Brain Inspired
Brain Inspired
BI 135 Elena Galea: The Stars of the Brain

Support the show to get full episodes and join the Discord community.

Check out my free video series about what’s missing in AI and Neuroscience

Brains are often conceived as consisting of neurons and “everything else.” As Elena discusses, the “everything else,” including glial cells and in particular astrocytes, have largely been ignored in neuroscience. That’s partly because the fast action potentials of neurons have been assumed to underlie computations in the brain, and because technology only recently afforded closer scrutiny of astrocyte activity. Now that we can record calcium signaling in astrocytes, it’s possible to relate how astrocyte signaling with each other and with neurons may complement the cognitive roles once thought the sole domain of neurons. Although the computational role of astrocytes remains unclear, it is clear that astrocytes interact with neurons and neural circuits in dynamic and interesting ways. We talk about the historical story of astrocytes, the emerging modern story, and Elena shares her views on the path forward to understand astrocyte function in cognition, disease, homeostasis, and – Elena’s favorite current hypothesis – their integrative role in negative feedback control.


Elena    00:00:04    The problem of the Astro field is try to prove that Astros are like errs. But now I think the time has arrived to think that Astros are not like errs. The conceptual breakthrough was the, um, understanding that those calciums responses in astrocytes were equivalent to action potential in neurons in the sense that astrocytes respond very precisely to neuro information by way of calcium and transcend The term KLIA has protected the glia fill and has helped the glia researcher to get together, to organize their own meetings and their own journals that has worked for a while. I mean, that has helped us produce data, but now it’s a highly damaging term.  

Speaker 3    00:01:00    This is brain inspired.  

Paul    00:01:14    That was the voice of Elena gal, a research professor running her lab at university TA de Barcelona. Hi everyone. I’m Paul. So when I say brain, if you’re like me and the majority of neuroscientists, you automatically think of neurons after all, since the early days of being able to record action, potentials from neurons, a ton of research has focused on how patterns of action potentials in single neurons and in populations of neurons relate to our cognitive functions and behaviors. And of course the deep learning models in modern AI are based on the idea that we can build intelligence by performing distributed computations among neuron like units, but neurons. Aren’t the only game in town in the brain. For example, there are many different types of glial cells, which is basically the name given to the plethora of cells that are not neurons in the brain.  

Paul    00:02:05    Our understanding of glial cell function lags that of neuron function, because as you’ll hear Elena say, the technology for recording certain types of glial cell activity, didn’t show up until 40 years after we were already recording neurons. One of those types of glial cells are astrocytes. I say one, but there are, uh, a handful of different types of astrocytes. And what we have understood about astrocytes is that they help metabolism and they clear out chemicals that are important for signaling between neurons and other housekeeping, like roles as Elena calls it. And while that remains true, a lot of researchers are starting to ask whether astrocytes might also play a role in the computations underlying our cognition, and now that we can record how they signal amongst themselves and to, and from neurons, we’re beginning to appreciate how they may also shape our cognition more deeply than traditionally thought.  

Paul    00:03:00    So that’s what we discuss in this episode, uh, Elena elaborates on the old story about astrocytes the new emerging story, Elena’s current favorite hypothesis that astrocytes play a key role in feedback control, sort of along the lines that Henry Yang talked about as an important brain principle in episode 119. And we talk about the future of astrocyte research, something Elena is calling computational Astro science. We talk about astrocytes emerging appreciation in disease states like Alzheimer’s and how we should stop trying to make astrocytes seem more like neurons and let them just be themselves. One last thing here before we start after we recorded Elena reflected a little bit and sent me this audio clip about how she thinks if, if you’re interested in astrocytes and their role in cognition, how you might go about studying them.  

Elena    00:03:51    If someone wants to enter the astrocyte field to study the role of exercise and cognition, I would recommend this person to avoid the clear field, to avoid clear journals, glia meetings, uh, even glia colleagues and join a computational neuroscience group where he or she can grow as a computational as scientist  

Paul    00:04:16    Show notes are at brain 135. Thank you for listening. Enjoy. So Elena, welcome. <laugh> thanks for finally coming on it. We’ve had to go back and forth a little bit to get you on the podcast, but I’m glad you’re here now.  

Elena    00:04:33    Yes. Well, thank you very much for having me on the possibility to share my signs with, with people.  

Paul    00:04:39    So neuroscience, the, uh, study of the brain, um, neuro means neuron, right? Because that’s, that’s really the only important thing in brains IST it  

Elena    00:04:50    <laugh>. Yes.  

Paul    00:04:52    So, but, but seriously neuroscience has been and well continues to be, uh, very neuro centric. It is through, you know, it is the seed of consciousness through all the computational aspects and glia, well, among all the glia, which means traditionally, which means etymologically glue, um, have been an afterthought. Right. Um, and, and we’re gonna talk specifically mostly about astrocytes here, but I, I just wanted say glia yes. In general, because astrocytes are not the only glial cells. So can you describe what the, um, traditional old story is about? Astrocytes and, and you can throw in other glia if, if you want and their functions within the brain,  

Elena    00:05:35    Right? The, the traditional view of the astrocytes beyond being glue, which I think, uh, even people start the articles with the notion that they used to be called glue, and they’re not, uh, glue any longer clear cells in general, but the traditional view and still true of as sizes that they perform housekeeping in the brain, unlike neurons, which are the cells that ENCO basically, and they are responsible for higher brain functions. Now Aroy carry basic functions, such as ch Clarine molecules by neurons, upon activation, such as depression and glutamate ACY participate in the re removal during sleep. This is a, a very important function that was discovered, uh, five years ago, aides also control the flax of waste molecules out of the brain. And aides are well adapted morphologically to the function of molecule removal, because they are like sponges that surround neurons and all their neuro rights.  

Elena    00:06:40    And, um, I want to emphasize architecture. I will mention architecture during this conversation because the, uh, name of ASTO, that means stars. It came from immuno stain. VACY with G F P um, this is a cytoskeleton protein, and it it’s like the human cytoskeleton for cells. And you only see like 10, 20% of the cell, and it has a star shape with the little body and then some processes, but the Astros are not like that. That’s only the skeleton Astro are bushy around. They’re very big. And they, their distal processes are highly divided, uh, to adapt to all the neuro to the neuroal elements, spine Saxon, somebodies. So this is architecture and the functions that the traditional view in the last, uh, say 50 years is housekeeping. And this is true. So they perform housekeeping and housekeeping is not a minor function. We have to say if they only did housekeeping, that’s okay. Now the question is whether take to something else, do it to computation. And that’s the new story. So if in addition to perform invasive functions, the Astro sites are, the aide are part of neural circuits. And whether in neural circuits, they shape the processing of information among neurons beyond elimination. And, and that’s the question we say in our jargon, whether ass are building blocks of neural circus, that’s the news story about ASNs?  

Paul    00:08:34    Well, so part, well building blocks of neural circuitry sounds like architecture, but  

Elena    00:08:41    It is architecture. Yes.  

Paul    00:08:42    Okay. Well, we’ll get into some of their computational roles also later because they’re also being appreciated more and more as being part of the computations underlying our cognition, right?  

Elena    00:08:55    No, not, I mean, not for now, it’s, it’s still a hypothesis. So while right. No returns, this is what we are really working on. So the, I, I can give a little bit of historical background as to how we are, how we came to think that exercise could be mobilizing and, and, and, and regulating cognition in addition to performing anesthetic functions. So we probably, you know, and the listeners know that any discovery is based upon new technologies and new conceptual developments. And the conceptual breakthrough in the nineteens was called imaging, which is the capacity to track calcium activity, NA with a microscope using flu and pros. That was in 1991. I actually, I started to do my post in, in New York at a time. So I lived through that very exciting PR uh, times.  

Paul    00:10:00    Were, were you already, sorry, I, I wanted to ask you about how you got interested in these things in the first place. So  

Elena    00:10:05    Accidentally, accidentally. Cause I, I went to New York city to do a postoc on vascular cells in, um, and stroke and revelation of, of vascular dynamics in, in, in hypoxia and, and in different traumas and someone in the lab, I was in the, in the department of neurobiology, in biology, in Cornell university. The director was, uh, late on Don Reese. And there, there was a very young assistant professor called Douglas Finestein who was working on astrocytes. And, uh, John told me if I wanted to work with him and I, I could work in his lab because, um, Dallas was starting a group in ass. And, and I said, why not? So I, I, I knew nothing about ass. It was just a matter of luck. And, uh, we started together working on ASTs and that was the time when ass, what nothing about the functional role of ass, whether regarding cognition or neural circus was mentioned or known, or even conceived and, uh, exercise where removing the passionate was known.  

Elena    00:11:20    And, uh, it was the time where the concept of new inflammation started. And, and later we had the chance. I will tell you what I think about inflammation, neuro inflammation now, which is not actually, I don’t have a good opinion of the term many longer, but there was a time when as, and my were involved in diseases in, in Alzheimer’s disease and in multiple. So it was there. And, uh, so I, it happened that it was, um, working with a person that knew how to grow ASTO as in culture, that was a very important technical development as well, to be able to grow ASTO science in culture, to, in cultures to study them. And this is how, how it happened. So I didn’t, it, it wasn’t a vocational. It wasn’t like, uh, I knew what I was doing.  

Paul    00:12:11    You weren’t following your passion for instance, right?  

Elena    00:12:15    No, well, I, I, I am a very intellectual person and I’m, and I’m very curious person. And it happened, I started to work in neuroscience because, um, there was a thesis position. I was a fellowship that I could get to work in a lab in Spain that studied the, the vascular cells in, in the rain and in did physiology neurophysiology. And I was happy with that, but I, it’s not that I looked for it, but now I love the field. So I, I, I love aides. It’s, uh, an amazing field to be in is neuroscience. We have a lot of work to do in this century, so I’m fine. I’m fine. With, like, I was very lucky serendipitously that I went into this field. I cannot tell stories that I was highly visionary that wanted to go into this field, but I, I think those are bullshit in general with exceptions probably, but Hey, uh, I’m very happy. Um, I love, I love science and my enthusiasm for things, uh, cerebral haven’t hasn’t diminished.  

Paul    00:13:30    Is it fair to say that because, uh, you mentioned like, like Alzheimer’s, and I know you still work on the role OFCY and Alzheimer’s was that also was, was disease an interest also, or did that  

Elena    00:13:42    Okay? Yes. Yes. Oh yes. I was because, um, I, I studied biology, but always I doubted when I finished high school, uh, I didn’t know if I could, uh, I would become a medical doctor or a physician or a, or study biology, cuz I, I got this course the marks to get both in medical school and, and in the, in the graduate school for biology. And finally I decided biology. I’m not sure why, but I decided biology, but I was always interested in disease. So eventually, um, coming back to diseases through as exercise and it’s a good position to be because we, we need to know basic the basic, uh, function of a cell in order to understand what happens when they get sick and vice versa. And sometimes when, when cells get sick or when the brain is altering diseases, we’ll learn a lot of how it works during, in healthy, in healthy situations. So I think I’m happy in combining both, both competitional as science and really clinical clinical research and, and I, I combine both fields very well and, and I think I should do it.  

Paul    00:14:58    Okay. So, well you were saying about the, um, the advent of some calcium imaging.  

Elena    00:15:02    Yeah. So in the, in the nineties we have calcium imaging and then the, um, in the conceptual breakthrough was the, um, understanding that those calcium responses in astrocytes were equivalent to action, potential and neurons in the sense that astrocytes respond very precisely to neuronal information by way of calcium and transients. So we say that neurons are excitable using changes in membrane potential that, uh, produce action potentials while assis are excitable yield changes in calcium concentration. So this is the way we see the calcium trends in astrocytes. Now it’s very important to understand the historical development of a given notion to understand why research in a given cell like ACY may be delayed over time. Um, neurons. We know that neurons conduct electrical impulses from the fifties, from the 1950s in the 20th century, when action potentials were discovered and calcium imaging ass appeared 40 years later.  

Elena    00:16:16    So research in astrocytes as possible excitable elements in circles is like 40 years with respect neuros and these kinda explaining parts why astrocytes and, and, and the role of astrocytes in condition is rather undeveloped as compare to neuros because time is a very important factor in research. I think people outside research don’t realize that research is a <inaudible> work done with our hands. And it takes a long time, not only to develop a mature concept, but also to gather enough data to, to support a new conceptual change. So this is the, the, the time when calcium imaging was possible in ostracized is actually relatively recent with respect to the, as compared to the, to the discovery of action potentials in, in neuron. So that explains in part why astrocytes and condition haven’t been, uh, related for, for a long time.  

Paul    00:17:21    So just stepping back and reflecting for a moment. I mean, that story kind of makes me think of two things. One it’s that, uh, the phrase, you know, when you’re a hammer, everything looks like a nail. And the other is the, the old story about someone who lost their keys and the only place they look is under the light. Right. Because that’s the only place that they can see. Yes. So, um, in some, so in some sense we haven’t appreciated astrocytes because we didn’t have the technology, the calcium, uh, imaging technology. Yes. But, um, the other way to look at that is that, so just like thinking about neurons only as spiking things, because that’s what we could measure, um, now, right. Is, is there a danger that we, we think of astrocytes as calcium signaling entities, because that’s what we can measure and we don’t, we, you know, maybe we don’t appreciate both neurons and astrocytes other facets simply because we don’t have the technology to measure those things yet.  

Elena    00:18:14    Yes, absolutely. This is absolutely correct. This is absolutely correct. Okay. Yeah. And, and this is, this is why it it’s very important to, we, we have to develop theory at the same time, as, as technology, both things have to go, go hand in hand. Cause sometimes even you may have a very amazing technology. As, as you say, you only look under the light. Exactly. You just don’t see anything else, but sometimes it’s also that the, the technology is just, it doesn’t allow you to, to detect anything cuz I I’m with microscopes the same thing. And now we have amazing microscopes that allow us to see interaction between organal live. We, or we have one single local R HC to, to do high throughput, molecular analysis. So they both go hand, hand in hand the conceptual development and the technical development. And we, we need to understand that we need to teach students. And we need to, to tell students that a given moment in science is determined by the con the concepts and the technology that are available on time. So you’re not, you’re not totally objective to actually depend on what you have. And, and this is, and you have to work with that and you have to move it forward, but you have to realize that your influence is highly determined by the time you’re doing your research.  

Paul    00:19:45    You just mentioned that the advent of calcium signaling and imaging has, you know, brought to the fore a larger role for astrocytes. And we’ll go deeper into that, but just some more kind of basics, um, about astrocytes in general, before we get into, you know, what’s being learned more about them, but, um, one thing in comparison, relative to neurons, the God of the brain, uh, so there are lots and lots of different types of neurons. And a lot of this is being learned through the high, high throughput of molecular techniques. Like you talked about, I think, you know, it’s over a hundred and counting and we don’t even know how many exact types of neurons there are, but there are very few types of astrocytes in the brain. Yes. Um, and one way to think about that is that that must mean that, um, that they have limited roles, right. That, uh, they they’re mostly housekeeping. Yes that’s. Is that, is that the right way to think about it? What does it suggest?  

Elena    00:20:37    I, I, I totally agree with you. This is a, a great, great question and issue to discuss, um, it important question because about fats, um, one of the current fats in, as in astrocytes or in brain research is the, is the, um, the data obtained with single nucleus and single cell iron sick. And that has brought up the notion of as ACY heterogene. And this is now a very fashionable topic to the point that some authors contend that aides are so heterogeneous that there are no typical aides and this is wrong. Um, single Nulo hernia sick is a great advance. I, I, I agree with it because it improves the solar resolution of molecular data, but we need to put the concept of ASTO heterogene perspective as suggested, um, comparing heterogene neurons with heterogene in ostracizes. And in short, there are thousands of different types of neurons, molecularly, distinct, and functionally distinct neurons.  

Elena    00:21:49    And there are most less 10 at most less than 10 molecular distinct astrocytes and whatever. We still dunno if those types that have been discovered recently are fake types that come from different progenitors or they are transient types that results from adaptations to local circuits. I support the, the second possibility. Cause I think that process are very plastic and they sort of adopt the flavor of the surface they serve. But yes, I agree with you. Uh, there are few ASTO and neurons, and these means the neurons are specialized to performance specific functions while ASTO perform general functions, be them esthetic or computational, but fear and general functions. That’s my idea. But I think people turn agree with me in that. So let’s see, time will prove that who is right, the Astros are genius and do 2000 things or they just do a couple of things.  

Paul    00:23:02    Well, like you said, there’s a lot of work to do. It’s  

Elena    00:23:04    A lot of work to do  

Paul    00:23:05    You, you said you, you kind of, um, supporting the latter of you, you see astrocytes taking the flavor sort of developmentally and, and plastically taking the flavor of the circuits that they’re involved in, but the other people, I don’t know if you’ve read and I think it’s Andrew CO’s book, the root of thought. There are people who are very, uh, pro who advocate the primacy of astrocytes that instead of taking the flavor of the circuit, they’re dictating the flavor of the circuit. Oh, But you, you, how do you, do you see this as a back and forth process or are astrocytes taking the lead or astrocytes listening more and adapting their function?  

Elena    00:23:44    I, that ASTO lead more. Oh, sorry. That neurons take the lead and ASTO listen. And they, if they compute, we, we can just get into the competitional part of ASTO. I think if they compute, they adapt to neurons. Okay. I think neurons finally, they are, um, I wouldn’t say the God of the brain, but they are the ones that encode information and, uh, as precise may modulate, this is my, my theory. Now we can just talk about hypothesis. So we don’t, we don’t know yet what they’re doing. I think we can, I, I can, I can tell you what is, what are the data we have. Sure. And what are the different options from, from that data?  

Paul    00:24:36    Yeah, that’d be great.  

Elena    00:24:37    So the, we, we spoke already about the architecture and this is very important. So as are not like neurons, they are huge. And each given as site has like includes in, in their, in their body, includes like 10 neurons, millions of synopsis and thousands of things rights. So they, they are sort of like, like a unit. This is, and this is very important to understand that. And we still don’t know what doesn’t mean if this is reflects that aides are able to, to integrate some, the information from all those neurons or, or is just some that kind of design has some type of advantage. We don’t know. Another feature of assize is that they’re territorial, neuros are highly PROMIS to they sort of mixed interactions with some direction and spines while neuro where ostracized, the, the EPIs process of one aide just doesn’t get into the territory of a, this process of another aide.  

Elena    00:25:42    And we, and we still don’t know what it means either. And when it comes to functions, as I said, there is no evidence yet the ASIC code, any cognitive variable. So if the person just mentioned, said that says lead, lead the process of, um, information transfer. I don’t think that, that this has been proven what we have learned in the last take. I, the last in the last 20 years, particularly because researchers have been using tools with increased spatial and temporal resolution, I have undertaken studies and complex behaviors. This is very important. So we, we know, I would say three things just to summarize the field in, in, in three key points. First that Astro S the activity circles that that’s will demonstrate in different  

Paul    00:26:39    Areas. That’s Al that’s almost like a saying they’re another housekeeping role for them, right?  

Elena    00:26:45    I would, yeah. I think I will get to that, but in bad, but computational  

Paul    00:26:49    Right. Computational housekeeping, but as opposed to metabolic housing.  

Elena    00:26:52    Exactly. That’s exactly. This is it. Poor as this is a great, thank you. Thank you for the, I I’m writing that down. I like that. I  

Paul    00:27:01    Don’t know, but it sounds like, but <laugh>  

Elena    00:27:03    Competition housekeeping. I, I like it. Thank you very much. Well, I will, I will use it,  

Paul    00:27:08    But then you, you know, because when you were talking about the housekeeping, traditional story, I kept thinking of, well, that’s like the traditional role of women in the United States, right. In Western culture. And now we’ve to appreciate <laugh> or give more respect to women and their computational role. Right. You know, or that’s a analogy, but now you’re saying they’re housekeepers again. <laugh>  

Elena    00:27:31    Well, it, I don’t think there is housekeeping, um, is by I, well, yeah, so, and, and, and the point with housekeeping is that it should be require that is an important function in general, meaning that it should be economically interesting and, and men should be encouraged to do housekeeping. Wow. That’s my vision of housekeeping. I, I, I think it’s, uh, it’s an important word, like taking care of kids and taking care of all people. So I think the point is just to, to give importance to housekeeping and in the rain, the same thing just let’s give importance to, to housekeeping by any type of self. I agree. So, yeah. So they’re military in all types of circus, in all different, in all regions of the rain that’s well demonstrated. And if someone wants to know the literature, just contact me because I can, I can forward them to really beautiful studies.  

Elena    00:28:24    Um, the second, the second point has been, uh, discovered in the last decade is, is that the culture responses in aide Isly mirror the neuro activity to the point that they are, it is predicted. And this means that one can tell the behavior of, uh, mouse in general, on animal rat or a mouse, very precisely just by looking at calcium transients in astrocyte. I mean, using the codes, this is another advance that we have been using tools that have been developed for neurons. Now we’re using them for astrocytes and using the codes with calcium activity. We can now predict what kind of decision an animal has taken, or whether the animal is thirsty or whether the animal is, has moved opposition.  

Paul    00:29:19    How, how do those predictions compare to, uh, predictions based on spiking, right. Decoded predictions,  

Elena    00:29:25    Um, equivalent equivalent they’re equivalent. Yes. So this is very, very important now, and now with optogenetics and chemos, which is the, the latest development in neuroscience. When we manipulate exercise, we change behavior. That’s proven to which further supports that aide are part of circles. And, um, we can talk about the problem, epigenetics and chemo, genetics that are not very physiological. In fact, we really don’t know how they work, and we have to be very careful to use tools to draw part on conclusions, but this is a fact. So we manipulate the astrocytes and we change behavior now. So the point is now to just the question, what does it mean? And, uh, the in field, the, we think either this means really sophisticated metabolic health keeping, for instance, that, uh, upon calcium increases ass release, uh, lactate to restore the metabol of neurons, or astrocyte connect the vascular, the blood flow to neuro activity in order to provide more glucose and oxygen to your neurons, or whether the Eide are activated to remove passion and put, and brain potential to other area of the brain.  

Elena    00:30:51    So they increasing calcium me some means that the aides are highly achieved to neuros, but it doesn’t improve whether they compute or whether they just do high, sophisticated housekeeping. So whether, and the question is we are talking about computation is either whether they do store information themselves, do they instruct neurons? Do they transmit information to neurons as one possibility, or secondly, whether even if they don’t store information themselves out, they don’t store the code. Part of the variables of the, of the brain code is they ate the activity of the circles, like for instance, controlling executory and inhibitory balance, controlling gain, um, providing and feedback. And I, I, and I think for me, this is where the field stands now, and this is the work we need to do in the, in the next decade. Let’s try to find exactly what are the Astros doing in the circuits and whether it’s anesthetic, hetic, housekeeping that I like it, or hypothetic competition or computational HEAs.  

Elena    00:32:17    This is basic, which is myself, the, the hypothesis. I favor, because as we mentioned, exercise is sort of basic. They’re not specialized. So probably they’re doing the same thing all over their brain. And then people, the engineers have to describe many basic functions that are necessary in circles and a could be doing those functions, or whether they’re really more sophisticated. And they’re really encode for some type of variable. Like they, they control ever. For instance, when, um, the, when you’re inspire differently that some, something has to put all the data together, perhaps exercise, computer, wherever, or they’re doing something that is more, more like coding for a given variable. We still don’t know that. And my recommendation for the field is just not to push things because the, so the, the problem of the Eide field, I think in the last 20 years is try to prove that ASTO are like, and this has been like the motivation of all the time.  

Elena    00:33:34    So let’s prove that ASTO so less proof that are very important in circuits. And this is fine because he has, he has produced under this framework. We have very interesting studies showing that investors has, can mobilize circuit activity in a very fine manner. But now I think the time has arrived to think that Astros are not like neurons. So we have to understand what they do and what are their very specific values that astrocyte as compared to neuros that are slow. So calcium signaling in the time, the, the time range is, is seconds or hundreds of milliseconds. And second while neurons there, some potentials working milliseconds, sometimes all track fast, uh, action potentials are less than one millisecond. So there, there are in different time zone.  

Paul    00:34:27    Well, even in a comparison like that, right? So I was gonna ask you about the timing of the signals. And you just said that Astros is lower, but that’s relative to neurons. So it’s hard to escape  

Elena    00:34:38    Relative  

Paul    00:34:38    Neurons about, it’s hard to escape a comparison to neurons since they’re sort of the standard of what’s  

Elena    00:34:45    Important, right. But which is fine, just fine too. Cause we learn a lot from comparing things. But I think that the, we have been forced Sy Aros to look like neurons. So people have been like once we described, uh, we found cosin signal aide, people were trying to prove that cosin signaling can be very fast and they were developing the increasingly more, more sensitive tools, imaging tools to detect signal any milliseconds or in 50 milliseconds. I think this is wrong. This is wrong because we have to think what kind of activities are important in the brain in the second scale. And there are very many, so the brain works in many different type scales, not only middle seconds. So we really have to refine our thinking with a forcing aide, into being something that they are not. And, and the time that of their responses is highly very important.  

Elena    00:35:45    And my, my impression now my current hypothesis is that what they, I do. So they’re, they’re, they’re computational HOAs is to control feedback. So the, and, and I have one article that I can, I recommend the you and the listeners to read by Hannah. I she’s in the department of biochemistry and biophysics. I UC Seth and his called it is published in cell two in last year in 2021. And his called biological feedback control, respect the loops. And he, she contains that organisms are an evolutionary masterpiece of Multiscale feedback control. And I think we should think about the brain also as a evolution masterpiece of Multiscale feedback control. And my idea is that ASTO participate in feedback loops at a, at a maker and master scale working in the second range. And this is the way they made it central to, to computational housekeeping. So it’s not that they code variables it’s that they, they control, they, they participate in feedback control, and we need to develop this idea and find what are the exact principles and feedback loop are relevant in the brain. And the aide may be maybe involved in,  

Paul    00:37:20    Well, I’m gonna have, uh, Matthew Larcom on the podcast soon. And one of the things that he has been working on a lot is in the, um, cortical in the cortex, the cortical mini circuit, the cortical column that there are in, in the upper layers. So let’s say <laugh> a LA a layer five parametal neuron has this, um, long ACAL ding, right. That reaches up into layer one. And that largely receives feedback signals. Whereas there’s a, a, a more localized dendritic branching tree, closer to the Soma, um, that shoots down essentially a little bit below the Soma, but, and that, that largely gets like bottom up signals, like incoming signals. So you’re describing that role for astrocytes made me think of this, because one of the things that he’s shown is that the, the ACAL din rights, when they’re those signals that they’re getting the feedback are electrically decoupled from the cell body. So they have like a very different role. So it would be very, it’d be interesting if astrocytes had a role in meaning mediating that feedback, but, but that’s a little bit adjacent to the kind of feedback control that you’re, that you’re talking about, I suppose.  

Elena    00:38:30    Yes. Yes. It’s a, so it’s in a given circuit, they may be maintaining the, uh, we, we Don know this is a very important thing to just isolate and identify what are the relevant circuits, whether it is one given astrocytes and, and, and the same as Percy is like a mini circuit. That includes several neurons, perhaps in quarters that has been described and actually courses Aros are different in different layers. So perhaps they, they engage in, in different, different feedback control with different neurons. But I think it’s very important to the relevant, the relevant circle, and try to analyze as size as exercise, as elements that provide feedback. They don’t, they don’t process information themselves. So this has to be proven or disproven, but that they, they provide feedback. And that feedback control allows for a finer performance of, of the whole circle. That’s, that’s the way, that’s the way I, I see now,  

Paul    00:39:35    Well, even homeostatic mechanisms are become, you know, even if that’s a homeostatic role homeostasis, it  

Elena    00:39:41    Is SATIC, but it’s, yeah, it is SATIC and that nothing wrong with it, nothing wrong with it. Right. Nothing wrong with it. And we should really be very humble. We don’t need AST don’t need to be like neurons. So yeah, if they do, if they don’t do it, that’s fine too. And if they, if they, what if the role is to do feedback control in the brain, in that that’s their SATIC, uh, rolling computation. That’s fine too. And that’s, that’s, that’s very, very important to, to find that as  

Paul    00:40:13    Well, before we move on, I, I, I keep having this question float through my head. I want to ask you about your general, uh, outlook or overview on the computational approach to thinking about cognition that dominates neuroscience these days, that it’s, we need to think about neuroactivity all in terms of the computations being performed, whether that’s in a population of neurons or in small circuits or at the single neuron level and things like a feedback control mechanism, like a homeostatic, um, I just said mechanism, uh, you know, role feedback role for astrocytes. Um, even that we kind of think of in terms of how it relates to the computational, um, function of neurons, which is the quote unquote important thing. Right. And I know what you’re saying is that it’s all important, but do you have a perspective on how you think about our cognition? Is it all computational? Do we need to, is it multi-scale multi-level that we need to somehow rejigger our thinking about how to think about cognition  

Elena    00:41:17    It’s well, is I think it’s the big question for the, for this century, how exactly the brain computes, whether the brain is digital or is more analog and certainly is multi-level so that, that we know and multi-skilled, but the question is to identify what is the relevant levels. Cause I think we are generating a lot of data that happens in general, in all the biology fields, not only neuroscience with all these new machines that we have and microscopes and, and molecular analysis, we, we are generating a lot of data and I’m not sure we are really, we are really developing knowledge very much. We are sort of have new choice and I’m not sure if we are using them correctly to generate new knowledge. It’s giving is giving jobs to many people and paying paychecks, but it will be creative. In addition to, to paying, to paying salaries, we could just actually advance knowledge in a meaningful manner.  

Elena    00:42:24    So I think this is the, the question in the future is, is what, what are the general principles that, uh, drive cognition and higher brain functions? And probably there are not very many. And also of course the, the brain codes, we still, we, there are some variables that we know that the brain computes like colors or notes or, uh, categories like being a dog or a, or a tree or emotions, but we still D know how the brain does it. And what, and that’s the tip of the iceberg. There probably very many variables that the brain computes and we are not aware of. And that’s the job to, to do this, our task, to understand those, those things in the next, I would say century, I would say, in the 21st century and in, when it comes to astrocytes, I think assis can do something for, for competition on our, for computational neuro science.  

Elena    00:43:26    It is to help understand the brain better. This is exactly using the, the importance of feedback control. So we may understand why brain does certain things. So study as science doesn’t remove or diminished importance of neurons. It just enhances how neurons can be modulated, because we introduce another element that brings feedback control. I mentioned that, and also something very important energy efficiency, one of the main features and amazing features of the brain that hasn’t been reproduced in, in artificial neural circuits is energy efficiency. So the brain uses like 12 to 20 jewels per second, which is like the energy that, that a light bulb refrigerator uses. And it’s not know why the brain is so ally efficient and in all these more functional designs that are now being developed very actively, you know, that you really produce the brain circus in a realistic manner. This is a main question.  

Elena    00:44:39    So why, why is it that the, the, uh, the brain manages to, to so many things with very little energy and perhaps other non neuronal cells ASCE, or perhaps NG two cells, other types of glial cell are the ones that, that account for the energy efficiency, perhaps the, the architectural design of ASCE and neurons response to a way to make neuro activity more sustainable using a modern term. And this is the kind of things that you, that ostracize non neuron can be used to understand, because until now we still, we don’t understand why the neurons can really perform highly demanding functions from an energetic point of view with using neurons. We cannot understand that there are some theories like spars coding and another theories, but perhaps we can introduce another element which are as precise, for example. So perhaps all that fever control the main goal and the main evolutionary goal has been to reduce the, um, the use of F energy or to make the use of energy in the brain more, more efficient. And this is, this is the kind of, this is an example of where using non neuronal cells or, or looking at other properties, uh, in cells that don’t have spikes can be very useful to understand the brain and to, and to develop artificial intelligence that has brain like properties.  

Paul    00:46:24    I want to ask you a little bit more about your thoughts on, I know that you’re not an artificial intelligence expert, but, but before I do that, I’m wondering, so as, as technology advances, just like we started our conversation saying that astrocytes have not been appreciated until, um, the technology kind of caught up to be able to see the, the calcium signaling and see how they’re, you know, kind of signaling to each other, where, where are we in? You know, the other thing with, with neurons, right, is that we can go in and, uh, I can electrically stimulate. <laugh> a little small area of the brain, although it’s really like a large area in terms of how many neurons it’s stimulating. Like it’s like hanging it with a baseball bat or something, but we can stimulate. And I I’ve done this, I’ve stimulated a part of the brain and made monkeys, eyes moves, move, you know, in certain directions. Um, we can also put little drugs in to, um, to reduce or ablate the activity of neurons. You know, you can, you, so you can be very specific. What I’m getting at is I’m wondering where we are. Technologically, can I, can I defi, can I look at it like a little part of the brain and then just, uh, silence astrocytes and see what that does to the cognition, or where are we in terms of being able to manipulate and, and stimulate, et cetera,  

Elena    00:47:40    In terms of manipulating is up to genetics and chemo, genetics, and that so far are activating astrocytes, I’m not sure if there are, uh, with optics, you can also inhibit neuronal activity, but I don’t think that there is, um, an epigenetic tool that inhibits astrocytes, but you can inhibit astrocyte using trans trans tools. So you can remove the, the, um, I, the calcium receptors and just up late, all the calcium activity and responses. The part of that is that it’s not very refined. So this is, I think the problem, the, the problem in the manipulation OFCY is that it’s not, it’s not very precise and it’s not military. So it’s just all or nothing stimulates a lot. And rather releasing with chemo, genetics, and epigenetics, a lot of pate that the aide never releases so much, or it’s just blocks completely the, the cultural response of ass.  

Elena    00:48:46    And that is not very, it’s actually very good because we, we see that things happen, but it’s not very, very elegant. Um, I’m not sure with neurons, we can do a little bit better, but, uh, with ass, we need to refine the temporal and, and the, and the amount of signal that we WEDU, or we, we manipulate. This is very, very important. Also the regional activation ass are huge. So it’s not the same to, to activate the whole asy as to activate a, a process or a quarter of an ACY cause the probably different areas that are, uh, activity in, in, in different times and with different intensities have different functional repercussions. So that’s a problem we need to be to refine the, the manipulation of the astrocytes. And when you mentioned the behavior, that’s also very important that we now, there are a few studies that have done it.  

Elena    00:49:53    We need to pair strictly Aroy activation with a given behavioral outcomes in, at the same time, in order to, to really determine whether astrocytes can modulate that behavior. This is difficult to experimentally, but this is very, very important, you know, to match astrocyte much ASTO activity with, uh, behavior. When it comes to recording. Calcium is still the, uh, the, the calcium imaging is the technique that we use and now can be done a high in temporal high in temporal special resolution. There are tools like a Aqua tool developed by KDA post cancer that allows to extract the signals from the aide with a very, very relevant, uh, resolution relevant to behavior. So we, we, we have tools, but probably we need to refine the circuits. So we need to understand what is the, uh, circuit where upon were upon Astro really are really active,  

Paul    00:51:08    Like the wiring diagram in terms of  

Elena    00:51:11    The wiring diagram, this is very important. And, and then we need to manipulate Astro better. And so, because if, if the role is feedback control by definition, feedback is, is, is motivation. Is there are cha very minor changes. We’re talking changes at 10, 20%. So we need to go down to the level of refinements in order to extract relevant data. And now, as I said, we are in the all or nothing type of manipulations. So this is, and of course modeling will help with that. So we can, this is another problem of the Astro side field. Unlike the neuro field is, is that, um, there is no much modeling in the Astro side field, which is modeling can help us to, to, to change, um, to model like asy properties in, in very small amounts that are relevant to, to the real brain. But, um, the modeling is, is, is not as advanced as it is for neurons. Another second component with the modeling is that the thes don’t work with experimental people, which is a big mistake because in computational neuroscience, you have the, the we science and you have also the computational neuroscientists working together. And this is very important, like in physics, like you have the theoretical physics and you have experimental physics and then one produces ideas, the others test them and they go, and a feedback in, in, in nationalize that doesn’t exist.  

Paul    00:52:54    I would say it’s still, it’s still pretty early on, even in neuroscience as well. There are more, that’s happening more in Europe, but it’s young, still  

Elena    00:53:01    More, but in Astro science, we don’t have that. And even the few people that do modeling, then they, and even they use aide like elements in artificial intelligence when that’s sounds very good. But when you read the studies, they have a different conception of  

Paul    00:53:18    As well, right? Because they, they have to like come up with a let’s invent, a computation that we think astrocytes might be doing. We’ll add it in, right.  

Elena    00:53:25    It’s it’s something know, usually you introduce an element there that’s that has the performance, some type of feedback, which is fine, as I said, but then you can have inhibitor feedback. You can have positive feedback, depend, you can act the feedback and be a different levels. So they don’t agree. So we need, now we actually, um, um, here in, in Spain, we’re trying to, to launch the field of competitional, Astro science, and, and we will, and, and I want to invite the, the, the silly researchers and the people that work in artificial intelligence to help them standardize their concept and approaches. And they find, when they say, ASTO like, just, just decide what, what is Astros like? And we should decide on exactly when we introduce Astros, like alums in a circle, what, what are we talking about? Cause we are not, we, we don’t agree on that. Very, very important point.  

Paul    00:54:25    It’s funny that, um, so, you know, like the, the, the foundations of artificial intelligence based on neurons, you know, but of course, neuron, spike, and hardly any one builds spiking neuro networks, that’s all rate based. And like their little point process neurons then without,  

Elena    00:54:41    They’re not very, they’re not very realistic. And some people have mentioned that this is a problem because the, even the way they, the model neurons is not very realistic. Not only because they’re not spiking neurons, but also because neurons are also very complex and they’re, they are competitional units in themselves because we, they, they, but model as a one element there, when that element has spines has interest Hass has. So, and each part, as you mentioned with, uh, regarding the work of, of that person that is going to take, uh, a podcast, it one given neuro is also competition unit and PR and integrates information from different, different parts. And that is not modeled in our, in artificial intelligence or artificial neural networks. So they’re not realistic at all.  

Paul    00:55:38    Well, and yet I could say, but, but look, how far AI and deep learning now has come with modeling it with, without considering the varieties of neurons and why, why should I consider astrocytes? Why is it just energy efficiency? Are they important for building intelligence?  

Elena    00:55:58    Well, perhaps they don’t, they don’t encode. So they’re, they are not crucial, but they are like, um, I think having a circle that is efficient from the point of view of energy, this is very important. I think so it’s like having, perhaps like just, just presenting an analogy in the body, perhaps it’s like exercise, like the lungs and the hearts and the liver of, of, of, of the body. So, but the brain doesn’t work if the heart doesn’t work or if the lungs don’t work. So perhaps the same kind of analog is housekeeping. So it is, it doesn’t look very sophisticated, but, but as, as I said, when they want to do artificial intelligence, that really can mimic human intelligence, they, they can, they need one of the reasons they need to use very complex computational devices, huge, enormous. And the rain does it in a very small device. There is some efficiency in the, in the, both the functional design of the brain and this architectural, the design of the brain probably is important in order to, to having this kind of efficiency and artificial intelligence now in two, in 2021, still for the time being doesn’t do  

Paul    00:57:25    It, Elena it’s it’s 2022. We’re in 20 22 0 20 22,  

Elena    00:57:29    Sorry. In 29. I’m still, you know, with, with this I’m, I’m, I’m a little bit  

Paul    00:57:36    Love time wise, get it.  

Elena    00:57:39    I’m a little lost at 2022. Yeah.  

Paul    00:57:41    Your housekeeping is all messed up.  

Elena    00:57:43    <laugh> my housekeeping is terrible, but, uh, yeah, it is. But, um, yeah, I, I, I think that in, in order to be intelligent, perhaps yeah, feedback control, energy efficiency, and, um, ass may be relevant for that. And that for me, I, I am happy with that. I’m, I’m happy with truth. So if, if we, we in computational, as science, eventually we find that Astros are not relevant for computational issues. That’s fine too. I think we, we have to be comfortable with that, but then, but perhaps in, in, in, in, in this century, we’ll find out that they, we find, we discover properties and principles of brain activity that in which astrocytes are an essential elements in regulating those circuits. So perhaps we we’ll describe that. Some people have said that astrocytes are some, and the reason, a very interesting researcher in Russia Liva who produces in silicons. And I always like to read her papers. Well, she says that aides are the ones that make the brain analog. The neuros are, are only digital. And with that, we cannot have an intelligence and aides because perhaps they’re, they are imperfect, uh, in providing feedback to neuros, make neuronal circuits to behave in an analog manner, which I think it’s an interesting,  

Paul    00:59:24    That’s interesting.  

Elena    00:59:26    So, yeah,  

Paul    00:59:26    Just as an aside, what, what, what do you, um, think of the hype of artificial intelligence these days, right. Is it overhyped? Are you impressed by it? Are you, do you find it, um, lacking? Or do you have any thoughts on it?  

Elena    00:59:41    I mean, I, I I’m impressed because actually I, as a tool to process information, not just in the brain too, we, we do, um, a mix, a lot of our mix in the lab. Yeah. And, uh, in addition to do, um, ontology and all these mining of data using platforms and then, and other statistical tools that we use, we have used artificial intelligence to fill the gaps and to draw predictions and it works. So I’m impressed, I’m impressed by the capacity of artificial intelligence to, to improve knowledge. And we, and it’s a little bit scary because we, we actually dunno how it does it, cuz you know, you know, the input and you know the output, but you actually Don know what happens there, but it’s a fact and it’s the, it is going it’s happening already. And it will be the in 500 years, it will be probably the, um, the way the humans will live with artificial intelligence. Either human species is still still around. Then. I, I have a colleague who works in artificial intelligence that he, he thinks that artificial and he’s very optimistic about artificial intelligence.  

Paul    01:01:11    Sounds like it. I, I know what’s coming, go ahead. Say it <laugh>.  

Elena    01:01:14    And he says that in division of humanity, that artificial intelligence going to be much better than human intelligence and that the, the, that humans use artificial intelligence, they’re going to live in a better society. He thinks that, and that he predicts that let’s say, uh, I don’t know, 500 years from now, 1000 years from now the vision of, of the 20th century and the, and the 21st century will be of the man that produced artificial intelligence. The man. So now in the same way that we see all the 15th century was, uh, Renaissance, uh, 20th century was of, uh, all the developments in physics and content mechanics. Then we will see the 21st century was artificial intelligence and his prediction is that it’s going to be much better. And that artificial intelligence is going to work is going to work much better than, than the brain intelligence, etc.  

Paul    01:02:17    But the, the physics let’s, let’s say like the golden era of physics right in the 20th century, that happened pretty quickly. But this is 500 years is a lot slower. Is it just a harder problem? Or why would it,  

Elena    01:02:30    Oh, I, well, 500 years is just the number. I general know. Cause actually, yeah, it could be 200 years of 100 years. Cause actually what people tell me, people that work in artificial intelligence tell, tell me, is that they, there are important developments like every week in artificial intelligence.  

Paul    01:02:49    Yeah. Well what well that, okay. Quote unquote important. I’m not sure you have to judge what’s important.  

Elena    01:02:55    Well relevant. So they, that they feel is, is very fast. Yeah. So moving, it’s going very fast to the point that they don’t even bother to publish things. They have this big tree where they, they connect to dots and they, they, they show the data from different labs and in that, and people say that it’s really moving on very, very fast in all realms in of course, for technology and for Google Facebook and for the clinical research, it’s moving very, very fast. So it’s, it’s out there, it’s out there. I think we cannot stop it and most we can control it and, and, and demand, um, that it, that it should be transparent, which as I say, the problem is like, like the, like the brain, we don’t really know how it works. And artificial inte is exactly the same thing, but I, I am, I’m curious about it and I think we can stop it. And in my lab we have used it to, to complement our manual analysis. And I have to say that very successfully.  

Paul    01:04:05    Well, I think a lot of young people are curious about it too. And a lot of people in academia, I’m not sure if this, if we mentioned this during the episode or before when we were talking, but people are, seem to be leaving academia in droves these days, or at least threatening to or complaining about et cetera. Yes. Um, why would you two questions? Why would you stay in neuroscience? Why would you stay in academia? Or why would you go into neuroscience as opposed to going to work for Google or some company that we can’t stop because it’s, they’re producing so much data and it’s so important every week, our is 10,000 new important things. So the, so the second question would be, um, if I were to gonna go into computational neuroscience, why should I study astrocytes should I study astrocytes? Is this the right time? You said, I think you said, it’s gonna take a hundred years until we, you know, fully know the  

Elena    01:04:59    Well, but then you do, you will contribute to that development. If you’re, if you’re a curious person, I think you can study competitional as science and, and then just contribute to the development of the field. But one very, I is not to be isolated from the people, which is another problem and a big problem Thelia fields. And I’m, I’m actually against the use of the term glia because it is, um, it has historically important term that was coin, uh, 100 years ago because there were the researchers like aha and others, they saw neurons that were obviously relevant to the performance of their brain and other cells, other cells.  

Paul    01:05:51    Yeah.  

Elena    01:05:51    Everything other cell types were clear and her and, and that’s, and that should be, and, and at that time it was the correct thinking. But now that is too simplistic. And in the, the term Clea has protected the glia fill and has helped the glia researcher to get together, to organize their own meetings and their own journals. And we review each other and that has worked for a while. I mean, that has helped us to produce data, but now it’s a highly damaging term, I think, because what happens is that when you don’t people take a look into aides, they don’t buy any of what we do.  

Paul    01:06:36    You have to make it sexier. You have to make it, uh,  

Elena    01:06:39    It does. So they, they think that they are primitive cells and they, they, there are that research in neurons is highly more relevant and there are priorities in neurons that need to be solved. Before we just ask questions about whether these primitive, um, huge cells are doing something in Al circuits. And that’s a problem because we, and, um, because we need the neuro people, we need their theories, we need their, their technical tools. We need a statistical analysis and, uh, we need to grow together. And my, my intention with the computational as sciences field is just not to develop it in isolation, within neuro field, but to, in Spain to grow it in the computational neuroscience field. And I think the students do the, the question to go into is not a question to, into academia or to go into goal in the question is whether do you do want to do research or do you want yeah.  

Elena    01:07:41    To do something else? That’s my question. And then you want to do research and, um, cuz you are, you’re a curious person because you want to ask big questions and the brain neuroscience is still a big, big question or whether you want to solve diseases. And I think computational neuroscience, science or science has the chance of curing diseases in, in perhaps not now, but in the next, I would say 10 years, because once we understand, for instance, the role of, as in, in synchronization of neurons, we can manipulate ATRAC size in order to restore synchronization. Or we can engineer neuronal surface to, to adapt their activity in very fine ways and control autism schizophrenia. I think in, in the, in 10 or 20 years, these neuro computational neuroscience going to deliver clinically relevant results. And if people, students that, that want to cure diseases, then do research, cure diseases. This is a fine field too. And they can do research in academia and they can do research in companies as well. There are many companies now there are that do very, very fine research. So I think the point is whether you want to do research and then where, or whether you want to do something else with your biological, biological degrees or medical degrees. That’s the question. But yeah, I think research is, uh, a magnificent endeavor and it’s very fulfilling in the long term.  

Paul    01:09:28    Oh it’s and it’s never frustrating. Right? Never frustrating. It’s always, <laugh>  

Elena    01:09:32    Never frustrating. Never, ever, if anyone tells you their research, frustrating, they’re lying. They’re lying. They’re lying. <laugh>  

Paul    01:09:43    So computational Astro science is the term.  

Elena    01:09:47    Yes. Computational science that will work on Omni study housekeeping.  

Paul    01:09:52    <laugh> right, right. That’s the first one. But, but, but then that also, um, not to belabor this point, but then that also kind of keeps it separate from computational neuroscience. I know that the phrase mirrors computational neuroscience, but what would be a, a term you wouldn’t say neuro Astro or Astro neuro or  

Elena    01:10:10    Yeah, those are, yeah. We, we, we, we play with those terms and it’s not an easy, it’s hard. It’s very, it’s a hard one because it’s very terms die hard and, and neuroscience will not die.  

Paul    01:10:25    What about just systems? The problem is neuro is always in the world. Right? Which focuses  

Elena    01:10:31    Neuros always in the world. That’s the problem. But yes. That’s, I dunno. We can perhaps, uh,  

Paul    01:10:38    Systems, brain science.  

Elena    01:10:39    Yeah. Systems, brain, but brain is not brain because the spinal well is the peripheral nervous system. It’s also part of the brain. And then the spinal core is on the  

Paul    01:10:49    Brain systems, brains, the universe, and everything sounds about  

Elena    01:10:53    Exactly.  

Paul    01:10:54    Uh,  

Elena    01:10:54    Neuro neuro things. Yeah.  

Paul    01:10:56    Let’s, let’s let’s end on this because it’s kind of coming full circle. So even during our conversation, so, so we live in an era where words, uh, have become, I don’t know, more people put more. Let’s see, how do I say this? It’s words are sensitive to people now. Right? You, they can trigger people. They can be violent, et cetera. I, you know, even during our conversation I have in the back of my head talking about, you know, the sexism of what housekeeping has been traditionally, blah, blah, blah. I have to be careful what I say and thinking about the term. Yes. Which is the term neuron and glia and, and how to move forward and, and how things are named quote incorrectly. Do words matter. Do, do the names of things matter.  

Elena    01:11:43    Yeah. Do words. Oh yes. There’s the word that they do matter a lot because they shape our thinking. Sometimes it’s is interesting. Sometimes words, the linguists say that sometimes you, you need words to be a little bit ambiguous because it’s very, if it’s very strict, then basically you don’t communicate. Oh. But yeah. But some other times you need words to be very precise and you need to, and then you need words to convey, to convey the right, the right concept. So w words are strictly related to, to conceptual refinement. So we need to refine our leic in order to refine our research as well. And I think this is one of the reason with that neuron ASTO sites. And it perhaps in the long term, I already said we shouldn’t use the term GIA because this is like saying that the nature is divided in, in water air fire.  

Elena    01:12:42    And I don’t remember the, for filament, we should stop saying glia because that cures the resolution, the functional resolution of glial cells. And perhaps we should stop saying neurons Andes as well. And, and describe the brain in terms of circuits, relevant circuits and relevant functions that are some of them we know. And some, some of them are still to be discovered. So perhaps in the 100 years, the, our, our vision of neuroscience will be based on relevant circuits and, and relevant building blocks in that circuit instead of, of neurons as precise. Because even we, we said that already there has, and different types of many different types of neurons. And why should, should they be collapsed in, in the notion of a cell that spikes  

Paul    01:13:36    Well, a hundred years is a, is a long enough time. I think that that might happen because like people, you know, linguists would also say that words are difficult to change. Uh, you know, you can’t, that’s not the way that language really develops unless you are really good at making memes or something. Right. And, um, so it’s hard to like force a change and somehow it has to happen over gradually over time and in a somewhat natural manner. But so, so I don’t imagine that you think that you could force the change.  

Elena    01:14:07    Well, yeah. Some, yeah, it’s very because these, the, the people that use those words retire.  

Speaker 0    01:14:15    Okay.  

Elena    01:14:16    Yeah. And, and, and new, this is like that. So, and new people can sing and they have different ways of, of, of describing things. But we, we should really encourage people to use better words and to describe systems in a more refined manner. Absolutely. And, and particularly as you say, is like trying to take it personally, cuz the problem now that words are loaded, emotionally loaded and particularly in United States in Europe is not like that yet. Oh,  

Paul    01:14:48    You think it’s you think it’s coming?  

Elena    01:14:50    It’s coming well, we’re watching, we’re  

Paul    01:14:53    Watching, don’t stop watching us stop doing anything that we’re doing. We’re  

Speaker 0    01:14:56    <laugh>,  

Elena    01:14:57    We’re watching. But I think it started crazy because you know, we’re, we we’re suggest words. So we kind of just be so of, you know, continuously authentic, offended by, by certain words. I, I don’t care. I just, in, in the, in the case of science, I think that we should, we, we can describe things better. And in personal terms, I, I, I’m not offended at all by many words and I don’t care, but I think that’s, we really have to be a little, I dunno, this is the, the way that the world is going. And sometimes it’s a little excessive in this sort of over reaction to, to name some words and getting offended by yes, basically. And, and that’s a yes. Another reflection.  

Paul    01:15:42    Yeah. Maybe the world is going through a rough patch, but I I’ve enjoyed, I think that our, our little time here has been like an Oasis within that rough patch. So I’ve  

Elena    01:15:50    Well, you can, you can use any work with me. I don’t care you.  

Paul    01:15:54    Oh, you better watch out. I’ll I’ll uh, now don’t get me going. So <laugh> okay, Elena, this has been fun. Thank you for the time. And uh, thank you very much. Continued success.  

Elena    01:16:03    Thank you Cole, for, for your time and your questions. It’s been very, very pleasant to talk to you.  

0:00 – Intro
5:23 – The changing story of astrocytes
14:58 – Astrocyte research lags neuroscience
19:45 – Types of astrocytes
23:06 – Astrocytes vs neurons
26:08 – Computational roles of astrocytes
35:45 – Feedback control
43:37 – Energy efficiency
46:25 – Current technology
52:58 – Computational astroscience
1:10:57 – Do names for things matter