 All right, it's my distinct pleasure to welcome our next presenter. Dr. Susan Buchheimer holds the Joaquin Fuster Chair in Cognitive Neuroscience and as a professor in the Department of Psychiatry and Behavioral Sciences, David Guest in School of Medicine at UCLA in the Department of Psychology, where she is director of the Staglin IMHRO Center for Cognitive Neuroscience. Try saying that three times fast. She is a clinical neuropsychologist whose work has spanned both basic research and clinical practice. Dr. Buchheimer's autism research program uses functional magnetic resonance imaging, or fMRI, to try to understand what differences in brain function give rise to the major symptoms of autism, especially in language, social communication, joint attention and emotion, and integrates imaging with genetics to understand how autism risks genes all to the trajectory of abnormal brain development in autism. Dr. Buchheimer is the principal investigator of the UCLA Autism Center of Excellence. Dr. Buchheimer received her PhD from Wayne State University in 1988. Please welcome Dr. Buchheimer. Thank you so much. Thank you for that very kind introduction and for the invitation here. And thanks for the family for putting together this incredible series of symposia. It's really remarkable and it's a great pleasure to be here. And thanks to the board for also choosing me. Am I loud enough? Can you folks hear me? A little louder. Can we get the, do you jack up the volume? I'll talk pretty loud, but not quite that loud. Okay. So a little change of course. We're going to be talking about mostly brain, brain and genes. I have to, is that going to, it is not working. Can you see why I'm not advancing? It sucks. There we go. Sorry. If I swear during the presentation, I'm going to apologize in advance. I actually had a series of concussion about a year ago. And although I'm much better, the major side effect is that I swear like a pirate. I'm trying really hard not to, but it just comes out that way. Okay. I have no disclosures. I don't work for drug companies or accept money from drug companies or anything. This is, advanced or is so not working. Can I just ask you to advance? Because I don't know how else I can get this work. Okay. All right. So I'm going to be speaking to you from a brain perspective. And from a brain perspective, autism is a disorder of developmental disconnection. I'm going to try to explain what that is and why that is. And so I would describe, therefore, autism as an autism spectrum disorders as arising from differences, not disorders, but differences in the trajectories of brain development that affects the formation of brain pathways that are essential for developing these core skills in social communication, joint attention, social communication, et cetera. Now, the differences in brain development, as we know, are governed by numerous genes. I mean, many, many thousands upon tens of thousands of genes. But I think it's important for us always to recognize that experience modifies the expression of genes in a very direct way, and it modifies the development of our brains. And so it's important not to think of having, talking about genes as being somehow a stamp of permanence. That's not the case at all. And I'm going to show you some data to support that. But the main point I want to make in the overall message here is that in our view, understanding the brain in autism empowers us to intervene at a more fundamental level than we do. Just use this as a really quick example. One of the deficits that we typically see in individuals with autism is face processing. They have problems in face recognition and face processing. Well, why is that? You see the behavior problems in face processing. But how many ways can you get that behavior? You could have a primary deficit in visual processing such that one can't process configurable information in only visual details. You could also have another primary deficit in visual processing in visual association areas where the brain's center for face processing is not specialized the way it's supposed to be. Or you could have a problem with the motor system that directs eye gaze. And if your eye gaze is not directly, is not appropriately directed where it should be, you're not going to be able to see faces and track eyes appropriately, and therefore faces may not be so interesting. Or you could have a deficit in the reward system, which I'll talk about in a little bit, where there aren't, you're not getting, the brain is not receiving reward signals from looking at faces. Therefore faces are not interesting and therefore they don't look at them. Or it could be that looking at faces creates anxiety. Looking at faces creates a little bit of anxiety for all of us, but maybe there's such a sense of anxiety that the children seek out away from the face and therefore they don't look at faces and therefore they don't develop expertise in face processing. So I've just described five and I can actually come up with half a dozen other different ways that we can see the exact same behavior. But what will be happening in the brain is extremely different and consequently what we'd want to do to intervene would be completely different because how we're going to intervene on a visual problem is extremely different than how we're going to intervene on a reward system deficit, okay? So that's why I think it's so important to study the brain. And let's see, can we go to the next slide? Okay, great. So I'm going to be talking a little bit about finding genes for autism. I'm going to talk more generally about translational research, which is something I hope that most of you have heard a lot about, the NIH's funding, translational research and why we do it and how we're trying to use these kinds of models to link genes, models in a system and an animal system to what is happening into a human in order to make it possible for us to learn fundamentally what is going wrong with the brain and hopefully develop fundamental interventions that will affect whole systems rather than individual behaviors. That's the ultimate goal. So I'll talk about how we do that, about brain connectivity and biomarkers of early intervention. Okay, and so here's my first video. I'd like to start with a video of a child. This is a high-functioning child. And let's just take a look at this really cute little guy. See how smart he is. This is his mom. That's with him. They're interacting together. And she's giving him nice rewards. And he's paying attention to his objects. As you can see, he said a little word at the end there. He couldn't quite make out what it was. How many people can see this child as having autism? Okay, so it's not very obvious, but this child does have autism. He's very object-focused. He's not interacting with his mother. He's maintained no eye contact, never looked at his mother once during this whole episode. He's very, very bright. And you can see how good his visual skills are. He can put all those blocks in so nicely. But there's no social interaction here. There's no social communication. So these can be very, very subtle problems. And yet, over the course of development, if you're not interacting, it's going to affect every aspect of development, including the trajectory of brain development. Okay, back one. Back one. Did I go back one? Okay, well, I'll just do here. As the previous speaker mentioned, autism is one of the most genetic of complex behavioral disorders. At least 75% of the variants is accounted for by genetics. But the genetics are very complex. There are two different ways that we can look at genetics to study disorders such as autism. One is the common variant approach, and the second is the rare variant approach. In the common variant approach, we take advantage of gigantic studies, huge gigantic studies, tens of thousands of people, maybe family association studies, and look at which genes are found in individuals with autism compared with those who do not have autism. And these approaches have been successful in identifying over 50 and probably several hundred genes that confer a risk. But they're not causative. They confer an itty-bitty risk, an almost immeasurable risk. All of us in this room have these risk genes. Then there's a rare variant approach. It turns out that genetic accidents, mistakes, duplications, deletions, things like that account for, we used to say 5%, and now we believe it's going to be up to 20% as our genetics gets better. So this is kind of like the same thing that you would see in Down's syndrome where there's a clear genetic accident that occurred. We believe that this probably occurs in now close to 20% of autism, and it just depends on where this accident occurs in the genome to see what kind of features they have. But what's really interesting, though, is even when you have these syndromes like Rett syndrome, Fragile X, Timothy's syndrome, et cetera, where there's a high incidence of autism, it's not 100%. It's not even 50%. It's like 30%. So even when you know you've got a gene that's hit right in an area that's very high risk for autism, there must be an awful lot more that is going on for that gene not to have this dramatic effect, to have so much variation in this. Oops. And so a model like this, and is this a pointer that works? No. I think this is out of batteries, actually. So in a model like this, you can see that we're not looking for genes that cause autism. Rather, we're looking at genes that may increase the risk of autism and there's not one, there are a whole bunch of them. And if you look carefully, these genes can have very different kinds of effects. For example, you could have simple additive effects, or you could have multiplicative effects where having one gene is a little bit bad, having another gene is a little bit bad, but having them both together is really, really bad. You could have genes that have no effect at all unless they're together. And then we also have another wealth of genes that are often ignored that are resilience and repair genes. So that even if you have some of these risk genes, if you have the right kind of resilience and repair genes, those genes may never be expressed. Because you can have a gene that's a high risk gene and never have it actually do its job. A gene actually has to be expressed. It doesn't just sit there, it makes proteins. And there are many factors that determine whether a gene is actually going to make its proteins, including some environmental factors which we're trying to identify. Okay, I'm looking for the right angle to do this. So in translational research, what we tend to do is take either of these two genetic approaches that I just described, and then develop an animal model. We have great skills to do that. We can find a gene, take the gene, just knock the gene right out of the genome. We can knock it out, knock it back in, do all kinds of really cool things with individual genes. And then you can see what the gene actually does. Where is it expressed in the brain? What protein does it make? When does it have its effect? And really understand what the genes are actually doing. This will allow us to then look at animal models, see if we can replicate some of the behaviors in animals, and then find ways of measuring these effects in what we call biomarkers. It could be substances that occur in the blood. It could be changes that occur in the brain. And this in total will allow us to go in and understand the root causes and the whole pathway from the risk gene to what's happening in making proteins to where it goes in the brain, to how it affects brain development, to how that ultimately affects behavior. That's our ultimate goal in translational research, to go from gene to disorder to treatment. So that the treatments will be not only informed of the cause of the problem, but that someday we'll be able to get treatments that actually affect the expression of the gene before it's ever expressed. Our future goal is to prevent autism from ever occurring. So, okay, whoops, twice. Okay, there we go. One more. Okay, so this is a long list of genes. You don't want to memorize this list of genes. This is from a review study that was done a few years ago. The list is actually about three times this size now. These are autism risk genes that are sort of ordered by the extent to which there's evidence that they are autism risk genes. And so when you see this many genes and there are now several hundred of them, it might make you despair. You think, oh my gosh, how are we ever going to find out about autism if there are hundreds of genes? We're going to look at every single one. But it turns out that if you look particularly at the really high-risk genes and look at what they do, and these are descriptions of what they do, I won't bore you with the descriptions of the FoxP1 transcription and the mTOR pathway and all that stuff unless you really care. But what we've noticed is that in these high-risk genes, the genes actually share common pathways so that most of these genes line up along the same biochemical pathways. And if you look at their functions, the functions are very, very specific. They are virtually all having to do with cell adhesion, neural signaling, synaptogenesis, neural migration, dendritic growth. To summarize that in non-neurobiological terms, these autism risk genes share the common feature of being involved in making neural connections early in development, including prenatal development, making connections. And so making connections is really the theme of what this research that I'm going to be talking about is all about. Let me go back to my neurons. So connections. This is a newborn brain, and these are neurons. So these little things right here are neurons. And these are the projections that are starting to make connections in the brain. As you can see, you've got neurons, but not a lot of connection. Here is a brain of a six-year-old. Same number of neurons. Look at all those connections. It's like this gigantic web of connections that have been formed. There's incredible growth of these white matter processes that connect neuron to neuron all throughout the brain. And then look at happens at the age of 14. It's thinned out. So there's this very interesting developmental trajectory where we start out with just neurons, and there's this explosion of forming connections all throughout the brain. And then those connections become pruned and honed by experience into modules that are functional modules so that we now have more efficiency in the brain. That is, connections that aren't needed, that aren't used, they're going to fade away. The connections that are used, they're going to stay. So this is how experience modifies the brain. So if you just look at the difference in what we see here and what we see here, this is at birth. This is what genetics gave us at birth. This is much later. You can see just how much what happens during development affects brain. And that tells us that what we do in interventions is crucially important. So I'm going to just show you a little bit of animal model work on how these connections are formed in these cool experiments where we can see them. And then I'm going to go over to human brain. So how do we do that? Here's a little mouse, and that's a little mouse brain. And here's a close-up of there's a little mouse brain, there's a little mouse skull. And so what they can do now is they put a little tiny, they take a hole in the mouse skull and shine it onto the mouse brain with a camera. This is a really teeny, teeny camera. It fits right into the mouse's skull. It shines down into their brain. And due to some other really fancy things that we can do with something called optogenetic imaging, they secure these cameras to the skull, the mouse is still alive and wandering around, and they take pictures of neural development. So we can actually watch the brain growing in real time in the first, in this case, eight weeks of life. And in this case, what they are doing is using a knockout model. So this is an autism risk gene. It's called P10, but that doesn't really matter. P10 is a gene that regulates neural growth. It prevents overgrowth of dendrites. So as these processes start and they grow out and they try to find their connections, it has to be regulated so that there's not too much. When this gene is very dysregulated and it can be so in adulthood, it can cause brain tumors. In development, however, it helps to regulate the overgrowth of these processes. So this is a eight-week postnatal. This is a 12-week postnatal. This is the difference in what happens in just those four weeks in dendritic connections. And... Okay. This shows us the difference between the knockout and the what we call wild type. So this is a control. So here we've been watching the neural growth in this autism risk model, okay, an animal model of autism of dendritic growth. And this is what the knockout looks like and this is what the control looks like. And what you can see is that the sprouting in this part of the neural tree is huge. It's over-sprouted right over here. It's not over-sprouted in the bottom. So this is called the basal dendrite. This is called the apical dendrite. But what's important is that here there's tremendous overgrowth in what is going to turn out to be local circuits in the cortex. Whereas here on these long-range connections, there isn't overgrowth. In fact, if anything, there's a little bit of undergrowth. And I think that you can look at this and you can see that this is not as well-organized as this, right? This looks nice and neat and well-organized. This is not. So what we're seeing is a particular pattern in this animal model that suggests that there's an overgrowth of these white matter connections in these local circuits in the cortex and that there is an undergrowth or a disorganization of circuits that are long-range that go far away. So local overgrowth, long-range undergrowth. So that's the animal model. What do we see in the human? I haven't found my angle yet. Okay. All right, so thus far I've shown you that there are many, many autism risk genes but they share these common properties that they signal neurons to find their connections. And in animals, we can see the effects of altering these genes and how they affect the brain and they affect it by changing brain connectivity. And in a very specific way, they do it by increasing local connectivity and either having the same or less long-range connectivity. So human brain. Okay. So we can image the human brain in vivo using two techniques, structural MRI and functional MRI in children. It's completely safe for children, thank goodness. In structural MRI, it's what I think many of you have had MRI scans, it's the typical MRI. We can just take pictures of the brain and measure the anatomy. We can measure the size of structures so that would be the gray matter, the neurons in the brain. We can measure the connections between the structures and that would be the white matter of the brain. And we can use this other technique called functional MRI which I'll talk to you about, which measures the brain in action and measures the brain while it's actually working. I'll show you how we do that. And so we do this in tasks that individuals with autism find difficult to perform. One of the first findings in structural MRI was the finding of brain overgrowth. And this is a dramatic example from Erica Shane's group of a case with autism and this is the average normal brain of the same age and you can see that's a really, really big brain. Now, on the other hand, if you just look at the MRI scans, you can't see any difference between a child's, the brain of a child with autism and a typically developing child. They look the same. I've seen hundreds, thousands of them. They all look alike. However, if you look across a group of subjects and you look at the trajectory of growth, we do see some differences and the main difference is that there is brain overgrowth early in life but there's a slowed growth trajectory. There are also some differences in the sizes of different structures and it's not so clear how they're related to autism behaviors but let's see what we can learn about that. Okay. This is a brain size picture from Erica Shane's group. This is the average size brain. This is the brain of a group of children with autism. So mostly they're close to average. There are a few outliers here. The brains are bigger but how are they bigger? Why are they bigger? This other structure here is the amygdala and many of you have heard of the amygdala. It's a very important structure. It's our primary fear center. It processes emotion at a subcortical level and these are growth trajectories of the amygdala in kids with autism and control groups and these are severity groups. This is the control which shows that this particular structure increases in size but in the individuals with autism, particularly the lower functioning children, it either doesn't increase in size or the trajectory is lower. So the trajectory of brain development is less but initially it's bigger. So we have this pattern of there's too much growth very early in life but the trajectory is not normal. So we have two problems. Now, what is the cause of the overgrowth? What is the cause of these enlarged brains? Well, this is a great study by the Boston group, Martha Herbert and she looked at the white matter of the brain. So these are the connecting tissues of the brain and she divided up into a whole bunch of different subcompartments but in general, she looked at these local areas in cortex and these long range connections that connected across the front to the back or the side to side of the brain. All right. What I'm going to show you is a boring graph. Sorry. Developmentally delayed control subjects by the way but the main thing that we can see is that white matter is overgrown in individuals with autism as well as some developmentally disabled but it's very specific to these superficial parts of the brain the local circuits and in fact there is less development in these long circuits. So that's the same thing that I showed you in the animal model an overgrowth of local circuits and a depletion of long range circuits and it's particularly true by the way in the frontal lobe and the frontal lobe is where there are a lot of very key skills that are impaired in autism all of those executive function theory of mind things that we talked about earlier. Okay. Now what's really interesting most of those studies are done on children a little bit older but Joe Piven's group has looked at the growth of white matter in infants before those are diagnosis of autism so we can't diagnose autism reliably until children are about two years old but what we can do is look at children who are at high risk for developing autism by virtue of having a sibling with autism the chance of getting autism if you have a sibling with autism increases quite substantially about 20 times. So what he did was he looked at children who had or had not a family history of autism in an older sibling and looked at the growth trajectory of these white matter fibers and this happens to be the corpus callosum and what you can see is even though there's a lot of variance across individuals starting as early as six months you can see that there's a difference initially and a difference in the growth trajectories these are the kids who are at high risk for developing autism and these are the typically developing children so what this tells us is that as early as we can measure probably from birth there is an overgrowth of white matter in local circuits and that the developmental trajectory decreases later on in the lifespan so how do these differences in brain size and things like that relate to behavior and a lot of it is actually related to IQ and if you control for IQ many of these differences go away the differences tend to be small and there aren't really clear relationships to autism symptoms and these differences are so small that as I said you can't just look at a brain and say well this is a brain with autism so we've been shifting our focus to look at brain function rather than brain structure how the brain is actually working with specific behaviors in autism and we do this with a technique called functional MRI I think many of you have heard about functional MRI now because it makes the news a lot it's a technique that maps the brain at work more specifically it measures where blood flow increases during task performance and it's very cool because it's non-invasive there are no needles, there are no drugs there's no radiation and that means it's safe for children it's safe for babies, it's safe for repeated scanning so increased activity during a task in comparison to a controlled task that's kind of the downside because it means that the kids have to be able to participate and that will limit what kind of children we can see this is what it looks like, this is a picture of one of our scanners and that's the daughter of my colleague Dr. Marela de Predo she's much bigger right now a lot of people ask me how on earth we get kids with autism into an MRI scanner so we get them into an MRI scanner and the magic of these goggles that are virtual reality goggles they have miniature TVs that fit right over the eyes and when you've got them on you can't see anything around the environment all you see is this huge big 3D environment and so we put them on and then you can put the kids anywhere they don't even know that they're inside it too after that and so they've got headphones they've got these goggles they can watch movies and then when we're ready we just shift it over to a computer and we make them tasks and we try to make the tasks wherever we can kind of like games it's noisy but they can usually do it so there are a number of things that we've studied and many groups have studied and there's no way I could talk about all of them but I'm going to talk about a few primarily related to emotional understanding and emotional responsiveness reward processing language and communication and sensory responsiveness and sensory hypersensitivity and cognitive skills I just don't have time to go into all of them so let's look at face processing studies so faces as you know are very special and very important in social communication they convey very important information about the emotional states of others we develop bonds through mutual gaze particularly with smiles so we have a lot of questions we want to ask and I raised some of these earlier is the perception of faces the same is it the response to facial emotions that is altered what about the gaze the gaze perception in individual with autism so I'm going to show you studies of all of these things okay so this is a very simple study where this is a group in Boston showed pictures of faces and pictures of objects to kids with autism and typically developing kids and this is a picture of the bottom of the brain along a region called the fusiform gyrus which is an area of the brain that processes visual information and it has a face specific area in fact it's so uniform in people we call the fusiform face area and in controls it's located on the outside of the fusiform gyrus and it's bigger in the right hemisphere than it is in the left hemisphere and then if you look at faces versus objects and I drew a region around the area so you can kind of compare that faces are on the outside and objects are on the inside in the typically developing brain and this is very very regular across individuals in individuals with autism you can see that they've got some face area activity but not as much and you can see a lot more activity outside this face area and the object area is pretty much exactly the same and so this led some people to think well geez maybe the organization of these perception areas in the recognition centers in the fusiform gyrus are wrong or somehow working incorrectly in kids with autism and that was the first idea that there really is a visual information processing deficit well it turns out not to be true so this is a great study by Karen Pierce in San Diego and what she did was she took the same faces but she showed kids also faces of people she knew well like their parents so they looked at the differences between stranger faces and the more salient family faces so what you can see here is if you look at the stranger faces there's a big difference between the kids with the autism and the typically developing kids but next slide oops I did not turn out can we go back one okay um if you look at the familiar faces however they're very similar so it couldn't be that the problem is that the face area is not working rather that the stranger faces were not very salient and not very important so they didn't engage and once you started showing them faces that were important they did engage so now we know it's not that the face area is broken there's something else that's involved processing that's important so now we have just ruled out a problem with visual perception which is very important because there was in fact a group that developed an entire treatment system based on visual perception training and autism and it failed and that's why so what about face emotions um here's a very simple task that we would give to children with autism we show them faces like this that have strong emotions and they are to match the face that has the same emotion not the same person the same emotion and in a related task they would see a facial emotion but this time they'd have to say well what it is they'd have to analyze that emotion and this is a simple control task so these are matching the emotions figuring out what the emotions are identifying them in a control task so one of the things that is found particularly in typically developing individuals is this by the way is the face area and this is the amygdala that's our emotion area and then individuals see these fearful faces their emotional areas their fear center reacts and here it is lighting up in typically developing kids and not in kids with autism so this was consistent with a model that another group had posed that the amygdala was broken in individuals with autism and that they didn't respond to the kinds of emotional stimuli that were out there that kids did and that's why they had problems with face processing and social problems but if you look at the difference between processing during the label task versus processing at just the observed task you can see a very different pattern typically developing kids down regulate their amygdala when they are analyzing emotions their frontal lobes kick in and they adjust it that doesn't happen with kids with autism and so in fact in this condition the label condition the kids with autism actually have more activity in the amygdala so again the problem is not that the amygdala is not working it's just not working in response to the same kinds of conditions now specifically it's not very as reactive in looking at faces and it's not regulated by the frontal lobe the way it should be it's not down regulated when it's trying to do analysis so it's not that the amygdala is not working but there's some problems in the connectivity with other areas of the brain in the context so we ruled out another problem okay have people heard about the concept of mirror neurons here yet many of you imitation is a very critical skill for learning language for learning most social skills and mirror neurons are very important in this process there are ones that fire during the performance as well as the observance the observation of a meaningful motor behavior and they're thought to be important for analyzing and understanding the intentions of others how do you understand the intentions of others well we do it by internalizing them and experiencing them as if we're doing it ourselves so when we see somebody else smile we kind of feel good inside we experience their emotion we feel it internally ourselves and that's what mirroring is very important function so by colleague Morella de Pretto did this study where she showed kids with autism and typically developing kids faces conveying different emotions and she would tell them either just look at the face or imitate the facial expression and just to make sure that we knew what the kids with autism were doing we actually videotaped them making expressions and they had some blind raters going in to see if they knew how to make the expression see if they were actually doing the same thing and they do and the raters couldn't tell which kids were actually had autism which kids were not just based on their ability to imitate the facial expression so they actually can do this task but how do they do it how does the brain do it so here's some great brain pictures I love the brain alright so this is the outside of the brain this is the left hemisphere this is the frontal lobe this is the occipital lobe and these are where the visual areas are this is the temporal lobe so the visual areas are here in the back these are motor areas too by the way right in here so when you're imitating somebody's facial expression you're moving your face so obviously your motor areas are going to be lit up so typically developing kids they have activity in the visual cortex the activity in these motor areas and they have this additional activity in the frontal lobe and in this particular part of the frontal lobe is if you're a frontal gyrus on the right side it's a mirror neuron area on the left side it is not quite as much it's also a language area here are the kids with autism they're seeing things fine they're moving their faces just like the typically developing kids but they're not mirroring what's happening emotionally and this is the difference between the groups in other words they're able to look at the face they're able to imitate the facial expressions but they're not internalizing the meaning of that facial expression they're not feeling it themselves they're not mirroring that and that's something that we have found consistently this is on the outside of the brain if we look on the inside of the brain you probably haven't memorized all these brain regions this is the amygdala our fear center and these are areas that are greater in typically developing kids compared to autism and what we found is that in observing fearful faces the kids typically developing kids show a lot of activity in the fear centers the kids with autism not nearly so much they don't automatically experience fear when they see a fearful expression whereas you and I when we see a fearful facial expression we feel it you get it inside these are reward centers which I'll talk about a little bit more in what we call the ventral striatum in response to happy faces and so when we see happy face we get a nice warm feeling inside these are differences between groups so the kids with autism do not get that warm feeling inside and this is an area called the anterior insula that's important in salience which I'm not going to talk about now but I will talk about later what is really interesting however is that the less activity in this mirror area the worse the symptoms of autism this is a correlation between our ADAS and our ADI scores and activity in this area so the less the brain mirrors is able to have this mirroring functioning of internalizing the affect of somebody else the worse the outcome is in terms of autism severity so this tells us that we do have an area of the brain that appears to be critically important in autism and that is important in understanding some of the symptoms that we've been talking about and it's much closer to the problems that we're having in face processing than those other things that I'd already talked about including perception gaze processing this is from a former student of ours and being able to monitor somebody's gaze monitor their emotional expressions and make sense of their gaze is extremely important and it's something that is impaired in autism and if you look at eye tracking patterns of individuals with autism they are not tracking for example the eyes and other things that are very critical in what they're viewing they tend to look at other places so let me do that so we put together this functional MRI study with direct versus averted gaze and we did it in this very particular way so we have different facial emotions and then we just went in and photoshopped grab the eyes and just shifted them just a little tiny bit over just a couple of millimeters the faces are exactly the same the only thing that's different in these faces is an artificial little tiny shift of the gaze direction when you look at this one does it look like he's looking over to one side does it look like his face is turned it almost looks like he's turned his face his face is not turned there's a misperception on our part because the gaze is looking over there and so we interpret that he's already turning to look over there that in fact is not happening anger a mad face like that has a very different meaning if it's looking straight at you then if it's looking at the person over there if somebody is angry at that person it's like if they're looking at you with that face that's a real problem and of course smiling directly at you that's a very nice warm feeling and we process that internally so these are typically developing kids so this is a typical pattern and you will see our old friends in the frontal lobe here right same visual regions as before because they're looking at faces but the frontal lobe is active when they're looking directly at these faces same faces tiny aversion look what happens in the brain you're seeing the exact same thing with just a couple of millimeter shift in the eye gaze pattern and the entire brain changes its activity pattern because what the brain is signaling to the typical developing kids is this is not relevant to me it's not important I don't have to internalize this whereas when you're looking straight at somebody in the eye it's very important and all of these mirroring areas and theory of mind areas start coming on board they come to play they say this is relevant to me I have to look at this when we look at the kids with autism direct or averted it doesn't make a difference they are simply not processing the significance of the direct gaze okay and we actually controlled for gaze controlled for where their eyes were in this experiment to make sure that they were looking straight at the eye region so they're looking at the eye region they're simply not processing the meaning or significance of it so that's interesting because this is an automatic process we don't have to tell ourselves oh think about what that person is thinking right now this goes on it's our natural brain's way to process this information to process the significance of what we see in our emotional and social world without having to think about it it just goes on automatically the automatic aspect of that is not happening in the individuals with autism how does that affect language language as you know is a core deficit in autism it's no longer on the DSM but we all know it's still a problem and what's interesting is that early language development is implicit it's learned automatically and it is learned with the help of eye gaze cues from the parents so if you've got a difficulty in following these joint attention eye gaze patterns then you're going to have difficulty in developing language so how do you test that in older kids who've already developed what language they're going to do well we create artificial languages and look at how the brain is processing language information so let me just show you this task can everybody see this okay it's kind of hard to see what we do is we present in three different conditions a bunch of sounds made up words so the sounds will go something like this this is what the kids hear it's like really boring they just sit there and listen to this but unbeknownst to the kids in some cases we repeat the same three syllables in a row several times over the course of the experiment whereas in the control condition it's just random so without their knowing it they're hearing words and they're learning new words just because by chance these three sounds are always going together but they don't know it it's just happening in an unconscious level and so how do we make kids do this because it really is boring we call it the alien languages task and so they see pictures of these little aliens and we say just you're going to see these different aliens and just listen to their language and so they listen to these sounds and they don't have to do anything so we could do this with pretty low functioning children and very young children and just see how was the brain responding when they simply hear sounds that are repeated again and again in extracting language information simply based on the fact that they tend to occur together these sounds so this kind of implicit learning is governed by a network again same area in the frontal lobe we've now seen this three times same area in the frontal lobe as well as these subcortical structures and this is the basal ganglia this happens to be the dorsal striatum part of the basal ganglia it's in a very important area frontal striatal connections are very important in language learning and what we do is we model an increase over time in how these areas are active this picture is a picture of the difference between typically developing and ASD children so what we're saying is that the typically developing kids show much more activity in the frontal lobe the temporal lobe and in the striatum during this learning process and as these sounds are repeated the activity gets greater and greater and greater so the brain is learning a language even though the kids don't know that they're learning a language but this is not happening in the children with autism so they're not having this implicit learning system going on so you're going to hear me talk about implicit a lot so we also talked about rewards the possibility that maybe the reason why kids have trouble processing faces and interacting is that they don't receive the natural reward that we get from looking at another person social rewards are primary reinforcers in development neonates, brand new newborn kids orient preferentially to smiling faces they do that almost right out of the womb I think most of you know that if you've had children and reduce orientation to these faces is seen very very early it's one of the first signs of autism actually so one theory is that reward processing deficits could reduce social motivation what if the reward areas of the brain are not signaling responses when they see faces wouldn't that impact your willingness your motivation to look at faces to engage in social contact interestingly these reward areas are active for any kind of reward it's their general reward areas they're active for romantic love when they see attractiveness happy faces they're active during eye contact they all show reward like processes these areas are active for drug addicts when they see drug cues and there was one study that showed these areas are active when you successfully seek revenge on someone who's harmed you so these are general reward areas okay so we can measure response to reward reward is a very subcortical it's not a thought related process this happens implicitly under the under covers there are a lot of ways you can do that this is a way we can do it behaviorally just by looking at pupil dilation this is one of our former students pupil dilation is a measure of pleasure of reward you look at a pretty face of pupils dilate in fact in the 1800s women used to take a drug belladonna because belladonna artificially dilates expands your pupils right and so they would be judged as more attractive and if you actually just manually manipulate a photograph to make the pupils bigger you'll look better this is a little cue if you want to get a good glam shot you know just go in there and a little photo shop there so Lee looked at kids with autism and to be developing kids looking at faces and she did it this way so and you can do this in very very young children so the moms would sit here with the babies on their lap and they have a face like that and the face would be either neutral or the face would be smiling or the face would be angry and they have an eye tractor so the camera video camera here an eye tracking camera here so that they can see where the child's eyes are looking and you can find out if they're looking here and they're looking you know what part of the face they're looking at and then so what you can do is pull out all the time periods where they're looking at the eyes and then measure their pupil diameters with the camera okay so it's a really technically tricky thing to do and this is what she found that and these are very very young children in typically developing kids when they look at the happy face their pupils dilate okay and they look at any of the other faces neutral or fearful or angry there's no change and as you can see there's a very very big difference okay now we already know that everybody's looking at the faces because we've measured their eye movements so what's what's happening well a reward response is simply not being triggered in the brain even though they're seeing the same thing the reward response is not being triggered so we can look at that a little bit more directly looking in the brain with a reward learning task and this is one of our former students Ashley Scott Van Zealand looked at a bunch of children with ASD and IQ matched boys and I'm just going to go straight to the task just in the interest of getting straight to it in two conditions in a monetary reward condition what the kids would do is they'd see this random pattern and they and we were told they were told press a button one or two that goes with this and they'd be told if they were right or wrong now there's nothing about this pattern that says one or two they just have to guess and we had a bunch of these but we set it up so that the probability was 60% 66% actually if they said one they'd be correct and 33% they'd be incorrect and they'd reverse for some other other designs in other words they couldn't memorize the relationship between this and an answer there wasn't a perfect relationship it was only probabilistic and the probabilities were low enough that you stop trying to memorize and you just do it by gut and so it turns out that the kids learn these associations but they don't know that they're learning it it's all implicit learning if they get it right we give them money and if they get it wrong they don't get the money and that's the monetary condition and in the social condition if they get it right they see the smiling face of one of our former graduate students who's very cute and if they get it wrong they see the pouty face okay now you might think that our jaded LA teenagers would like go for the money but their response is greater for the smiley face of our former graduate student than it is for money and that's because smiles are a primary reinforcer money is a secondary reinforcer so this is a cross section through the brain right here to show you the reward centers which are located deep inside the brain and this is the brain's response to rewards it's an area called the ventral striatum it's our primary reward center and in typically developing kids you can see look at that whopping activity in the reward center and they're really responding to these rewards these are the kids with autism and that's the difference between groups it's not that there's no reward response it's just reduced and doesn't meet threshold so this supported that notion that the reward center itself is not fully functional this automatic reward response is not working why is that important but there are indications that can affect reward responsiveness oxytocin is one of them and it's safe the only problem is that it's nasally administered and it only lasts 20 minutes so all day long like a cocaine addict now what is really important is that those rewards affect learning and this is all implicit learning so this is not a matter of having a deficit in memorization like that the typically developing kids learn with those rewards the kids with autism do not learn so if you could imagine therefore if the reward system is not working well how profoundly is that going to affect learning of all kinds I'm going to skip that just in the interest of time what is also interesting and now I'm showing you just typically developing kids give them a questionnaire of the social responsiveness scale which is a measure of how social the kids are these are all in the normal range but there's quite a bit of variability and it turns out that how much activity is in the reward center even in typically developing kids relates to how social the kids are the more socially responsive or the fewer deficits in fact that they have in social responsiveness the less activity the more activity in this area is associated with better social skills less activity in this area is associated with poorer social skills and so it's not simply that this is just true in autism this is true for all of us those of us who have got a good reward response to social cues have better social skills so many aspects of language are very subtle and irony is one of them and this is the kind of thing that individuals with autism have particular difficulty with sarcasm irony where what you say the words you say do not match the intention of what you say so one of our very productive former students over here Ting Wang did this cool study on on irony and I have to play you the sound so you can get this so for sound please Brian and Dina are blowing up balloons okay and then the next one Brian keeps blowing his till it's huge Dina says nice going and the next Brian keeps blowing his till it pops Dina says nice going okay so you can see that there are cues about what Dina really means okay she's got a vocal cue in her prosody and she's got a face cue there now we don't have to tell typically developing kids that they need to pay attention to that tone of voice to get that meaning out of it they get it right away it just comes automatically but we did an experiment with kids with autism where we either just didn't give them any instructions or we told them attend to the voice or attend to the voice pay attention to those cues and then see if you can do it so here we find that this is an important area by the way this is an important area in the medial frontal cortex that's involved in theory of mind and if we just say just pay attention generally but we don't tell them what to pay attention to the typically developing kids get this no problem the kids with autism don't but if we say explicitly to the children with autism pay attention to the face cues or we say pay attention to the tone of voice then that area which is involved in understanding the intention of others becomes active so again it's not a matter of this area not working it's a matter of it not coming online automatically it's an implicit processing deficit you have to actually bring in explicit cues in order to get the area to work does that make sense and again when we look at this area and relate it to symptoms severity we can find again that there is a significant correlation and this is true in typically developing kids as well more activity you have in this area the better your social skills the less activity you have in this area the worse your social skills so we're able to relate some of these brain areas very directly to symptoms of autism ok and this is the last I'm going to talk about and then I'm just going to talk a little bit about combining the genetics with the brain imaging so sensory over responsiveness was not considered a core deficit but as the last speaker demonstrated it's very common it's founded up to 75 percent of kids with autism and for parents will report this is one of the worst problems to deal with it's also associated with anxiety and there are multiple forms but the one that I see most frequently is sensitivity to touch and clothing you know they don't want to wear itchy fabrics so I have kids who wear the same damn shirt every day right and loud noises too so this is a bunch of kids let me just go straight to the data ok this is a little bit hard to make out but I'm going to show you some brain areas these are the auditory regions the sound processing areas of the brain these are the touch processing areas of the brain these are typically developing kids these are kids with autism these are the difference between groups autism greater than typically developing and so in all the other studies I've shown you there's been less activity for the kids with autism but here we have more in the sensory areas the kids with autism have greater activity in auditory cortex when they're getting this is by the way this is not just regular sound these are annoying sounds you know like screeching icky sounds and the touch is a rough fabric like it we've got a an old piece of scratchy wool sweater and rubbed it against them they hate it they actually show more activity in these sensory areas than typically developing kids so basically the it's just another way of looking at the same data scratchy fabric they show more activity in the bad emotion fear center more activity in auditory cortex more activity in somatosensory cortex so this for the touch this for the loud noises and what's interesting is that this increased activity is associated with the severity of their symptoms of sensory over responsiveness so really what's happening in the brain is that these sensory areas primary sensory areas are not being appropriately filtered or regulated and this sensory over responsiveness is very associated with anxiety and no surprise because we get all this activity in this fear center so you might think oh well anxiety lots of anxiety lots of fear lots of you know bad experiences let's do exposure therapy right isn't that what we usually use for kids who are having problems with annoying things that are getting them upset let's do exposure therapy exposure therapy is supposed to result in habituation of the response over time so this is a habituation curve in kids with autism and typically developing kids here are the emotional centers as you can see these emotion centers are simply not habituating in the children with autism they're habituating very nicely in the typically developing kids nobody likes these obnoxious sounds the habituation rates are present but slower in auditory cortex and in somatosensory cortex much much slower so what does this mean this tells us that if we were to use the traditional treatment for anxiety producing stimuli for kids with autism we would do a very bad job in fact we might make the problem worse we might sensitize them because the habituation rates themselves are not normal so I like to use this as an example of why it's so important to understand what is going on in the brain if we understand what's going on in the brain then we can start to design treatments if we're just going to look at the behavior we may not only put in the wrong treatments we may make these children worse okay so obviously there are a lot of implications for treatment I'm going to go ahead to the last section we're running out of time I'm probably already out of time right how much time do I have five minutes okay so one of the things I want to just review real quick is some of the commonalities in the anatomy the brain anatomy we've talked about we've consistently seen networks involved in parts connecting the frontal lobe particularly the inferior frontal lobe and the medial frontal lobe with these subcortical more primitive brain regions are involved in these primitive emotions and reward systems and these regions are very important in selecting out what is important receiving positive and negative emotional signals processing and learning from them now can we relate this to the gene functions that we discussed earlier so we can again where we're talking about autism as a connectivity disorder when we talked about genetics we can examine connectivity in the context of genetics in humans in vivo using this approach called functional connectivity what functional connectivity tells us when one brain area is active how closely is that linked to another brain area okay so in this particular case these two areas are highly connected functionally so when one goes up the other goes up when that goes down the other goes down so these are connected regions that means there must be a lot of white matter connecting these two to make those very efficient pathways the brain is filled with these connecting networks which I'll talk about in a second but can we link them to some of these genes so these are my three favorite autism breast genes this is CNT NAP2 which we affectionately call catnap and then there's a met gene and oxytocin gene we talked a little bit about oxytocin so I'm going to show you very quickly the relationship between those genes and connectivity I'm going to go to the next one this one is not a gene okay so catnap is a gene that is related to language learning differences between TD and ASD children it's actually also implicated in selective language impairment this is our gene expression maps in a fetal brain it's just to show you that they are expressed that this gene is expressed because different genes are expressed in different parts of the brain this gene is expressed in the basal ganglia would you vote that deficit in autism during language learning and in the frontal lobe which is the other area that we saw as a deficit in language learning and we were very struck by the close relationship between human activation and brain expression so what we do in the connectivity analysis as we pull out one of these regions of the brain we look at its activation pattern over time we correlate it with every other area of the brain and we try to see what's connected to it okay and what we found in this case was that we found the correlates of the risk allele so having catnap too and this is for both ASD and to be developing children we're in the ventral striatum the reward center that I just showed you before and the frontal cortex that same medial frontal cortex area that I showed you before this is a gene that expresses in two of the regions that we know are important in autism and we know to be impaired so when we looked at the connectivity between those areas of difference what we found was this really interesting pattern and the pattern is that these are the risk, this is the non-risk, this is the summary of it this is the best way to look at it is that when we look at local regions there is more connectivity in carriers and when we looked at long range connections there is less connectivity for risk gene carriers in other words just like we were showing you earlier when we looked at the gene pathways we see increased local connectivity and decreased long range connectivity in carriers of this risk gene but what's really interesting is that these are not necessarily kids with autism with this risk gene this is anybody with this risk gene so basically what's happening is that the risk gene is biasing the brain to a certain pattern of neural connectivity and it's this pattern of increased local connectivity and decreased long range connectivity now, we wouldn't have genes like this unless they did something good otherwise they would have been out of our system ages ago, what do they do that's good well local connectivity allows you to focus on one thing and really focus on it and really process it and become very expert at it like a lot of our kids with autism who have selective interest and become very very well focused on one thing that's a plus because we do need to have detailed understanding of certain aspects of our information processing we also need to connect to the rest of the world so that was one gene this is the second gene the oxytocin gene so oxytocin is a neuro peptide actually that is expressed in the reward center of the brain and it's been associated with reward and social affiliation so there are very it's an autism risk gene the oxytocin receptor gene it's an autism risk gene and if you've got the wrong gene you have an increased risk of autism and actually there are about six variations on this receptor gene and so what Leanna Hernandez one of our students did was she picked a brain region this anterior insula that was important in a certain brain network called the salience network it's a network that is important for figuring out what's important in the world and it's connected to some of our best friends the amygdala and the medial frontal lobe and so forth and what she found again was that risk carriers so these are non-risk carriers this is connectivity within the salience network for non-risk carriers connectivity with risk carriers and here's the difference between the groups and overall the pattern that you can clearly see is that there is much more connectivity throughout the brain in individuals who don't have the risk gene and there is less connectivity throughout the brain with individuals who do have the risk gene and I'm just going to pull out one of these regions which is that frontal region that we keep seeing and you can see non-risk, intermediate risk because that means I've got only one copy of the gene and risk and you can see the connectivity decreases depending on your genotype and you can see that you see the same effects for both typically developing and ASD children okay one last gene and this is the MET gene this is a gene that we like because it's densely expressed in the face areas that we talked about before and so we looked at connectivity within a network that's also very very implicated in autism, I'm not going to go into details on it, it's called the default mode network and this is the risk group, this is the non-risk group and this is the difference, non-risk is greater than risk and again what we can see is that there's just greater connectivity along these long range networks in those who do not have a risk gene compared to those who do so we've looked at three different genes now and all three of them show the same pattern of abnormal connectivity across the brain particularly involving the frontal lobe and subcortical regions and it's often characterized by too much connectivity in local circuits but when you look across the brain a breakdown in connectivity so what is the significance of this aberrant connectivity well there are ways that we can look at connectivity structures throughout the brain and look at patterns of connectivity that are associated with a well-connected brain with a brain that works well a well-connected brain is not a brain where things are connected randomly or where everything is connected to everything a well-connected brain has very specific properties and one of them is modularity that is there's certain areas of the brain that are clustered together they form a module like the language system like the visual system like the auditory system and the regions within those modules are highly connected to each other okay I'm just going to do those two because those are the most important areas this kind of thing happens in normal development this formation of these well-integrated networks what happens is within a connectivity network so this is within one network there is increased integration with age this is adults and this is children so over time within a functional network the different areas within the network all start really talking to each other very closely and that happens over time the other thing that happens is that different networks become more segregated from each other so here's one network here's another network this happens to be the dorsal attention this is the ventral attention network and they're all highly connected together within each other but they're separated from each other in adults and in kids you can see that they're starting to connect up but they're still not segregated so these are two essential properties of a well-constructed network in the brain so we did this in autism by looking at the brain at rest and what was connected to what looking across all these regions of the brain where you see white spots extracting the activity and then looking visually at how the brain is connected so I'm just going to do a quick summary of just a couple of our major findings one is clustering how well within a network do the regions that are within a single network cluster together we call it clustering coefficient and we look at it across I'm not going to describe network sparsity for you but I think that you can appreciate that there is greater clustering for typically developing kids compared to autism all across the board so that means locally there is less efficiency in the network but I think what's more critically important is modularity the ability to form segregated modules again is greater in TD kids compared to ASD kids but I want to show you a way of visualizing it that I think will make more sense to you I think yes, okay so two of the major networks in the brain are the task positive and the task negative network this means that when we're doing something we want to focus, we've got our attention we know what's important and we've got a network of regions that are trying to engage in the task we've got a bunch of other things at the same time that we want to shut down and have them be quiet while we're focusing on something and then when we're at rest the opposite happens these other areas become more active one of them is our memory center it allows us to remember what we're just doing and consolidate that and the task positive areas go down so there's this inverse relationship and so we can graph it so this is a graph of all these different regions in the brain but in typically developing kids and what you can see is that within the task positive network all of these regions communicate very nicely with each other they form a great module and in the task negative network they do the same they all are nicely clustered they talk to each other but to go from here to here they go through a single node one node, a switching node that allows you to switch between this and this okay so they're segregated beautifully segregated networks with autism, same networks well what you can see is they do have the networks they're not as well integrated within the network but here's the real killer is that the networks are not segregated from each other there are many many points of contact between these two so imagine cognitively what that's like where you're supposed to be focused on one thing but you've got all this intrusion it's like there's noise going on all the time in the background it's always intruding in you don't have segregation from here you're here the sensory information is pouring in you've got information coming in from the auditory system you've got information coming in from the visual system and it's all intruding on your ability to focus on what you want to be focused on so you can see how the network dynamic how this functional connectivity can really impact cognition so just to wrap everything up in the context of a model of translational research environmental factors are important at every stage of the way because environmental factors will impact how a gene is expressed those genes will be responsible for starting neural migration synapse formation, circuit formation but environmental factors how much you use those circuits is going to strongly impact what is saved and what is not we can look at that at a large scale we know that ultimately it will go into specific behavioral anomalies we know that depending on what is happening early in the circuitry that it's going to affect behavior but the more that behavior, abnormal behavior takes place the more the individual will have an atypically biased behavioral experience so for example if you've got a kid who's focused on a single repetitive behavior and they keep doing it and doing it and doing it those circuits are being modified they're being strengthened and strengthened and other circuits that might interfere with that are being weakened and weakened and this is why we say that early intervention is so critical because we can modify brain function at an early age once those synapses are all formed and all together we don't have a lot to work with fortunately they don't get fully formed until much later in life but it really emphasizes the importance of intervening so that we don't reinforce and enhance abnormal circuitry because we can intervene particularly if we can find people very very young so finding people very very young I'm just going to get through these to get to the pictures so we are scanning babies who are at high risk for developing autism by virtue of having a sibling with autism and this is a network connectivity map of an eight week old infant and I think what is really striking to me we were shocked actually when we saw these data the auditory network is already there the somatosensory the touch network is already there the visual network is already there and even the salience network of what's important and what's not these networks have already formed they are already starting their formation at eight weeks of age functional how functional are these networks well to test that we played language in this case this is a six week old baby we played language to a sleeping baby that was either in their native language English or in Japanese and we compare them and you know a six week old this is the difference between a six week old infant could tell the difference between their native language and the non-native language their brains know the difference already so what this tells us is that if we are going to start intervening to prevent a disorder from occurring we are really having to start we can't wait until they are three years old we can't wait until they are four years old we want to start now we want to start then because these networks are already starting to establish their connections they want to establish the best connections possible and so of course that is where we link genes and brain and behavior if we understand those genes and how they work then we can start to intervene at a place that's way before the disorders ever manifest and of course the ultimate goal then is to prevent autism from ever occurring so with that I'm going to show you this last video this is our little boy we saw before after by the way two months of parent training and joint attention so this is different from ABA I want you to look at the difference in this little kid go ahead that was good you looked so nicely at me circle, good showing oval yeah oval oval oval, good showing yeah so you can just see how dramatic that kind of early intervention that's in this case is based on trying to since we already know that the child has problems with joint attention not a visual problem right an attention salience problem teaching them what is salient exaggerating the rewards exaggerating the eye contact making eye contact essential for everything exaggerating the emotional response and in a very short time this kid is speaking functionally smiling at his mom and this is all parent based by the way so I like to show that because it shows that if we interview we can actually make a really big difference in these lives so with that I'm going to thank my collaborators and thank all of you for your attention