Josh Huang Transcript
SPEAKER 1never pointer. Mhm. Do I need a quick use the arrow. Okay, um, so it's a true pleasure and honor to
SPEAKER 0
be invited by the Brandeis graduate students, you know, Um, so this visit definitely brings back lots of Precious memories. Not the one that just a better stories. Michael just doesn't remember. Uh, okay, so I have to say that I'm I was extremely lucky to be scientifically born and raised BRANDEIS a time. Literally. Uh, you see, the stars were aligned, and I have to say the stars are still aligned, and I Okay, so, um, it was here that I really established my scientific roots and long term interest to explore the biological basis of behavior and also was really bestowed by several powerful approaches, of course. Genetic analysis in just Drosphilia flies pioneered by Seymour Benzer and brought here by Jeff by implemented Michael, Kalpana, Charlie and Leslie. And it's a fantastic fly community. But in addition eyes, really, this, uh, influenced by this powerful comparative approaching invertebrates on, exemplified by Eve on the time Irving was working on a pleasure and John was the nameless photoreceptors. And of course, this really powerful and wisdom of evolutionary perspective. And who would think that you know, the security clock mechanism? The court mechanism in the fly would have preserved human and down to the Sun Slam mode and in addition to the people who directly taught me. And there are people that I really benefitted on that maybe don't know to them. But from seminars and the discussions Larry and Chris and Jim Haber many was on my thesis defense Ron Johnson, who has left on, uh, in addition, great graduate student body that I cannot lead to many people. But, uh, reaching low was two years ahead of me, and he picked me up at the train station when I first arrived on Hong cuisine was two years after me on she did rotation with me Paul Michaels assignment and in the end, also not listed here was undergraduate senior who became my wife. So things really couldn't get better. You could see by the time I was leading towards the end of 94 I think I was really instilled and blessed with really a remarkable I would say heritage. The two influence really the rest of my career until today. So I have to say I still have doubts after looking at what people have done here. You know, Michael, even you mentioned it. Whether it was it was the right decision to venture into really the jungles of the of the cortex. But as Michael would put it, you know, life is end of one and you know there's no controls and going back. So the new context is where you all know is, uh really consists of a constellation of functional areas. Form representation map of the external internal world to integrate multi sensory information to with internal goals. Make decision intelligent behavior. So, for well over a long time, generations of your scientists have explored the functional organization of the cortex across scales and levels, using the tools that are available to them and have, along the way, discovered the coin, the words and the concepts such as a cortical areas, layers, columns including cell types down to synapse and molecules. And this is a tradition that is a different from the one that I grew up was here. You know, the genetic analysis comparative approaching invertebrates. Here is the tradition of anatomy, from now to physiology Penfield, human reason and psychology. But the enduring challenge is really with the immense sort of cellular diversity and warren complexity to decipher really the cellular basis to construct this increasing, higher level function. Architectural. It's still challenge, and so the question is, how to do it. And, of course, my BRANDEIS training taught me. Genetics Try that. So the key point is that it's not using the genes to do a knockout, but really to use the incredibly rich information in the genes to get themselves, including cell types. To understand circles. That's easily said, uh, but the inspiration really is also, you know, this is a sort of rather radical statement. But you say that to some extent, that has to be true that several years back A to the Centenary of volunteering Sydney Brother wrote, Fundamental theory in biology is concerned principally with using millions of living organism. It's the only part of the natural world. So this is more than just the biological world physical world that contained internal description of themselves. That is why the whole biology must be rooted in DNA. Again, our task still is to discover how is DNA sequence rose in evolution. How they are interpreted. So I think this has to be true because what is stated and it's true not only for other organs, but also for the complex organs. So the complex, the relationship is very convoluted and complex, but it has to be there and again. I don't need to deliver this at all except to say that symbol. Banzer, you know, more than 40 years ago said that genes can be a scaffold on genetic Moses, um, to study the brain organization except not at the level of cell types and circuits. But really, all we do is genetic Moses. This is incredibly, I think, insightful. And the question is, can we apply this to, you know, higher organisms, including those ones with millions of neurons? So my sort of overarching sort of mindset is that despite the cellular complexity, we know that the cortex, including the human cortex, emerged during evolution from its president. And there are basic conservations across and the keys to recognize those elements that actually build increasingly higher level organization. So we know the areas in the mouth there about 40 or so, and they are located in roughly the same location if you don't. If you consider the rapid expansion of the frontal cortex and so forth and their connectivity is largely conserved and certainly this tried descending output, the cortical spinal critical strategy than on a loop and including the local service templates. And I think these are known, but sort of more modern view, in my opinion, is that all of these local circuits global templates has to be constructed by a large set of what we would call Cardinal neuron types. So let if I the cortical, uh, spinal neurons, including the chandelier cells. And so these are reliably generated during development and they are almost identical all of this. So there has to be a developmental program. It has to be somehow encrypted in the genome and also conserved across species. So with this in mind, the idea is to really to capture or to catch those cardinal types which are likely the fundamental elements. And so probably there are only two types the group detergent neurons that form the local skeleton, the all the critical processing streams and all the output channels, and it is massively underestimated. In, uh, Sasha has put it out quite some time ago. And also, the garbage neurons regulate the early the temporal dynamics, balance and organization of parameter ensembles. Question again is beyond these diagrams. How do we actually get to those cells? How many of those? What are those? So this really again, sort of, I think, came back to my, uh, education here, and I actually had some training before, even in But your physiology and this idea of you can identify that you're actually numbered on and put your electrodes, you know, in a neuron with pigmented cell one of anymore after another really has generated a very deep insight. And the other is that I remember so well that this is a paper from John Muir who was in Jeff Halls lab and also Michael Rosbash. At some point we identified, okay, identified a set of so called lateral neurons. There were nine of them at the time. I remember now there are many more, and these are the cells that actually presumably constitute the circadian clock. But the icing was mentioned interesting to mention. So I was entering the lab, remember? There was a graduate student who want to SC, where the per m I was oscillating skating time and she wound up the whole flight but had very mixed results. So later the postdoc Paul Harding, who taught the Arnie's protection just use the head of the flies and found this is one of the fundamental discovery off the circadian mechanism, and it's kind of absurd to now you grind up the head of the fly and you actually can discover you know, this clock mechanism. But that is the power and the beauty of the simpler system. And for many years on, the Sacha one of the pioneer, actually, to see how we can begin to grab onto cells and the sub populations, even with at the time very limited mouse lines and actually measure there are any content. So that's the bottleneck. And I have to say, you know, for almost half of my effort is really trying to just get to the point where you know many of the people invertebrate and your neurobiologists take it for granted. And so I just make one slide to summarize this effort. Basically, one is very important to Houston. Targeting Sl's. The transgenic is uninformative or not misleading the other is that not only to use so called markers in these different cell population, but to really, uh, engaged this very powerful developmental genetic mechanism where they are informations in the progenitors in the linear progression in the time of birth. And these have been demonstrated in whatever is also and to use intersection. So not only to use markers, but I think now are my sort of conceptualization is actually to target genetic program. When is the lineage program one is first timing there There's some kind of molecules that represents some of the appreciation program. If we did this. It's actually encouraging that even with two or three genes with the time with the virus, you can actually get quite specific population or even cell types such as, uh, chandelier cells and so we have made various progress, and this is a literally In the next few weeks, we finally we'll put this out major progress in targeting different parameters neuron projection types. So the overall goal is to achieve both. Specifically, this is a appropriate for spender, sending circuits not to just divide for the case of division. The other is to have many of them so that we can begin to clean their organization. And there's the example from the fly project. And I have to say that this was again Sasha was quite a visionary. That actually pushed me back to start this project, which later lead to one and another. And now we are part of this major brain initiative brain cells, Census network. So beyond this experimental access, equally important, if not even more challenging problem is the conceptual problem of how do we think about define cell types? Uh, because, you know, they're inherently model from authority to physiology gene expression. Developmental. So we all know here the question the is A. If you classify these according to different modalities, will they come together in the end? And the currently the answer is not too encouraging. For example, you know, in this very important work from the islands actually again organized by Home Quiz, um uh, who also came here. So the smartest adding cells can be either 15 or 13 types, depending on the parameters that you use for clustering. It's hard to think that this is the biological truth, So the question is literally with this. Seems, you know, endless variation off multiple domains. Uh, is cell type real biological entity in a million
SPEAKER 1
brain, or it's going to be some kind of, you know, committee decision or, you know, some kind of operational classifications.
SPEAKER 0
And everybody wants to know how many cell types are in their circuits but our mentality again. In order to get to that, we need to understand, or at least you know, um, have some insight on how to define a type and how finely should we define them. So our approach to this problem is to use the tools that we have built using combinatorial targeting to define these six population off what we called cardinal neuron types of you know, typically defined types. Some are true, we think Bonafini types, for example, the chandelier cells or the long projecting GABA neurons other at least sort of restrict to, you know, to one or 22 or three population that are largely non overlapping, and then to have a very high resolution image profile off the molecular content. So rather than using connects and others that give you a relatively fuzzy picture using a method that animal power in the lab sort of improved upon to be able to detect on the order of almost 10,000 genes. per cells. So that's still one of the highest resolution. And then to ask the question that given these, uh, sort of anatomically and physiologically distinct cell population, what are the key molecular profiles that actually tell them apart and explain some of the result. In fact even predict some of the properties so this is published, I just want to say that the conclusion that we begin to do is, to be exciting. So it turns out that there is a relatively small number of genes and gene families that are especially discriminative and these four into about six gene families and ultimately is in contributing either to the connectivity which are the multiple families of salvation molecules, all the synaptic input output. So these include synaptic these machineries ion channels, a set of signaling molecules and then transcription factors. So our overall sort of a way of thinking about this is that although there are many genes, there's a small set of genes that is really working to shape the synaptic communications stop how they talk to and listen. To neurons and to perform this input output transmission and this can be achieved by over a similar molecular, scaffold for singling, but then treating different components on these come from gene families with different members of different physical chemical properties, and you can take those to gradually customized these various properties. But in the end, it is, this synaptic input output transmitter. So this is a idea that we're now also testing in other cell population, and the result has been quite encouraging on the overall idea is that this sort of neuron communication element, we think can capture some of the other more traditional description. For example, mythology is likely a reflection of activity. Whether the dendrite looks like this or not is because trying to capture input, axons trying to reach certain targets, and many of the intrinsic properties, the synaptic properties and we all have to do with how the cell is trying to called transform input to certain form of output. And then the fact all of these actually can be read out from the transcript home. And I should not should say not transcript on per se, but certain key features. We call them functional transcription signatures that is shaping these properties. This suggests also that those have to come from commenting on, perhaps from developmental process and were indeed testing these ideas using other data sets. So for now, I just want to spend a few minutes on one of our favorite cell type so, this so called chandelier Interneuron sonic cells. And this cell has the most distinct specificity in, synapse onto the Exxon issues settlement of parameter neurons on the only parameter neurons. And a number of years ago, when Francis Crick was in his later years, I was told that he wrote many notes about some of the important questions that people should study. One of these notes is what is special about the prominent neurons that are innovated by a single shot of
SPEAKER 1
yourself because this kind of yourself presumably has a very influential power on the far firing on these private neurons.
SPEAKER 0
So for a long time it's hard to study these cells because they are very rare. But a number of years ago, now former postal Hiroki Taniguchi, using genetic fit map so there was no markers and now we have markers. But in those days, actually, this sort of test by the power of developmental genetics is that there is a lineage program and birth order. And these, uh, chandelier cells are wrong. Towards the end of the mg lineage, using this transcription factor induce ful, he found that the chandelier cells are born at this time and they take a very long route elaborate migration, but very stereotyped to end at three, uh, mostly three layers. And they target these, eccentric segment of the provident neuron, which is labeled by a molecular marker in green here and already there are different subtypes. And one of the key experiment he did was actually to do the transplantation experiment to label these a newborn, which on the liver cells or the late stage, put them in a unlabeled mouse. And after 34 weeks, these cells that our topic actually migrate to the right lemina location and assume I mean elaborate this the problem pathology and innovate eccentric segments. So this suggests that there's information in the cell when they are born, to tell them you know where to go and who to lead with not to say that they will meet, you know, the right circuits the right strength. But there is a sort of inherent identity there in the in their political program. So subsequently, we have made various progress. One is the finding off, unfortunately, subtypes of chandelier cells just based on pathology. So here is ah movie to show that using a genetic labeling with virus on the in collaboration with chiming law slide in China in Wuhan were able to achieve this high resolution whole brain imaging registered to another channel, with all the cells being positioned and this allowed this reconstruction off relatively modest set of chandelier cells in the three critical areas. But even here, you can see that you know, cells in layer 2, layer 5, layer 6. They have different authority, but more importantly, they're dendrites Are either in layer one or their within layer five, and they're exxon's can be either in later, 2 to 3 or later, two on their five so and so forth. Because these Exxon's are really targeting the parameter neurons, this'll allow us to in for at least that they actually might receive different inputs. For example, the ones that have have dendrites in layer one versus the one that you are later five and also the Exxon's. So this, uh, propriety hypothesis at least that there actually might be different subsets than regulate different aspect of private in your population. And finally, you may say that this, uh, further diversity is that a stochastic event that somehow when they were born, the same and just through the migration and other events that they end up in a different morphology and connectivity or they are actually, sort of know that information, when they were born. So our answer favored a second. This is a very complex slide. I only want to say that we found that there are actually two waves of chandelier cell generation from the so called really placed directly diversity in the intermediate presenters. And this is the one to look at. They always go from inside to out, which is odd from the most of the GABA neurons, which are inside out. So the upper layer chandeliers born first and then the deep layers in the first wave. And then the second also a letter to 300 so. This suggests to us again that these laminar success are not randomly generated. But there is again in order and exactly how that is implemented. We don't know yet. Finally, I just want to say that way. Have the question off for a particular chandelier cells that has Exxon's in layer to layer three. Do they innovate parameter neurons indiscriminately, or they actually further find substantial parameter neurons on? The answer seems to be the later. So in this case, what we need is to be able to identify prominent neurons parameter neuron success, in this case defined by their projection targets. So in the prefrontal cortex, in layers to three, they are parameter neurons that project to either face the lateral Magdala in green here or to the controversial context. Again, this is published with a long story short, that there is actually rather striking specifically that the chandelier cells in layer to three. At least the ones that are born late preferentially target the Biela, projecting prominent neurons. On the other hand, they mainly receive input from the country like project immunized, and this directionality is not only local. When we look at the global connectivity using release tracing again. The chandelier cells is mainly receiving their input from, this network of private neurons that are connected bilaterally. But they do not receive input from another network that is communicating between the cortex on the B l A. So it's really sitting in a very key positions, both locally and globally, to regulate presumably information flow. But this is slice many slides study. So only recently, using the information from a single cell expression, we begin to use interception on this. Allow us to really, you know, purify these channel yourself. And this is, now that we can really begin to study their functional connectivity in a in a way that make a difference on I don't have the result to tell you today. Yet for for the reason that now that we know the development of origin and we know something about this end point on, this is relatively rare in the cortex, where there is one of the sort of still most stereotyped module of your presumed module that we hope this give us a system to see how does this build and how it functions? However, to understand the functional chandelier cells. It is very clear you cannot do that by studying chandelier cells alone because the chandelier cell only is meaningful. If it is, regulating through the prominent role would actually is carrying the information input output. And this brings the problem of you know, our poor understanding of parameter neurons. And it is possible truly with the information streams, different areas of the brain regions, there are more prominent around
SPEAKER 1
types GABA neurons. If you just think about how many critical areas, how many, sub cortical regions and harmony to be wired. These are all from through around in your arms and
SPEAKER 0
from the Allen Institute, even the last count. And this number has increased significantly that there are at least 60 transcription transcript atomic types. The question is again how to relate these punitive molecular types with anatomy. With function on that we need again genetic tools. So we have can use a similar strategy and, in this case, both using mature markers but also using what is has learned from the molecular genetic program from the work of Jeff Necklace and call it a lotta here, to name a few, uh, to combine this lineage related the mechanism and birth order. A secrete driver in useful creed driver with the other molecular markers is a flip driver. And then through this interaction intersection, we can convert using a reporter to a viral access. So at that time, things viral factor already can integrate developmental information of molecular information that allows us to engage in anatomy. And this actually is quite powerful in. Then begin to dissect neuron Citrus layer to cortical, cortical or verify cortical structure of neurons. Show the end. So here's a couple of examples to show This is, ah, layer five or six cortical thalamic neurons. they're five, the critical, feudal neurons and their five using, viral based the targeting. We begin to map their connectivity production patterns as part of the B. I. C. C N Project. But I have to say, ultimately, the single cell resolution is the is the is the ground truth on. Even though it's very demanding, it is really necessary to really understand that these population level patterns are can be dissociated into grandchildren. So all of these are fine. But ultimately, I think, you know, cell typing is not for the sake of cell typing, at least in my mind, is to understand how the how the circuit in the brain works. And I do buy into this, statement Not only that, well, no one, but also the one that is posed in the brain report that I think he was part of this. So a question at the time is with the tools coming together from, intern rounds and the primary round what might be a good system to to explore. And again, I spent more than two years, but Adam Catfish was also a graduate here, actually was important in making this decision is to study movement and in part for this the statement that I like that for, You know, most of the brain function. Ultimately, movement is the one that actually made the difference. So although there have been a lot of progress in understanding the simpler movement such as locomotion, CPG, spinal circuits, including the ones now being explored at Bernstein. But the roll of heart level structures from the basal ganglia, especially the cortex, is very much debated. In fact, when I was thinking about this, I was very scared because rather In contrast to what's known from the sensory coding from getting back to the work of human weasel, there really is not a framework in understanding how model control is influenced by the cortex or it is influenced, and the work can go at least back to Penfield from the stimulation in humans and this mortal combat Oculus to efforts to others, including the now sort of influential, dynamic systems models. But what is going on the level of circuits is unexplored, and this is not surprising because most of these work has been done in primates, so the genetic tools in mice is clearly very powerful and there's no doubt in my mind the question is, how good is the mouth system in studying movement is virtually falling movement. And again, there has been a long tradition in humans and primates. In fact, there are, you know, relatively short realistic views, and there are different levels of shooting is what I want to show you is that one I think influential thinking from Young Wisher and others is that this driving force for the following movement has a lot to with food manipulation in the city. And you can see from here this brief history, you know, from the from lampreys that has not even draw swim future to the jaws. For all these years, to carnivore that kill and the bite and only eat meat and really is from rodents and up that we begin to eat are not we. But you know, the animals begin to eat many different things and they have to eat many different things that are not readily edible. But I don't know how many people know how they eat crickets, but what I told was told is that they actually catch the cricket, the crickets, and then they turn them upside down, and then they take out the legs and wings and then they bite in the middle. So this means that there is significant manual dexterity and you can appreciate in this movie.
SPEAKER 1
So here is how a monkey will peel an orange with the coordinated movement of the hands of the mouth. And here is in our lab. If you feed the mice with, cereal, it'll crack it open and eat. So this presumably has driven many interesting falling behavior.
SPEAKER 0
And again, the question and some of these are clearly not just motor. When you eat something, you know you get a state with response, you have prediction of what you're sending to your mouth. You have a prediction of once that is confirmed. You make your next movement. So there is this continous sensory motor prediction feedback organization. In order to eat, which is a very complex movement. There is eating program, but every piece of food is different and it requires a continuous monitoring. So how do we approach this on again? Here is my BRANDEIS training. Is this incredible power of screening if you're able to do that on. So what we have done here is a very, very modest screen in three cents. One is with the generation of again to dozens also create lines. We can look at these different parameter neuron projection population one by one. And she spearheaded this project on screen. I think eight of them and then zoomed into three of them. On the second is, how do you screen? So in this case is activation screen inactivation screen on the engine. So this is like in the genetic scale function loss of function and expression screen basically, so you can we can put the mouse had fixed with so the skull cleared. You can basically not even opening the skull and then use the laser light to. Activate these different cell population one at the time in a three by six millimeter area but every 300 microns. So this is important because it is not constrained by the traditional idea of what is a primary motor cortex and second motor cortex, because all of those are defined by various caveats. And then to characterize behavior, which is by no means easy in the beginning, falling behavior, but later also our official behavior. So here's the example. You know, this has to be awake animals, and so this is a cortical feudal cell population labeled by transcription factor offensive to it has many targets from the spinal cord. To pass so and so forth. Uh, piers. This is the so called grandma area, and then so. There's a sort of a topographic organization, almost of different movement trajectories and the synergy, so these correspond to different colors. And then there are many other squares that are not shown here. So, for example, in this media, lateral media posterior region, you induce this lateral on a posterior movement with open arm open pomp and then this green area with a repetitive movement that we don't know quite interpret. But the sign region is becoming more interesting. Is it? May you can see rebounds number, some kind of reaching type movement with a partial closure off the palm. And then the red area is very intriguing because it is a different kind of movement where the mouse bring the shoulder the elbow on the risk of digit and also with the cemetery is a movement of the mouth and tongue so, you know, resembling eating. And so although they are, you can think that there are many ways of moving. But there are several trajectories and the synergies. We can quantify these using these movement vectors, so there's a there's a sort of the topographic map of movement, if you will, and the important thing is that we can then look at different cell populations. And that's where the sort of even more interesting findings came. So this is a flexing D one is a critical, critical, critical struggle population is categorically different from the output. Population is a population that immediate intra cortical on the struggle processing at least in the head fixed condition we don't see much movement in other regions doesn't mean that it doesn't do anything. But in this spot you see this really remarkable bilateral movement of the shoulder, the elbow, the digit on the official of the horse it resembles eating. And this is activating this, you know, about 300 microns. Region on the light effect can spread a little more for people who follow this. And I don't expect too many people in the mouse motor cortex. There is the so called kado following areas the roster column areas and then now increasingly famous hln from cars, a boat and others. So this is the area that is a relatively new We call them our fourth for Roscoe following or official area. Okay, and then s O. This is a population that is a cortical thalamic population that doesn't actually go beyond the thalamus. She didn't know what to expect, but nevertheless, he also found, actually, map on. I just show you one spot that this actually again, including a complex movement off the following and our official more jittery compared to the plexi on the more shifted towards the complex. I mean offensive to map and over currently, hypothesis that this might be intermediate than perhaps is more sensory aspect of the movement. That is, then send information to the modern area and the through the first two immediate these movement. And we have some evidence support that and this is the area we call the code or following or official area. And it actually is in the sort of sensory part of the traditional map. This is the or official in the secondary following cortex. But for many people who are in the modern field and influenced by the work of primates with these action map, what we have found is that these action map, if there is one, is really implemented by different private neuron projection types. With this critical feudal population, there is a more topographic organization of different actions. But then, at the higher level hierarchy, there are, elections and the others that perhaps reflect this critical strider thalamic organization off the basic architecture of, super hemisphere and for the last just to show you one image. So the head fixed is very convenient and powerful with being able to manipulate using light. But it's limited by, you know, just head fixed so what happens in the free moving condition. In this case, she first mapped the spot in the plexid one animal and then implanted the optic fiber. And what he saw was quite remarkable. Basically a stimulus. The insert a eating posturing sequence in the middle of uh so he was raising. Then the light turns out, you can see it stands back lower. The head raised the hand. And so and then even here, there's another example where this mouse is about to scratch his head with the Hei ling and then the light turns out and then again, posture gets into the teacher posture and then when the lights off, it goes back to scratched ahead. So somewhere, somehow, this population neurons really have a very powerful road to induce this behavior program. And the question, of course, many questions one is whether these cells are involved in this during this behavior. So going back to the head fixed the preparation we can now put camps in these different cell populations and observe in a behavior. And this behavior is shown here. Here here is also using the deep cut to track so The food palette is delivered by a motor belt, and then the animal will use the time to lick. And then because the food pellet, that's right, a big so it will always take a bite transferred to the hand and then wrote it and then bite until it finishes on. There is, uh, there are various ways of eating different kinds of food, and this is a relatively stereotyped program. But again, every eating event is different. We can then track his hand lift on the palate in the mouth, and then the manipulation episode on to sort of framed the Marissa brain activity. So here is a wide field causing imaging of the first of two. This is the critical Hugo output neurons. On. This is a more than 15 second video. So many activities, many patterns going on on the but there are key events. So here the palate is coming and then basically, you know, then this we don't know what to expect. They could be motivation, it could be trying to reach. But these two activity patterns in the proud of cortex in the frontal lobe and in the motor cortex is the one that really very nicely college with the hand lived, with a ton speaking basically to reach the food and then to raise the hand to manipulate the food. And so this is the so called pride or frontal or model loop that actually has been shown in monkeys for intention to move and to the preparation aspect. And we can see this nicely correlated with both the ton looking part, but also the uplifting off the funny. And there are various components. One is a related with peaking with time and with aren't lip, the other is the one that has a continuous activity during this episode. So another, again the power is the comparison. So for this cortical stride off and the critical critical, you know, it's a different pattern, and again there are many activities. But what? I want you to pay attention to these two spots so these two spots are activated when the mouse is using the mouth hand through this synergistic movement of manipulation, and it's quite even during tree. But whenever the mouse begin to manipulate using this coordination, those patterns and lo and behold, those are the two regions that when we stimulate generate is very complex movement. One is R4 the other is the CFO so suggesting that these two areas, actually maybe, in fact active during this moment on? Of course, I said to Eve Don't scold me way have to use after genetics, but this is an interesting behavior. So so she build a mouse restaurant. It's mainly for only dexterous mouth because they can start the trial from this lobby and then go down this narrow corridor. So in here, then the food, different kinds of food is delivered automatically to this region and there three cameras to capture how they eat the food. And the trick is to give them different kinds of food. If you give them just food pellets. If even if you disrupt, you know, cut out part of the cortex, there are hungry, so they will try to eat one way or the other. But you know, this is a pastor. This is a shelled seeds, and this is a simple sunflower seed with half of the side is covered with plastics on. Then you see a lot of interesting issues, but also these are difficult to track. So he actually build this, uh, dexterity array where these are 3D printed the cubes with the hole in the middle, and you can fill in the peanut butter and they like to pick up and lick. But then there are letters along the way so you can actually track how they manipulate this. And so and then there's a microphone to see how they bite. Here. You can see that this is again using the people have cut. He sort of a framed 6000 elevated 6000 frames, and there's actually works. So these red are the bites off each, Pastor, the pastor is a very good food item because it is difficult to eat, but rather the stereotype. So what, he found is that very interesting in this case, the bite. How always happens when these animals is pulling? You can see this reliable, uh, almost rhythmic behavior. And then with that, so this has been characterized by young wish on others. This eating program, it's complex. But there is a there is a organization in both movement of the mouth on the the hands on in terms of manipulation. So the first question the first experiment is relatively crude is just to put, uh, musical in this area to inactivate.