Subscribe
Share
Share
Embed
Welcome to UCL Brain Stories, the monthly podcast series from the UCL Neuroscience Domain presented by Caswell Barry (UCL Division of Biosciences), Steve Fleming (UCL Division of Psychology & Language Sciences) and Selina Wray (UCL Queen Square Institute of Neurology). UCL Brain Stories aims to showcase the best of UCL Neuroscience, highlighting the wide range of cutting-edge research going on within the Neuroscience Domain as well as bringing you the people behind the research to share their journey of how they ended up here. Each month we’ll be joined by a leading neuroscientist to offer their perspective on the big questions and challenges in Neuroscience research, to find out what stimulated their fascination with the brain and hear how they ended up becoming part of the UCL Neuroscience community.
For more information and to access the transcript: https://www.ucl.ac.uk/research/domains/neuroscience/brain-stories-podcast
00:03
Hello and welcome to Brain Stories.
00:04
I'm Selina Wray and I'm here with my co-host Steve Fleming.
00:09
On Brain Stories, we aim to provide a behind-the-scenes profile of the latest and greatest work in neuroscience, highlighting the stories and the scientists who are making this field tick.
00:19
We don't just ask about the science, we ask how the scientists got to where they are today and where they think their field is going in the future.
00:28
And today, we are joined by some special guests in an unusual Brain Stories episode.
00:34
So we have Brain Stories' very own Professor Caswell Barry, who is Professor of Neuroscience and AI at UCL, and normally one of the hosts of this podcast, but today is in the role of guest.
00:49
And he is with Lucy Unwin, who was formerly a music journalist for BBC Radio 6 Music, and is now a children's author.
00:57
And that brings us neatly onto the reason for having you both on today, which is that Caswell and Lucy have just published a book.
01:06
Thank you.
01:06
So congratulations, guys.
01:08
Thank you very much.
01:08
Very exciting.
01:11
Their book is Inside Your Brain, 10 Discoveries That Reveal How the Brain Works. So we thought we would start this conversation by asking Caswell to say a little bit about your day job as a neuroscientist and then we'll move on to find out how you both started working together on this exciting project. So Caswell, can you say a little bit about what you do day-to-day in the lab?
01:38
It's really nice to be here on the other side of the microphone. I'm really enjoying this already. So my day job is a mix of two things, really. So my lab sort of sits across the split between experimental and theoretical work.
The thing that pulls those two streams together is that we're interested in how the brain represents space so that you can store memories about those spaces, but also how you can make decisions about how you navigate through that spaces, how you get from, say, if I walk in the front door of my house, I'm able to, for example, point up to the left where my bedroom is. And that's a really weird thing to be able to do because I can't just fly up left towards where my bedroom is. But I have this sort of weird, innate sort of feeling of space. I know where it is.
And that's the basis of you being able to do things like take shortcuts, for instance, which are really important. They're important for humans. They go wrong in some diseases, for example, Alzheimer's.
And also we'd like AI to have abilities like that so that you could have, say, I don't know, an autonomous vehicle that could take shortcuts. And so we really want to know how those things work in the brain.
02:43
So we get at that by doing essentially these two sets of things. We do theory. Some people in my lab build computational models or train AIs to do these sorts of things.Then we examine how they're doing them. And some people in my lab work with rodents, rats and mice, and we get them to do tasks like navigate around in a maze, a bit like your traditional idea of what a brain scientist should look like.
We look at what neurons, place cells, grid cells and a whole bunch of other things like that are active and what they've got to do with decision making.
03:11
And you've also got a particular interest, and it's in the title of your job title is it this intersection between neuroscience and AI, which is an increasingly growing area, but perhaps
03:23
unusual in the sense that you try and position yourself at that boundary.
03:28
So can you say a bit more about how that came about and what the goal is there?
03:33
Yeah, sure. So yeah, the relationship between neuroscience and AI is a really old one.
03:40
And we can forget that sometimes. The first people working on AI were neuroscientists, indeed.
03:45
If you look at the people who've won Nobel prizes in the last round, like Jeff Hinton originally, a neuroscientist, computational neuroscientist, Demis Hassabis, did his PhD at UCL at the same time as me in neuroscience.
03:58
And historically, since the early days of AI, there's been sort of interchange between the two fields. So initially, they're very, very close together. We took inspiration from the brain to build better AI. And those two worlds slightly drifted apart, I guess, over the intervening 50 years.
04:14
But there was a sort of realisation from a whole bunch of people, I guess, 10, 15 years ago, that there was still a lot of common fertile ground between them.
04:22
We can still take inspiration from the brain, as in to figure out better ways to do AI or come up with sort of high-level ideas about how to do AI. But what also really changed, people started to realise that we could use AI as a model of the brain. So almost all data-driven fields are going to benefit from having machine learning tools that sort through your data.
04:42
But it's special for neuroscience because we can do things like, for example, train an artificial agent to say, navigate like a rat.
04:51
There's a classic rat task, which is the Morris water maze. We put a rat in a swimming pool and it finds a little platform it can stand on.
04:57
If you put it in a swimming pool again, even though it can't see the platform because it's just under the water, it's really good at swimming towards it.
05:02
It's quite a hard task to do, actually. It's quite complex.
05:05
You can train an AI to do that, put it in over and over again, and eventually it will learn how to do that.
05:11
And if you then go and examine the sorts of neural representations the AI is forming, they have a huge degree of similarity with those sorts of things that are present in the mammalian brain when it's solving those tasks. People notice this over and over again.
05:24
I mean, one of the things we highlighted in the book was actually, if you go and look at the visual cortex, one of the
key things about it is it builds up this sort of complexity from simple lines all the way up through to highly complex representations like faces.
05:37
If you train simple convolutional networks to recognise images, I don't know, pictures of dogs, cats, the sort of stuff you find on the internet, you say you find the same sort of increasing complexities and move through them.
05:49
So there's a lot of sort of specific and functional similarities between artificial networks and the brain.
05:55
And, the artificial networks are actually just much easier to do experiments on.
05:59
Brains are a right nightmare. They're inside skulls, they're alive, they die when you do things to them.
06:03
Whereas you can use the artificial networks as a simple model for your, in place for your experiments.
06:11
And so... This whole new field called Neuro AI was born, where the belief was that we could, take inspiration from the brain to do better AI, but also take our knowledge of AI to do better neuroscience, and there'd be this sort of symbiotic relationship where both sides would benefit.
06:26
I guess I was lucky enough to be around in the early days of that as we sort of started to move that sort of field forwards and, yeah, did some fun things.
06:34
So how do you see this evolving in the future, Caswell, then?
06:37
If the AI is a really kind of good representation of these neural networks, do you see those actually becoming more dominant in the field than perhaps the kind of real brain experiments?
06:49
Is there good kind of validation between them?
06:52
Could one, you know, eventually replace the other?
06:55
It's a really interesting question.
06:56
And depending on what day you ask me, I have a different view on this.
07:00
I'll give you both of them to say.
07:03
Part of me thinks actually the two worlds are drifting apart a lot faster now than they were. There's sort of huge successes in the last three or four years of things like everyone's seen ChatGPT and sort of things like Claude and all the huge successful large language models.
07:17
And those things are driven by transformers, which was a machine learning framework that was discovered about 10 years ago now.
07:25
Transformers are probably the most different thing from the brain that AI has been using.
07:30
The sort of earlier things like convolutional networks were really quite similar to the way the brain works, which transformers are increasingly different.
07:37
And so it seems we may have hit a point where machine learning's got enough sort of self-sustaining inertia that it doesn't need as many sort of points of inspiration from the brain as we had previously.
07:48
So the two fields could continue to diverge. That's the negative view. There is still a huge amount of common ground.
07:55
We can still, in our labs, we don't have to use the latest, fanciest models to sort of explain, to build sort of artificial networks. They're going to work as models of the brain.
08:07
And actually, some people think even these hugely complex things like ChatGPT, which is like some giant transformer stacked together, trained in a horrifying way, incredibly complex. So people think the interesting thing about it is that it's complex, and it doesn't look that much like a brain, but it's trained on human data.
08:25
And so it's actually sort of distilling out some of the essence of data we used to train it is human conversations and human-written text, and it's capturing the essence of something about us.
08:36
And so even that is useful to look at, and you can sort of infer things about human decision-making, other stuff you're interested in, Steve.
08:43
So there's still a lot of fertile ground.
08:45
I think there's more to come. It's just going to change a bit.
08:48
The style of experiments are going to change a bit.
08:50
So we should come to your book.
08:51
So Lucy, can you say a bit about how this project originated? What was the genesis of Inside Your Brain?
09:00
Oh, it's evolved over such a long period of time.
09:02
We've been working on it for actual years, haven't we, Caswell?
09:04
So, I mean, me and Cas are old friends. We went to school together. So we've known each other for decades and decades and decades. And basically, he's the only person I know who's a scientist. So whenever I'm writing books and there's one book in particular that I was writing and I was trying to link chapters to animal behavior.
09:18
And I was like, oh, I need science, I need science help. So I just call Cas and ask him science questions. And he would tell me all the answers.
09:26
Essentially, I'm like, give me an animal that represents calm and its behavior, things like that. So, both of my fiction books that are published feature animals quite strongly. And he's not even an animal person, but literally I know no science people and that's as close as I get.
09:41
So Cas would be on the phone giving me very patient advice about things. And then eventually he said, if you are interested in this, I do know quite a bit about brains. So maybe we should do a brain book together. Maybe we could talk about something I actually study and work on.
09:56
It was post-COVID. I've written down in my book, I was looking through our emails this morning to try and work out when we started working on it.
And we opened that document in February 21, which is quite a long time ago, isn't it?
10:07
So I think we were just bored and wanted to do something. And you were really inspired, weren't you, by like the paucity of science books for children.
10:17
And I was too, you know, like, my kids are reading books and they're interested in science. They're not massively obsessed with it, but they're interested enough.
10:27
But everything they buy or pick up seems to skim the surface in very similar ways with just different illustrations on and never seems to go into any one particular field in any depth.
10:37
And we were just wondering why.
10:39
Is it too complicated for children to understand? And we didn't really think so.
10:43
Or is it just because publishing's too cautious about things, which is probably the real reason.
10:50
Or maybe just people hadn't done it yet. And obviously, the brain books, there's not a huge variety of choices out there that go any further than just saying, this bit of your brain does this and this bit of your brain does that.
11:02
And they're all really important. Look after them. And yeah, so there's just a bit of a gap in the market, really. And we just wanted to have something to chat about, really, didn't we, Cas?
11:10
It's true, though, right?
11:11
We were both frustrated.
11:12
You look at science books in general, it's often the same things in every one.
11:16
You're like, oh, come on, this is boring.
11:18
I know the kids haven't seen it before, but I just thought neuroscience is full of so many cool things.
11:23
Like, the experiments you read about, you're like, that is so cool.
11:27
Like, and it's fundamentally not that complex.
11:29
It's just the language that gets in the way.
11:30
And we could tell kids about, you know, crazy stuff, monkeys with robot arms and, split brain patients, things you don't always get.
11:38
Just such fascinating, topics to cover that I didn't really know anything about.
11:42
It's been really interesting for me to work on it because I started from a point of knowing absolutely nothing.
11:49
So it was a good way of filtering what was interesting and what wasn't, because, I was totally naive and full of wonder and awe and was like, wow, cool, your brain does that, everything he said.
11:59
So it was pretty...
12:00
So conceptually, the book is around ten specific experiments or stories around the brain. And I just wondered, can you talk us maybe a bit through the process of how you selected what goes in the book?
12:15
Because neuroscience is huge, it's vast.
12:18
As you just said, Caswell, there's so much cool stuff out there. So how did you distill that down what made it into the book?
12:26
We started just trying to, like, you just... vomited onto the page every single thing you could possibly think of. And we kept trying to organise it into different patterns, didn't we?
12:37
And the first thing, but we kept like leaning towards textbooky, like every direction we went in. If we tried to break it down into blocks about different things the brain did, it just automatically just gravitated towards feeling too textbooky.
12:49
And so we really needed to cut through that. And obviously stories are how.
12:55
I'm sure you can tell me what, I feel like human brains work better when they have stories around things, right?
13:01
And kids certainly respond really well to stories. So putting humans into it and their lives and their stories and definitely, I think that appeals to publishers a bit more.
13:10
And I think it just makes things more possible.
13:14
So once we decided on doing that, it was just a matter of choosing the experiments.
13:18
We didn't have a lot more, did we? that we didn't put in.
13:22
You were pretty determined that these were the ones to go in.
13:25
Well, I remember thinking we were trying to strike a balance between, like you said, the first iterations did just seem like a toned down textbook that was really boring and dry.
13:35
So it had to be sort of stories where I knew about, you know, some of the gossip or the people involved or something like this, or like random stories, you know, these things you pick up about.
13:45
What was then?
13:46
There was nearly one that made it in there, which was Golgi, or maybe it's Cajal. I think it's Cajal, how he got arrested in, any time he came to the UK, he got arrested in Oxford because he's out.
13:57
He liked drawing, you know, so Cajal famous for like drawing diagrams of bits of brains, but also liked to draw buildings. And so he was out, he'd come to some conference in Oxford in, I don't know, 1900 or whenever he was alive and was out drawing pictures of buildings at like 6 A.m. in the morning. Didn't really speak English, got arrested by the police.
14:14
Apparently this is true.
14:15
I think it's Cajal I should have, I should have checked with that.
14:17
And so, but anyway, we were looking for things like where there was, we knew, so he nearly made it in because of that.
14:21
But we were looking for things where there was like a human story attached to it.
14:26
But also where... so that was one constraint, but also getting from nothing to like modern neuroscience.
14:31
We had to do some of the historic things.
14:32
So we started with the Egyptians, because as far as we could tell, the first written occurrence of brain is in an old papyrus.
14:40
So we started there, sort of rapidly jumped forwards through some of the classics like Phineas Gage.
14:45
But I was super keen to get to sort of the modern, you know, like the last 50 years, which just doesn't seem to make it into kids' books.
14:51
So all the sort of developmental neuroscience and Sally-Anne, stuff like recent recordings from hippocampus, stuff like that.
15:01
And I guess with those, we're more likely to know some of the people involved and the sort of scurrilous rumours that we could hear.
15:07
I mean, it was called originally the brain chain, wasn't it?
15:09
Because part of the thing we wanted to communicate was that, all of our learning builds on previous learning. And if the Egyptians discovered, the thing that kids seem to respond to really well is that thing that what the Egyptians discovered about the brain was that if you can see it, that's bad.
15:25
And first principles, brains are really important. You should look after them.
15:31
So starting from there and taking it through to obviously what Cas is saying, like really interesting things that are happening in labs at the moment.
15:38
But, we wanted to show that you couldn't do that until you'd done this, and you couldn't do that until you'd done the thing before.
15:44
And the fact that each thing kind of pushed everything forwards, nudged everything forward a tiny little bit to show that progression, which is something that is important for kids to understand about science, that it takes teams and teams of people doing different things and building on each other and chipping away at stuff and just nudging things forward incrementally, bit by bit.
16:04
So that was the kind of concept.
16:06
So some things were really fascinating. So it needed a human story.
16:09
It needed to fit in a basic chronology of time and it needed to show some progress on our understanding of the brain that then could nudge into the next chapter and build.
16:22
And each thing is kind of, so the way we've kind of formatted the book is that we have a discovery, say Phineas Gage getting the rod through his head, and then what actually happened and then what we learned from it, and then brain science building blocks at the back of that.
16:37
So the brain science building blocks of Phineas Gage is the regions of the brain and how they all do different things.
16:41
The brain science building blocks for Luigi Galvani discovering electricity, having an, you know, working in neurons was that, you know, how what neurons are and how they, what different parts of them do.
16:52
Is that right, Cas? I'm saying science words and this is bad.
16:55
What I found really lovely about the book, and unusual for a children's book, is that you're not only positioning these stories and facts at the level of
17:07
the biology, but you're also engaging with quite deep philosophical questions about the mind-body problem, the page on the Egyptians thinking that the mind was in the heart, not the brain, and then moving forward to brain-computer interfaces and how that might change how we think about what the mind is and how it controls the world.
17:32
I can imagine
17:35
And I've been reading this with my six-year-old son and he loves it and responds just so, it feels like very thoughtfully to it.
17:45
And I imagine lots of kids will be responding to it in a similar way.
17:49
And I just wonder how you think about... how that might change how kids think about themselves in quite a deep way.
17:56
And it's quite an unusual thing to get across to a child.
17:59
It's made me think differently about myself.
18:01
There's a question.
18:02
It's like a pop-up.
18:02
Professor Caswell says in one of the chapters that I really like, but honestly, I'm still thinking about it.
18:08
I find split brain patients.
18:10
And when someone has got a split brain and Caswell's like, are they really one person or is it actually two people inside your skull?
18:16
No, I can't quite process that.
18:17
And it still blows my mind.
18:19
I'm still thinking about it.
18:20
I need a whole book on that.
18:21
What is consciousness. Yeah, I'm having an existential crisis because of the children's book I've written.
18:27
Well, there you go.
18:28
Next book, Consciousness for Kids.
18:30
Yeah.
18:31
That's you, Steve.
18:32
That's not what you were.
18:34
But it's true.
18:35
It's really interesting.
18:36
I think I'd be really happy if we can land one of those moments for some number of kids.
18:42
We really changed the way the world works.
18:44
I can remember. I remember watching Blue Peter as a child and them saying, Oh, look, if you cover one eye, you've got a dominant eye, and the other eye is not your dominant, and you get this switch in perspective.
18:53
I remember spending the rest of the day just going around the garden doing this but then starting to wonder, Which one's me?
18:58
If the perspective can shift, which one's actually me?
19:01
And it really had a strangely profound impact on everything I subsequently thought.
19:06
I'd love it if we'd even managed to land something like that for a small number of kids.
19:13
It's super powerful.
19:14
It sort of can change the way you perceive subsequent things.
19:17
And it's something that Lucy touched on that I really hoped would do with this book is, you know, often, even up to like, A-level, science is often presented as, here's the thing, some smart scientists, a smart scientist, figure this out and this is the way it worked.
19:32
Obviously, that's not true, because first of all, there was a whole bunch of people piling stuff on top of another.
19:36
There were a huge number of dead ends, wrong-footed, like someone comes up with an idea, pushes it really hard, and that's just wrong, you know, which is something you only realise later, I just wanted to try and get across some of the incremental sort of bit-by-bit nature of science, and it's like this group endeavor, and like what an awesome thing it is.
19:54
It's not perfect, but it's the best we've got at the moment, and so yeah, we just hope to get that across somewhere.
20:00
Continuing that point a little bit about kind of the big questions and how it really makes you question who we are, I guess for somebody discovering this for the first time, that is a natural response to it.
20:12
But how did it affect you writing this book Caswell?
20:14
Did it kind of, because I guess as research scientists we're quite focused on very specific questions, very specific details, and all of a sudden, you would have found yourself kind of taking this step back.
Did it kind of make you revisit any old concepts in a different way or kind of change your thinking on anything?
20:34
So somewhat worryingly, considering I'm a professional neuroscientist, I learned quite a lot writing this children's book.
20:42
That's actually reassuring because I was feeling inadequate that, wow, Caswell's just got all this knowledge and this story and I know nothing. So that's actually reassuring.
20:52
Because it's true, right?
20:53
Because we're all we're all so specialists, by the time you've sort of become a PhD student and a postdoc and beyond that, like you're just so ridiculously specialist.
21:02
I mean, I didn't do neuroscience degrees weren't really that, I don't think they were a thing actually when I was an undergraduate. I did biology degree. And so by the time I switched to neuroscience, I'd already switched to one brain area and basically focusing on the hippocampus and spatial memory and things tangentially related to that.
21:18
And there's all bits of the brain that are a total mystery to me.
21:21
Indeed, I look at PhD students, post-docs in my labs who did actual neuroscience degrees and have like this hugely well-rounded knowledge of things that I just don't even know exist or have sort of vague, vague feelings, you know, brain stem, what's that?
21:35
Doesn't sound very interesting, probably doesn't do anything important. Like things like that, That's not true, by the way. It does do an important thing, it keeps you alive.
21:44
And so being forced to go back and engage with some of the history was, first of all, really interesting and actually sort of filled in some of those gaps.
21:52
And so for some of these things, I went and looked at, because I was, interestingly, the other thing was I was paranoidly worried about having mistakes in this, like far beyond, say, if we were sending a manuscript to Nature or something like that.
22:05
For some reason, much, much more worried about this.
22:09
And I guess, why is that?
22:10
I guess it feels more important not to get something wrong for the kids.
22:15
And maybe, we just accept that all science that's coming out of the lab, there's probably going to be some wrinkles and misinterpretations, et cetera.
22:23
So it's like, you know, obviously if I'm submitting a paper to Nature, I should say, I do check it very, very carefully.
22:28
And I don't submit papers to Nature every day, by the way, as a secondary point.
22:33
But so I'd go back and read some of the original papers and or even sort of for example the description of Phineas Gage. There's not that much information about what happened to him. There's one-, the doctor who came along half an hour later wrote up his best guesses about what happened. But actually, we all take for granted that he was tamping this thing down, it went through his head.
22:51
No one really knows. No one was actually paying that much attention.
22:54
We just like, they sort of heard a bang and saw him on the floor with bits of brain coming out of his head. And so the rest is sort of pieced together about what had happened.
23:02
So it's quite interesting going back to the basics and sort of being like, oh, this is actually what we do know.
23:06
And there's a huge amount of interpretation knocking around on top of that. It's not always totally founded.
23:12
But it's really interesting because actually now, with my new knowledge of neuroscience, which obviously is very, very specifically related to this one book, I can see other children's books, and some really glaring errors.
23:24
So I think it's amazing that you have the respect for children, that you want it to be so specific.
23:29
Like every single, I'm so confident that every single thing in this book is completely accurate.
23:34
Back to first sources, it's been checked and checked and checked.
23:36
And then I read other books and no, you know, they obviously do their job and there's different books for different kids and stuff.
23:43
But, I don't think that level of total accuracy is actually there in some other kids' science books.
23:51
Can I say that?
23:51
Yeah.
23:52
God, tell us the names.
23:53
No, don't do that.
23:54
No, I was a bit shocked.
23:56
There's like some really very well-selling books that are about bits of brains and I think people skim over.
24:02
They're not, most of them aren't written by scientists, frankly.
24:05
So towards the end of the book, you have a page on predictions for the future, where you see this going.
24:13
And I found that very interesting because it kind of reveals something, I guess, more about your personal opinions because it hasn't happened yet.
24:22
And it got quite sci-fi in some ways, like about mind uploading and how this is going to interact with how we think about AI and so on.
24:30
So can you say a bit about that?
24:32
Like how did you come up with the predictions for the future?
24:36
How do you, because I guess we don't usually get the opportunity to really think long range like that.
24:41
So it's quite hard, right?
24:42
I mean, when Lucy said, oh, we should put a prediction in the future, I was like, oh my God, like we don't, you know, those are almost always wrong, right?
24:47
Like it's the classic thing, like we totally underestimate how quick the near future will approach and totally overestimate how rapidly the next 50 years will approach.
24:55
It's probably going to be the same as now.
24:56
So. I was quite aware of this.
24:59
Interestingly, what it reveals is where probably where my original interest in neuroscience came from.
25:06
I remember, I did an undergraduate degree, came in biology, came to London and worked in some.com stuff.
25:14
I was like traveling backs and forwards on the tube, often wondering what I was meant to be doing in my life, reading William Gibson's Neuromancer, which, awesome book.
25:22
if you haven't read it, but it's basically set in the future. It's probably now, actually, because it's quite an old book.
25:27
It's written in the 80s.
25:28
You know, it's all about uploading your mind to things and AI and stuff like that, like awesome stuff.
25:33
And that sort of really sort of more than anything, pushed me back towards neuroscience.
25:38
I did like neuroscience. A few lectures we had in neuroscience in a biology course, so that was pretty cool. I should go and get into that.
25:45
So a mixture of, I guess, that history plus, my work does sort of bridge this gap between brains and machines quite a lot.
25:53
Indeed, we touch on some of that in the book. One of the things we try and do is, for projects we're doing at the moment, is trying to replicate specific circuits.
26:02
Like if you could see the activity going through a bit of a brain, have you got enough information to distill the essence of that into an artificial network?
26:10
That has been really dominating my thinking recently. Because if you can, on a small level, in principle, it means you can on the whole brain level, maybe with a worm first or then a mouse, but in principle, you could distill the essence of the human brain into a machine.
26:25
I mean, it'd be a technically incredibly demanding task. The amount of information alone would be beyond the limits of what we can deal with now.
26:33
But one thing we've got good at, technically speaking, is increasing the amount of data we can deal with every few years, and we can double it.
26:41
So that could, you know, it could just be an engineering challenge that comes within reach relatively soon.
26:46
And so that's kind of where I saw the future going.
26:49
I mean, I think we are-- there are a whole bunch of people pushing for a closer technical link between brains and machines. For better or worse, it's probably coming.
27:00
So can I just ask you a quick follow-up about this?
27:02
Because I mean, this is a huge area, a fascinating area of how we see the link between what brains
27:10
do and what computers and machines are capable of.
27:13
And in my field of consciousness science, this has become a live issue recently because people have been, I think partly because of the development, the such rapid development of AI and these emergent linguistic and cognitive capabilities, people are naturally now saying, well, could these systems be conscious?
27:35
The debate has kind of been circling around this notion of computational functionalism, which is essentially what you were describing there, Caswell, that if you can extract out the relevant detail of what a brain circuit is doing and implement it in an artificial system, then you're essentially done.
27:53
But there's been some pushback against that in the sense of maybe brains are not computational in a deep sense, that maybe you do need something about biology to actually get you over the line. And I don't know really how to think about that, because I never really see the argument for why it shouldn't be computational in the way we think about the brain and mind.
28:20
But I'm wondering whether, you know, do you feel, have you experienced the same resistance when you pursued this line of work?
28:29
I wouldn't say, I guess I'm at a lower level, so the resistance would be less clear-cut.
28:34
I mean, almost all of modern theoretical neuroscience, at least systems level, is built upon the assumption that you can approximate the biological machinery with equations or computational code.
28:48
You need, even to approximate a single neuron, you'd need a vastly huge amount of compute, almost beyond what we can manage or that people have attempted. But essentially core to that idea is it should be computationally tractable.
29:04
There's no sort of magic sauce that makes a brain conscious. If you could, you know, and indeed a term that we were using for a while was like the functional connector. Like in neuroscience, there's somewhat of an obsession by trying to scan physical brains and get the physical connecting.
29:20
But that's less interesting. What we're interested in is the functional connector, like the functions that brain, individual neurons, circuits, and whole brains instantiate.
29:29
And I do strongly believe that you can copy that, if you can get the functional connector and then copy it into artificial machinery, a computer or whatever, then in essence, you have a copy of all of the things that brain performs.
29:42
And that would presumably include consciousness if the circuit copied was conscious.
29:48
I mean, there's a big gulf there, right? It's almost like you basically quickly get down to the, what is consciousness?
29:53
Where does it come from? Is it just some sort of... a phenomena that arises from computational complexity.
30:00
And that's kind of hard to study.
30:01
I'm glad I don't do it, Steve.
30:03
I don't even know where I'd start.
30:04
I prefer to pretend it doesn't exist.
30:07
There will be lots of things to say there, but that's going to take us off on a tangent.
30:11
But no, fascinating area. And I mean, brilliant that it's triggered these thoughts on the basis of these pages of your book.
30:20
And hopefully, you know, it's triggering similar, it's planting similar seeds. in the children who are also encountering these pages.
30:30
A little bit cynically, also, we wanted kids to end this book desperately interested in neuroscience, right? We wanted them to shut the book thinking, I want to be a neuroscientist.
30:38
So slightly we wanted the coolest possible imaginable things to go in the future to kind of as a little teaser and a little temptation for kids to think they want to know more.
30:47
So I think that's probably influenced slightly the direction your future predictions went in, maybe.
30:53
I genuinely think, I don't know, actually, I genuinely, that's the future I think we're going to get. I genuinely think that.
31:01
So I wanted to ask a question about that. You will have met some of the kids who have read your book, and I'm sure that they possibly had questions for you. So what sort of seeds have you been planting? What kind of questions are they coming to you with, Caswell?
31:16
Is there anything there that you've never thought about that somebody's kind of said, actually, I read your book? How does this work? What will happen here?
31:25
It's true.
31:26
So kids come up with brilliant questions, right? As the first thing to say, like in some ways, much harder to answer questions than say if you give a standard academic talk where the questions will be, you know, like very closely linked to what you've done, and hopefully not some terrible error you've made in analysing your data.
31:43
Kids are sort of slightly less burdened by the dogma, right? And so they just come out with these, what can superficially appear to be wild ideas, but actually sometimes are tapped in some deep truth.
31:57
I was talking to someone at work, one of my postdocs, who had given the book to her niece. And the niece was talking about the chapter on HMs. HMs, this patient had both hippocampi removed, which revealed to us that the hippocampus is important for memory. And she said, oh, how can he smell?
32:16
Initially I was like, that's a bit of a weird question, isn't it? And so I started thinking about it. I mean, so in a sense she's right, like those whole bits of the brain, like hippocampus, entorhinal cortex, et cetera, are quite closely connected to the rhinal cortex, important for smelling. I wasn't sure where this question had come from.
32:32
And actually I ended up having to look through some old papers to be like, was HM's sense of smell affected by this? Apparently it was to a certain extent. That's like, this is wild, like I never thought about this.
32:44
obviously there's sort of this sort of common trope that smells very evocative of memories and there's evidence for and against that, which I guess is where this question had come from. And it basically sparked like about half an half of conversation lab about what is the role of smell in evoking memories and how does it get integrated and what would happen in these patients?
33:08
What if you'd lost your memory? Does that affect your ability to remember smells and recognise smells?
33:13
And yeah, some of these things, like kids just come up with the wildest, coolest questions.
33:18
And sometimes see things in such a new light that you're just not ready for it.
33:21
It's kind of cool.
33:22
I mean, a lot of the questions you get at the moment, aren't they?
33:23
We're doing events to promote the book, so they've not actually read the book yet. So you get their unfiltered basic neuroscience questions, don't you? Which is generally, yeah, what happens when we... Everyone wants to know about dreams, don't they? That seems to come up quite a lot.
33:37
Yeah, it's true.
33:38
What happens when you sleep is your brain doing nothing can...
33:41
Like, no, it's doing lots of what was, so we gave it to, we very excitedly, we gave a talk, in fact, two talks, same talk twice at the Royal Institution on the weekend, a Family Science Day, but we got to stand behind the big Faraday desk, which was somewhat intimidating.
33:57
But yeah, there were some awesome questions, weren't there?
34:00
Like, you know, people were like, Do you sleep with half your brain at a time? And sometimes, one of the things we do when we talk about it is we bust some of the brain myths.
34:08
People are always like, oh, you only use 10% of your brain. I'm like, no, it's not true.
34:12
It's just not true. Where did that come from?
34:14
But yeah, the kids come up with cool questions.
34:17
It's very hard to get a presentation.
34:18
And also, one of the reasons why I think it was hard getting a neuroscience book published is because...
34:22
So much of kids' science is aimed towards experiments and demos and physical representations of science, which obviously with neuroscience, you just can't do, can you?
34:30
So I think that's probably one of the big barriers to getting neuroscience books out there.
34:34
And also, it's certainly a challenge when you're in the Royal Institution and everyone else is setting things on fire and creating, you know, they show a video beforehand of like the kind of cool experiments they have in this space.
34:45
And then we're there going, well, this is a bowl of porridge. It looks a bit like a brain. Which is pretty much as tactile as you can get, isn't it?
34:53
But there you go.
34:54
You had an actual bowl of porridge?
34:56
Oh, yeah.
34:56
A big cauldron of porridge, which we spoon into plastic bags and colour with a bit of food dye and pass them around so everyone can squish them and see, you know, have a more of a tactile experience of what the brain might be like in a family safe way.
35:11
I mean, what I find really wonderful about this project is that often on this podcast, we've had guests who said, you know what, I didn't discover neuroscience as a subject until reasonably late in life, like in certainly into teenage years, maybe even through university. It was certainly the case for me.
35:35
I just essentially didn't know this field existed until after I'd finished my A-levels and was casting around for what to do at university and then kind of started reading popular science books on psychology and the brain.
35:53
And it could well be transformational for kids to start realising that this is a field much earlier in life. So it's very exciting.
36:06
So which is also why I'm now going to ask you what's the next one?
36:09
Yeah, Lucy, what is it?
36:10
What are we going to do?
36:13
I don't know.
36:14
We were talking about it quite a lot, aren't we?
36:16
So we're thinking, are there another ten discoveries that are just as interesting as these that we could go on?
36:20
And obviously there are.
36:22
I hope there are.
36:23
I hope there are.
36:25
But do they have the same kind of human story?
36:27
Do they have, you know?
36:30
There's only ten things, right?
36:32
That's it.
36:32
That's the end of neuroscience.
36:33
We've covered it all.
36:34
It's neuroscience done.
36:38
Or we could go in a different, we're thinking about doing an AI one, but the trouble with AI is the speed of progress, obviously.
36:45
Yeah, no, it's up in the air.
36:46
Do you remember, when we were first pitching this to editors, what, two years ago, I was like, ha ha ha, we could get AI to write like the last chapter, wouldn't that be funny? And obviously by the time they got back to us, it would become a bit less funny because you're like, oh yeah, that would happen and put us all out of a job.
37:05
Which makes us question your last chapter of future predictions.
37:08
I can't see anything, see, you know.
37:11
Okay, so I'd love to hear from you both, actually.
37:14
What are your favourite bits of the book?
37:17
Lucy, maybe we can start with you.
37:19
So I think, I've kind of got two, my favourite bit to tell kids about, because I was doing a lot of assemblies before the book came out for my other books, my fiction books, and I wanted to kind of tease ahead that this one was coming at the end. And my favorite bit to get the response that you get out of children when you mention it, which I think lots of them have heard of Phineas Gage and the rod going through his head and the fact that it was that.
37:38
But just something we found when we're doing the, or maybe you just told me, Caswell, maybe it was something that we found in the research, was that after the rod had gone through his head, he sat down and vomited.
37:49
And when he vomited, some of his brain oozed out of the hole at the top. And it was described as being a half a teacup's worth of brain. And there's just something so vivid and so it just makes it so real. And if I tell a kid about a half a teacup's worth of brain oozing out the top of the hole in her head, like I've seen like meerkats, kids in the audience sitting up and being like, I need that book.
38:12
I don't know what it is about the teacup thing.
38:15
I'm so glad I ate breakfast just before we recorded this.
38:19
That's a lovely image.
38:20
Thank you, Lucy.
38:21
It's so vivid, right? It's so vivid.
38:25
But my favorite bit, for myself, was learning the very specific cells. When we're talking about it, we do place cells and grid cells in the book, but they've got like the speed ones and the edge ones.
38:39
And sadly, it got cut, didn't it? But we wanted to put in back Jennifer Aniston cells and Greg Pasty cells. And just learning about that, I thought was really fascinating.
38:46
But I love the grid cells and also the illustrations and your future predictions.
38:49
Sorry, I've got too many favourite bits.
38:50
I love the future predictions because I found that really, really, from a science fiction kind of way, as you were saying, like really fascinating and fills you with ideas about the future.
38:59
And I kind of love that.
39:00
And I just think the illustrations are amazing. Like our illustrator is Maria Jesus Contreras, and she's a Chilean illustrator. I was a fan of hers before she got hooked up in the book, because she does weird and wonderful things.
39:10
And as we were saying, it's hard to do experiments for the brain. It's hard to make it physical. And it's really, really hard to illustrate.
39:17
And we had to do a bit of pushback about, let's draw another picture of the brain.
39:23
But she was really great. And the design team, they're really great at making it nice little lateral illustrations. There's dancing kittens all across a page and like a really hippie spread. And, she draws monkeys in a way that just makes them look adorable and sad that we've been experimenting on them.
39:41
And yeah, there's just, I just love her illustrations and I just think the physicality of the book is my favourite bit.
39:47
Yeah, and really brings those stories to life, right, when you've got that real nice imagery to go along with it.
39:54
Caswell, how about you? What was your favorite?
39:56
I mean, come on, for me, it's grid sells all the way and Steve's nodding because he knows what you might not know is I failed to recruit Steve to the grid cell field when he was a PhD student, which is still my biggest failure.
40:08
It's like, so I mean, I do, I just love, I mean, I think grid cells are amazing. The fact that I remember when the first paper came out in 2005 about this, it's relatively upstate neuroscience and indeed it's not even the last chapter in this book.
40:25
I just think they're amazing. You can look in the brain and see how it's organized space.
40:29
And it didn't go for square grid lines, it went for hexagonal grid lines. It still blows my mind when I see those.
40:35
And it's a real shame 'cause one of the things related to that, there was a whole, one point in the book, we had like the whole zoo of cells. So we do mention place cells which are related.
40:45
I can't remember if head direction cells stayed in, but I've been talking to a UCL researcher, Dan Bush, so there's the famous...
40:53
About 20 years ago, someone famously recorded Jennifer Aniston cells. So you know, when you, they've got human patients with electrodes in the brain, when they looked at pictures of Jenna Aniston or indeed her name written down, you find cells that responded. And they're very, very badly understood by some of the public.
41:08
People think, oh, you've got a cell right there ready to represent Jennifer Aniston, but actually it's much more complex than that. But a researcher at UCL had been doing a very English version of this work with patients, Dan Bush. And we were going to include it, asked him if we could, but sadly there wasn't space.
41:22
He'd just been using things like pictures of the Chuckle Brothers, pictures of Greg sausage rolls that have found, basically found the Greg sausage rolls cell. But it didn't make it to the final version, which I'm kind of sad about.
41:35
Just the best comment on British neuroscience they could find.
41:40
Volume 2 of the book, the Greg's sausage roll cell.
41:43
Absolutely.
41:44
I wasn't sure it was the vegan version or the carnivore version, but you know, we can still do a follow-up study.
41:50
Just before we wrap up, can I say a bit about the creative process here?
41:54
So in terms of, you describe something like the grid cells work, and obviously this is something you, it's just been part of your life, Caswell, for years, like you just steeped in this area of neuroscience.
42:07
But then so how do you translate that together with Lucy over to the page, in sentences, descriptions that children can gel with.
42:20
How did that process of iterating between the two of you to take that very broad topic and distill it down into a few key elements for the book.
42:33
So, this is where I got, I mean, this is where I was so lucky. So, I think Lucy thinks, she's like, Oh, Caswell you had all the knowledge, I couldn't do this without you, but actually that's not the case. It's exactly the opposite.
42:43
The crazy process was, once we'd figured out what was going to be in it, I'd basically write down all the things I knew about a given field. And actually, in some ways, that was worse for the fields I knew more about, like you say, for grid cells.
42:52
We'd just write down all this stuff. Some of it totally superfluous, like grid cells have phase procession. I'd be like, I'll explain phase procession. That's like the esoteric firing of neurons.
43:01
And then I'd give these like pages of ramblings to Lucy, who would do the hard work of being like, trying to condense that down, be like, look, we've got 800 words in this chunk here.
43:10
And you've given me like some... the 4,000 word ramblings of a madman who's obsessed by this type of cell, and that's not okay for the kids to read.
43:19
And so Lucy did the hard work of basically translating it down to the core essence of things that mattered and told the story and brought it out.
43:28
And actually, yeah, Lucy doesn't let on that she knows a lot more about neuroscience than she claims. You know a lot. It's quite impressive.
43:38
I think you're underselling your role there. I mean, when we first started out, I did think that was my role.
43:43
I thought that you talked to me in pure neuroscience, and I magically translated it into human. And then I'd ask follow-up questions and realise all the stuff that you'd already filtered out. So I realised you'd already, because you're an amazing science communicator, and I think you'd already distilled it down into human, and then I just made it.
44:00
I just put jokes in, tried to make it into kid, nudged it from human to kid.
44:04
But I mean, it was hard to pick those individual things. And also our editor had a huge role in it because we would get it down to, it did go backwards and forwards quite a lot.
44:11
She was like, no, I don't, I still don't get it. Try again, try again, try again. And I mean, as we went through the book, towards the end of the book where it gets to more modern neuroscience, the process on those chapters was a lot harder.
44:24
The beginning ones where there's less information available, obviously much easier and the stories were much more specific. But the more modern it got, the trickier it was to distill it, as you say, to get it, like to find those key elements and draw them out.
44:38
But I mean, it's, yeah, I think the process was just iteration and going over it again and again and again and like filtering out and filtering out and filtering out until we got down to the to the core of it and the minimum thing we could say that would in any way communicate what we were trying to say.
44:55
You know, the bare, bare bones.
44:57
So we are almost out of time.
45:01
Thank you both, Caswell and Lucy, for coming on and sharing the backstory of producing this wonderful book.
45:08
Just a reminder that the book is Inside Your Brain: Ten Discoveries That Reveal How the Brain Works, available everywhere. Do go and get yourselves a copy. Kids everywhere will thank you for it.
45:22
But before we do wrap up, we ask all our Brain Stories guests the same question. And given you have a book full of facts about the brain, hopefully this will be relatively straightforward for you both.
45:37
But the question is, what is your favourite fact about the brain? So perhaps we can start with you, Lucy, and then go to Caswell.
45:44
I think the favourite fact I learned is how rubbish brains can be, like the making of fake memories. I was really inspired by that. Because as an outsider, I just have been in awe of the brain.
45:54
So to realise all the places where it gets things wrong was really interesting.
45:58
And the idea of how memories are formed and how that can be derailed slightly and make fake memories. I found that really fascinating.
46:06
Nice.
46:07
So for me, the thing that always blows my mind is this. So you've got a lot of neurons in your brain. I can never remember the exact amount, but you know, getting towards 100 billion.
46:17
So it's a large number of neurons. But actually, the amount of information coming in, the number of neurons that project in, it's tiny.
46:24
So if you look, take your sense of vision, you look around, you think you see everything. Actually, you've only just got over about a million neurons, not each optic nerve coming into your brain.
46:33
Almost everything you experience or think you experience, you aren't really. It's all the internal model creating it. It's just, it's a fiction and you're just dropping these little bits of information in it to keep it like as good enough to keep going. And that just, yeah, blows my mind.
46:48
Wow, that's something for our listeners to take away and think about. Mull upon until our next episode. It's been a really fascinating discussion.
46:57
Thank you so much, Caswell, for joining us as a guest and bringing along Lucy too. It's been really lovely to meet you and hear about the story of this book.
47:05
And for all our listeners, we look forward to seeing you next time.
47:09
We'd like to thank Patrick Robinson and the UCL Digital Education for editing and mixing, and UCL's Neuroscience Domain for funding the podcast.
47:19
Follow us on Blue Sky and LinkedIn at UCL Brain Stories for updates and information about forthcoming episodes.