Ep141 "What do brains and weather systems have in common?" with Nicole Rust

Does brain science need a new grand plan? Is the brain less like an assembly line and more like a weather system? And if so, what does this mean for how we might go about understanding how to think about it, and how might AI help us in the near future? And what does this have to do with how the drug riddle In got its name. Today we'll speak with scientists Nicole Rust who's been thinking about these issues. So get ready for a great brain stretch. Welcome to Inner

I was hearing concerns from the heads of funding agencies and elsewhere that while researchers had been discovering a lot of things about the brain, those discoveries hadn't been moving the needle in helping individuals with certain classes of disorders.

So you know, one of the textbooks in our field is Principles of Neuroscience. That it just keeps getting fatter over the years, absolutely, and it always has struck us that if it really were principles, it should be getting thinner. But what we just keep doing is a dated dump of all the information we're getting. But your point is we're not seeing, Ah, here's the clear pathway to solving certain problems exactly.

For certain conditions.

So for some conditions we have been moving the needle quite effectively. And so those include things like new drugs from ingrained headache or insomnia, epilepsy and pain. But there are other classes of conditions that we've been more frustrated with.

And so yeah, that's the big question.

So one of the arguments you make in your new book is that many of the pharmaceutical treatments that we have, for example, were discovered by accidents, So things like pain or ADHD or in some cases depression. So tell us about that. What's the story there?

Yes, absolutely, so those stories are wonderful, the serendipitous discoveries that happened long ago before we knew much about the brain at all. One example is the first antidepressant, which was discovered during clinical trials for the lung infecting bacteria tuberculosis. So their clinical trials for the drug for TB and what they found was the patients were joyous. There's even a picture of light in Life magazine of them dancing around.

And what was the name of that drug.

Ipronia is it?

And so that was totally an accident.

It was totally an accident.

And interestingly, you know the history of medical science is shot through with these sorts of accidents, really is Yeah, tell us about pain medications.

Pain medications? So are opioid drugs? Those come from ancient Mesopotamia where the Mesopotamians were harvesting opium from the poppy plants. And our drugs today, like oxycodone, are just a slow release form of that drug that we harvested from opium in the early nineteen hundreds.

How do they end up ingesting that?

I don't know. That's a great question.

That's a great question.

How did they figure it out?

Yeah? Yeah, yeah, okay. So and adhd M.

That's another great one. Riddlin.

So, Ridlin was developed in the nineteen forties by a chemist who was Swiss, and he was using a technique that we call try it and see what happens. We don't do that much anymore. But so he synthesized the drug. He liked it.

He gave some to his wife. She liked it too, because.

It improved her tennis game, and so he named it after her. Her name was Rita, and that's why we call it Rita Lynn. So another great story of as a drug that happened long before we understood much about the brain at all and certainly wasn't based on some big discovery about the brain that led to a new breakthrough. So there are a lot of discoveries like these.

Yeah.

Great. So your argument is that several of the drugs that we have were totally accidental. And when it comes to things that involve science as we typically do it, where we say hey, look here's the gene, here's the chemical involved, and so on, it's an enormous undertaking. So give us a sense of let's say, for insomnia.

Yes, yes, you're right, when a new discovery leads to a new drug, those discovery stories are absolutely epic. So one example of that. A drug for insomnia is subarexcent, so superreccent. The way it works is it blocks chemicals in our brain that actually keep us awake. And so the discovery of superreccent dates back to nineteen ninety eight when brain researchers discovered these chemicals in our brain the first time. They were then linked later to insomnia via

The day, and this was the chemical erectionin.

These chemical ereccin exactly.

And yeah, so they figured out this was a problem in the erecxin pathway in the brain. It was then linked to human narcilepsy. And once researchers discovered that there are these chemicals in our brain that exists to keep us awake, the assumption was that at least some of us have insomnia because these chemicals are too active. So the pharmaceutical industry went wild trying to find chemicals to block the effectiveness of these keep you awake, the erecxins

And what kind of money is involved in that?

It was about a billion dollars.

Yeah, and that's about as quick as has ever happened from a big discovery to a new therapy.

Yeah. So it's absolutely epic.

Got it. So many discoveries are accidental. Ones that aren't accidental are epic in terms of the amount of time and money they take. So where does that put us in modern neuroscience research. Let's jump to nineteen ninety eight when Eric Candell wrote a paper suggesting, look, here's the framework by which we should think about these things.

Yeah.

So in Eric Kendall's nineteen ninety eight paper, he was really channeling the ethos of an era of brain research that followed on excitement around two big new technologies, our ability to sequence genes and image the human brain non invasively with techniques such as functional nandecoresonance imaging.

And Yeah, he laid out.

A proposal of the new intellectual framework, as he called it. So, in Kendell's framework, it all begins with genes. Our genes are the code that is used to make our brain cells, which are wired into these circuits, and it's the activation of those circuits that give rise to all of mental function and in term behavior, Kendell suggested that there's one big feedback loop, so our behavior, our interactions with the world feedback to shape how our brains are wired up.

That's learning.

And Kendell focused on this big arrow from how the brain gives rise to the mind as the great challenge for psychologists and biologists to delineate the relationship between those two things.

And the arrow is pointing from genes, two circuits exsolately, Yes, experience and behavior.

Okay, so yeah, to summarize this idea about the brain and the type of thing it is, it's really set up as a big chain of causes that lead to effects. And the notion then is that when the brain becomes dysfunctional, when you have some type of disorder, it's a broken.

Link in the chain.

It might be a mutated gene that leads to a disorder that you might want to target with a drug, or it might be a part of the brain has aberrant activity which you could then target with stimulation.

So this era of brain research I.

Like to call find the broken link in the chain so we can go in and target it for a fix. And that example that we just talked about super excent it was very much of that type of find the broken link in the chain target it for a fixed type of approach that led to that big discovery.

Right, So sometimes that works, and that probably felt like real progress. I'm sure when Eric Candell no Bel laureate wrote this paper in ninety eight, he felt like, Hey, we're really simplifying this and getting this straight how one thing leads to another. But when you take a look at what's going on in the field, you think that's somehow not sufficient.

Absolutely so, there are certain classes of disorders that have really proven to be somewhat impenetrable using that type of find the broken link in a chain approach. What's then example, So they include our psychiatric conditions like depression and anxiety and schizophrenia. So those are all cases in which we do have therapies, but they don't work for everyone. And many of those therapies date to pre date our understanding

In other words, when we look at all these disorders, we think, wow, this is really somehow more complicated. And why because when we look for, let's say, a gene for schizophrenia, what do we find.

Absolutely so, in the case of schizophrenia, it's very rare to have a single gene variation or mutation that leads to the disorder. More likely, well, now that we've sequenced lots of genes, we know that it's variation in hundreds of genes that are tied to the condition. So if one identical twin has schizophrenia, the chances of the other identical twin having schizophrenia they're fifty percent. It's not one

Play, and these intertwine in ways that are super complex. As a side note, you know, the first gene pulled for a major disease was for hunting tins and it was a gene and if you have that gene, you're going to die of hunting tins unless you dive something else first, Yes, and everyone thought this is great, We're going to figure out the gene that goes with every disease, and it turned out to be much more complicated.

Yeah, And even now, thirty years later, we still don't have an effect of treatment for Huntingtons, although fingers crossed, it looks like maybe there might be one on the way in clinical trials, but it's taken over thirty years to get there, even when we knew exactly what the problem.

Was, right, Okay, So but for something like schizphrenia or major depressive disorder, we're looking at something that's much more complicated. We can't even find a single gene for it. As you said, we find hundreds of genes. So where does that put us?

Yeah, so I think researchers are definitely waking up to the idea that this idea about the brain as a big, long chain is probably oversimplified. The human brain is often held up as the most complex thing in the entire universe, and this chain of causes the lead to effects, it's not very complicated. So we might ask ourselves, well, what's so complicated about the brain and what are we missing in this chain?

Like idea.

And by the way, it still might be causes leading to effects, right, but it's the feedback loops at every stage exactly.

That's the complexity that we're beginning to embrace. So causes that lead to effects that feed back on themselves as causes. The brain is not like a chain in this type of idea. It's a whole different type of thing.

Right. So your analogy that you make in the book, which is wonderful, is like a weather system, right, So unpack that for us.

Yes, so you can think about these complex systems that have many interdependent parts, and the weather is a terrific example of that, and you can think about when a complex system goes awry, it's a lot like a weather breaking out into a storm.

Yeah.

We know a lot about systems like these because we've studied them quite extensively. And one thing we know about them is they're very, very hard to perturb in ways that you would want to shift them out of their storms, which in the case of the brain, would be shifting the brain from a less healthy to a more healthy state.

And one second tangent in your book, you tell us a wonderful story about Johnny von Neuman and weather I had no idea tell us then.

Yeah.

So in the nineteen forties, the end goal of weather research really was to control the weather. Researchers wanted to not just dissipate hurricanes, which is a worthy goal in and of itself, but they also even wanted to weaponize the weathers.

This was the US government that, This was.

The US government.

Yeah, so they were very interested in funding weather research with that end goal. Part of von Neumann's development of the first computers was explicitly in the end goal, first predict the weather, then learn how to control it.

And so researchers tried that out and it didn't work out.

So well, and it still hasn't worked out.

It still hasn't worked out.

And why it's because the weather's so complicated.

It's so complicated, and it is a system because you have these big feedback loops, right, any type of intervention you try to do will reverberate in unexpected ways, and so that's what makes these systems really really difficult to control.

So when we look at something like major depressive disorder, the temptation to say, look, if we could just find the gene. There is no theugen, but if we could just do this pharmaceutical here there, we can solve this, and that has proved to be ineffective precisely because of the complexity of the system here.

Yes, absolutely, and so.

One example that you talked about in the book was emotions research and the one hundred years' war that's happened there. So explain to us what that is.

Yes, So, our ability to understand what's happening in many of the psychiatric conditions comes down to wanting to be able to measure an emotion, say, in the brain, and that's proven to be very difficult to try to do.

So.

Researchers for over one hundred years have been arguing about what types of things are emotions in the brain and how are they organized. It might be that different emotions like fear and disgust and happiness, they might be organized in kind of their own little compartments in the brain, kind of like our sensory systems where we have one part of our brain for vision and another part for hearing, Or might be that they're much more intermingled in the

Yeah, and so that's led to this one hundred years war because there are people on both sides of this argument. It's either separate or it's spectral.

Absolutely, yeah.

And so we're looking at things like emotions and we all want an explanation for it. But the question is what will it take for us to be able to answer something like that.

It will take an appreciation that the way that emotions manifest in the brain is not going to be simple and straightforward. It's not going to be an individual neuron that's activated when we feel aggression or sadness.

It's going to take.

Embracing these ideas that in the brain, if brain area a sense information to b be send information back to A again, and so we expect emotions to be reflected in ways that kind of reverberate and dynamically evolve in the brain as opposed to snapshots.

That you could take a picture of.

That's a really simple way, and in fact, it might not even be that we can talk about in let's cut two one hundred years from now, that we can talk about area A and area B, right, because in a sense, the whole brain is spectral in the sense of you've got eighty six billion neurons that are all doing their things, but they don't have border walls between them. So what we do as neuroscientists as we say, oh, this area seems to be involved in blah, but boy, these things are spread out.

How do you think about that?

When you think about the brain in terms of how compartmentalized is it? How much is everything everywhere all at once? What's your take on that? You've thought a lot about it?

I know, I mean, you know, so starting with the experiments of Carl Lashly whatever last century, as you all know, Lashley was trying to figure out where is a memory stored? So he trained little mice to run a maze and then he would cut parts of the brain to see, Okay, where is that memory stored? So if I take all these rats and I cut different parts. Where can I find the memory store? And what he found is that none of the experiments yielded anything because the memory is

Absolutely, Yeah, I completely agree with that. Yeah, but I think we're also ready for the first time in history to take on the complexity like we've never been able to do this before. So it is an exciting era for brain research to build on this oversimplification.

That's right. And so you've been looking at other systems and other scientific voices from the last fifty years that have suggested things. So what do you see as possible ways forward there.

Yes, so.

There's been a long thread through brain research. It's been more of an undercurrent than the most dominant idea that the way we should be thinking about the brain is something much more akin to the weather, a dynamical system where we're interested in how it evolves in time in terms of things like it's patterns of activity and how it is structured, not just as a computer, but something

And tell us about cybernetics.

Cybernetics was this idea that the brain exists to control the body and interact with the environment and a big feedback loop.

That was the gist of cybernetics.

Yeah, and so that was Norbert Veener and other people have built on that idea of having dynamic systems, lots of feedback loops and so where do you see that moving forward. So if we think today, okay, look, let's think of the brain as a very complicated system with lots of feedback. How do you tackle something like that.

Well, there are a couple of different things are really important. One is because these types of systems are so integrated, you have to measure all their parts at the same time. You can't measure their parts one at a time, And for the first time in history, we're able to do that. Twenty years ago, when I was recording from brain cells and looking at their activity, I was able to look at one at a time. Today we can record from one million brain cells simultaneously in a mouse.

It's remarkable.

That's exactly the type of data that you need in order to understand how all these brain cells are interacting with one another. We also have to build these really complicated models to make.

Sense of these dynamical systems.

Again, causes lead to effects that feedback on themselves as causes. These are not things you can think through and try to reason through. You need computers in order to do this, and for the first time in history, we have artificial intelligence of a type that can actually help us sift through and make sense of this data. And build computer programs that rival something as complicated as the types of

Give us a sense of, for example, David Anderson's lab at Caltech, how he looks at this giant data and figures out, hey, here's a way to capture what's going on.

Yeah, that's a great example, and it's so relevant to a problem that we've really been struggling with, and that is how do we measure an emotion in the brain. So in David's lab, he is really interested in the evolutionarily ancient emotions like aggression, fighting, or feeding, and he looks into a part of the brain that we know is involved, the hypothalamus. And we know it's involved because if you naturally, if you put two male mice together.

They'll fight. They're aggressive.

If you stimulate the hypothalamus of a mouse, even if they're all alone, it will cause that type of aggression. And if a mouse has damage to their hypothalamus, they won't be aggressive anymore. So we know the hypothalamus is

In other words, it's not like the cells turn on and then turn off.

Okay, yep, yeah, It's just not an obvious answer. And so these researchers in this group they started to shift to this new way of thinking about the brain, not as a big chain, but again as one of these

So you're measuring all the cells and you're representing it as a point on the landscape.

Yes, that's right.

So at any one point in time, the activity and the hypothalamus will be somewhere on this landscape, and where it ends up falling in the valley along this long line determines how aggressive the mouse will be. At one end of the valley, that will translate into a mouse that's not going to be aggressive, perhaps because what they've seen is maybe a female mouse or not a mouse at all.

On the other end.

Of the valley, that's where the population ends up sitting, that will cause the mouse to be aggressive. And they could see that this was true, not just by doing observational work where they observe what's happening in the hypothalamus, but they actually could use this new generation of tools where they could causally perturb the system and confirm that that indeed was causing the mice to be aggressive.

Amazing. So instead of looking at a particular cell or a group of cells and trying to think through it, you have to take all the cells and collapse that high dimensional activity onto a point on a landscape, and then you can start describing what that landscape is doing.

Absolutely, and the big shift here is that that landscape can't be shaped by a big chain of causes.

It lead to effects.

The formation of the landscape depends on thinking about the brain as having these big feedback loops in it.

Yeah, you know, it's funny. Even in any neuroscience textbook, you know you have sell A talks to sell B. And of course, so we know that every cell in the cortex is talking to you about ten thousand of its neighbors, and lots of these are very complicated feedback loops, and of course you have excitatory and inhibitory neurons, and so straight away, I think any clever student looks at this and says, wait a minute, something is something is

That's a really important question. Back in the nineteen forties and fifties, we used to have an ecology food chains, and then at some point they became food webs because

Yeah, And even if dynamical systems science as we understand it now turns out not to be the full picture, at least we're getting closer.

Absolutely.

Yeah. And Nicole, despite the limitations and where neuroscience research has gone, you're very optimistic.

Tell us why absolutely.

When I started to write this book, I actually wasn't sure where it would lead, and I started from a place of kind of confusion and even a little bit of pessimism because I could see that there were these certain conditions for which we were get a little bit stuck. On the other side of writing the book, I'm unequivocally optimistic about the future of our field for the conditions like the psychiatric conditions and their degenerative conditions.

And why it's.

Because I see that the changes that need to happen are already happening in our fields. Right we were oversimplifying the brain. We were treating it like this chain of causes that lead to effects, and it was just a massive oversimplification of the most complex thing in the entire known universe.

But now researchers are starting to embrace.

This important type of complexity that we can again for the first time in history, because we have new biotechnology, we have artificial intelligence. For the first time, we're really to study the brain in this way, and I am very excited about the idea that that will be the key to unlocking progress for all of the millions, billions actually of individuals who are suffering from these conditions.

That was my interview with Nicole Rust. This conversation circled around the idea that the brain may not be the kind of object we once hoped it was. For a long time, neuroscience advanced under a parsimonious assumption that if we can you could just identify the right pieces, the right links in the chain, the story would come into focus. Genes lead to proteins. Proteins built cells, cells form circuits, Circuits generate thoughts and motions and behavior fix the broken link,