’s pretty rare. And I think it’s also, if, usually when people change their minds the way I I’ve done and they’re older, it’s a sign of sil impending senility. Oh,  

Paul    00:03:38    We’ll figure that out. Yeah.  

David    00:03:39    Yeah. So, uh, that gives me pause when I think about that. So, uh, you’re not being, you  

Paul    00:03:46    Know, I, that’s not serious. You’re not being serious, are you?  

David    00:03:49    <laugh>? No. Well,  

Paul    00:03:52    I mean, I’m getting older. Lemme start,  

David    00:03:54    Lemme say that I would, I’m perfect. I would perfectly believe that my colleagues would take that view.  

Paul    00:04:03    Okay.  

David    00:04:04    I, I would be surprised if some of them didn’t take that view.  

Paul    00:04:08    Well, do you have colleagues that, uh, from before you changed your mind that give you a hard time with regard to after you changed  

David    00:04:16    Your mind? Uh, I not, not so much. Uh, I mean, some people just say, um, you know, they just kind of rolled their eyes. So some of my colleagues and the people who’ve known me for the longest time, they, they don’t agree with me, but they take it seriously in the sense because they know me and they know the work I’ve done. And, you know, I’ve done pretty good work in the field of synaptic plasticity and learning memory. And so, you know, I think people have to respect that. So at least they know that I’m coming. It’s not like I don’t know anything about synaptic plasticity, <laugh>, let’s put it that way. I know quite a bit about synaptic plasticity and how that relates to learning and memory in a pretty rigorous way, because I work on a model organism in which you can make very rigorous connections between neural activity, synaptic change, and learning memory. I mean, it is probably the organism that is, um, that gives you the most rigorous ab uh, ability, uh, ability to make the most rigorous connection. So, I mean, I know something about this, and so I think people, people who know me very well, uh, they may not agree with me, but I would say people who, you know, take it seriously.  

Paul    00:05:42    So the organism that you’re talking about is, is one of two organisms that you are studying these days, which is the Sia Yes,  

David    00:05:48    Correct.  

Paul    00:05:48    The C snail. Yeah. And which was one of the, um, an organism in which some of the first plasticity, synaptic plasticity studies were done by Eric Kendell, of course. Uh, and, and others. But, okay. So, um, maybe let’s back up, and I will have said this in the, Eric was, go  

David    00:06:05    Ahead. Eric was my postdoctoral mentor. I mean, I started working on ple his lab. What  

Paul    00:06:11    Does he think? What, what, what, what has he thought? Uh,  

David    00:06:14    It, you know, I, uh, honestly, I don’t talk to Eric much <laugh> anymore. Okay. So, I don’t know what he  

Paul    00:06:20    Thinks, honestly. Alright. Um, okay. Well, let’s back up. I I, I would’ve said this in the, I will have said this in the introduction, but we’ve been talking about you changing your mind. Um, and there’s, there’s a before and after. And, um, do you wanna just summarize, you might be just so sick of summarizing this, but what is it that you changed your mind about?  

David    00:06:38    Um, so, uh, we did an experiment, uh, where we, I, well, to back up. Um, I’ve long been fascinated, uh, by the work of Karine Nader, and I always thought it would be fantastic to be able to look at memory reconsolidation on a synaptic level and to see later on if you induced reconsolidation, would the synapses that, uh, were strengthened or grew, uh, would they retract? I mean, that, that seemed to me the logical, um, that would be the logical outcome of that. And for a long time, I wanted to do that experiment, but I didn’t have the people or the resources to do it. And when I say do that experiment, um, in plegia, you can take the sensory and motor neurons that mediate the defensive withdrawal reflex, which is the reflux that’s been the center of most of the investigations of learning memory and plegia.  

David    00:07:56    You can take those sensory neurons and motor neurons out of the animal, and you can put them into cell cultural. They will reform their synaptic connections. And somewhat remarkably, those synaptic connections in the dish have all of the forms of synaptic plasticity, learning related synaptic plasticity that you see in the animal. So that, that means that you can do studies of synaptic plasticity in the dish and have confidence that what you’re seeing is what’s going on in the animal. And there’ve been a lot of studies that show that, that that’s the case. So the experiment that I wanted to do was induce synaptic memory, reconsolidation of synapses and culture, and then look to see if the, when you blocked memory reconsolidation by treating the cultures with a protein synthesis inhibitor and reactivating the synapses, um, would the synapses that grew during the memory, uh, be erased?  

David    00:09:13    So that’s what I wanted to do. And the way you train a synapse and culture is that you apply five pulses, five minute pulses of serotonin to the sensory motor, synapse and culture. The reason for using serotonin is because due to the work of CandE and his colleagues over, you know, several decades, it’s known that serotonin is the neurotransmitter that’s released during behavioral sensitization. And that is, uh, you know, it plays a major role in sensitization. So, and the reason for using space pulses is because that gives you long term, and by long term, I mean 24 hours or more memory, if you put mass pulses of serotonin, you don’t get memory that lasts, uh, 24 hours. And in this case, the memory is the strengthening, the long-term strengthening of the synapse. And you can measure that electrode with electrodes. You can measure the strength of synapse.  

David    00:10:27    So, um, the experiment we did was we induced long-term the synaptic analog of long-term sensitization with serotonin. The, the synaptic analog is called long-term facilitation. So the synapse becomes facilitated, and that facilitation can last 24 hours or more. It’s in the dish and then in the dish. And then 24 hours later, we gave a single pulse of serotonin to some of the cult cell culture dish. And that was the reminder stimulus that reactivated the synaptic memory. That was the idea. And then in some of the synapses that got the reminder pulse of serotonin, we followed that with, um, application of a protein synthesis inhibitor to the cell culture dish. So that’s anis. So that is the reconsolidation blockade protocol. And what we reported was that when we tested the synapse, 48 hours after the original training, just as Kareem Nader, uh, had shown in rats, and other people had shown the synaptic memory was gone.  

David    00:11:49    So the synapse had reverted to its original, uh, strength. So then we wanted to do the next experiment, which is to actually look at the synapses and see if they disappear when we do the reconsolidation blocking. So we did morphological logical analogs of this experiment. So, uh, we would fill the motor sensory and motor neuron with fluorescent dyes, and then using, uh, you know, just imaging the neurons in culture, we could count the number of synaptic connections. And the way the, the standin for synaptic connection we used were these pre-synaptic varicosities that are on the sensory neurons. Uh, people who work on mammals call ’em pre-synaptic batons. Mm-hmm. <affirmative>, we call ’em varicosities bulges. Don’t ask me. Yeah. Bulges. Don’t ask me why the difference in terminology, <laugh>. I never, I never found out. But anyway, so, so we counted, we imaged the culture, counted the number of varicosities, then we gave the five times, uh, s we gave the five pulses of serotonin.  

David    00:13:07    24 hours later we counted the varicosities. And then again, and then some of the varicosities, we, uh, some of the synapses, we gave the reconsolidation blockade protocol. So we gave a pulse of serotonin, followed by a niso myosin. And then at 48 hours, we recounted the varicosities. And, you know, to my great satisfaction, what we found was that indeed, the number of these pre-synaptic varicosities, which reflect synapses, they strengthened, uh, at the 24 hours after the serotonin treatment. And then for the synapses that got the reconsolidation blockade treatment, they went down to their original value. Now, it’s important to note that they didn’t disappear. You didn’t wipe out all the synaptic connections. You, they just reverted to their original number. So there you have reconsolidation blockade addition. So, you know, usually it’s a good idea to stop when you’re ahead. But we didn’t do that.  

David    00:14:18    So we said, because, because we could, we asked the question, well, do the varicosities that get eliminated, are they the same ones that grew in response to the serotonin ’cause of memories contained at the synapse? You, that’s what you’d predict. So we were able to look and follow individual varicosities in cell culture over time. And what we found was that it was totally random. So, so the vari some of the varicosities that grew in response to the serotonin retracted, but so did some of the original varicosities. And it looked like the synapses were totally uncoated with respect to memory. The only thing that mattered was the number. So I, at first, I didn’t believe this result, frankly. And so I made my research associate repeat, repeat it, you know, just to make sure it was right. And this, these were very tedious experiments. So anyway, she, she did that.  

David    00:15:27    And then I got convinced. So then I thought, well, what does this mean? And I thought, well, you know, it sounds like the synapses aren’t where the memory is. The memory is back in the cell body somewhere, and the synapses are simply a reflection of the memory back in the cell body. So we’d also been doing some experiments because I was very interested in the idea that epigenetic changes are involved in memory. And so we did experiments in which we tested the effects of D n a methylation inhibitors on, again, on long-term facilitation in the dish. But we also looked at long, we did behavioral, um, we did behavioral corollary, behavioral experiments, looking at the effect of inhibition, inhibiting D n a methylation on long-term sensitization. And both sets of studies gave us the same result. If you train an animal and then immediately inject an inhibitor of D n A methylation, the memory goes away and we can’t bring it back. We there if we do, one of the things we found was that if you do reconsolidation blockade and test the animal 24 hours later, it looks like the memory’s gone. But you can bring the memory back with, uh, reduced training, training That by itself doesn’t produce long-term memory. But, uh, but in trained animals in which you induced reconsolidation blockade, this truncated training gives you long-term memory. You  

Paul    00:17:15    Just, it’s kind of a reminder to the organism that this is what you used to do. Yep. Yeah.  

David    00:17:19    And so to the, uh, in putting all these things together, I thought, well, it’s, it looks like what’s going on is the nucleus is remembering, and the synapse is a reflection of that memory. It’s a functional reflection of the memory. So that, that was what I started thinking. And then I thought to myself, well, what could be inducing the memory? And it was known that one of the things that’s is known that, uh, non-coding r n a can, um, induce, uh, epigenetic change. Hmm. So I thought, well, maybe what happens when you train the animal, it induces the ex, it causes the expression of non-coding, r n a, the non-coding r n a, uh, produce epigenetic changes. And that’s the long-term memory. So I was aware from my background in psychology that there were experiments done in the 1960s showing that you could transfer memory from one, uh, animal to another by injecting r n a from a trained animal into a naive animal. Now, those experiments are notorious, uh, and, uh, they basically, if you read the introductory psychology textbooks or neuroscience textbooks, they always talk, talk about this as a wrong term in Okay, behavioral neuroscience,  

Paul    00:19:02    These are aria, is that, uh,  

David    00:19:04    Well, they were done initially in Planaria, but then people did similar experiments in goldfish and mice and rats. Yeah. And the problem was some people reported success and other people reported failure. And eventually the idea died out, and the synaptic idea of memory became predominant. So I decided, okay, well, look, we’re gonna train in plegia. We’re gonna sensitize a whole animal. We’re gonna extract the r n A out of that animal, and we’re gonna inject it into a naive animal and see if they sensitize. So I got really excited about this experiment, and I went to my lab members of my laboratory and told ’em, let’s do this experiment. And they didn’t <laugh>. They were really, why not?  

Paul    00:19:54    What was wrong with  

David    00:19:54    It? Well, they just, they thought it was a waste of time, frankly. Huh. Okay. So, uh, anyway, so eventually though, they came around and, you know, they said, okay, we’ll get rid a try. So they did it, and it worked.  

Paul    00:20:09    Uh, so we published, let, let me just summarize like what that was. Right. So you trained some SIA to do a task, uh, or to, to be sensitized, what is it called? A to  

David    00:20:18    Be sensit? Yeah. The, the, uh, siphon withdrawal reflex. The sip. They had  

Paul    00:20:22    This withdrawal  

David    00:20:23    Re called it. Yeah,  

Paul    00:20:23    Yeah, yeah. Okay. So it’s, it’s trained, it’s sensitized to that. Um, and then you extract, its R n a, the, the trained, uh, conditioned trained organism, sia. Correct. And then you inject that into, um, other, uh, naive animals that have not been trained. And all of a sudden they show the same sensitization.  

David    00:20:43    Well, we tested them 24 hours after the injection.  

Paul    00:20:46    Well, all Yes. And all of a sudden, yeah, sorry. Yeah. Right. 24 hours later, they, they show the same sensitization. Correct. Whereas when you inject, uh, saline or a, a, uh, control basically into, uh, other untrained sia, they do not show that sensitization.  

David    00:21:00    Yeah. One of the controls we use, uh, you know, physiological solutions, one of the controls, but one of our controls was R n A from naive, untrained sia. That was the other control.  

Paul    00:21:12    Right, right. Okay. So, so the take home there is that somehow the r n a that was injected into the untrained naive SIA induced this sensitization behavior. Yes. And then the bigger claim would be that it, it was a memory of that behavior. Yes.  

David    00:21:30    Yes.  

Paul    00:21:31    Okay. So we’ll, we’ll come to it. Like I, I mentioned to you beforehand, I’m all confused about what memory means now. Ah, you know, I see thinking, thinking about these things. Yeah. But we’ll come to that. Okay. But I interrupted you ’cause I just wanted to give sort of a layman summary of, of those experiments.  

David    00:21:44    No, that’s, that’s very, that’s basically correct. Now, one of the things that we also did is that we did a cellular analog of these experiments. So, um, this was done by a, a, a sinus who was a postdoc in my lab, a French sinus named Alexi Beta. And what he did were two different experiments. One, he took sensory neurons, isolated sensory neurons in cell culture, and treated them with either saline, uh, r n a from trained plegia or r n a from untrained Ecclesia. And what he found is when he tested the animals the next day, they were hyper excitable, meaning that the amount of current that you needed to inject into the neurons to trigger an action potential was, um, lower. And also, uh, sensory neurons, it’s a little technical, but sensory neurons, when you give them a long depolarizing current pulse, apia, sensory neurons fire at the start of the pulse.  

David    00:23:00    But then they remain depolarized, but they don’t fire. And in sensitized animals, if you put a depolarizing current pulse, they fire throughout the current pulse. And we found that the sensory neurons that were treated with the trained R N A, but not the control, R N a, showed that same effect when they were, as long as they were depolarized, they fired action potentials. Now. So that was a cellular analog of behavioral sensitization that you see, um, jack burn and others, uh, had shown that you see this in sensitized animal. So then we also look to see if we would get long-term synaptic facilitation using the trained r n a. And there our results were ambiguous. So what happened was some of the synapses showed large changes and other synapses showed no change. And the mean was no change, but the variance was much larger, significantly larger than we saw in the control animals. Mm-hmm. <affirmative>. So we didn’t quite know what to make of that. Um, but we, you know, basically we weren’t able to replicate the synaptic effect of sensitization. But something was going on because, you know, as I said, the variance was, was much larger in the synapses that got the treated R N A.  

Paul    00:24:34    So, and, and, uh, the, I believe the effect was specific to sensory neurons and did not, and was not in motor. That’s correct. Yes. Correct. That’s  

David    00:24:42    Correct.  

Paul    00:24:42    So what does that tell you?  

David    00:24:45    Well, in general, you don’t see motor neurons, uh, increase their excitability with sensitization. Okay. So that’s similar to what you see in the animal.  

Paul    00:24:55    So it’s more like a quicker pipeline to the response via the sensory neurons or something.  

David    00:24:59    Yes, yes. Yeah.  

Paul    00:25:02    Okay. Um, so, uh, let’s see, where do I start? <laugh>, where do I start? So, so I, you know, kind of zooming out then. So you’ve obtained all this evidence that, um, learning and memory that the synapse is not where the action is, that it must be Yeah. Intracellular and to your liking. It’s, it’s within the nucleus, within Yes. Epigenetic, um, d n a methylation changes.  

David    00:25:30    It could also be genetic. In other words, there’s a type of genetic change called retro transposition. And we’re investigating now whether learning in plegia involves, uh, retro trans position. So it actually could be, so retro transposition is when, uh, I mean, if you’re familiar with it, typically it’s, it, I mean, it’s when, uh, a non-coding r n a is expressed and, uh, and then it’s, uh, reverse transcribed by reverse transcript dates, and then inserted back into the D N A. Typically it’s, you know, it’s, oh, it’s inserted into a different place. Often it’s inserted from motor regions of genes. So this is something that, uh, has been, um, documented in the central nervous system. Fred Gage at U C Ss d the, sorry, at the SOC Institute has shown that, uh, miles pups that are maternal deprived, there’s an increase in retro transposition in neurons at the hippocampus. And we think the same thing is happening during learning. And we’re looking at that right now. So you could have two types of changes. You could have epigenetic changes, but you could also have genetic changes, and those could encode the memory.  

Paul    00:26:57    And, uh, just to be clear, we’re talking about long term memories here, correct?  

David    00:27:01    Yes, we’re, yes. Yes. Because that not short-term memory.  

Paul    00:27:04    Yeah, yeah. That, that encoding takes time. And the idea is that yes, it does. You need a more permanent kind of storage for this long-term kind of memory. Correct,  

David    00:27:12    Correct.  

Paul    00:27:12    Uh, um, okay. So, well, I mean, how, just in a very general broad sense, how has this changed your worldview, uh, about learning and memory? I mean, I, I know that there’s the old sort of dogma story about synaptic learning and memory. Yeah. And in fact, when I was, um, uh, when I was interviewing for graduate school, one of, one of the interviewees just sat me down and said, so what do you think is, uh, plasticity important for learning and memory? And of course, you know, like, like any good student, I, well, I hope so. According to the textbooks, you know, <laugh>. Yeah. Um, yeah. Well, you know, um, well, let’s, let’s first talk about how, how it’s kind of just changed your overarching worldview of, of learning and memory and just cognition as a whole. So,  

David    00:27:57    Um, it’s, it’s interesting to think about that this was a view of many prominent neuroscientists back in the sixties. Yeah. So many prominent neuroscientists, not just McConnell. He is the one who’s the best known, but many prominent neuroscientists were very taken with the idea that memory is encoded molecularly as, as a molecule. And, um, that idea didn’t survive. Um, but it’s starting to make a comeback. Well,  

Paul    00:28:42    That, that’s, it didn’t survive because of the plasticity stuff in the seventies, which came to the fore. And then Yes. Arguments, evidence for and against the molecular memory mechanisms. Right? Yeah.  

David    00:28:53    And, but the, I think, uh, there’s a larger problem. There’s a, uh, a scientist neuro, uh, biologist by the name of Gunter Stent who wrote an article on scientific prematurity. And he, he discussed several examples of ideas that had come along in science, but didn’t survive. And they didn’t survive because the scientific groundwork Yeah. Uh, wasn’t there. Yeah. And if you think about this molecular encoding idea, many people suspected that it was r n a, but back in the sixties, nobody knew about non-coding r n a. Yeah. Uh, nobody knew about retro transposition people to the extent that people thought that r a might be inod memory. They thought that memory was encoded by the, uh, base pairs of the R n a, just by analogy with genetic encoding and D N A. And that idea is almost certainly wrong. Uh, so you, you say  

Paul    00:29:55    Now, but who knows,  

David    00:29:56    Right? No, I, I don’t, I don’t, I don’t think that idea is correct. Okay. <laugh>. But, you know, I, you know, it could be Sure. I mean, I, I, I don’t know. ’cause I don’t know what the actual mechanism is. And so if you compare that with the synaptic idea, the synaptic idea had all the elements in place, not only did they have, uh, not only did they have, uh, the sufficient knowledge to envision what a synaptic change would be like, but they also had the tools and the technology to record changes in synaptic strength. So if you compare that to the molecular folks, they didn’t have RNA C. So how could they even, uh, how could they figure it out? So I think it was partly, uh, I think it was scientific. Uh, this is an example of scientific premature, and now we have the tools, and we have some, we have a much richer idea of, of, uh, mo molecules that could encode memory, and particularly R N A, or at least maybe r n a might not encode it, but it’s, it induces the changes that do encode it.  

David    00:31:12    And so I think it’s, the time is right for a reinvestigation of this idea. And people now are reporting, uh, changes in, uh, me inherited people are reporting inherited memory in C elegance and mice. Yeah. And in several cases, those, uh, their r n a is involved in the inherited memory. It’s one of the mechanisms for the inherited memory. So the idea is an organism learns something that induces the expression of non-coding r n a that gets into the blood, and that produces epigenetic changes in, for example, the sperm. And so that’s been documented. And so I think that’s, uh, I think that’s another, um, I think that’s another case where this idea about molecular encoding of memory is coming back. So this is, I, you know, frankly, you asked how this has changed my worldview. I almost never read articles on synaptic plasticity. Oh, I’m total, I’m totally bored with them. And I find myself reading articles on R n A and the effects of r n A on memory and, um, retro transposition. And, and that’s, and all those things.  

Paul    00:32:49    Are you bored with those synaptic plasticity articles or disgusted by them, or there are no use? No,  

David    00:32:54    I, I just, I just think they’re, I just think unlike, I think Randy gal and I disagree on, uh, we agree on on many things, but I think one of the things we disagree with is whether or not synaptic plasticity has any role in memory. Right. And, and he just doesn’t think that that it does. Yeah. And to me, it’s inconceivable that it doesn’t play a role that synaptic plasticity doesn’t play a role in memory. I mean, we can show in aia, I mean, it’s, it was, it was shown by Eric Handel back in the seventies that when animals learned there were synaptic changes, the synaptic changes correlated with the memory. And they were, uh, you know, that it’s just, there’s no way that those are simply, uh, phenomenologically non-functional changes. Mm-hmm. <affirmative>, uh, the di where I differ from most of my colleagues is I don’t believe that that’s the encoding mechanism.  

Paul    00:34:08    Okay. The, uh, okay. So, okay, well, let, let’s, but, but then you have to appreciate that there is a role for syn synaptic plasticity. The curious thing to me is that you’re just, you know, uninterested in the synaptic plasticity story now, because at, at this point, I think I’ve, even, you, you’ve written about this idea, well, you know, how are we gonna connect plasticity then to these intracellular mechanisms? Yes.  

David    00:34:35    Yeah. It’s not, look, you have to <laugh> you have to understand, it’s not that I, it, it’s not that I don’t think they play a role. And I do think that it’s really important to understand, uh, you know, if memory is encoded molecularly as I believe back a, the nucleus, then there has to be a mechanism whereby the nucleus is regulating the synaptic strength. There has to be some mechanism. So that, I’m very interested in that. I’m very interested in. But first we have to figure out what, in order to answer that question, we first have to figure out what the encoding is, because we don’t know right now. Hmm. So it’s, it’s, so that’s what I’m focused on, because that’s why I say synaptic plasticity Bos me, because the people who are doing synaptic plasticity studies, they’re not interested in this, uh, communication between the cell body and the nucleus and the synapse. They’re simply interested in the synaptic change as synaptic change. Yeah.  

Paul    00:35:48    Okay. I’m gonna, I’m gonna, I was gonna ask you this later, but I’m gonna, I’ll ask it now, but, because when we talk about, excuse me, like a synaptic change, let’s say a memory, right? Or, or the encoding of a memory, we think about like one synapse that’s kind of like a, you know. Yeah. And then there must be some, so then the, there’s a pre-synaptic activity. It goes to the posts synapse, and then, then according to your story, according to the r n a or nuclear or genetic story, then somehow there’s an encoding of that association or that activity associated with that memory into the genome. But in reality, um, so this is just one, um, slice of the complexity of our brains, right? Yeah. There are tons of different neurons all connecting, right? Yeah. And of course, we’re just in constant motion throughout the day, learning different ways of moving, et cetera. Yeah. And so presumably then you would have to have, you know, you kind of picture it as like one, one mechanism. All right. I’ll remember that, that cell over there, I’ll remember that, um, axon over there. ’cause I need to Yeah. Yeah. But really what it would have to be is I’ll remember these 10,000, uh, axons, and I’ll have to encode all of that genomic gen in, in my genome or epigenetically or something. I mean, is that, does that bother you? Is that feasible? Is that  

David    00:36:59    <laugh>? Uh, it definitely bothers me. Uh, is it feasible? I don’t think anyone knows. I mean, I think in principle, yes, I think it is feasible. And also the other thing to remember is you don’t, you know, people say, well, each neuron has 10,000 synapses, so the nucleus is gonna have to know 10,000. And, and really the reality is probably more like hundreds.  

Paul    00:37:25    That’s, that’s  

David    00:37:25    A lots. That’s still a lot. Yeah. And so my,  

Paul    00:37:30    I know that that’s not beyond the capacity with, with the Yes. The capacity of computation with, with that’s right. Nucleotides, but  

David    00:37:37    That’s right. No one’s, no one’s shown that yet. Okay. And I think to me, that’s something that’s, that would be, you know, really important to show that, that, uh, what I’d like to do the experiment, but unfortunately I don’t, right now, I don’t have the resource to do that experiment, is to repeat our experiment, but just simply show that two different synapses onto the same neuron can, uh, be altered by, uh, yeah.  

Paul    00:38:07    So kind of push and pull. Yeah.  

David    00:38:08    Correct.  

Paul    00:38:09    Using those mechanisms  

David    00:38:10    And that, and then that would make it feasible. I mean, my, the problem is my colleagues just dismiss this possibility outta him. Mm-hmm. Many of them do. And I take it very seriously because I think that the cell really is the cognitive unit.  

Paul    00:38:31    So people talk  

David    00:38:32    About neural, yeah, yeah. People talk about neural networks, and I, you know, I, I agree there are neural networks, but I actually think the cell, uh, is a, is a cognitive unit.  

Paul    00:38:45    I know plenty of people who would still agree with you despite the story of modern neuroscience. Yeah. Well, okay. So there, there’s, the thing is, there’s so many different levels, right? Of the, of organization and the, the more we discover, let’s say even about the structure of dendritic trees or something, right? Oh, it’s more complex than we thought. Um, yeah. The structure of networks, oh, they’re more complex than we thought. And then the structure, the, the networks within a cell, oh, it’s more complex than we thought. And there are the, all of these levels of organization. Um, and then, okay, so then I’m just gonna immediately jump back to what the hell is a memory, because there, there’s memory in the sense of like a subjective kind of, you know, the, the experience of recalling something, right? Yeah. Let’s say, yeah. But then there’s the storage of memory, there’s the retrieval of memory. The expression of memory. Yes. Yes. And so often when I’m reading work that is talking about these encoding mechanisms on the, um, intracellular level, and they call it a memory, does it need to be called a memory? Or is it, it’s like that bit is part of a memory, perhaps? Or like, where, where do we stop talking? Where do we start talking about a memory? Like at what level? Yeah.  

David    00:39:53    Yeah. That, that’s a very good question. I think that’s a, that’s a, the answer to that is really complicated. Um, let me, let me go, uh, before I get to that, let me go back to this idea of the cells, the cognitive <crosstalk>.  

Paul    00:40:10    Yeah. Okay. You bet.  

David    00:40:11    And, and, and, and, uh,  

Paul    00:40:13    How I, I like this idea, by the way.  

David    00:40:14    Yeah. Okay. So if you think about it, cells were around for billions of years before the nervous system came along. And the question is whether non neural organisms have memory. And people have, uh, at our, at our meeting, there were people who, uh, talked about memory like phenomenon and plants. And, um, Wade Marshall at U C S F is looking at memory in, um, centor, which is a unicellular organism. He’s shown obit situation, which is memory. I mean, you know, uh, I think that, well, that’s another issue we can go into, but it’s memory. Um, and so if single celled organisms, if non neural organisms can have forms of learning that, you know, we, everyone agrees our learning, like, you know, sensitization, uh, habituation and, uh, classical conditioning. Well, when the metazoans came along and nervous systems appeared, did they, they just jettison those ancient mechanisms of memory.  

David    00:41:38    That just seems impossible to me. You know, I just don’t believe that that’s, that’s not the way I think evolution works. And so somehow the, uh, the synaptic mechanisms of memory became integrated with these nons synaptic mechanisms in some way. And the question is, how, to me, that’s the question. How, and if we knew that, then we would know, we would be a long way towards answering this question. You ask is how the nucleus, you know, how does the nucleus know what’s going on at the synapse? Somehow the synaptic, the nervous system, mechanisms of memory, neural mechanisms of memory, if you will, became integrated with the non neural mechanisms. And I think that that’s persisted to this day in us. And so if we knew how that happened, then we could answer a lot of these questions, functional questions about memory. Now, the question you ask about what is memory? That’s, that is a really difficult question. And you’ll get different answers from different people. So some people will only say it’s memory if you can show it, uh, if it’s associative, and you can show it in a mammal <laugh>, and then, and then other people will say, it’s like a crease in a piece of paper is memory, you know? Right. It’s a persistence of, so other people,  

Paul    00:43:14    <crosstalk> molecules in a rock are memory. Yeah.  

David    00:43:16    Right? Right. So to, you know, to me, to answer that question, I guess I would say if it’s, if it’s functionally related to a form of learning, then I would say it’s memory.  

Paul    00:43:38    Well, then you have to ask what learning is also <laugh>.  

David    00:43:41    Well, okay. So with learning, I, I, I’m pretty much of a classicist, so I look at, you know, the standard ideas about learning. So, you know, habitation, sensitization, classical conditioning, canarian conditioning, and then, uh, um, um, uh, other forms of declarative memory, which, you know, that’s, those are quite complicated. Obviously.  

Paul    00:44:07    Well go going back to the idea of the, you know, a single cell as a unit of cognition. Yes. And I, you know, I, I leapt and said, oh, I really like that idea, which I do. Yeah. Um, however, um, I think that’s true, and it’s also not true at a different level. Right. So, um, my whole organism, my brain and whole organism learns to play baseball and can remember how to swing a bat. And my, my, that single cell doesn’t know anything about it. Right? Yes. Yes. So that’s a different level of organization and memory, which, so to con, I don’t know. I, I just, sometimes when I’m reading the literature, we all do this, use words like learning and, and intelligence and things like that. Sure. So, um, that’s a unit of cognition. Right. But, so, uh, am I right, some sort of unit of cognitions and how to reconcile those two things while using the same term memory is,  

David    00:45:00    Well, you have to understand, when you say a single cell doesn’t know how to hit a baseball, that’s true. But you don’t know what role the single cell has in hitting a baseball. And that’s because there’s a lot that your consciousness is not aware of that’s going on. Yeah. And so, you know, these days, I have to say that I think the idea that there is, uh, a person with a unique personality, uh, with coherent behavior called David Glanzman, I think that’s a kind of useful fiction that my brain, uh, tells my body in order to get coherent behavior. Mm-hmm. <affirmative>. So I really think you, you’re basically a colony of billions of little conscious units. Yeah. And somehow that produces coherent behavior. And I would not have said that 10 years ago. So that’s another way in which my worldview has changed. No.  

Paul    00:46:17    Do  

David    00:46:17    You know who Somebody after I set, I made this statement at this meeting, we just had pushing the boundaries, and two people came up to me afterwards and said, you know, that’s what the Buddhists believe, the Buddhists believe the same thing. And I’m going, oh, I didn’t know that. Because, you know, I’ve never, I never paid any attention to religion.  

Paul    00:46:35    Yeah. Everyone who studies consciousness ends up in Buddhism, apparently. Yeah,  

David    00:46:39    I guess so. Yeah. I can’t even if you’re not aware that you’re ending up in Buddhism.  

Paul    00:46:44    Right, right. Well, I mean, you said consciousness, which is, let’s not go off on that tangent, I guess. Yeah, sure. Right. But, uh, uh, do you know William B. Miller is, um, he’s a cell? No, I don’t, but he has the same view of like the, he really elevates the status of the, the cell as an individual unit that has lots of intelligence. Right? Yeah. Um, and I think it’s really hard  

David    00:47:05    To, I’m gonna re I’ll look at his stuff because I’ll send you a, that’s exactly the, yeah, okay. That’s exactly the viewpoint I’m moving towards. Yes.  

Paul    00:47:13    I, I, I just don’t, you know, it’s, I think it’s difficult to talk about because we’re using these single terms and we’re really talking about different levels of organization. Um, yes. So when I say I, so when you say memory, like to me it means like, uh, you know, lots, let’s say it’s, let’s say it’s encoded in the nucleus, right? Yeah. Like the Yeah. The synaptic, um, structures are encoded in the nucleus, the tags Yes. With the ID IDs. Yes. Yeah. Well, the memory is not the thing that’s encoded, that’s like the IDs and, and like the little things that are required to get the mi to, to Yeah. Express the memory. Does that make sense?  

David    00:47:48    Yes. Yes. That does, that makes perfect sense. I mean, in a way, I guess, I mean, you could, couldn’t you say that about any mechanistic change? Yes, yes. Couldn’t you say the synaptic change? You could say exactly the same thing about there’s nothing special from the viewpoint you’re looking at to, because, so the synaptic change, when people say, well, they’re more den, you know, they’re more, yeah. Uh, dendrites and, you know, it’s the same thing. And so the, the memory, then you would have to be an outside observer and say, oh, that’s that epi phenomenon, that summary of all those tiny, uh, functional mechanistic changes. That’s memory. Right? I mean, there wouldn’t be anything. I mean, I’m trying to think, is there anything that you could look at and say, oh, that’s memory, without knowing what the functional, uh, product of those changes were.  

Paul    00:48:49    Yeah. I mean, it’s all behavior. You, you always have to express the behavior.  

David    00:48:53    Yeah, I think so.  

Paul    00:48:54    Well, okay, let’s go down this road then. So you, you’ve made a conscious decision or a conscious decision, Hey, decision you’re trillions of cells made a decision that <laugh>  

David    00:49:04    I would say a professionally, somewhat suicidal decision.  

Paul    00:49:09    Well, that’s another thing I was gonna ask. If your favorite poem is, um, what is it? Um, Robert Frost, 2 2, 2 Paths diverge I the Road Less Traveled, right? Is that the poem? Yes. Yeah. It must be  

David    00:49:20    Your favorite. Unfortunately, if you wanna be a scientist, you have to get funding for your research.  

Paul    00:49:28    How’s that going for you?  

David    00:49:29    That’s going not really. Well, so why, so I, yeah, I know. So I tell this story, I told this story at the meeting. So our paper is, uh, do you know Alt Metric scores of papers? No. Do you know what that is? No. No. Uh oh, I’m surprised. Okay. So now these days, alt metric is a way of keeping, uh, an index of the social attention that your papers are getting. So it scores, tweets, blogs, news articles, and there’s some algorithm and it comes up with a score for you for your paper. Yay. So this is <laugh>.  

Paul    00:50:10    Yeah.  

David    00:50:11    So, yes. And so you will, I want you to cite my paper ’cause that’s gonna boost my whole metrics.  

Paul    00:50:16    Alright. Alright.  

David    00:50:17    But anyway, my AL score for that 18 memory transfer, it, it, it varies on each given day, but for the last five years, more or less, it’s been the number as the number one alt metric score for any paper published in EUR as it’s in the top 1% of papers published in the field. And at that time, and it’s in the top 5% of all scientific papers, period. Okay. Now you, you’d think that that would translate into some kind of funding. One would, okay.  

Paul    00:51:00    Yeah. Well, that, okay. That’s a, that’s a good point. Yeah.  

David    00:51:03    But it absolutely had the opposite effect on my funding. So, oh. So when I published in 2018, my lab had, I had two RO ones and an N S O, and I went from that to zero. Uh, I, I went to zero funding. And so eventually I, and this, I have to give N SS F credit for this. They have given me a couple of grants to look at R N A and memory mm-hmm. <affirmative> and also this memory transfer. And that’s to their credit. Um, ’cause there’s no way I could, it’s clear to me that I could not get this funded by N I H. The problem is that you N S F grants are smaller than N I H grants. So my lab has contracted quite a bit.  

Paul    00:52:00    Oh, oh, I’m sorry to hear that. Yeah. Hmm.  

David    00:52:05    You know,  

Paul    00:52:05    And this, this is also the kind of work that you actually need a lot of, um  

David    00:52:10    Yes,  

Paul    00:52:10    Yes. Person hours, et cetera. Yeah.  

David    00:52:13    Yeah. So, I mean, I’m totally grateful Dennis sf. It’s just that I couldn’t keep my le the lab that I have going with N SS F funding lab.  

Paul    00:52:22    So you’re gonna try to bring your, um, that metric down, that social media metric. I think you should try to bring it down <laugh>, you try to get canceled, get canceled in social media, and that’ll  

David    00:52:34    <laugh>  

Paul    00:52:36    They can’t be, they, they can’t be related though. Right. But there, there must be like some  

David    00:52:40    No, no, no, it’s not, it’s not related. No, no, no. There are plenty of people who have high, whose papers have high alt metric scores that are well funded. Sure. It’s just, it’s just kind of, I just consider it somewhat ironic that a lot of people can consider this paper to be important. And n i h just doesn’t consider it to be important at all. So no, not important enough to put money in it, into it.  

Paul    00:53:09    What are your current thoughts? So I was gonna ask you this later. I’m gonna, I’m gonna ask you now, like then since we’re on the topic, so, so I,  

David    00:53:14    But I wanna say something about this because there’s, um, I don’t know if you are familiar with this, but there was a paper that came out in nature in January and it plotted the ratio of so-called disruptive scientific publications. Okay. These are publications that eventually change the field, and they use a metric, an algorithm to decide whether paper’s disruptive. And typically it’s disruptive if it starts out with a very low citation rate and then gradually gets stronger and stronger. Yeah. Okay. The ratio of that over the last 50 years has declined dramatic  

Paul    00:54:02    The ratio of that to non-disruptive.  

David    00:54:05    Yes. But  

Paul    00:54:06    Is it normalized by just the sheer volume of publications as well? Well,  

David    00:54:10    That’s the point. Okay. So it turns out the number of disruptive papers has more or less remained. The absolute number remained constant in a field, but the ratio has, as, as a percentage of the publication has declined dramatically. Also, the number of new patents has declined dramatically in various fields.  

Paul    00:54:33    One way to interpret that is that there’s too much money being put into science.  

David    00:54:37    Yeah. That’s one way to interpret it. Uh, the other way to interpret it is say, well, look, there’s a fixed number of creative people in the world, and those drive the field and there’s a whole bunch of other people who are doing stuff that’s just really unnecessary Yeah. To the field. And as the fields can remain healthy, as long as you have this fixed number of creative people. So some people have said, well, there’s no problem here. Right. Some, you know, there’s always gonna be new ideas coming along, but a lot of people think they’re, it, it’s potential, it’s really is a problem. And part of this, and of course people, you know, people are trying to figure out what’s the reason for this. But to me, it’s almost certainly the case that some of it is due to the funding mechanisms that are available for science. You can’t do science without money. And the peer review system is a fantastic system. It’s resulted in America becoming the greatest scientific enterprise in the world, but it’s not very good at supporting radically new ideas. You know, in a sense, you can think, well, if you have a radically new idea in the field, you’re not, you don’t have any peers anymore. <laugh> Yeah. You’re outside your peers. And so that’s not peer review is gonna fail at that point.  

Paul    00:56:10    Are you worried that you’re gonna become a starving artist? You  

David    00:56:14    Know, I started out as a star man. That’s  

Paul    00:56:17    What I am right now.  

David    00:56:19    And I may end up as a star man. You know, when I was a young man, I lived in a tenement, literally a tenement apartment on the Lower East side in New York. And I was working as a messenger boy for a, uh, film company. My, I paid $66 a month rent for my apartment, and it was probably wasn’t even worse. 66 a month.  

Paul    00:56:42    I don’t know, maybe the rats thought so  

David    00:56:44    <laugh>. Yeah. Well, the cockroaches did. And I even had bedbugs, so, oh, it’s pretty, but  

Paul    00:56:50    Know. Well, but, but seriously, I mean, does this worry you about, you know, con continuing a lab moving forward, or, well,  

David    00:56:55    It, well, continuing a lab, definitely. But, you know, you, you also have to understand that I’m in a kind of privileged position, you know, I’m a full professor at a major research university and, you know, I have tenure and so at least I’ll be able to, you know, go to work every day. And so, but I do, I, you know, yes. It bothers me a lot. Uh, the frank answer to that is yes, it bothers me a lot. I mean,  

Paul    00:57:24    There, there’s the other issue of just momentum in science, because if you don’t get funding, right, yeah. Things kind of come to a, they slow down. Yeah. And then you kinda lose the thread of what you’re doing and it’s hard to restart. Right.  

David    00:57:35    Well, it’s not only that. I mean, people, one of the, uh, metrics that people use when they’re deciding whether or not to fund you is your rate of publication.  

Paul    00:57:44    Sure. Yeah. Yeah. You gotta start writing a bunch of opinion pieces,  

David    00:57:49    <laugh>. Yeah.  

Paul    00:57:51    Um, yeah. Well, I’m sorry to hear that. Uh, the funding’s not going that do, do you have advice then for, you know, I, I’ve often had people, uh, I remember like Paul Chek saying, you know, ’cause he, he, he’s a cognitive science person, neuroscience person, and he’s kind of going off on this tangent in his career about how following the path, reverse engineering evolution, evolutionary stages to help Uhhuh help us understand cognition and what brain areas we’re doing and how they led to what we’re doing now. Yeah. Et et cetera. Yeah. And he’s super interested in it, but he gave a word of warning, you know, to like, well, you can’t go too far off the beaten path because you need feedback from, uh, your cohorts. Um, and then you might not make it back. Right. Um, but in your case, uh, this is the path. This is the Yeah. And you feel confident about it. Well,  

David    00:58:40    Yeah, I feel confident. And the other thing is that there are people out there who are moving in the same direction. Yeah. It seems, yeah. Uh, Sam Gershman, uh, you know, Randy Gallow, I don’t, uh, you know, he’s doesn’t have a lab, uh, as far as I know, but he is still, you know, he is still very active, um, there, and Colleen Murphy is working on memory transfer. Yeah. Jeanette, uh, inherited memory, but she also showed, she also showed lateral transfer of memory and C elegance, you know, that, that she showed by R n a mm-hmm. <affirmative> <affirmative>. So, um, so there are people moving in this direction. And so that excites me. So I don’t feel like I’m this person who’s shouting in the wilderness. It’s just that I have new colleagues now. So that’s, you know, I, I don’t, you know, if I, I just have different colleagues. Yeah. And, and it’s not as big a number obviously, because there aren’t a lot of people in this moving in this direction, but the idea that there’s learning and single cells, Wade Marshall’s work and people who are looking at learning, and <inaudible> d who’s French, who works on learning and slime mold, she’s shown really quite clearly that slime mold’s habituate. Yeah. So, you know, there are people who are in this area, and to me, that’s really exciting. So what I’m trying to do is, is promote that area as much as possible.  

Paul    01:00:20    It’ll come around again. Right.  

David    01:00:22    It, it always does. Yeah. That’s, yeah, it always does.  

Paul    01:00:26    A moment ago you mentioned how, um, when you’re doing, when, when you’re off on your own doing something or, or when it’s a, when it’s a radical new idea. Right. But, but this is not, and then before that, uh, you mentioned that this, there’s historical precedent for this. Yes. Yes. And that it got re replaced, quote, unquote, by the synaptic plasticity hypothesis. Right? Yes. Why, um, and, and you mentioned that in textbooks, you know, they maybe mentioned that and say, well, this was the wrong idea, this was the wrong path. And I don’t remember, I don’t even remember this when, when I was learning about learning and memory. Yeah. And, and maybe, maybe there was something like that, but there are so many ideas that seem to, um, come back when, like you said, the technology is available for actually studying it in, in a reasonable, intractable manner. Why aren’t ideas like that presented as alternatives in the beginning to students who are, you know, because we’re, we’re just taught dogma most frequently, right? Yeah.  

David    01:01:21    Yeah.  

Paul    01:01:21    So why isn’t it given some sort of, at least in, in the parthenon of ideas and Yeah. And given some historical context that, that the story has not always been just septic plasticity. Right.  

David    01:01:32    Yeah. That’s, uh, I, I think that’s a difficult question to answer. Uh, you know, there’s a, there’s a saying I heard, uh, once that the problem with having to open a mind is that your brains might fall out. Um, good <laugh>, you know, I think scientists are very keyed in to figuring out the truth. And so in order to do that, you have to discard a lot of non ideas that you think are non-truth. And I don’t, I don’t think it’s, I don’t think you can have to open a mind when you’re trying to push a field forward. Mm-hmm. <affirmative>, you know, I, I, um, I, you know, scientists don’t generally say, oh, I think it’s this, but you know, it could be this, or it could be that, or it could be that, you know, that’s not the way scientific papers are written. Right. And to a large extent, that’s not the way textbooks are written. Although there, you know, you have a certain amount of freedom to speculate.  

Paul    01:02:39    Yeah. Depending. Yeah. Yeah. Yeah.  

David    01:02:43    I think that’s, that’s the re uh, you know, I think that’s the, you know, do you know ign alweiss? You know who he was? Have you ever heard him?  

Paul    01:02:53    I believe, uh, uh, uh, jogged my memory if I, maybe not, maybe not. Yeah. So  

David    01:02:58    I, OTSIs was a obstetrician at the, the major hospital in Vienna during the 1850s. And at that time, uh, women died a lot of childbirth of fever infection during childbirth. And somehow some of us came up with the idea that this was caused by put faction that was under the fingernails of surgeons, because they would dissect corpses and then they would go up and deliver babies.  

Paul    01:03:35    I do know who this is. He noticed, I don’t know the name. I know the story. Yeah.  

David    01:03:38    Yeah. So, and he noticed that his fingers smelled badly when he finished dissecting a corpse. So he experimented with solutions that would get rid of that smell. And what he came up with was a weak solution of carbolic acid. So thereafter, whenever he did dissections before he delivered babies, he would wash his hands. Yeah. And he insisted that all the other surgeons on obstetricians on his ward do the same thing. And the rate of death of, uh, women on his maternity ward decreased to almost zero. So some of us presented his findings to the Viennese Medical Society, and they were outraged. Right. Because he, they, he was claiming that they were killing their patients.  

Paul    01:04:36    Yeah. Yeah. So  

David    01:04:37    They basically didn’t renew his contract. He had to go back to his hometown of Budapest, and he continued to, to document the effect of watch, simply washing your hands. Now, the problem was, this was before the germ theory. Yeah. So nobody had a mechanism. All he had was the data. So in order to convince people, he published a monograph, and it was semi hysterical. And he basically accused all the doctors in Europe of being murdered.  

Paul    01:05:12    Well, there’s a poli Yeah. Political aspect to this. Yeah.  

David    01:05:15    So this was not, this didn’t make him popular. And his wife, who was a socialite was very unhappy with him. And she thought he was exhibiting, you know, unstable behavior, which he probably was. And so she tricked him into going to a psychiatric Oh man institute. And when he realized that he was in a psychiatric institute and tried to get out, and he was beaten by the warders, and he died, he was beat effectively beaten to death. And then of course, 20 years later, Liston, uh, pastor Lister pastor came along and they came up with a mechanism for what he had found. But there’s no question that the science that he was Right. Absolutely. Right. So, like, whenever I think about my situation, I always think, well, it could be worse. It could, it could be semi white  

Paul    01:06:06    <laugh>. I know. You could be, the headline would be Dr. Glanzman dies by, uh, dies, trying to escape the asylum after claiming <laugh>, all synaptic plasticity hypothesis advocates are murderers.  

David    01:06:18    Yeah. <laugh> thought murderers. Thought  

Paul    01:06:20    Murderers. Oh, that’s good. Oh, yeah. I hope, I hope it doesn’t come to that. Of course. Um, <laugh>. Yeah. Let, let’s, um, I, you know, I, I still have so much to ask you, I, I want get back, let, let’s get back a little bit to the, the science of it all. Um, yeah. And, and thinking about the scale. Um, right. So, so you, you, uh, on purpose, you chose simple organisms to study. Yeah. Yeah. And this has the, been the dominant, um, organism type simple quote unquote, with, um,  

Paul    01:06:52    You know, brains or nervous systems that it com are comprised of less than billions of neurons, for instance. Yes. Well, you know, thinking about the scale of like, what this would apply to, what it would imply with regards to higher cognitive, uh, organisms, right? Like, if I, I have a buddy who loves Tolstoy. I love Dostoevsky. We slid our arms and became blood brothers, and I still don’t know Tolstoy, right? So what am I doing wrong? Mm-hmm. <affirmative> is, is the question. Like, but seriously with, you know, so, so it’s these behaviors and these organisms that get transferred that which we’re calling memory are, are fairly sensitive or fairly, um, simple behaviors, right. Sensitization, uh, reflexes, et cetera. Yeah. Like where do, do you see this scaling up? Like what is, what is your vision for like, how,  

David    01:07:40    Yeah. Okay. Okay. Yeah. So, so people, you know, people have asked me that, and, and they say, okay, look, I’m willing to buy that. Uh, you know, maybe obituary, sensitization, classical conditioning and operant conditioning can be encoded molecular. Mm-hmm. <affirmative>, I’m willing to buy that, but I’m not willing to buy that. You know, my memory of Shakespeare Shakespeare’s plays, or my memory of the gift I got for my 10th birthday is encoded as a molecule. I’m just not willing to buy that. And I, you know, I think that’s fair enough because I don’t, uh, you know, I can’t, I personally cannot imagine how that, you know, that kind of memory could be encoded molecularly. Having said that, um, I don’t see how it can be ruled out. I mean, just because I can’t imagine, it doesn’t mean that you can rule it out if the capacity, you know, if, if molecules can indeed encode, you know, the complex information, then you, I don’t think you can rule it out.  

David    01:08:53    Now, I, frankly, I personally don’t spend a lot of time thinking about complex cognition. Mm-hmm. <affirmative>, I mean, I, I’ve always been a reductionist in my approach. So I’m not a reductionist in terms of the application of my knowledge. So, um, what I learn about simple organisms, I mean, so I’m very focused on trying to figure out memory and so-called simple organisms. But it’s important, and this is something that people I don’t think in the field sufficiently grasp, is we do not have a satisfactory neurobiological explanation for any form of memory in any organism. Now, think about that. Think about that. Okay. We cannot say how sea elegance habituates. I mean, we knows a lot about it, but we don’t, there’s a lot, we dunno.  

Paul    01:09:59    Famously we know its entire, um, connectome.  

David    01:10:02    Yes. And that was why, that’s why this connectome enterprise, which is getting tons of money and attention murderers, it’s  

Paul    01:10:11    Just kidding. No,  

David    01:10:12    No, no. It’s not to say that it’s not important. I mean, one of the major advantages of plegia was we had a primitive sort of connectome of the nervous system. We didn’t know everything, but we knew enough. It’s, but I think many of the people working on that think that the solutions to cognition and memory are gonna fall out of knowledge of connect. Mm-hmm. <affirmative>. Mm-hmm. <affirmative>. That’s certainly incorrect. That’s certainly incorrect. Um, you know, i, I kind of, you know, this will get me in a lot of trouble, but I, I sometimes think that people who work on connectomes are doing, are really smart people who do that in lieu of coming up with really creative, profound ideas about how nervous systems work that now that’s  

Paul    01:10:57    Gonna get, I’ll leave that alone. That’s  

David    01:10:59    Good. That’s gonna get me in a lot of trouble. But, but, um, so if you realize that then for me, it seems that the driving ambition should be to try to figure out how me a form of memory works in some organism. And once you know that, then you can try to build your way up. So that’s the way in which I’m a reduction. It’s not to say that I think that the mechanism that we find for o ation and c elegance is gonna be, you know, important for all forms of cognition and, you know, declarative memory in humans. I don’t believe that at all. It’s just, uh, if, if you take what I’m saying seriously, then you realize how little we truly understand about the brain.  

Paul    01:11:51    Yeah. It’s a depressing thing.  

David    01:11:52    I mean, if we can’t, if we can’t figure out how learn, if we not, we can figure it out. I, I’m convinced. But if we don’t know how learning works in sea elegance an animal that has 325 neurons in its central nervous system, and we don’t know that, you know, how long is it gonna take before we understand cognition in humans? Just how long, just so you know, just try to figure it out.  

Paul    01:12:16    Yeah, yeah. I’m, I’m preparing a job talk for, uh, tomorrow actually. And, um, it’s, it’s the same old story, right? You, you, you think, oh gosh, we know nothing about the brain still and and cognition. But then, but then you look back at your, I’m preparing this talk and having to go through Yeah. And prepare figures. What, what I, what I’ve done, and I’ve done so much and I still, and, and, and you learn, you learn so much. And then mostly what you learn is how much we don’t know. And I mean, this, it’s just this catch when two, yeah. It keeps going.  

David    01:12:47    When I was in my twenties, I completely bel I was convinced that synaptic plasticity explained learning. And I thought, okay, look, this is not gonna be hard. I’m just gonna figure out everything there is to know about synaptic plasticity, <laugh>, and then I’m gonna understand learning. And I thought I would finish that by the, certainly by the time I was 50.  

Paul    01:13:10    Yeah.  

David    01:13:10    <laugh>.  

Paul    01:13:11    Wait, what was it? Was it, was it nader’s work that, what was it that sparked the change? Uh, the beginning of the, uh,  

David    01:13:17    Nate Kareem thinking about Kareem Nader’s work really, really was the thing that, I mean, it didn’t, premier’s work didn’t make me think that synaptic plasticity wasn’t the answer, but it drove me down the path to doing the experiments that led me to where I’m now.  

Paul    01:13:34    Okay. Um, but offline, before we start talking, I asked if you had any thoughts about artificial intelligence, because the Yeah. Memory mechanisms that you’re interested in are far, far different than Yes. Neural networks. Yes.  

David    01:13:48    And yes, and I think, I think that, um, you know, deep learning is, is going to, is a, a it’s not, it’s not a bad way to teach machines. And obviously they will learn and they will learn pretty interesting things, but, uh, that’s not the way the brain learns. Okay. I’m convinced of that. And that’s not the way the brain encodes memory. I mean, I always say, if you could show me a computer that’s as cognitively sophisticated as my snail, I’ll be impressed. And so far, there is no computer that’s as sophisticated cognitively as my snail. In other words, if you put a deep learning computer out in the ocean and it had, you know, let’s say you kept it from rusting and it had to do deal with all the things, a snail deals with a marine, it couldn’t survive. It could not survive. So, so until we understand the real mechanisms of encoding memory, the we’re not, you know, people, everybody gets all excited about, you know, chat G p T and a generative ai. And I, I just say, look, this is not the way humans learn. This is not the way they remember. If this is not the way they remember, it’s unlikely it’s gonna replace us anytime soon.  

Paul    01:15:18    Yeah. Well, there’s the replacing aspect, right? Which I, it’s just silly to me. But, but it is interesting to see how many people are worried about it. But then there’s the, what is it, you know, it is impressive as a tool and, uh, um, yes. It’s, so what does that say’s, what does that say about our, about our cogniti? Or what does it say about what we deem to be cognitively awesome like chess and like language? I mean, does that re, does that reef configure how we should be thinking about, uh, intelligence and cognition?  

David    01:15:50    I honestly don’t think so, because I don’t think that that’s the way we do it. I mean, I would rather than, look there, there’s a, a functional value to generative AI that that’s undeniable. I mean, it’s gonna be able to increase our capacities enormously, but if you really wanna figure out how humans think and remember, this is not how they think. And remember, this is just not how the, in my opinion, this is not how they do it. It’s a very artificial system, and the kinds of things it can do are increasingly sophisticated, but they’re still fairly artificial. Mm-hmm. <affirmative>, in my opinion. Um, I don’t, you’re, you probably are too young, but in the eighties, I think, or maybe nineties, uh, when Deep blue beat, um, I think it  

Paul    01:16:51    Was 97, right? Maybe. Yeah.  

David    01:16:53    Okay. So yeah. Beat Gary Casper off on in chess, you know, everybody said, okay, well now computers are smarter than people. Yeah. Okay. So then, uh, I think it was Nike came out with an ad where they, a series of ads where they had deep blue in a gym playing against David Robinson.  

Paul    01:17:13    Oh, nice.  

David    01:17:14    And, and David Robinson was dunking over deep blue and he was stealing the ball. You know, deep blue is just a box in gym.  

Paul    01:17:23    Your kids out there, David Robinson was one of the best basketball players in that area, and the deep blue area. Yes. You know, professional basketball  

David    01:17:29    In the deep blue, the admiral. So then they, they interviewed David Robinson, you know, after he, after he played the, what do you think? And he goes, well, I think he’s gotta work on his ball control. You  

Paul    01:17:41    Know? Yeah. That’s, that’s good. So,  

David    01:17:43    So I, that’s what I think of what I think of, of generative ai. Yeah. It can do things and increasingly impressive things that are not so rigid. You know, the, the, the formal demands are not as rigid as chess. They’re much more flexible mm-hmm. <affirmative>. And so it’s getting more impressive, but, you know, I just simply don’t believe that that’s how organisms learn.  

Paul    01:18:13    Well, alright, so, and remember, so, so you turn one of those off and you can, everything stays, um, unchanged, right? The weights stay unchanged and everything’s stored, quote unquote in the synapses. You turn it on a year later and it performs the exact same. Is the, is it just a matter of, um, adding some sort of nucleus to the units, you know, some sort of internal storage mechanism that tags the connection strengths and possibly, you know, and then you can, then you can sever things and it’d be okay, you know? Yeah.  

David    01:18:39    I, I guess that is true. But then you, then I think you have to get to the big bigger question. The one that you, um, alluded to is how do you get all these independent units to cooperate mm-hmm.  

Paul    01:18:55    <affirmative> mm-hmm.  

David    01:18:56    <affirmative>, how do they cooperate? How do they know? All right? Now you might think, oh, well that’s easy. They’re connected to each other, so it’s synapse. But it turns out that neurons have ways of communicating with each other. Nons synaptically, so I don’t know if you’re aware of the recent, uh, findings of Jason Shepherd and Vivian Budnick, but they looked at a well-known synaptic protein called arc. And ARC has been known to be involved in synaptic plasticity for 20 years, and it regulates AMPA receptor, uh, trafficking mm-hmm. <affirmative>, and it maintains, uh, synaptic strengthening. That’s one of the things it does. Well, what Jason and Vivian Budnick found is that the ARC protein ha is a capsid just like, um, just like h i v,  

Paul    01:19:58    Like a virus. Yeah.  

David    01:19:58    And it contain like a virus, and it contains r N A and it shuttled between neurons, nons, synaptically.  

Paul    01:20:10    Is it, but is is it at the synapse or is it actually extra synaptically? It’s from nons synaptic sites.  

David    01:20:16    It’s, um, that, I don’t know, it’s, it, it is at the synapse. Okay. But it can, I’m pretty sure it can be released at nons synaptic  

Paul    01:20:26    Sites. It seems like a, a, um, efficient transfer would be at the synapse. So  

David    01:20:30    Yeah. But it, but it, but it, but it’s not, it’s taken up po it’s taken up by the receiving cell. Yeah, yeah. By endocytosis. So it’s  

Paul    01:20:40    Released Yeah. It doesn’t have anything to do with the action potential is what you’re saying. Yeah. Right. Yeah.  

David    01:20:43    No, it, it does have something, the release does have something to do with action potentials, whether or not it can be released without action potentials. I don’t know if that’s clear.  

Paul    01:20:53    Yeah, I I meant like the synaptic strengthening and weakening at the, like the pre and post-synaptic mechanisms Right. That are purported to strengthen and weaken. Is that what you’re Yeah, it’s, it’s more like just a, here, here’s some information. Go. And the idea is like, yes,  

David    01:21:07    Yes.  

Paul    01:21:08    It’s, it’s just a messenger kind of.  

David    01:21:10    Yeah. Okay. That’s true. But it’s R N a. Yeah.  

Paul    01:21:13    Right, right.  

David    01:21:14    Think about that. So it’s R n a and what if that R N A gets transported back to the nucleus and how many of these proteins does the nervous system have? And if ARC is released extracellularly, then that means in principle it can go out into the extracellular medium and into the blood. Mm-hmm. <affirmative>. So now you have activity dependent release of R N A that’s getting outside the neuron itself and out in principle outside the brain. So what’s the function of that? What’s the cognitive function of that  

Paul    01:21:59    Open question? Okay.  

David    01:22:01    Yeah. So, but going back to single cells, if single cells existed at one point, why isn’t, why aren’t, why wouldn’t r n a be a mechanism of communication among, among single cells? And why couldn’t that be used to coordinate their behavior? And the answer is there’s no reason why it couldn’t. And it probably is, and it may well be the primitive before there were neural networks. There may be functional networks of individual cells whose behavior is coordinated by release of R N A and May, and probably other molecules as well. Yeah.  

Paul    01:22:48    Alright. Sticking with the single cell for a moment. Um, so the, the classic textbook story is D n A en codes life, right? D n A encodes, r n a encodes proteins. Yeah. And that’s, that’s the central dogma, which we know is there are lots of different, um, right. Ways that can go. But, but my question is, um, so still the central story and, and evolution as well is that d n a like controls, um, who we are and controls the cell. Right. Um, but I, I think a really ni i, I love these inversions, these thought experiments that are inversions. Um, one could say that it could inverse that and say it’s the cell that’s actually controlling, uh, the cell as a cognitive unit, let’s say that’s, that’s using the D N A, that’s controlling the D N A. Is that a, is that a healthy way to look at it from your vantage point? So,  

David    01:23:34    You know, this goes back to, um, evolution. So, you know, um, some of the, a couple of speakers at the, again, this pushing the boundaries meeting talked about the interaction of the environment on D N A and how the, um, the changes in D N A weren’t random. Weren’t really random. Okay. They were guided by some of these mechanisms.  

Paul    01:24:05    By environmental mechanisms.  

David    01:24:07    Yes. By environmental mechanisms. And so that suggests that the, the the, it’s actually more interactive. So the D N A isn’t this master switch, the D N A is acted upon by other factors. Yeah. And so I think that’s a more realistic view of  

Paul    01:24:30    The kinda a circular causality, um,  

David    01:24:33    Yeah. Right. Exactly. Yes. Yes.  

Paul    01:24:36    David, in, in the last few minutes, I, I just have a couple extra questions for you if you have time for them.  

David    01:24:41    Sure.  

Paul    01:24:42    We, we’ve talked about your struggles walking this lonely, solitary,  

David    01:24:47    Uh, it’s not so lonely or solitary.  

Paul    01:24:49    Well, okay. Alright.  

David    01:24:50    Yeah. I mean, it’s, it’s not exactly’s.  

Paul    01:24:53    Okay. Instead of 10,000 friends, you have 50 in the, something like that, right?  

David    01:24:57    That’s right. That’s right. That’s, that’s supposed  

Paul    01:25:00    What, what are some of the benefits though? Um, you know, do you, are you, do you full, well, I don’t wanna put words in your mouth. So like, what, what, what is it that you gain from, from  

David    01:25:09    Doing that? So the, the benefit is I’m meeting a whole new group of people and I’m very excited about their ideas. And this, um, you know, this keeps my juices flowing. You know, I really  

Paul    01:25:20    Is is fun. Is it fun to go against the grain also, or, or is that part of  

David    01:25:24    Fun? Well, that’s, for me, that’s, you know, this is something I have to discuss with my therapist, but, you know, I’ve, it’s always been difficult for me to, to take orders from people <laugh>. Yeah, me too. And so there’s, there’s probably something in me that is, you know, there’s some people who, who, you know, they just want to know where everybody else is going and they wanna do the same thing and, but just maybe be at the front of the line, but they want to, you know, they wanna keep doing the same thing.  

Paul    01:25:52    I don’t like  

David    01:25:53    Lines, you know, I’ve, yeah, I’ve never been that kind of person. So there probably is something, you know, I’m almost certain there’s something about my personality that would mean that I’m the kind of person who would go this way. Um, and also, you know, it takes a certain amount of self-confidence, which, you know, I’m, I have that as well. So that’s, uh, but you know, I just, you know, that’s set aside the funding issue because for me that is a serious issue because, you know, I’m a laboratory scientist, so I’d like to have a lab, but I just find I’m really grateful that, that I, that I am going in this direction. Because for me, it’s, it’s endlessly exciting and I’m learning stuff that I didn’t know. I mean, I, I, you know, this is kind of embarrassing, but I never took molecular biology in college because, you know, I was a psychology major and I thought, you know, what are molecules gonna tell you about thought, you know, I, I’m gonna study psychology, you know, and so everything comes back to bite you in the end. But, you know, I’m learning all this really interesting stuff and, um, I, I, it, for me, it’s really satisfying.  

Paul    01:27:15    Speaking of grants, when yours gets funded, what, what’s your vision for the future? What’s, what are the next steps for you once you, uh, can go <crosstalk>?  

David    01:27:21    Well, I think it’s pretty, it, it, it’s pretty obvious. I mean, what I’m gonna try to do is figure out the role of r n A in learning memory. And I wanna, I wanna figure out, is it, uh, you know, is, does it, is memory encoded epigenetically? Is it encoded also partly genetically, you know, what role does r n A play in the, in nuclear encoding of memory? And how does the changes in the nucleus, how, how did that, how does that structure the synaptic changes and keep them maintained? And for me, that’s what I’m interested in. Or re and I’m also  

Paul    01:28:07    Right when they’re not maintained  

David    01:28:08    Or reconstitute them, when they get destroyed, I mean, you know, we showed that, that when we did re when we had reconsolidation blockade and we eliminated the memory, supposedly we were able to bring it back with this truncated training. Well, the synapses came back too. So that means that the, that the me, you know, the synaptic structure can be reconstituted. And by the way, that would give me hope if I were working on Alzheimer’s. Because, you know, in principle, if you understood that process, you could reconstruct the synapses that are destroyed during memory, uh, during Alzheimer’s and re retrieve the memory, get the memory back. Now, of course, when the cell dies, that’s it. When the neurons die, that’s the end of the story. But certainly for years, the major, I think the major problem people have in, uh, with memory is that the synapse, there’s synaptic loss, but if you could reconstruct the, the appropriate synapses, you should be able to get the memory back. So the memory should still be there. So for me, that’s part of the future.  

Paul    01:29:21    Alright, David, I’m sorry to hear that you’re going senile, but I’ve really enjoyed my time here with, with you. I’m, I’m joking. Sorry I had to throw it in there.  

David    01:29:30    No, no, no, that’s, that’s totally fine.  

Paul    01:29:33    No, I really appreciate it. It’s a really fun conversation and I wish you the best in funding. And of course, uh, keep up the good work. Thank you.  

David    01:29:40    Thank you very much. And I’ve enjoyed this conversation immensely as well. Thank you for your questions. A really thoughtful, very thoughtful question.  

Paul    01:30:04    I alone produce brain inspired. If you value this podcast, consider supporting it through Patreon to access full versions of all the episodes and to join our Discord community. Or if you wanna learn more about the intersection of neuroscience and ai, consider signing up for my online course, neuro ai, the quest to explain intelligence. Go to brain inspired.co. To learn more, to get in touch with me, emailPaul@braininspired.co. You’re hearing music by the new year. Find them@thenewyear.net. Thank you. Thank you for your support. See you next time.