 As they said, I am from the Allen Institute in Seattle. We, what we do essentially is create gene atlases of brains. So we have atlases for mice and for macaque and for human. And so today I'm going to talk to you a little bit about that. But because it's atlases of the brain, I was asked to start off with some basic neuroanatomy. The understanding that a lot of people here come more from a math background or an informatics background as opposed to a neuroscience background. So we'll start off, I'll give you an introduction to the Allen Institute and sort of how we work and what our goals are. And then a basic neuroanatomy tutorial, essentially. And I apologize to those of you who do have degrees in biology. It's probably basic enough that there won't be any new information for you. And then the second session we'll move into actually discussing the different atlases and the types of data that we have available for use. All of our atlases are freely accessible. All of the data is freely accessible to the public. And then we'll circle back around and just a couple of quick slides on some specific anatomy that would be helpful for using those atlases. And then in the afternoon, we'll do some walkthroughs to look at some of the main data modalities, how to grab that data out of the web, and look at some visualization tools, and talk very briefly about the API. All right, so the Allen Institute for Brain Science is an institute with a goal of understanding how the human brain works in health and disease. So we use a team science approach. It's not quite like academia. We don't have a PI that defines what the project is. The projects are defined as an institute. And the entire institute's goal is seeing those projects come to fruition. We generate useful public resources, hopefully useful. We drive technological and analytical advances in the way that we display the data and use the data. And our hope is that we can discover some fundamental brain properties by integrating experiments, modeling, and theory. We're independent, nonprofit, medical research organizations, so we don't fund additional research. We do the research, and we make it available for everybody else with the hopes that we can cut back on some of the experiments that you need to do as pilot research for your projects. We're project-focused and milestone-driven. That means that for each of the projects that we have, we have definitive milestones and times for data release that we must meet. And again, an overriding goal of understanding the brain and health and disease. So why does this matter? These are stats from America. I'm going to assume that the numbers aren't going to change too awfully much throughout the world. But over 1 out of 4% of American adults have suffered from some sort of mental disorder, be it depression, bipolar, schizophrenia, autism, Alzheimer's disease. And the numbers are really astounding. So there are more than 1,000 disorders of the brain in the nervous system. And they result in more hospitalizations than any other disease category. And yet we know the least about the brain. So our goal is helping to understand how the brain functions so that disease models can flourish. All right. So oh, goodness, should have checked the contrast. We'll see how some of these pictures go. So our resources are used for advancing disease research in autism, Alzheimer's disease, epilepsy, obesity, multiple sclerosis. As I said, we accelerate the development of promising new research tools. We do things in a high throughput way. So the work that you do at the bench, we do on a very large scale. I want to say we are capable of processing ISH slides on an average. If we were to look at our daily capacity, could be around 600 slides a day. So if we were running at full capacity. And our resources also support classroom teaching. And as I said, our goal is to help you guys save time and resources by providing you data upfront. Our overreaching goal is in understanding these things down at the right consciousness, memory, behavior, thought, and how the brain works. And much like Jim said, starting with things like gene expression, electrophysiology, understanding the different cell types, the connections that they make, who talks to who in the brain that elicits consciousness, memory, and everything that makes us who we are. Yeah, we're really going to have to fix that. Yeah, I don't think that it's going to be a light thing so much as it is. Is it bright things? Oh, god, so cool. Yeah, well, they were working on that to get the pictures to work. So we like to say we use big science to tackle the big questions. We look at development to find out how things come to be. We look at the parts, the genes, the proteins, the synapses of different functional areas to see what's going on. We want to know how things are connected so we look at synapses or direct connections between cells. And we want to know how the parts receive, store, and act on information. And we also want to know what goes wrong in disease. So our tools are designed to answer these questions either in part and specifically at one of these levels or at multiple levels for a given project. What you can't see here is a brain. And these are data points showing gene expression in the human brain. So hopefully we'll get that fixed or else the pictures for the rest of this talk are going to be a little not understandable. And so before I go any farther, our team is currently up to about 225 people that work at the Institute. This is a picture from last year. There was probably about 150 people at this time last year. So we're in a huge expansion mode as we introduce some new projects. Yeah, let's see. OK. Oh, yes, that's much better. Yeah, look, it's a brain. We'll stick with this one. This is the one you're recording the slides, but. OK, but OK. You can probably. You can probably. Try to drill the camera so you can learn a bit more about the human brain. So that's your introduction to the Allen Institute and what we do. And we're going to, as I said, hit the atlases at the second session. But to understand brain atlases, you need to understand a little bit about the brain. So we're going to take a basic neuroanatomy walkthrough. Now, I'll admit two things. I'm an anatomist, not a bioinformatic person. And two, my focus is on the human brain. So although we do mouse brain stuff, a lot of this talk is going to be geared towards the human brain as well as the second one, looking at more of the human experiments. Mostly because that's what I work with on a daily basis. So I understand those tools a lot better. But some people like to think we're just big mice. I'm not sure if I quite fall for that. But it's going to be relatively ubiquitous across. So the basic makeup of the brain involves a cerebrum, this large part here. At the back of the brain is a cerebellum. And then the third component that you can think of is the hindbrain that consists primarily of midbrain pons and brainstem or medulla. If you were to cut this, you would see that what we have is essentially a cortex or this ribbon of cells that surrounds the brain. And then we have a subcortical region filled with nuclei that sits deep in the brain. And separating these are white matter tracts running to and from all the different structures. So the mouse brain has about 75 million neurons, about 4 million of which are in the cortex, cerebral cortex, and 100 billion synapses. The human brain is at a completely different scale. And you're looking at about 85 to 100 billion neurons with around 100 trillion synapses to account for. So human brain, incredibly complex. And of course, we all want to understand how it works, right? In the mammal, organization is relatively conserved, at least as far as the sensory motor cortices go. So what you find is that the olfactory cortex is in the front. Motor cortex tends to be towards the front, followed by a somatosensory cortex, touch, pain, perception, behind it. Auditory cortex is off to the side, and in the very back is the visual cortex. There's topographical organization to these cortices. So what you have here is the homunculus for the mouse. We have the sensory motor, or the sensory strip on this side, motor strip on this side. And as you come around the edge, what you see is that we have different areas devoted to different parts of the body. And for the human, we've got a large amount of our sensory cortex devoted to the face region and to the hand. And our motor strip, you have a similar mirrored homunculus with a lot of space devoted to the face, especially the mouth because we have to talk, as well as the hands, because we have a lot of fine motor skills with our hands. How this homunculus is set up is in parallel with those processes that we use a lot. So if you look at the mouse, a ghost bat, and a short-tailed possum, all of which have brains of roughly the same size. The mouse that gets most of its information through its whiskers and other touch sensors has an enlarged somatosensory cortex as compared to the visual and auditory cortex, whereas the ghost bat that uses echolocation has a much larger area devoted to the auditory cortex. And the short-tailed possum, which apparently has a really good vision, or so I'm told, has much larger area devoted to the visual cortex. So our brains are designed to handle what it is that we do. Or, contrary to that, we do really well what our brain is specialized for. There we go. For the human brain, there are essentially six lobes. We have a dividing central sulcus here. And in front of that is the frontal lobe. This contains that motor strip that we talked about. Behind it is the parietal lobe. Down at the bottom is the temporal lobe. And at the very back is the occipital lobe. These lobes fold over. And if you were to pry them apart, what you would see inside is called the insula. So there's a whole layer of cortex. And when you cut chronally, you can see the insula. And on the medial surface, there's this limbic lobe that's comprised of the cingulate gyrus and the perihippocampal gyrus become very important in disease scenarios. Now the bumps on the brain are called gyri. Cingular is gyrus. The divots in between those are the sulci or a sulcus. And each one, if you want to get down to the nitty gritty of gross anatomy, all have a specific name. And you'll see that in the atlases of the macaque and the human where everything is defined by their location. So in the frontal lobe, you'll have a superior frontal gyrus, a middle frontal gyrus, and an inferior frontal gyrus, a superior temporal and middle temporal and an inferior temporal gyrus. So the name gives you some indication of where you're at. Now the cortex is made up of layers. So for those of you who focus on more electrophysiology, you're well aware of this. In general cortex, there are six layers. And what you see here is Maureen Boyle has overlaid in pseudo-colored gene expression for genes that primarily exist in a given layer. And so what you can see very exquisitely is how well layered the mouse cortex is. The human cortex does the same thing. I don't have any pretty pseudo-colored images of the human cortex, at least not healthy human cortex. But what you have here is our sulcus. And we have a relatively cell sparse layer 1, a layer 2 that if you're used to looking at cortex, you might be able to make out. Layer 3, that's comprised primarily of pyramidal cells. Layer 4, which is small and granular. And layers 5 and 6, again composed of primarily pyramidal and multipolar cells. And each of these layers does relatively specific things. So layers 3 and 5 with the pyramidal cells are output layers. Layer 3, you can think of as primarily sending output to other cortical areas, either ipsilateral or adjacent cortices or cortex that it's going to talk to. Layers 2 and 4 are essentially input layers. Layer 4 gets input from the thalamus for the most part. Layers 5 and 6 output back to the sub-cortex, either to the thalamus or down into the brainstem. And these cells all talk to each other in essentially a columnar fashion. So the cells here are going to mainly interact, or I won't say mainly, but have a lot of interactions with cells in the same column as compared to cells outside of the column. So there's a vast network in the cortex where it's just exquisitely organized. And I think when we talk about the visual cortex, you'll probably be getting into a lot more of that. Not all cortex is the same. Just like all the areas are specialized, cortex is specialized. So here we have, essentially right from this part, a cut. And we have the central sulcus coming right through here. What you can see is that this cortex up here, the primary somatosensory cortex, very different from the motor cortex down below. Sensory cortex is going to have input from the thalamus so it has a well-defined layer four. Whereas the motor cortex is sending information out to the rest of the body. And parts of it are characterized by these incredibly large cells sitting in layer five. Those cells are sending axons all the way down to your tailbone to control leg movement. So that's pretty amazing for a single cell. So those guys are awfully big. And because the cortex looks different, people have spent a lot of time trying to parcelate out the cortex based on the arrangement of these cells. So over here you have the most popular cytoarchitectonic map done by Brodman in 1909. And overlaid on it is a separate way of looking at the cortex, which is by high order function. So, which have in blue are the primary sensory motor cortex. Yellow are unimodal association areas. The red or pink are high order association areas where a lot of thinking and computation goes on. And green is our limbic cortex, paralimbic regions that deal with emotion and memory types of things. Variable and not variable. So your motor cortex is almost always going to fall in front of your somatosensory cortex and it's always going to fall on roughly the same gyrus. The division between those cytoarchitectonic areas can vary quite a bit. So, the most variants that I've seen we actually had a case. One of the most prototypical things is this division at the central sulcus where you have motor cortex on one bank and just as you come underneath the bend of the sulcus changes to primary somatosensory cortex. And we had a case that actually motor cortex came all the way up to the top of the post-central gyrus where it had essentially pushed somatosensory way back. So, there is quite a bit of variance and especially when you get into areas like prefrontal cortex or the higher order association areas, those borders can move around a bit but you're never going to find your areas nine 46 of the prefrontal cortex back in the occipital lobe. So, the general pattern stays the same but there is variation and the same goes for the gyral structure of the human brain. Every brain is the same. Everyone's got a central sulcus unless you're diseased. A sylvian fissure that sets out your temporal cortex but every single brain is going to look different. And that's going to come into play when we talk about the atlases and the difference between doing a reference atlas for a mouse and how we have to deal with those discrepancies in the human. The Brodman map is not the only cytoarchitectonic map. There have been many cytoarchitectonic maps and part of what makes them different is probably the fact that they all looked at different cases but they also looked at different things and went to different levels. So, up here we have the Brodman map and down here is Vonokonimo which is probably the second most used and a lot of anatomists are pushing to go to the Vonokonimo map. Brodman identified 52 areas. Vonokonimo identified 107 and again, not much difference as far as how organization goes. So, a lot of it is a level of fine detail in what they were looking at. You can also parse out by myeloarchitectonic so you can look at the white matter tracks and how myelinated fibers are. This gives a very different pattern to the brain. You can separate out by functional network so you can organize by visual system and auditory system and high order functioning. You can separate out by neurotransmitter network so all of this comes from a wonderful chapter that Zillis and Ammons wrote in Nature. Ammons and Zillis are working on receptor autoradiography and using the different expression of receptors to further parcelate the brain. So, they'll go into what someone typically thinks of as a single Brodman area and separate this out into many more areas based on how the receptors populate it. So, as we move forward in time we're getting finer and finer and finer details of what's going on in the cortex. Each of those regions has a specific function but what really gives us the high level functions are the ability of the brain to talk to different regions. And so, that's all. This is a diffusion tensor imaging of the human brain. Looks at white matter tracks in the directions that they go. So, you can see this large track coming down here in blue which is the pyramidal track goes from motor cortex down to the spinal cord can make out the corpus callosum, all sorts of different tracks. So, what our behaviors are have to do with our functions and our networks, okay? And there's lots of networks in the brain, some of which people talk about more often than others. Probably the most talked about is language involves auditory cortex so that we can process what we're hearing and a connection to motor cortex so that we can move our mouth and produce speech that is understood by people. And you can damage this area, you can damage this area or you can damage the white matter connections in between and get what's called an aphasia where you either can't understand what's being said, can't produce words that are understandable or a combination of that. Whoops. There's a frontal parietal network that involves the ability for us to direct our attention either internally or externally to the world, right? So, in front of temporal dementias that damage the frontal part of this network people look internally, they don't really much care about the external world and you get some very strange behaviors like we had one woman who wore a cow outfit to work and was not Halloween. She saw nothing wrong with this, right? So, when you damage these networks you start having behavioral issues. The most studied network is probably the visual system and you're gonna hear a lot about that later in the week. Primary visual cortex is mostly on the medial surface of the brain but it does come around to this tail end here and spreads out and propagates across the brain. So, what we have are fibers coming from the eyes, going to the thalamus, the lateral geniculate nucleus actually and then coming back to the middle part of the brain. Primary visual cortex or V1 then sends processes to V2, to V3, to V4, to area MT, to a variety of areas in what are essentially two different paths. The ventral path is the what path and is involved with identification of objects, naming objects, types of things, whereas the dorsal path is the where path. So, again, this attentional portion of that frontopriatal network, now we're getting attentional where things are in space. Gets more complex. This is a diagram from Chun-Jin Wang who is now at the Institute heading up our Connectional Atlas. Here's your primary visual cortex in the mouse and these other areas are all V2. They're all subdivisions of V2 and they talk back and forth to V1. They talk to each other and in fact, this is essentially a network model of what's going on in the visual cortex and it's not even at the finest level of intricacy here. This is considered a relatively moderate level. I've got your visual cortex down here and a variety of cortical regions and you can see that there's a lot of crosstalk in those networks back and forth, not just within the visual system, but up to the frontal cortex of the frontal eye fields to area 46 in the prefrontal gyrus, area 36 in the temporal lobe that's involved in integrating with the limbic system. So, what's going on here in the brain is not a one-stop thing. What goes on here depends on what's going on elsewhere in the brain. There is some lateralization in the human brain, most often talked about with language, so we just showed the language network going from the auditory or vernachies area to the motor frontal brokas area and anatomically, you see this lateralization as well so this area of cortex tends to be larger on the left and sit at a different angle. So, the anatomy, the function, go hand in hand, but language is not just on the left hemisphere, spoken language tends to be on the left hemisphere, on the opposite hemisphere tends to be non-spoken language in a very similar way. So, in the back you have perception of things like facial expressions and emotional content of language and in the front you have the production of these same things. So, you can damage the right side, still be able to talk, but really not be able to know whether or not someone's coming up and telling you something humorous or whether they're serious or be able to interpret stuff like that. And like I said, the behavioral neuroanatomy of these networks follows along with disease. So, in Alzheimer's disease it's the medial temporal regions that control memory functions that are affected primarily and early on in the disease. Aphasia is the language system, you have disruption of that network between the temporal parietal region and that frontal region. Visual spatial disorders affect the regions of visual cortex back in the parietal and inferior lobes. Attentional disorders, ability to direct your attention throughout space as opposed to just one side shows up when you get damaged parietal lobes. And frontal temporal dementias that generally affect the prefrontal regions in the same medial temporal regions of Alzheimer's disease can give you a wide range of behavioral problems from wearing cow suits to doing stuff like hiking up your skirt to fix your hose or something like that out in the middle of public. You just don't really have this perception of the outside world watching you. Must be a big slide coming up. It seems to have locked up. All right, so that's the cortex. But as I mentioned, we've got a sub-cortex too. And so I'm gonna talk about the sub-cortex mostly so that as you're going through the atlases and you're seeing these words, if you're not familiar with neuroanatomy, you're not completely lost. Sub-cortex sits, as I said, deep inside the cerebrum for the cerebral sub-cortex. And you can see it progress through. And these are cells that are grouped in the clusters called nuclei. And the nuclei themselves cluster to form larger structures. So you have the thalamus made up of multiple nuclei that do specific things, hypothalamus, basal ganglia. And each of these nuclei are involved with different networks in the brain as well. So we'll start off with the claustrum, mostly because I would be remiss in skipping the claustrum or putting it anywhere else. Christoph Kott, who is our new chief science officer, has done a lot of work with Francis Crick in the realm of consciousness and understanding consciousness. And he works a lot with the claustrum. The claustrum is the sheet of cells that sits between the basal ganglia, which we'll talk about next, and the cortex. It tends to run into cortex and some people have wondered whether or not it's just a seventh layer of the insula. It integrates various modalities. We think it's involved in consciousness. And what Christoph likes to talk most about are his Marilyn Monroe cells. The claustrum have these very interesting cells that will fire and respond to a single concept. So you can have a cell that will respond when you see a picture of Marilyn Monroe, when you think about a movie of Marilyn Monroe, when you hear a sound clip from Marilyn Monroe, but it won't respond from Michael Jackson or anyone else. You've got a different cell that responds for that. So each of these cells is very finely tuned and is integrating data from a lot of different modalities. Striatum and basal ganglia are pretty much some of the largest structures subcortically. The striatum consists of the caudate and putamen, as well as the nucleus accumbens, which is considered the ventral striatum. When you add in the globus pallidus, which sits in here, you have the group of structures termed the basal ganglia. Sometimes the subthalamic nucleus and the substantia nigra are included in that. And these guys are involved in, there we go, quite a few different functions. So striatums involved in a motor loop, which is probably the most famous as this is the loop that is affected in Parkinson's disease. So there's been a lot of work with that. But it's also involved in an executive loop with the prefrontal cortex. So involved in executive functioning and day-to-day functioning of your life, as well as the limbic loop. So involvement with emotion and memory formations. The basal forebrain is this area sitting right in here. It consists of a couple of really small nuclei, primarily cholinergic. And the cholinergic innervation of these nuclei go throughout the cerebral cortex. It helps modulate activity within the cerebral cortex. So it's involved in a lot of memory and learning processes. It's one of the areas that has affected fairly early in Alzheimer's disease. And most of the therapeutics that are used right now for Alzheimer's disease that involve colon S-race inhibitors are designed to counteract the fact that the cells that send out the acetylcholine are in the process of dying and not sending out quite as much as they used to. So by altering the colon S-race activity in the rest of the brain, you can keep the neurotransmitter in the synaptic clef a little bit longer until these cells die. And then those medications don't work too well at all. The thalamus is the other big one in the cerebrum. It is the relay station of the brain. And with the exception of olfaction, all of your sensory processes go first through the thalamus before being divvied out to the cortex. And thalamus, as I said, is made up of quite a few smaller nuclei and each of those nuclei have individual subdivisions themselves. But each of these nuclei receive specific input of a specific modality and send it out to help integrate stuff. Motor directions also go to the thalamus before being shuttled out to the body for a lot of things. The hypothalamus sits ventral two and anterior to the thalamus. It's the other half of the diencephalon which is actually comprised of four components rather than just two. It links the nervous system to the endocrine system or the pituitary gland. And so it's involved with a lot of metabolic processes. It's secretes our neurohormones. It helps regulate body temperature, hunger, thirst, circadian rhythm, sleep. And it's interesting because there are nuclei in the hypothalamus that are sexually dimorphic, meaning that in males and females, it'll be different sizes or different densities of cells or nuclei that can fluctuate based on a variety of things. Hippocampus is technically, well, it's a cortical structure in my mind. There's always arguments about that. But people tend to talk about the hippocampus as an entity itself, so I'm gonna cover it with the subcortical stuff. You know, it's a continuation of the cerebral cortex. It is involved in memory function, input and retrieval. Memories are not stored in the hippocampus per se. They need the hippocampus to get memories in. You need the hippocampus to get memories back out. It's also the location of place cells which are studied quite a bit. And these are cells at fire when you're in a distinct position. So they do a lot of studies in mice thing where they fire when they're in a box. There is this layer of dense granule cells called the dentate gyrus. And it is involved in the hippocampus functioning, but also just underneath it in the subgranular layer of the dentate gyrus is where there is still some adult neurogenesis going on. So there are new cells being born in the hippocampus. It's a lot of studies going on with that. The hippocampus itself is divided into four main sectors, CA1, standing for Cornus Amonus or Ammon's horn, CA2, CA3 and CA4 that sits inside the dentate gyrus. Each of these sectors has multiple layers. Primarily the largest of these layers is composed of pyramidal cells. So it's called the pyramidal cell layer, but there's also layers of fibers running through there. There are thinner layers, the stratomorions. It contains very few cells and is essentially a layer that is a single cell thick. And then the final component of the hippocampal complex is the subiculum or the bed that the hippocampus sits on. And this is transition cortex into entorhinal cortex and the rest of the neocortex of the brain. As soon as it goes, it stops. All right, and then finally the last one we'll talk about and this is not in any way a complete list of all the subcortical nuclei. But the other big player tends to be the amygdala, which is the sister component to the hippocampus of the core limbic system. And it's involved in processing memory of emotional reactions. It's been shown in mice to be important in fear conditioning. So that's the cerebral subcortex. The cerebellum in the back. I think we've got a couple of lectures on the cerebellum as well this week. It's located here at the base of the brain and it, like the cerebrum, has a cortical and subcortical components. It's involved in fine motor control, so timing, precision, coordination. Damage to the cerebellum tends to give you different motor syndromes compared to, say, Parkinson's disease, where you have a constant shake. Damage to the cerebellum can give you issues with being able to fine tune at the very end of grasping types of things. While the cortex and the cerebrum is six-layered, for the most part, in the cerebellum, you have three main layers. You have a cell sparse molecular layer. You have a layer of large, bodied, purkinje cells that line up in a single layer and then you have this very densely packed granule cell layer made up of granule cells. The fact that this layer is so dense means that in the human brain of those 85 to 100 billion neurons, about 70 billion of them, reside just in the cerebellum. Here we have a nistle. You can see the large purkinje cells. This is an SMI-32 stain, which stains neurofilaments. These are actually some of the most beautiful cells in the brain. You've probably all seen pictures of them without knowing it. Beautiful tree-like arbor. Subcortex contains three to four nuclei, depending on how you name them, largest of which is the dentate nucleus, and these nuclei are involved in communication with the rest of the brain sending out to the cerebellum. The brainstem and hindbrain. There are different levels here. We've got midbrain, pons, and medulla. It's where the cranial nerve nuclei live. So there are 12 cranial nerves. With the exception of the first two, olfactory and the optic nerve, they all terminate back here in the hindbrain. So starting with the oculomotor, all the way down to nerve 12. They do integrative functions, so they're involved in cardiovascular control, respiratory control, alertness, awareness, consciousness, and damage to this is very bad for you when you can't have control of those functions. So it's a very important structure, and it's comprised of white matter tracks and individual nuclei, and there's quite a few of them. The only other thing to really touch on in preparation for the atlases is a bit of neurodevelopment because we do have developmental atlases that look at prenatal function. Developmentally, we start off with a neural groove. It folds into a neural tube. Got an aller and a basal plate on that. And then you can think of this tube as becoming segmented from one end to the other. And those segments go on to develop into different areas of the brain. And so essentially the prosimeric model is this segmentation and looking at how those different segments evolve during development. And so as you get farther along in development, those segments can be parsed out finer and finer and finer until we have individual nuclei. This is a little different than the way that we tend to talk about the adult brain. Additionally, because the cortex is still forming during development, we have structures that are considered transient structures, many of which go away either by birth or shortly after birth. Mainly the ventricular zone, which is where cells are born and the ganglionic eminences, the medial and lateral ganglionic eminences. These eminences are the birthplace for cells that are going to go become the striatum and become cortical inner neurons, whereas the ventricular zone is essentially the birthplace for cortical neurons and those cells are gonna migrate out and do what's called tangential migration around the cortex and then they're going to change their path and do radial migration of astrocytic processes until they reach the cortex. The cortex is born essentially in an inside-out manner so that first layer is formed and then the sixth layer and then as the cells are born, they layer in and push that sixth layer down. The result being that when you're looking at prenatal brains, we generally don't have a sixth layer cortex and so we have cortex layered differently. We start off with the ventricular zone. We've got a subventricular zone which can be divided into an inner and outer zone, an intermediate zone. There's a lot of trafficking of these cells as they migrate. There's a subplate zone where there's additional divisions going on and the cortical plate as the cells are moving up, differentiating and maturing, to the very outside we have a marginal zone and a subgranular zone. The marginal zone is likely to become layer one in the adult cortex. So the previous one was in the middle of the of the previous disease. The what? The previous one, the previous picture was in the middle of the pregnancy. Let me zoom back here. So this is, if I can. Of the other, but the previous one was around. This one? Yeah. So what you're looking at in the next picture is essentially right here. Okay, all right. So one of the questions that comes up and I'll touch back on this again when we talk about the atlases. Because there's so many different ways of organizing the brain, so many different ontologies people always ask, why can't you just have a single ontology so that I can go into the mouse brain and then look at it in the human or why'd that just call the same thing in the developmental stuff? And this is essentially what ends up happening. There's about 14 competing standards and someone comes along and says, well that's ridiculous, we need to combine them all and they make a system and now there's 15 competing standards. It's very, very difficult to get all these ontologies into a single entity and there are attempts going on. But especially when you throw that fourth dimension of time in there and you start dealing with structures that will no longer exist in the adult brain or you're looking at adult structures that just simply don't exist yet in the developmental brain, it becomes very, very difficult. Yes. I'm related to this big picture question of how we even approach this whole thing. This was throughout your very nice introduction that there's kind of like this tension between structure and function, right? And so there's a lot of description based on what different cell types do I see and then there's a lot based on, okay, if I knock out this part of the brain and it affects this capability. So I'm not in that area, but I have a sense that it's going more and more structural, is that right? You can tease apart more and more different cells. It really depends on which direction you're coming from, right? So the clinicians will continue to see syndromes which will indicate areas that are damaged. But if you're in looking at how cells themselves respond then you have to really look at the different cell types and see who's connected to who, which informs back to the clinical picture and what's going on. So it really is a combination of both and unfortunately those two camps until very recently haven't discussed much. So to speak. Maybe you head up the famous Vanessa and the thing of the usual cortex where it's like five to two million different areas with 50 million different things. And so I think there may not necessarily be a simple match of like structure to how a function is going to be done. And part of that is having to divide out function into cellular function and network function and behavior, right? Because those are all on three very different levels. And so even in talking to colleagues who for instance look at interneurons in the cortex primarily, to them if you alter the function of one interneuron now you've altered the circuit, right? From a behavioral perspective I'm not sure that one interneuron being altered is sufficient to affect behavior in the long run, right? So there is this question of at what point do we hit that tipping point of individual cells being altered to the point that they're not functioning correctly before it affects the circuit as a whole. So it is quite complicated. And again if you go back to the very beginning and where we're trying to get to in the end and the understanding of the brain you really have to take those other 10 things into account. Our initiatives in the future are moving towards more understanding how the cells connect to each other and the intracellular pathways that are going on in them. So we're kind of moving into the next step and the next phase for us as well in helping with this understanding. So hopefully within a couple of years the goal is to understand how every cell in the cortex works and who talks to who. And that's quite an endeavor. So the timeline is quite aggressive, I will say. And there's actually, if you wanna find out more about the future of that I'm gonna show you how to get to our YouTube channel and there is the press release and the press conference that was held as we started to get into this and you can hear Kristoff talk a little bit more about the objectives of this new program. In a few years in neuroscience doesn't necessarily mean two, so. But hopefully a lot earlier than we had. Just to press harder on this issue, don't you think that the Allen Institute is sort of uniquely situated to try to promulgate one or a small set of standards? Yeah. And many of us outside thought that some of the initial choices in the two systems, for example, were a bit idiosyncratic. Yeah. But regardless, be that as it may, it's sort of analogous to the early days of the genome where there were again competing standards and it's not like the adaptation is finished in perfect even for genomes. But still, I think there's room for it to be a lot better. And so this slide is kind of, I find too, you know, I mean, it's okay for us to say that, but for you guys to say that. It's not fair. Yeah, no, you know, our first thing that we are trying to address is not necessarily a universal ontology to use across everything, but ways to link them so that you can... There's more facility at moving between the different species and the different apocys. It is not a small endeavor by any account, but it is something that we have talked about and is kind of always in the back of our mind. Part of the problem is that for each of these projects, to do it to the best of that project's ability means that you need to take into account the limitations and structure the entire project around that. So for instance, with the human atlas, where there was a lot of talk about, you know, how deep do we go and what ontologies do we use, what atlas do you use. We had lots of problems with this. There's no major standard atlas that everyone accepts for the human brain. And yet for each of the major structures, there are gold standards that people will point to. And so we tried to pool those gold standards in for each of the structures. That doesn't include any sort of time reference. And when you move to the developmental atlas, since the focus is on time, you really need to address it in a time manner, which gives you a completely different ontology. So we have this sort of a, this is a great overview of the human brain. We obviously know a lot of it. There's also a lot of data that we don't yet have. So can you just think a little bit more about, you know, over the next four or five years, the types of data that we expect to have? Yeah, and some of that's in the second talk. So we are working on data for connectivity. For instance, we have a connectivity atlas that is in progress. It will be looking at injection sites in, I think it's like up to 500 different spots in the mouse that you can then look and trace back to see exactly where all the projections go to. That is going to go through the end of next year, I believe. And so that data is going to continue to come out for a while. We are adding on fetal time points to our macaque developmental atlas, which right now is all postnatal. And there'll be fetal time points to that. So kind of working with this developmental scheme and now being able to have a mouse development, a macaque development, a human development to move between. That's, well, yeah, so that will be microarray, I believe. So again, looking at all the genes and the ability to get a comprehensive picture. The new initiatives are more focused on things like electrophysiology and mapping that back in to different cell types. So looking at firing patterns and individual connectivity and synapses and finding a way to map that back to specific cell types. So in conjunction with our transgenics a lot to figure out how individual cells work and talk. So. That's quite labor intensive and the flexibility we're used to doing that. So how do you kind of pull up together with existing and how does that make sense to you? Yeah, I'm waiting to hear, you know, we've all got ideas but that initiative is just getting underway and so there is still a lot of discussion at this point on exactly which directions it's gonna go and how we're going to be able to integrate a lot of different things. There will be EM, there will be calcium signaling and stuff so hopefully and again a lot of work with these transgenics that have been developed so that we can look at very specific cell types that are highlighted by the transgenics to get a better understanding of individual cell properties. So fingers crossed. But yes, a very large initiative. They're looking at doing that in the mouse visual cortex as in mapping out all of V1 in that way. So and hopefully I haven't given away too much that's confidential and I don't think so. But I think that pretty much and now we're not too far behind now but other questions about the anatomy and again we'll start diving into the atlases after the coffee break so you can see what kinds of data we have. So we have some damage in some part of the brain we lose some functionality. So what is the plasticity for example? Is it possible, I mean, what are the cases when some other part takes over? Yeah, part of that depends a little bit on development. So for instance, you can take a small child that has epilepsy and needs a surgery to remove half the brain and you would expect this child to not be able to move one side of their body and they will develop into a fully functional individual if that surgery is done early enough. Later in age when you start looking at stroke that ability to reorganize is not nearly as strong. And so there are some studies that show that with therapies some of those areas that were affected surrounding areas may start to take over some of the functions but I don't think that that happens on a timescale that would necessarily get us to for instance recovery of stroke in old age type of thing but there is a lot of plasticity that goes on and it's very different to say change the plasticity between two cells that are connected to each other versus changing an entire network to do something else. So the plasticity is there. The capacity changes as we age. It lessens somewhat as we age but I don't know if we were to live till we were 200 maybe we would be able to see some of that but there are functional imaging studies that look at shifts in function in surrounding tissue to strokes and it does look like there is something that goes on there. Not quite clear what it is.