Full episode transcript below. Beware of typos!
Nick Jikomes
Professor Sam Wang, thank you for joining me.
Sam Wang 2:41
Thanks for having me on.
Nick Jikomes 2:43
Can you describe for everyone what you do and what you study in your lab?
Sam Wang 2:48
Yes, my name is Sam Wang, and I'm a professor of neuroscience at Princeton University. I'm here at Princeton right now. And in my laboratory, we study how sensory experience teaches the brain and adult life and also during development. And we focus on the cerebellum. The work has implications for autism.
Nick Jikomes 3:05
So for those who don't know, what is the cerebellum, part of the brain are we talking about here? Oh, well, the
Sam Wang 3:12
cerebellum is the little brain. And it's, it's in the back right here. And so it's right here at the base of the skull that occiput. Like, if I go like this, and it's this little wrinkled structure, that's about 1/7 of the brain in humans. It's actually about 1/7 of the brain, and almost all mammals, larger in certain mammals, which is kind of interesting by itself. But it's thought to be important for sensation, when it's damaged, movement becomes jerky speech becomes unnatural. And sometimes, cognitive processes go off track.
Nick Jikomes 3:49
So you mentioned it's about 1/7 of the brain is that in terms of just area or volume, of volume, so the brain's about three pounds in weight? What is that about?
Sam Wang 3:59
1300 cubic centimeters, and about? About 200? That is cerebellum, so about 1/7 of the volume of the brain is cerebellum.
Nick Jikomes 4:09
What if, what if you count based on the number of neurons? Oh, the
Sam Wang 4:14
cerebellum has got a lot of neurons. And so a typical human brain has about varies, but let's say about 80 billion neurons in a typical human brain. And over half of those about 50, or 60 billion of those neurons are in the cerebellum. So it's packed with itty bitty neurons called granule cells. So it's got some of the simplest cells in the brain. It's also got some of the more complex cells in the brain, there are these crazy tree like structures called Purkinje cells. And it's got those as well. So it's got both the simplest and most complex cells of the brain.
Nick Jikomes 4:44
So it's packed packed with neurons, even though you know 1/7 of the volume. And you mentioned some of the things that start to happen when this area gets damaged. So can you talk a little bit more about what we know about its function based on what happens when there's physical damage to this area?
Sam Wang 4:59
Yeah, some of it earliest knowledge that we have about what cerebellum is good for comes from lesion or injury experiments. And this goes back centuries. But a good example of this would be in World War One. So in World War One helmets were not very well designed, there are these kind of like, dish like helmets that go kind of on top of your head, and don't actually go down here. And so it turns out that, unfortunately, there were a lot of cases of people who basically got the cerebellum shot off. And there was this British physician named Gordon Holmes, who looked at them and said, you know, these people have a lot of things in common they, they survive, they can still walk, but their movements become jerky, they, their gait becomes unsteady, they they're not so good at balance, and all those things like a ballistic movements, where if you try to make a gentle movement you were when you're in your movement is out of control. And so that was the most obvious sign that came from cerebellar patients. And this has been known for centuries. But Holmes did an unusually good job of studying it, because Because warfare became advanced enough. And battlefield medicine became fast enough that people would get these injuries and survive. So yeah, jerky movement, also dysarthria, which is kind of halting and jerky speech, where someone will will will repeat themselves and become inarticulate. Like, what I just did that.
Nick Jikomes 6:24
So World War One, they didn't have helmets that covered all of the skull. So people were getting more or less selective cerebellum damage. So is it fair to say that this part of the brain is not necessary for living, it's not like the brainstem or something where you'll die if you're missing it.
Sam Wang 6:40
It is not absolutely necessary for life in the same way that prefrontal cortex is not necessary for life. But in under a wild situation, you would not survive all that long if your cerebellum was damaged. And so it's not, you know, you'll still have a heartbeat, you can still walk. But your biological fitness would not be superduper.
Nick Jikomes 7:01
Okay. And then what about? So a lot of this and historically, I remember learning about the cerebellum as being this, this area that was very, very much about motor stuff. Is that mainly my understanding today? Or does it do other things as well.
Sam Wang 7:17
Um, and the most obvious signs are motor signs, but, but as I mentioned, speech becomes halting. And there's a syndrome that emerges when you, when you get damaged when you get injured in certain parts of your cerebellum, specifically, the posterior cerebellum, which is kind of the part kind of towards your spinal cord, and also sort of towards the sides and the hemispheres. There are cognitive problems. So for instance, if I asked you to draw a complex figure just to copy a complex figure with pencil and paper, maybe that figure you'll be able to draw the parts of it, but you won't be able to assemble it into a coherent whole. Another example would be injury at birth, injury to the cerebellum at birth, is leads to a dramatically enhanced likelihood of autism. And so they're seemingly other things that it does that are not just controlling movement. And, and that that's there's been more and more evidence in the last 1015 years that these other functions are also there, but just harder to notice. Because the thing you see in front of you when you have a patient is is the jerky movement. But But neurologist, if you if you ask neurologist, is there something else about them? They'll say things like oh yeah, like halting speech, or disrupted. For instance, you can look up a thing called cerebellar, cognitive affective syndrome. And there's a neurologist at Harvard, Jeremy schmahmann, who's made a career out of looking into that particular syndrome, and cataloguing exactly what goes wrong. When people get cerebellar injury in the cognitive and emotional domains.
Nick Jikomes 8:55
Yeah, you mentioned autism as well. So I want to get there, but I want to paint a bit more of a picture of the structure for people. So how would you describe the the cellular level wiring diagram for this brain structure? And what does that tell us about the kind of computations that might be happening here?
Sam Wang 9:14
Um, you know, it's this pretty remarkable structure, it resembles this thing that that computational neuros computational scientists have called a perceptron. So basically, you know what I might, I might go and take you up on the offer of doing screen sharing since you said that we can do it. Yeah. Let's see if I can do that. Hold on for a second. So let me let me have the permission to do that. Okay, go for it. Okay, so if I do this, and I see if I can do this, this is going to be a little. There we go. Okay. So does this look like okay, here? Yeah. Okay. And so if you look here, the right hand is a cross section of a cerebellum. And you can see here that it's got this really characteristic layered structure where it's folded like, you know, We're used to seeing folds on the surface of the human brain. Most mammals have a folded cerebellum. And it's got the structure where it's got most of the input comes into this kind of purplish layer, which is called the granule layer. And those granule, that granule layer is where those 50 billion neurons are. So most of the neurons of the brain are in this purple layer. And then they form synapses, onto Purkinje cells, which is this green layer on the right, and that green layer is the Purkinje cell layer. And just one example of a Purkinje cell is this crazy Rococo structure on the left, which is one Purkinje cell, where you can see from its branching structure, it has the capacity to get lots and lots of inputs. And so a single Purkinje cell might have several 100,000 inputs. And so basically, the cerebellum structure, the wiring diagram of the cerebellum is one in which information comes from all over the brain goes to the granule cells, 50 billion neurons in us, and then those 50 billion neurons, then their output converges onto a few 10s of millions of Purkinje cells, which have this complicated structure that's shown on the left hand side of the screen. And so it's this funnel in for information that gets narrower and narrower, and then these 10 million guys on the left, or whatever number it is, then they send information back out towards the rest of the brain, along this axon, that you can kind of see, the bottom of this image here has a pipette, which I use to fill this neuron, and then the little wiggly thing there is, is a is an axon coming out of the neuron.
Nick Jikomes 11:26
Okay, so for those just listening on the right, we've got a cross section of the cerebellum and the cerebellum, how would you describe it visually, people, I think people describe it as almost like a cauliflower or something like that,
Sam Wang 11:37
it looks like the cross section of a column. What we have here on the right, is a, let's call it a fluorescent multicolored cauliflower here. So imagine a cauliflower and cross section. And imagine the STEMI bits of the cauliflower are purple. And that's where fibers come into the cerebellum. And then there's a layer of deep purple, and that's where the granule cells are. And then kind of the rind of the cauliflower all along the edge is where the Purkinje cells are. And so if you were to visualize this cauliflower section, for those of you listening on audio, that that very folded Ryan on the outside, is where the Purkinje cells set.
Nick Jikomes 12:14
And these Purkinje cells, you've got an image of one here on the left, and you know, there's a cell body with many, many, many, many elaborate branches. So these are where the inputs come into the cell. And so what you've told us, if I understood correctly, is that there are neurons, brain cells in the brain from all over, more or less, that are sending inputs to the cerebellum. And so there's this massive sort of convergence or funneling of all of those inputs into the cerebellum and ultimately onto these things called Purkinje. cells, which I'm guessing can accept many, many, many inputs. Yeah, these Purkinje
Sam Wang 12:47
cells are crazy. So it's, for those of you who are not looking at video, you got to look these things up. It's spelled pu r k i n g. And these neurons, they're like when Kahal the great neuroanatomist study them, he said that they resembled the espalier trees that one sometimes finds in the south of France. And what he meant was, when people grow trees in southern France, they will sometimes prune them so that they grow in a plane. And this plane looks like a hand I'm holding my hand for for people who are watching. And then the in the whole thing is about 200 microns wide, the breadth of the hand, if you turn sideways, it's just about 1015 microns thick. And so the whole thing looks like this flat Harbor. And this flat Arbor is incredibly Baroque. I mean, I would say, I love saying the word Rococo. So it's like, it's not just Baroque and Rococo. It's like this crazy curlicued structure, where all the dendrites have like little hairy things on them that are the dendritic spines, where the synaptic inputs come in. And the whole thing is just remarkably complex. And it receives input from those 50 billion neurons that we were talking about a minute ago.
Nick Jikomes 13:51
So there's a lot a lot a lot of neurons in the cerebellum, they're receiving this convergence of inputs from all over the brain. And you mentioned that something about this architecture is similar to what computational scientists call the perceptron. So what exactly is that?
Sam Wang 14:09
The perceptron is this old idea about how you would ever recognize a pattern. And so it's a pattern recognizer? And in some ways, the perceptron is, it's sort of the original ancestor. Now when we do things like machine learning, and we build a simple neural network that recognizes things like whether it be Google Translate, or whatever it is that we do, using our agents, that that perceptron The idea is that Purkinje cells, these very branch things, learn to recognize patterns by picking inputs from this vast array of granule cells that come in. And so each Purkinje cell has a couple 100,000 inputs. And the idea is that by picking just a few 100,000 of the inputs, a few of the 100,000 inputs to come into the Purkinje cell. Those patterns, then are What the cerebellum can learn. And by changing the weights of those, those connections through a process called synaptic plasticity, by changing the strength of those connections, then Purkinje cells are able to learn from the environment and learn combinations of input that that can help coordinate some motor process or cognitive process. And the idea, the idea would be that it does it in real time. And so in the context of biology, it can do it in real time, over and over and over just incident after incident after incident, constantly responding to the world.
Nick Jikomes 15:37
I see. So so it's not, it's not simply that the cells are learning associations that will necessarily remain stable for a long time, but they're sort of learning them on the fly.
Sam Wang 15:47
Yeah, yeah. So the cerebellum gets, it's basically this big, with the rest of the brain. So if you look at the way the cerebellum is arranged, compared with rest of the brain, it actually is in a loop. And I'm changing the diagram, which we can talk about in a moment. But the output of the cerebellum goes out into the brainstem, it can go down to the spinal cord, and can also go up to the forebrain to the neocortex. And the neocortex is one of the major sources of input that comes in on those two those 50 billion cells. And so the whole thing is like this loop, where it's a two way conversation, where information goes up from the cerebellum to the neocortex comes down from the neocortex through the pons, back into the cerebellum. And so it's this constant closed loop of activity, almost like a feedback circuit.
Nick Jikomes 16:37
And can you talk a little bit about how the structure of the circuit might relate to Now you mentioned, you mentioned what happens in adults, when you damage the cerebellum? Can you contrast that with what might happen developmentally here? Like what is the cerebellum doing anything over the course of development to sort of teach or learn from other parts of the brain?
Sam Wang 16:59
Yeah, so the brain generally speaking, mammalian brains generally develop in a back to front manner, so that the brain stem and the cerebellum develop earlier than, say, thalamus, and neocortex. And so the general pattern of mammalian development in all mammals is back to front. And so consequently, the cerebellum and brainstem are ready pretty early. So for example, if you encounter a newborn baby, you know, the baby has certain reflexes, and she can do things. And that and at some level, what you're encountering, when you meet new newborn babies, you're encountering this little brainstem that is capable of doing brainstem and some cerebellar things. And as babies mature, in the first year of life, there's this proliferation of connections in the in the, in the court in the cerebral cortex. And that proliferation is like 10s of millions of synapses per second forming in the first year of life. And so one possibility for what the cerebellum may do in development, is that it might act as a guide to help shape all that circuitry. So the thing, let me let me rewind and say, the thing we know for sure, is that the cerebral cortex is proliferating at a mad pace, 10s of millions of connections per second around age one, the density of synapses peaks, and then there's a pruning process where the connections get start getting trimmed away. And that whole process of proliferation and pruning is how our brains become shaped, to do all the things that we do, you know, to be adapted to whatever environment we grew up in. So we know that that happens. And then to get into the speculative, what my lab thinks, we believe that the cerebellum might play a role in shaping that printing, the idea is that imagine like a topiary where all these connections grow, and then somebody has to prune those back. And what we imagine is that the cerebellum might be part of the process by which that topiary gets trimmed back.
Nick Jikomes 19:00
And and why do you say that is, is? If you take out the cerebellum experimentally, or you mess with its function during development? Do you see abnormalities in what gets pruned elsewhere in the brain?
Sam Wang 19:14
Yeah, there's a few things that happen. So one thing first off, let's start with the human symptom, which is that if you look at cases where there's a difficult birth, and the mom has a difficult birth, and there's a bleed in the cerebellum, there have been case studies of cerebellar bleeds in neonates, you know, kids at birth, and those kids about half the time by the time they reach age two are autistic, they get scored as being on the spectrum. And that's work done by Katherine Lim propolis at Children's National Hospital in Washington, DC. So that's, that's a clinical observation. But then if you look at animals that have unusual, you know, genetic or other alterations to cerebellum, my lab has some evidence that when you look in mice that have cerebellum specific alterations, their dendrites are have more connections. And so we have some evidence where we count the dendrites on the count the synapses on the dendrites of neocortical neurons, we can see that there are denser connections, I see.
Nick Jikomes 20:20
So the idea is there's an abnormality in the cerebellum. And a result of that seems to be that there are actually more connections, perhaps too many connections elsewhere in the brain.
Sam Wang 20:30
Yeah, and this, this could account for something that people who work on autism have known for a while, which is that there seems to be something that doesn't get fully. There's some kind of pruning or or refinement process, that doesn't happen fully in kids with autism. The original observation was kids with autism at age one seemed to have kind of large heads. And that seemed like a pretty crude observation. But then, if you look at MRI, they're there, their quarter the thickness of their gray matter in their cerebral cortical sheet, the thickness seems to increase a little bit faster in in, get trimmed down slower in kids with autism. And so there's something about the growth process of the cerebral cortex of autistic kids, that that seems to be overgrowth, or under pruning. And I would say that that's an observation that's been around for for a few years now.
Nick Jikomes 21:28
Yeah, I believe I've read that there's at least one major hypothesis around autism, which has to do with the autistic brain more or less having many more synapses than a normal brain. It sounds like perhaps that's true.
Sam Wang 21:42
Yes, I think I would say that there's pretty good evidence for that at this point. And just to back away to get off of the topic of cerebellum for a bit, it's generally known that that pruning process depends on experience and activity, one of the major principles of how our brains develop is that we go through this period of a sensitive period, what's called a sensitive period of development, where our brains are receptive to experience, and are able to change your response, whether it be the sounds of a language, or other environmental events, like say, Yeah, learning a language was would be a perfect example. And we are receptive to that experience for a certain period of life. And that at some point, our brains become less receptive. And it becomes harder to change in response to that. And the classic example of this is work done by Hubel and Wiesel, back in the 60s, I believe, I think in the 70s, where they studied visual development, and they found that the visual part of the cerebral cortex was strongly dependent on visual input, and that official input did not come in, then the circuitry did not get refined properly.
Nick Jikomes 22:52
And so we've already been talking about autism a bit. How would you? How would you define what autism is for a non neuroscientist.
Sam Wang 23:01
Autism is so interesting, it was discovered in I think, 1943, or is defined in the 1943 by a psychiatrist named Leo Connor, K nn, er. And autism is defined by, let's say, aloneness, and sameness and so and so if we think about what what kids with autism are, like, they, they're alone. So they have difficulties with language, and so they have difficulty communicating. They also have difficulty getting inside the heads of other people like understanding other people's motivations, other people's states of mind, and so that they're alone in that way as well. And then there's a sameness, where autistic kids really want things to be the same. And so if you meet a child who might be at risk for autism, maybe, let's say me to a child who only likes one toy, and that's the toy that he likes. Or maybe he's been say, pushed to a preschool in the morning, and he gets upset if any other route is taken to the school. And so autistic kids often have a real liking for sameness. And and when you put these two things together, aloneness and sameness, these are criteria that that are used to, to discover children who are who have what's called Autism Spectrum Disorder.
Nick Jikomes 24:22
And so is it truly, is it truly a spectrum here? And at some point, the diagnostic criteria diagnosis is made that we're gonna call this child autistic, or is it truly a spectrum? I guess that's the question.
Sam Wang 24:37
Yeah, so there's a saying in the autism community, if you've met one kid with autism, you've met one kid with autism. And the idea there is that kids are just different from each other. So maybe one kid might have a real sensory sensitivity where she might not like particular sounds. Another kid might not like the touch of new clothing. Another kid might might actually have those issues but in fact, be able to speak quite coherently in life. And when we have a child who has the signs of autism, but actually can speak, that's called Asperger's syndrome. And so all these different kinds of things are, are broadly, if you look in the DSM, the Diagnostic and Statistical Manual, they are together, lump does a single thing autism spectrum disorder.
Nick Jikomes 25:22
And so there's an aspect here, it seems to me have something to do with how sensory processing is happening. You mentioned and others have mentioned before, that children with autism often have some sort of sensory deficit, but but it's actually it tends to be in the direction of being hypersensitive to sensory inputs. Is that true? And is that related to the observation that there's actually an excess of synapses in the brain?
Sam Wang 25:51
You're correct that hypersensitivity is a major feature of autism. So, you know, noted no to kids or with autism are exactly like, but it is commonly observed, that they are hypersensitive to sensory input. So for example, like I said, I mean, there's some pretty well known essays by by people with autism I, I've written a popular book, welcome to your Brain, in which my co author Sandra, and I, Sandra, and I write about the kinds of hypersensitivities that occur, like maybe a plain sounding, passing overhead sound so loud, that the kid, the person has to hold their ears because it's just so loud, or, or maybe just a door closing or something sounds like nail chop fingernails on a chalkboard. And so, sensory hypersensitivity seems to be quite common. And you can imagine, there's this, there are these two psychologists, cognitive scientists, friendship, Francesca Hoppy, and Buddha Frith, and they have this idea. I mean, they're not the only ones who have this idea. But I think they've been real leaders in forming the idea that all this sensitivity to the world just makes you really immediately focused on right now what's happening in front of you in the world, you can't focus on things that happened before, you can't focus on things that are in some past or other context. And you get so stuck on that, that you become very focused on the present. And you just have difficulties with broader context. And you can imagine that if the world becomes sensory, you know, extreme like, we're intense, that it might shape the way your brain develops. And so this has been a this is a, I would call it a major hypothesis for where autism comes from.
Nick Jikomes 27:31
So sort of on this theme of where it comes from, you mentioned something interesting earlier that had to do with cerebellar damage that happens during the birth process, sometimes, can you give us an idea of the current state of knowledge for the causes of autism, and in particular, the relative contributions between genetics or in utero? Causes or environmental factors during or after birth?
Sam Wang 27:56
Oh, yeah. Glad to talk about that. Let's see, somebody just show you another slide. Let's see if I can get this done correctly. I hope I can do this, you know, it's just so zoom is so fraught, but hopefully this shows up on the screen. And so if you look at one way to think about factors that may cause autism, is to think about how much a factor in the environment or in the genetics of a person, increase the odds of autism. And so the top factor by far of, for the risk of autism is genetic. If a kid has an identical twin who's on the spectrum, then it's a better than 5050 chance, then that kid himself is also on the spectrum. And so having an identical twin, in other words, sharing all your genome with somebody is a powerful predisposition for autism, even sharing half your genome increases the odds of autism, say in a fraternal twin, by tenfold. And so these, I'm showing a slide right now, for anyone who's watching on video. The light blue bars here on the slide, are factors that are predominantly transmitted through DNA, where it's either sharing all your genome and identical twin sharing half your genome or fraternal twin. Also, things like say having a parent with mental illness, or having a parent who's over the age of 40. Those are smaller risk ratios, but they also have a risk ratio that's significantly greater than one. And so really, genetics is a powerful predispose to autism. And then there are also environmental influences. We talked about cerebellar injury cerebellar injury doesn't happen very often. But when it does happen, it's a pretty large risk factor 30 or 40 fold. And there's other risk factors that have to do with adverse experience in early life. Being brought up in an orphanage where you get no social input, or mom getting stuck in a hurricane strike zone or other kind of very stressful situation during pregnancy. These are also stress factors that stresses that may increase the odds of autism.
Nick Jikomes 30:01
And so dwelling for a moment on the cerebellar injury at birth. So for those just listening, we're seeing this really interesting bar graph showing the things that might put you at higher risk for autism. Number two is cerebellar injury at birth. And this can have a big impact is what the graph is showing us. Is that typically, that we noticed at birth that the cerebellum is abnormal in some way, or is there actually injury that happens during birth, and the cerebellum is somehow more exposed than other areas of the brain?
Sam Wang 30:33
The cerebellar injuries at birth are not that common, if you look through a whole bunch of case studies of difficult births, and so on cerebellar injuries are only a small fraction of all births. And so So one has to actually do a research study to find this out. And there have been several groups that have done this one I mentioned before this, this woman, Katherine Lynn propolis, and Andre do Plessy. They did it. There are others who have approached it as well. There are other ways that the cerebellum can can be damaged, for instance, congenital congenitally. There are certain kinds of malformations of the brain. One is called dandy Walker syndrome, there's another one called Jubair syndrome. But when you look these things up, they also lead to cerebellar. malformation, and lead to autism diagnoseable by age two. And so the idea is that, if you notice, in cases where the cerebellar abnormalities are noticed at around birth, then if you track those kids, by age two, they're quite often, they quite often show developmental differences that that add up to autism.
Nick Jikomes 31:38
And so what this graph is showing us overall, and what you've told us so far, is that genetics play a big role in the development of autism. Are there any deeper connections you make with the particular types of genes? And whether expressed, for example, are are there certain genes that you can associate directly with the cerebellum? Are these genes that are involved in things like plasticity and development?
Sam Wang 32:02
Yes, this is great. This is great question. There, by now, researchers who work on human genetics have found hundreds of genes that whose variation, increase the risk for autism. And so there are a few of those genes that are called syndromic, where just one gene dramatically increases the risk of autism. And then there's lots and lots, as I said, hundreds that are that, that increase the risk of autism by a modest amount. And these are called common variants, because in fact, we all have them, or not all of us, but many, many of us have these variants. And so in any given population, you know, like, it could be that something like half the population has these common variants that in most people do not have an effect, but increase the risk of autism. And the idea is that these are not say, you know, a protein getting messed up entirely. But maybe there'll be a variant where more of the protein gets made or less the protein, or the protein gets regulated a little bit differently. And many of these proteins are set are expressed at synapses. And so one common theme that emerges in a lot of autism, genetics, is the fact that many of these, these genes are in fact, synaptic proteins, which suggests that there's something about synaptic I mean, you know, the way I'm saying it, it sounds kind of obvious, the brain is composed of synapses and does things with its synapses. And so maybe we shouldn't be superduper surprised that genes that regulate synaptic proteins could play a role in autism. But anyway, that's, that's what's been discovered so far.
Nick Jikomes 33:33
Interesting. And so there's this other, you know, we mentioned cerebellar damage during the birth process. There's a lot of things on this graph that we've talked about sides of genetics that imply a developmental component to this, a couple that I think are worth dwelling on. Well, one that's worth dwelling on is two of these risk factors here that I'm looking at. They're a bit more modest, but still a father or a mother that are older. So what's going on there? Is anything interesting happening that has to do with the parents age? Yeah, so
Sam Wang 34:07
just to orient people who can't see the graph, this is a graph showing that the risk factor, let's say that the risk of autism in the general population is 1x. What this graph shows is the risk associated if some other factor is present. And so for example, just to orient you, a genetic risk might be anywhere between 10x and 50x. And so those are big risks. So as you say, Nick, the red there are some smaller wrists on here. And we hear about these because there's something that we there are risks that we can do something about. When if if mom waits until after the age of 35, to to have a baby, then there's a very modest risk gets 1.3x relative to the general population of 1.0x. Or if the father is over 40 years of age, then the risk is 1.4x. Again, relative to the general population have 1.0x. And so, you know, these get talked about right? It'll be you, it'll be something like, you know, honey, you haven't had a kid yet. Did you know I saw in the newspaper that it'll be something about this risk factor. So what's going what's what could be going on here is that as we get older for various reasons that are different in men and women, as we get older, there can be alterations in DNA in DNA that accumulate over time. In the case of the Father, since sperm are generated throughout life, those sperm may be generated to end have new, you know, de novo mutations. In the case of the mother, since girls are born with all the eggs they're ever going to have, that those eggs may acquire cumulative DNA damage. And so the idea would be one possible explanation for this is that that cumulative alteration to DNA, as we get older, slightly increases the likelihood of autism in the kid.
Nick Jikomes 36:06
And some of the things that I'm seeing on this graph for risk factors have an interesting pattern, I think. So if you're born premature more than nine weeks inter birth interval less than a year, so mom has two kids within a year, that's a risk factor. You know, we just mentioned the age thing, if, if you have an immigrant a mom that's emigrating, or is in a hurricane strike zone, those things would seem to imply that perhaps, mom is under more stress. And these are things where my first guess is will have something to do with the the stress in the mother causing the in utero environment to be different. Do we know if that's the case, or what some of those environmental in utero factors might be?
Sam Wang 36:50
It is not known what the mechanism is for these risk factors. But it is true that the thing that they have in common is severe stress. And so in the case of inter birth interval, well, let's see, let's go with emigrating while pregnant. In Scandinavian countries, there are birth registries where every birth is recorded, and there's follow up for virtually every birth. And so it's been found in Scandinavian countries that for women who are emigrate, who come to Sweden, while pregnant, that, that if they, if they come to Sweden, while pregnant, there's about a 2.3 fold increase 2.3 fold risk of autism. And furthermore, that risk is larger if they come from far away. So for instance, if they come from another country, in Europe, then the risk is smaller, if they come from Sub Saharan Africa, then the risk is larger. And you know, if you think about it, if you if you grew up in Sub Saharan Africa, and you move to Sweden, I mean, that is a really different place, you show up and it's cold, and it's like, people get these weird food and all kinds of things are going on in Sweden. And so if you're from, from Africa, this is like, Wow, a very different place. And so, similarly, if you if you're caught in a hurricane strike zone, and this is not just a bad storm, but caught in a severe storm, where say, you lose your home, and you get displaced, that leads to a tripling of autism or escalates, it's been observed in a in a study of these mom's. So it, I think it's a parsimonious explanation for these is that there's something about stress, you might expect that it might be sustained stress hormone signaling, but I will say that that is not known. You can imagine studying it, it would be great to, I gotta say, this could be studied, you can imagine studying moms who get in a hurricane strike zone, and then say you could take samples of hair because cortisol gets gets, you know, accumulated in hair, as you can imagine taking a hair sample, and finding out how stressed mom got and compare that with outcomes. And so it could be studied. To my knowledge, I think that's not been done. But, but you know, I think stress is a pretty plausible hypothesis for how these things happen.
Nick Jikomes 39:04
And so the graph that we're looking at for those on the audio, part of the point of this graph is to show what some of the risk factors are for autism, the things that increase your odds of getting autism compared to baseline. And so we're looking at things that are predominantly above baseline that increase your chances of getting autism. There is one thing on this graph that is right at or just below baseline, and I'm wondering if you could unpack why that's there. And what we're looking at.
Sam Wang 39:32
Oh, yeah. So this is a this is a figure that's mainly meant to show what does cause autism. It's largely genetic, some environmental, but there's one I stuck in there where I'm basically trolling people, and there's one risk factor that is slightly less than 1.0x and is the risk associated with with the baby getting the MMR vaccine. So, as I'm sure most of your listeners know, there is is a widespread belief that vaccination has something to do with autism. And so it turns out that this has been studied about a dozen times where people have taken large populations of kids who are vaccinated and compare them with a population of kids who are not vaccinated. And it turns out, there is virtually no difference in the rate of autism, between kids who are vaccinated for this MMR vaccine, and kids who are not vaccinated. And ironically, the risk of autism is slightly lower. It's not it is not statistically significant. But it is slightly lower in kids who get the vaccine. So now I don't actually think the vaccination does anything to the risk of autism. But I love the fact that the risk factor here is 0.9x, as therefore very slightly smaller than the general population.
Nick Jikomes 40:50
Interesting. So basically, if you just look at children who do and do not get the vaccine, there's no increased propensity for those who get the MMR vaccine to develop autism.
Sam Wang 41:00
None, none. And, and the way I would describe this whole slide, just synthesizing the whole thing is that it appears that any, any risks that in the environment that do increase the risk of autism, appear to be in the second half of pregnancy, up to birth, but not after birth. And it appears that, that, you know, there's all kinds of things that happen to us after birth. But for the most part, the only thing that can really lead a kid to end up being on the spectrum after birth is severe deprivation, like really severe deprivation, and so autism, so vaccines are not part of the story.
Nick Jikomes 41:34
I see. So that's why I like the Romanian orphanage bar is up there. That's what you mean by severe deprivation.
Sam Wang 41:39
Oh, yeah, like back in the church, Heskey regime was terrible. Ceausescu, the dictator, in charge at the time, rightly said that a great at the nation's greatest resources, its people, so that was good. But then he also said, Therefore, everyone must have as many babies as possible. And so they're all these babies born in Romania. But there were there were not enough resources to take care of those babies. And so there was a, there were a lot of abandoned babies. And there are these orphanages that housed the babies that fed them, but didn't give them any social care, no, very little human interaction. And these terrible orphanages were the site of a natural experiment, where were researchers went and looked at those kids, and found that there was an eightfold increase in the rate of autism, and among those kids, especially among kids who were, who lived in those orphanages past the age of four, it was just very hard for them to recover afterwards. And so it supports the idea that there's some critical period of development, that that is necessary for neural wiring refinement to take place.
Nick Jikomes 42:49
So if I had to summarize, all of this so far would be that in terms of developing autism, the primary risk factor, the primary cause of autism is genetic. There are various in utero factors at play. So what's going on during certain parts of pregnancy, especially, are super important, but we don't really understand the details there. And there aren't a whole lot of risk factors that happen after birth, but those that are there tend to be extremely severe forms of deprivation, like the orphanage example.
Sam Wang 43:20
That's right. That's, that's, that's a perfect summary. Thank you.
Nick Jikomes 43:24
Um, I'll do my before we move on, I'll do my best to play devil's advocate here. Because this is such a hot button issue. So there are a lot of people that appears in the world who really, really believe in this idea that the vaccines have something to do with autism. Is it conceivable that someone could argue that, okay, maybe overall, there's no difference between the vaccine group and the no vaccine group? But maybe there's some sort of interaction where like, a subset of people with some kind of genetic predisposition, who also get the vaccine have something happen? Yeah. Yeah,
Sam Wang 44:03
I guess you could come up with Okay, so the the excellent, the alternative you're suggesting is, vaccines has some causative role, but they do not appear when you look at the entire population. But part of that statement, I think, to my knowledge, I mean, to my knowledge, that now we're into a zone where there's no evidence at all. So if you say to me, I have a hypothesis for which there is no evidence at all. I just don't know what to call that. I mean, at that point, those things could be true. But generally, in order for someone to really to entertain such an idea and take it seriously, there needs to be some kind of evidence that holds up to peer review that holds up to statistical scrutiny. And to my knowledge, there just isn't any evidence like that. So I, I guess it could be true. But, but if there's no evidence for it, I don't really know what the reason is for, for pursuing it. And the other thing is, I think there's an opportunity cost which is look It's really important to help kids with autism lead the fullest lives they can. And time is finite. And research resources are finite. And if we spend our time going after something for which there is no evidence that takes time away from other things, it takes time away from developing therapies that might help kids, you know, get along in everyday life better, it takes time away from neuroscience research that, say, determines whether neocortical circuitry can be rescued. There's all these things that we could do with research to go off after this thing for which there's no evidence just seems to me. You know, to be counterproductive, I mean, I think it'd be nice to help these kids. And I think it'd be helpful to go in directions where the risks are largest. And so the reason I constructed this slide, which I'll take down now, the reason I constructed this slide is just to show people what the biggest risk factors are. And I think I think we have a responsibility as, as caregivers, as parents, as researchers, to go after the big things first. And, you know, I look, vaccination is like, it's kind of terrible if you if you ever taken your kid in for a vaccination. You know, sometimes the kid cries and hates it. And ah, you know, everything was great. And I took my kid, and this is the one thing I ever did for my kid who to like, make her cry. And it just, you know, that, that doesn't feel good, right, who wants to make their kid cry? And so you can kind of see why people would feel that way. But, you know, the, it doesn't change the fact that there's just not support for it.
Nick Jikomes 46:33
Yeah, I mean, it strikes me that that what you were just showing us in the data was that, you know, very clearly the biggest risk factors are either genetic, genetic, or pre birth factors. And so if there was some sort of post birth factor like a vaccine, it would have to be so so strong, it's a fact that you would think that we would have seen some unmistakable signal for it by now.
Sam Wang 46:57
Yeah, I think there's just not and I and given the preponderance of evidence, the the pattern of all the significant risk factors. The overall pattern has to do with events before birth, and has to do with stressful events. And with genetics. And given that overall pattern, it strikes me that as we start to build a framework for autism, it's it's a little bit more like, our genome sets the stage for how we process information in the world. It sets the stage for how we learn from sensory experience. And that program of development can be driven off track by stress. And that whatever the outcome of that is, leads to a developing brain that then learns from the world from say, age zero to two. And I think that storyline that I just told, accounts for most of what we know about autism. And then there are, you know, many interesting details about what causes autism. But broadly speaking, I think that story I just told, captures a lot of it.
Nick Jikomes 48:03
So as we mentioned, genetics play a big role. The number one risk factor by a healthy margin, was having an identical twin, someone with your same genome that is autistic. Nonetheless, there was something interesting in one of your papers that I want to read here. And then I want to start sort of connecting the autism story to some of the brain stuff a little bit more. So in one of the papers, it says, quote, most autistic children have to neurotypical parents. first degree relatives of persons with autism spectrum disorder often show distinctive mental traits, including unusual social and emotional characteristics and interest in technical subjects, indicating that these disorders, there's risk genes that may drive variations in outcome within the normal range. Can you unpack that a little bit for us? So even though genetics plays a big role, most autistic individuals, the majority have to neurotypical parents, and yet there are these associations with their first degree relatives. So what's what are some of those things?
Sam Wang 49:08
Oh, yeah, this is such an interesting topic. So let's see. So there's this puzzle that might come up in one's mind, which is that if autism is genetic, then why is it that we do not see autism running in families? And it turns out that I would say a lot of the causation of, of autism is consistent with individual genes not having much effect, but combinations of genes leading to autism. And so for instance, in cancer biology, we had there's an idea that people talk about called the to hit hypothesis where two genetic hits, or one genetic hit and one environmental hit can lead to a higher risk of cancer. So one possibility is that maybe these genes, like imagine human development as like a cloud of outcomes. We're all different from each other and we're all different in different ways. To be somebody who's better at math or more interested in art, or more interested in sports, or what have you, and imagine this cloud of diversity that makes up who we are. And now imagine if we get up far enough out on that cloud, then we get called autistic, or ADHD, or developmentally delayed or what have you. So these genes in combinations might lead to autism, but maybe they do something individually. And what my, one of my students and I found was, we surveyed a bunch of people. And we found that it was quite often the case that autism was found in first degree relatives of people with an interest in science and engineering. And so if you looked in people who are who are college freshmen, with an interest in science, technology, engineering, or math, or STEM disciplines, those freshmen were twice as likely to have siblings or even a parent on the autism spectrum than their classmates. And so that suggests the possibility that there might be some gene or multiple genes that increase the likelihood of interest in science, and in certain circumstances, increase the likelihood of autism. So the idea would be that we have a diversity, you know that some people are interested in science, some people would love to listen to this podcast. And then other people I don't know, they would just assume, listen to a podcast on say, Renaissance painting or something like that. And so different people might have different interests. And the survey results suggest the possibility that, that science interest in autism might run together and families, we weren't the first to do this, it turns out that there's this guy in the UK, Simon Baron Cohen, and Simon Baron Cohen also found the same thing in a different population of people. I should say that we also found something like this in populations of people who are non college students. And so we found it not just in college students, but also people who didn't go to college.
Nick Jikomes 52:05
So so the idea is basically that the genetic predisposition here is really playing with minor variations in a variety of genes, such that you're, you're sort of playing at the boundaries of what the normal, the normal distribution of traits are among people, right? So
Sam Wang 52:23
if so, imagine a normal range where you know, someone is unusually interested in science, I would characterize myself as being unusually unusually interested in science enough to pick it up to do it for a living. And say, Here I am, like at the end of the spectrum of science interest. Likewise, you can imagine people with, you know, autism are at the extreme of how they interact with other people. One observation that's been around for a while, is an old result, that, that fathers of kids with autism are unusually or unusually less interested in, in other people, and, and so on, there's a thing called the broader autism phenotype. And if you do a Google Scholar, Scholar search on broader autism phenotype, there's a survey that has questions on it, like, I have trouble telling when people are not interested in what I'm saying. Or I tend to have single interests to the exclusion of other topics. And people who score high on that inventory are more likely to to have siblings or children with autism. So there's something there's something this broader autism phenotype might coincide, there appears to be reasonably good evidence that it coincides with with the causes of autism.
Nick Jikomes 53:40
And then you commonly hear that there are sex differences here that autism is more common in males than females. Is that true? And if so, what is that telling us?
Sam Wang 53:52
Yes, autism is currently diagnosed about four times as frequently in boys than it is in girls. That's something that's true in multiple countries. It's not true. For extreme cases, if you look at extreme autism, the sex ratio of extreme autism is about one to one equal between boys and girls. So yeah, so what's going on there? It's possible that there's something about the genetic background or the developmental background of boys versus girls, that pushes kids towards different kinds of outcomes. If you think about small children's tendencies when they're very small. Even from a very early age, girls, more often than boys have a tendency to be interested in doll type toys, and toys that involve human interaction. And boys are, on average, are more likely to be interested in tool toys like pushing a truck or a hammer or whatever it might be. And so it could be that because of this sex difference that emerges very early in life, that might be one source of difference among humans, and then add on top of that, specific genes that might To increase the risk of autism, you could easily imagine that this that this naturally occurring, predisposition between boys and girls might be a starting point from which autism could can progress.
Nick Jikomes 55:13
So if I had to sort of summarize at a very high level so far, some of the things we've been talking about, it's the autism has a very large genetic component. The genes and the proteins involved here very often have something to do with plasticity in the brain. And very often, these individuals have a kind of hyper sensitivity to the sensory environment. And this hypersensitivity, perhaps, is a cause or a driver of why such an individual will often want to focus in on a narrow range of things. So another way of saying that might be that, you know, if I have autism, and the world is overwhelming to me, and there's this sort of sensory overload, one way to cope with that is to sort of zoom in and focus on one thing, so that I don't have to deal with all of that overload. And at that point, if that's the habit that your brain is sort of incentivizing you to have very early in life, that's then going to affect the subsequent trajectory of your brain development. So it brings to mind the concept for me of developmental cannibalization. And I wonder here, if this is a good place to think about how you and others think about treatment options, because it would seem to me that the earlier on you could treat something like the the excess number of synapses are something, the more likely it would be to actually be a beneficial therapy. So how do we think about therapy here? And when when we would want to think about therapeutic interventions? Oh,
Sam Wang 56:43
this is very interesting. Right? So what can we do for kids who are heading down the path towards autism? Like what can we do to maximize the quality of their lives later on. So the one therapy that I know of that seems to have a reasonably high likelihood of success is something called Applied Behavioral Analysis. And it's just one of a category of cognitive therapies where you work with kids, many hours a day, and you just teach them very slowly, you slow down the world, and you just show them that a voice goes with a reward. And so therefore, they should like the voice where you show them once event leading to a social consequence. And as far as I can tell, a lot of applied behavioral analysis consists of just slowing down the world. And then just teaching kids stepwise. event A, leads to event B, and then rewarding them for it and doing it just for many hours a day, day after day. This is an intensive therapy. And so and so the idea would be that if it's difficult to make sense of the world, if the world is very intense, if it's hard to make predictions about the world, maybe it'd be nice to just give the kid an environment where these events happen more slowly, they can form more easily formed prior views about the world. There are some cognitive scientists who call autism, possibly a disorder of hypo priors, where it's just really difficult to have prior beliefs about the world. And if you teach them slowly enough, then you teach them priors, and they can learn about the world. So that's the therapy that works. And, and you know, one could easily imagine, maybe someday we will be able to come up with therapies that can that can accentuate that process, maybe we can improve synaptic pruning. If cerebellum is a cause of autism, you can imagine maybe putting a stimulator I'm okay, now we're getting very speculative here. But you can imagine something that you do say to the cerebellum, that that would somehow help with that process. I have no idea what that something would be. But you can imagine that stimulation or inhibition of some kind, could be assistive and maybe enhance the likelihood of, of cognitive therapy being helpful for kids. And then the other thing is, to your point, I think the one thing that we do, that is believed fairly widely among people who work with kids with autism, based on the research is that therapy is best given in the first few years of life. So between age one and six is when those synapses are at the highest rate of being pruned and eliminated. And so there's a belief that really given these therapies should happen as early as possible in small children.
Nick Jikomes 59:25
There's something interesting here that's coming to mind. So if the basic idea with why an intervention like this works is you know, you're essentially making this overwhelming sensory environment less overwhelming. And by doing that early, the development of the child will be affected such that they'll, they'll develop less extreme autistic phenotypes. So the sensory world is over overwhelming because of these inborn things in the brain. And that is causing this overwhelming environment to lead to this sort of narrow focus which is then going to impact development further, and you're gonna get autistic phenotypes more strongly. So the idea is simplify the external environment and thereby help treat the autistic individual's brain that way. So is the flip side of this, I wonder if this actually connects to why we're why we've seen an increase in autism diagnosis over time, is it because the environment itself, the background environment for humans is becoming more overwhelming overall. Um,
Sam Wang 1:00:30
I don't think there's good support for that. Um, it is true that the reported rate of autism has gone up over time. But epidemiologists are of the opinion that the reason for this is not that the true rate of autism is going up. But that are the quality of care that we give to kids, or the availability of diagnostic resources. And in options has gone up. So for example, in New Jersey, where I am, the reported rate of autism is something like one in 40, which is quite high. It's three times as high in New Jersey, as it is in Alabama. And so that's funny, like, why would the rate of autism in Alabama be 1/3? But it is in New Jersey, it seems unusual, since New Jersey and Alabama are the same country. But a plausible explanation might be? Well, the the approaches to diagnosing kids and the approaches to early childhood care are pretty different in New Jersey and in Alabama. And you can imagine that, in fact, there are kids who if they were brought up in New Jersey, might go to the pediatrician. And the pediatrician would say, yeah, we've noticed something here. And you, you know, we wanted to flag it for your attention. And so you can imagine that the availability of health care and the approach to diagnostics, could have a pretty large influence. I'll give you one example of another example of why I think this in the last few decades, the reported rate of autism has gone up, but the reported rate of mental retardation has gone down. So why is that? Like autism going up mental retardation going down? There's a phenomenon that epidemiologists called diagnostic substitution. And so the idea there is that, you know, 20 years ago, we would have called a kid, mentally retarded. But now there are different treatment options. Now there are things that can be done, where we would if we were to help a kid with Down syndrome, or we were to help a kid with developmental delay, or we were to help a kid with autism spectrum disorder, we might do different things to help those kids. And the fact that the treatment options are different for those kids might actually be a major driver for what box we put kids in, because we're trying to find ways to help kids as optimally as possible. And so I think there are good reasons to suspect that the true rate of autism is not changing.
Nick Jikomes 1:02:48
Interesting. So the true rate of autism is not changing. It's just sort of our ability to recognize and treat it. Is
Sam Wang 1:02:57
that over time. Yeah. And I should say that, I think there's more pressure to that, on that in the future. One, if it's true, that early intervention can help. One thing that we may look to in the future is finding ways to identify kids who could benefit from that attention as early as possible. So you can imagine that if we could say develop, you know, a mobile app that could help diagnose possible risk, and make that available to parents in Alabama, that might be helpful to parents in Alabama, because in fact, if they don't have access to a pediatrician who can tell them, they can at least try, you know, finding out some of these other things, and see whether there's a risk, and just maybe think of ways to be helpful to those kids as they're as they're growing and getting bigger.
Nick Jikomes 1:03:47
Interesting. One of the other things I wanted to talk to you Well, is there anything that you want to say before we move away from from Brain Stuff? Is there anything you want to discuss further on the topic of either the cerebellum and or autism, or some sort of just overall takeaway that you think you might want to give people?
Sam Wang 1:04:05
Well, only that these are, you know, on their surface, pretty, pretty different topics. But my laboratory is basic neuroscience research lab. And we're very interested in the circuit basis of how the cerebellum might account for all these things that we've been talking about at a high level and at a level of generality. In my lab, we're very interested in how the cerebellum might be a predictor of the next split second, where it takes all that input makes predictions about the future, and guides action or guides thought, or guides development. And, and that's something that we're super excited about. And I would say that although we've spent a lot of time talking about autism, my hope is that by understanding the basic neuroscience of it, that'll be a way in to really knowing what autism is, and also what it means to be neurotypical at a pretty deep level. And I think that I think the neuroscience is at a point where it's possible to start connecting the dots Between cellular and circuit mechanisms, and what it means to be diverse in our, in our neural in our neural systems. Hmm,
Nick Jikomes 1:05:10
interesting. And what's also interesting about you is you seem to have and I didn't do too much reading here, but you seem to have this sort of other side to your professional life. So you're not just running the science lab and doing the hardcore science stuff. You're also using data science and some of the tools of science to address things that have to do with Democ. Our democracy. So you have something called the electoral innovation lab. So what is this? And why are why are you also doing all this stuff?
Sam Wang 1:05:38
Oh, boy. Okay, that is a totally different topic. So let's see. So let's pause and say that the thing that's shared is that my backgrounds in physics, I, these days often call myself a data scientist, because that phrase is entered the language in the last few years. And so fundamentally, a lot of what we've talked about is me and my, in my lab members, using data analysis methods, whether it be statistics, or modeling or machine learning, or machine vision, to all to understand cerebellar function. And so that's just a general thing that we do. And yes, there is this other thing I do, that's, I would say, pretty, entirely different, which is to use data and science, to understand our democracy here in the United States, and to find ways to make it work better, to help rescue it, repair it, and make it work a little bit better. But yes, I have an entirely different group. That is that is parallel and separate from my neuroscience group. And what we do there is we work on basically the science of democracy.
Nick Jikomes 1:06:41
And and what is that? What does that mean to even talk about the science of democracy?
Sam Wang 1:06:46
It means a few different things. So let's see. So an example is something that's very much on people's minds as we're recording this, which is redistricting and gerrymandering, every 10 years in the United States. Basically, all the electoral districts of the United States get redrawn, because we have a system in the US of legislators being drawn elected from a single district, where it'd be a congressional district, or whether it be a legislative district where even county districts, and so these districts elect individual legislators, they are redrawn every decade, because the law requires that each district have equal population within a state. And because of that redrawing process. There's this, let's call it a creative process of figuring out which people go with which go together in a district. And it turns out that that is a highly technical and mathematical subject. And it also turns out that that is a process that can be manipulated by the people who are who are themselves elected. In most states in the United States. Legislators themselves are in charge of drawing the districts, which is kind of mind blowing. And, and so what my team works on, among other things, is we have a project to prevent abuses of that process. And abuses of that process are called gerrymandering, after a governor of Massachusetts, who who the processes name, the practices is named after Elbridge, Gary. So the idea is that math and data go into committing the offense of drawing advantageous districts and go into the process of politicians picking their voters instead of what should be the case, which is voters picking their politicians. So anyway, to the point, my team and I have been working on mathematical standards to identify gerrymandering, we've been working on data tools that the citizens can use to diagnose gerrymandering, we've come up with a report card, showing different ways that a map can be bad for citizens. And these are tools that we were giving people so they can fight back and give input. So journalists can tell stories about it. So citizens can weigh in. So that that that Democrats and Republicans can either fight one another or work together to create a better map. And so the idea is that all this requires data. And I've got a whole team of people working on that. Pretty hard right now.
Nick Jikomes 1:09:05
Interesting. So let's take it piece by piece. So yeah, so first, let's, let's give a peek. Let's give people a sense for the magnitude of the gerrymandering problem. Can you connect that to something like the the likelihood of an incumbent winning an election and how much that's tied to this?
Sam Wang 1:09:28
Yeah, so let's see. So, um, because of geographic, geographic factors, a lot of congressional seats in the United States are not competitive, just because, you know, maybe a lot of Democrats live together in a city, or a lot of Republicans live near one another in a rural area. And broadly speaking, about one in seven congressional seats, naturally would be naturally competitive, if you just drew lines without regard to partisanship. But if by being creative in the drawing of those lines, you can make it district 55% Republican or 55% democratic. And if you do that you think data, it becomes possible to eliminate virtually all the competition. So what little competition there as can be eliminated by creative drawing of districts, which basically is an Employment Guarantee for the legislators who represent those districts. Furthermore, you can actually create these massive inequities. I'll give you an example. North Carolina is a state that has a 14 seat congressional delegation, and it's a closely divided state. So you can imagine that a fair outcome would be to say, have seven Democrats and seven Republicans represent North Carolina and Congress, or maybe eight Democrats or eight Republicans, the actual map that's being drawn right now is likely to send 10 Republicans and four Democrats to Congress. And that is a massive asymmetry where basically, something like half the seats of a state are up for grabs, not based on how they vote. But based on how the lines are drawn. And so now we have a weird situation in which members of Congress become impervious to the will of the voters. And it doesn't matter how people vote, the only thing that matters is who drew the lines. And so this, this huge inequity, has been terrible in places like North Carolina, Ohio, those are places where Republicans have been in charge of drawing the lines. It's pretty terrible in Illinois, where Democrats are in charge of drawing the lines. But in either case, competition is suppressed. And it becomes almost impossible for a challenger to win an election.
Nick Jikomes 1:11:38
Okay. So it's a pretty big problem, I think, a good sense of that. Now, let's go back to like, first principles and the data. So starting from first principles, how do you use data to construct how would you use or are you guys using data to construct ideal objective nonpartisan maps.
Sam Wang 1:11:58
So the way it works is this, it used to be that election data was hard to get held privately, you'd use proprietary software, and you'd have to have a lot of expertise to draw those maps. And so what we're doing is we and others around the country are doing is the following. We're gathering public data about how people vote, whether they vote for Democrats and Republicans, we gather that on a precinct by precinct basis precinct is is, is a district a little zone where people cast votes, and they're counted together, we make that data publicly available. There's public software that allows people to use that data. One such famous piece of software is called Dave's redistricting app. And these software's are free and available to the public. We ourselves on my team have have come up with standard statistical standards that show when a map has gotten lopsided. So we have statistical standards that identify rigorously when a map is out of whack. Or another thing we do is we have a computer assisted algorithm draw a million different alternatives for a state. And we can find out whether a map is extreme relative to those million automatically generated alternatives. And if it's extreme relative to those million alternatives, then we can argue to a court or to a journalist or to a legislator, Hey, you did something out of line here. So the general idea is to put these tools in the hands of citizens, of journalists, of legislators, of redistricting commissioners, and we even have something that's a simple thing, a fairness report card. And if people want to read about it, they can go to gerrymander.princeton.edu, and read about the simple report card. So this is just an example something we're working on pretty hard right now, to basically find a way to blow up in the process, and make it as transparent as possible, basically, using the math and computational tools, which often overlap with my neuroscience research.
Nick Jikomes 1:13:55
So there's an obvious tension here. I mean, we are talking about politics after all. So you're using data and science to build tools to empower all various types, journalists, voters themselves, etc. With the ability to put pressure on the people that are drawing these lopsided districts, but the people doing that are themselves doing it to retain power. So it's a direct threat to their power. How do you have any good examples of success here so far?
Sam Wang 1:14:30
This is a hard fight. And so there are several examples of this. One example is there's been an outburst of reform across the country. So there are states that have passed reform that have taken the power of redistricting out of the hands of legislators and put it in the hands of citizen commissions. That's happened in Virginia. It's happened in Michigan, Arizona, California, New Mexico. These are all states where citizen Commission's are in charge of drawing the lines and so that would be one example of a case where there has been there, there has been improvements. And then several of those cases, it's new since like a decade ago was the worst round of partisan gerrymandering ever in the United States. That was 10 years ago. And many of those offenses will not occur again, because of these conditions. And so I'd say Michigan and Virginia are examples of, of significant improvements. Colorado is another example of an improved process.
Nick Jikomes 1:15:22
So no lines were redrawn there, and that that should that had something probably major to do with the outcome of subsequent elections.
Sam Wang 1:15:30
Yes, those are states. There are states where the majority of voters voted for one party, and the majority of legislators came from the other party, Wisconsin is a case where something like two thirds of the legislators are Republican, despite the fact that even if more people vote for the Democratic candidates in Wisconsin, something like 63 out of 99 state legislators are Republican. So that's a case where that happens. Michigan's another example, North Carolina is another example. All cases where competition is reduced. And representation is out of whack with how people vote.
Nick Jikomes 1:16:06
Interesting. So are there any are there any particular states right now, where you guys are focused on like, are there some really dates in terms of gerrymandering, where you're trying to move the needle? We are focused
Sam Wang 1:16:19
on different states. And so when it comes to gerrymandering, I mean, we, this 50 states, this report card that I described, we're deploying in all 50 states. But there are certain states where there is more possibility of getting leverage more possibility of either getting a better outcome, more possibility of casting a light on the process. It's some of the states I've mentioned before we were we're very interested in Michigan, Wisconsin, Virginia, Georgia, Florida, North Carolina, Texas. These are all states where we're super interested. And there's other states we haven't even talked about where we're interested in other ways to improve democracy as well, where maybe it's possible to change the way that people vote, using tools like ranked choice voting, and there's other states like Maine, and Alaska, where alternative voting rules might might trim off the extremes and make elected politicians less extreme relative to how they would be under existing rules.
Nick Jikomes 1:17:19
And where can people go to see these report cards and see what you guys are doing?
Sam Wang 1:17:24
Well, we there's two places they can look. gerrymander.princeton.edu that's g RR y, Ma, N D r.princeton.edu. And we also work on democracy we are interested in using all the tools of cognitive science, political science, computer science and math to improve democracy. And that one, you can look up@democracy.princeton.edu
Nick Jikomes 1:17:48
Great. Well, it's been a pleasure talking to you so far. Do you have any final thoughts overall that you want to leave people with?
Sam Wang 1:17:56
No, I'm very impressed that we talked about this whole range of topics I I was not necessarily expecting that but it is an impressive range. And I gotta say, it fits with what I know about the podcast, but but seeing it in action is another thing entirely.
Nick Jikomes 1:18:10
Great. Well, Professor Sam long thank you for your time.
Sam Wang 1:18:13
Thank you.
Comments