The following is a conversation with Paola Arlotta.

She's a professor of stem cell and regenerative biology at Harvard University and is interested in understanding the molecular laws that govern the birth, differentiation, and assembly of the human brain's cerebral cortex.

She explores the complexity of the brain by studying and engineering elements of how the brain develops.

This was a fascinating conversation to me.

It's part of the Artificial Intelligence Podcast.

If you enjoy it, subscribe on YouTube, give it 5 stars on iTunes, support it on Patreon, or simply connect with me on Twitter at Lex Friedman, spelled F-R-I-D-M-A-N.

And I'd like to give a special thank you to Amy Jeffress for her support of the podcast on Patreon.

She's an artist, and you should definitely check out her Instagram at lovetruthgood.

Your support means a lot and inspires me to keep the series going.

And now, here's my conversation with Paola Arlotta.

You studied the development of the human brain for many years, so let me ask you an out of the box question first.

How likely is it that there's intelligent life out there in the universe, outside of Earth, with something like the human brain?

So I could put it another way, how unlikely is the human brain?

How difficult is it to build a thing through the evolutionary process?

Well it has happened here, right, on this planet.

So, that simply tells you that it could, of course, happen again other places.

What the probability that you would get a brain like the ones that we have, like the human brain.

So how difficult is it to make the human brain?

But most importantly, I guess we know very little about how this process really happens.

And there is a reason for that, actually multiple reasons for that.

Most of what we know about how the mammalian brains or the brain of mammals develop comes from studying in labs other brains, not our own brain, the brain of mice, for example.

But if I showed you a picture of a mouse brain and then you put it next to a picture of a human brain, they don't look at all like each other.

And therefore there is a limit to what you can learn about how the human brain is made by studying the mouse brain.

There is a huge value in studying the mouse brain.

There are many things that we have learned, but it's not the same thing.

So in having studied the human brain or through the mouse and through other methodologies that we'll talk about, do you have a sense, I mean, you're 1 of the experts in the world, how much do you feel you know about the brain and how often do you find yourself in awe of this mysterious thing?

Yeah, you pretty much find yourself in awe all the time.

It's a process by which, by means that we don't fully understand, at the very beginning of embryogenesis, the structure called the neural tube literally self-assembles.

And it happens in an embryo, and it can happen also from stem cells in a dish.

And then from there, these stem cells that are present within the neural tube give rise to all of the thousands and thousands of different cell types that are present in the brain through time, right?

With the interesting, very intriguing, interesting observation is that the time that it takes for the human brain to be made, it's human time.

Meaning that for me and you, it took almost 9 months of gestation to build a brain and then another 20 years of learning postnatally to get the brain we have today that allows us to this conversation.

And so the brain is built in a much shorter period of time.

And the beauty of it is that if you take mouse stem cells and you put them in a culture dish, the brain organoid that you get from a mouse is formed faster than if you took human stem cells and put them in the dish and let them make a human brain organoid.

Which means it's on purpose, it's not accidental, or there is something in that temporal

Exactly, that is very important for us to get the brain we have, and we can speculate for why that is.

You know, it takes us a long time as human beings after we're born to learn all the things that we have to learn to have the adult brain.

From when a baby is born to when a teenager goes through puberty to adults.

Do you think you can maybe talk through the first few months and then on through the first 20 years and then for the rest of their lives, what is the development of the human brain look like?

Yeah, at the beginning you have to build a brain, right?

What's the very beginning, which beginning are we talking about?

As the embryo is developing in the womb, in addition to making all of the other tissues of the embryo, the muscle, the heart, the blood, the embryo is also building the brain.

And it builds from a very simple structure called the neural tube, which is basically nothing but a tube of cells that spans sort of the length of the embryo from the head all the way to the tail, let's say, of the embryo.

And then over in human beings, over many months of gestation from that neural tube, which contains stem cell-like cells of the brain, you will make many, many other building blocks of the brain.

So all of the other cell types, because there are many, many different types of cells in the brain, that will form specific structures of the brain.

So you can think about embryonic development of the brain as just the time in which you are making the building blocks, the cells.

Are the stem cells relatively homogeneous, like uniform, or are they all different types?

You start with a more homogeneous, perhaps more multipotent type of stem cell.

That multipotent means that it can, it has the potential to make many, many different types of other cells.

And then with time, these progenitors become more heterogeneous, which means more diverse.

There are going to be many different types of these stem cells.

And also they will give rise to progeny, to other cells that are not stem cells, that are specific cells of the brain, that are very different from the mother stem cell.

And now you think about this process of making cells from the stem cells over many, many months of development for humans.

And what you're doing, you're building the cells that physically make the brain, and then you arrange them in specific structures that are present in the final brain.

So you can think about the embryonic development of the brain as the time where you're building the bricks, you're putting the bricks together to form buildings, structures, regions of the brain, and where you make the connections between these many different type of cells, especially nerve cells, neurons, right, that transmit action potentials and electricity.

I've heard you also say somewhere, I think, correct me if I'm wrong, that the order of the way this builds matters.

If you are an engineer and you think about development, you can think of it as, well, I could also take all the cells and print them all together into a brain in the end.

So the cells are made in a very specific order that subserve the final product that you need to get.

And so for example, all of the nerve cells, the neurons are made first, and all of the supportive cells of the neurons, like the glia, is made later.

And there is a reason for that, because they have to assemble together in specific ways.

But you also may say, well, why don't we just put them all together in the end?

It's because as they develop next to each other, they influence their own development.

So it's a different thing for a glia to be made alone in a dish than a glia cell be made in a developing embryo with all these other cells around it that produce all these other signals.

First of all, that's mind blowing, that this development process.

From my perspective in artificial intelligence, you often think of how incredible the final product is.

But you're making me realize that the final product is just, the beautiful thing is the actual development process.

Do we know the code that drives that development?

First of all, thank you for saying that it's really the formation of the brain.

It's this incredibly choreographed dance that happens the same way every time each 1 of us builds the brain, right?

And that builds an organ that allows us to do what we're doing today, right?

And this is why developmental neurobiologists never get tired of studying that.

Well, it's millions of years of evolution, of really fine-tuning gene expression programs that allow certain cells to be made at a certain time and to become a certain cell type, but also mechanical forces of pressure, bending.

This embryo is not just, it will not stay a tube, this brain for very long.

At some point, this tube in the front of the embryo will expand to make the primordium of the brain, right?

Now the forces that control, that the cells feel, and this is another beautiful thing, the very force that they feel, which is different from a week before, a week ago, will tell the cell, oh, you're being squished in a certain way.

Begin to produce these new genes because now you are at the corner or you are in a stretch of cells or whatever it is.

And so that mechanical physical force shapes the fate of the cell as well.

So, from my perspective, biology is this incredibly complex mess, gooey mess.

How different is a computer or any kind of mechanical machine that we humans build and the biological systems.

Have you been, because you've worked a lot with biological systems, are they as much of a mess as it seems from a perspective of an engineer, a mechanical engineer?

Yeah, they are much more prone to taking alternative routes, right?

So if you, we go back to printing a brain versus developing a brain.

Of course, if you print a brain, given that you start with the same building blocks, the same cells, you could potentially print it the same way every time.

But that final brain may not work the same way as a brain built during development does because the very same building blocks that you're using developed in a completely different environment, right?

Therefore, they're going to be different just by definition.

So if you instead use development to build, let's say, a brain organoid, which maybe we will be talking about in a

So, if you use processes of development, then when you watch it, you can see that sometimes things can go wrong in some organoids, and by wrong, I mean different 1 organoid from the next.

While if you think about that embryo, it always goes right.

So this development, for as complex as it is, every time a baby is born has, with very few exceptions, the brain is like the next baby.

But it's not the same if you develop it in a dish.

And first of all, we don't even develop a brain, you develop something much simpler in the dish, but there are more options for building things differently, which really tells you that evolution has played a really tight game here for how in the end the brain is built in vivo.

So just a quick, maybe dumb question, but it seems like this is not, the building process is not a dictatorship.

It seems like there's not a centralized, like high level mechanism that says, okay, this cell built itself the wrong way.

It seems like there's a really strong distributed mechanism.

And if you think about, for example, different species building their brain, each brain is a little bit different.

So the brain of a lizard is very different from that of a chicken, from that of a, you know, 1 of us, and so on and so forth, and still is a brain, but it was built differently, starting from stem cells that pretty much had the same potential, but in the end, evolution builds different brains in different species because that serves in a way the purpose of that species and the wellbeing of that organism.

And so there are many possibilities, but then there is a way, and you were talking about a code.

Nobody knows what the entire code of development is.

We know bits and pieces of very specific aspects of development of the brain, what genes are involved to make a certain cell types, how those 2 cells interact to make the next level structure, that we might know, but the entirety of it, how it's so well controlled, it's really mind blowing.

So then the first 2 months in the embryo or whatever the first few weeks, months, months.

So yeah, the building blocks are constructed.

The actual, the different regions of the brain, I guess, and the nervous system.

Well, this continues way longer than just the first few months.

So over the very first few months, you build a lot of these cells, but then there is continuous building of new cell types all the way through birth.

And then even postnatally, I don't know if you've ever heard of myelin.

Myelin is this sort of insulation that is built around the cables of the neurons so that the electricity can go really fast from-

And so as human beings, we myelinate ourselves postnatally.

A kid, you know, a six-year-old kid has barely started the process of making the mature oligodendrocytes, which are the cells that then eventually will wrap the axons into myelin.

And this will continue, believe it or not, until we are about 25, 30 years old.

So there is a continuous process of maturation and tweaking and additions, and also in response to what we do.

I remember taking AP biology in high school and in the textbook, it said that, I'm going by memory here, that scientists disagree on the purpose of myelin in the brain.

So like, I guess it speeds up the, okay, I might be wrong here, but I guess it speeds up the electricity traveling down the axon or something.

Yeah, So that's the most sort of canonical and definitely that's the case.

So you have to imagine an axon and you can think about it as a cable of some type with electricity going through.

And what myelin does by insulating the outside, I should say there are tracts of myelin and pieces of axons that are naked without myelin.

And so by having the insulation, the electricity, instead of going straight through the cable, it will jump over a piece of myelin, right?

To the next naked little piece and jump again.

And therefore, you know, that's the idea that you go faster.

And it was always thought that in order to build a big brain, a big nervous system, in order to have a nervous system that can do very complex type of things, then you need a lot of myelin because you wanna go fast with this information from point A to point B.

Well, a few years ago, maybe 5 years ago or so, we discovered that some of the most evolved, which means the newest type of neurons that we have as non-human primates, as human beings in the top of our cerebral cortex, which should be the neurons that do some of the most complex things that we do, well, those have axons that have very little myelin.

And they have very interesting ways in which they put this myelin on their axons, you know, a little piece here, then a long track with no myelin, another chunk there, and some don't have myelin at all.

So now you have to explain where we're going with evolution.

And if you think about it, perhaps as an electrical engineer, When I looked at it, I initially thought, and I'm a developmental neurobiologist, I thought maybe this is what we see now, but if we give evolution another few million years, we'll see a lot of myelin on these neurons too.

But I actually think now that that's instead the future of the brain.

Less myelin, my allow for more flexibility on what you do with your axons and therefore more complicated and unpredictable type of functions, which is also a bit mind-blowing.

So it seems like it's controlling the timing of the signal.

So in the timing, you can encode a lot of information.

The timing, the chemistry of that little piece of axon, perhaps it's a dynamic process where the myelin can move.

Now you see how many layers of variability you can add, and that's actually really good if you're trying to come up with a new function or a new capability or something unpredictable in a way.

So we're gonna jump around a little bit, but the old question of how much is nature and how much is nurture in terms of this incredible thing after the development is over.

We seem to be kind of somewhat smart, intelligent.

Cognition, consciousness, all of these things are just incredible ability to reason and so on emerge.

In your sense, how much is in the hardware, in the nature, and how much is in the nurtures, learned through with our parents, through interacting with the environment and so on?

So we are born with a brain as babies that has most of its cells and most of its structures and that will take a few years to grow, to add more, to be better.

But really, then we have this 20 years of interacting with the environment around us.

And so what that brain that was so perfectly built or imperfectly built due to our genetic cues, will then be used to incorporate the environment in its further maturation and development.

And so, your experiences do shape your brain.

Like if, you know, you and I may have had a different childhood or a different, we have been going to different schools, we have been learning different things, and our brain is a little bit different because of that.

And so, especially postnatally, experience is extremely important.

What that means is a brain that is able to change in response to stimuli.

So perhaps some of the most illuminating studies that were done were studies in which the sensory organs were not working, right?

Like if you are born with eyes that don't work, then your very brain, the piece of the brain that normally would process vision, the visual cortex, develops postnatally differently, and it might be used to do something different, right?

The plasticity of the brain, I guess, is the magic hardware that it, and then It's flexibility in all forms is what enables the learning postnatally.

And how can you use them to help us understand the brain and the development of the brain?

So the first thing I like to say, and please keep this in the video.

The first thing I like to say is that an organoid, a brain organoid is not the same as a brain.

It's a system, a cellular system that 1 can develop in the culture dish starting from stem cells that will mimic some aspects of the development of the brain, but not all of it.

They are very small, maximum they become about, you know, 4 to 5 millimeters in diameters.

They are much simpler than our brain, of course, but yet they are the only system where we can literally watch a process of human brain development unfold.

Remember when I told you that we can't understand everything about development in our own brain by studying a mouse?

Well, we can't study the actual process of development of the human brain because it all happens in utero.

So we will never have access to that process, ever.

And therefore, this is our next best thing.

Like a bunch of stem cells that can be coaxed into starting a process of neural tube formation.

Remember that tube that is made by the embryo early on?

And from there, a lot of the cell types that are present within the brain.

And you can simply watch it and study it, but you can also think about diseases where development of the brain does not proceed normally, right, properly.

There are also many different types of autism.

So there you could take a stem cell, which really means either a sample of blood or a sample of skin from the patient, make a stem cell, and then with that stem cell, watch a process of formation of a brain organoid of that person.

With that genetics, with that genetic code in it.

And you can ask, what is this genetic code doing to some aspects of development of the brain?

And for the first time, you may come to solutions like what cells are involved in autism,

So if you take this human stem cell for that particular person with that genetic code, how, and you try to build an organoid, how often will it look similar?

Yes, or how much variability is the flip side of that.

So there is much more variability in building organoids than there is in building brain.

It's really true that the majority of us, when we are born as babies, our brains look a lot like each other.

This is the magic that the embryo does, where it builds a brain in the context of a body, and there is very little variability there.

There is disease, of course, but in general, a little variability.

When you build an organoid, you know, we don't have the full code for how this is done.

And so, in part, the organoid somewhat builds itself because there are some structures of the brain that the cells know how to make.

And another part comes from the investigator, the scientist, adding to the media factors that we know in the mouse, for example, would foster a certain step of development.

And so as a result, the kind of product you get in the end is much more reductionist, is much more simple than what you get in vivo.

It mimics early events of development as of today, and it doesn't build very complex type of anatomy and structure, does not as of today, which happens instead in vivo.

And also the variability that you see 1 organ or to the next tends to be higher than when you compare an embryo to the next.

So, okay, then the next question is how hard and maybe another flip side of that expensive is it to go from 1 stem cell to an organoid?

How many can you build in like, cause it sounds very complicated.

But you can really grow a very high number of these organoids.

You know, can go perhaps, I told you the maximum they become about 5 millimeters in diameter.

So this is about the size of a tiny, tiny, you know, raisin, or perhaps the seed of an apple.

And so you can grow 50 to 100 of those inside 1 big bioreactors, which are these flasks where the media provides nutrients for the organoids.

So the problem is not to grow more or less of them.

It's really to figure out how to grow them in a way that they are more and more reproducible, for example, organoid to organoid so they can be used to study a biological process.

Because if you have too much variability, then you never know if what you see is just an exception or really the rule.

Are there different neurons already emerging?

Is there, you know, well first, can you tell me what kind of neurons are there?

And how much of that variance, if any, can exist in organoids?

So you could grow, I told you that the brain has different parts.

So the cerebral cortex is on top, the top part of the brain, but there is another region called the striatum that is below the cortex and so on and so forth.

All of these regions have different types of cells in the actual brain, okay?

And so scientists have been able to grow organoids that may mimic some aspects of development of these different regions of the brain.

And so we are very interested in the cerebral cortex.

We wouldn't be here talking if we didn't have a cerebral cortex.

It's also, I like to think, the part of the brain that really truly makes us human, the most evolved in recent evolution.

And so in the attempt to make the cerebral cortex and by figuring out a way to have these organoids continue to grow and develop for extended periods of times, much like it happens in the real embryo, months and months in culture, then you can see that many different types of neurons of the cortex appear.

And at some point also the astrocytes, so the glia cells of the cerebral cortex also appear.

astrocytes are not neurons, so they're not nerve cells, but they play very important roles.

1 important role is to support the neuron, but of course they have much more active type of roles that are very important, for example, to make the synapses, which are the point of contact and communication between 2 neurons.

So all that chemistry fun happens in the synapses happens because of these cells?

Happens because you're making the cells and they have certain properties, including the ability to make neurotransmitters, which are the chemicals that are secreted through the synapses, including the ability of making these axons grow with their growth cones and so on and so forth.

And then you have other cells around it that release chemicals or touch the neurons or interact with them in different ways to really foster this perfect process in this case of synaptogenesis.

So you- So the mechanical and the chemical stuff happens.

See all Lex Fridman transcripts on Youtube

Paola Arlotta: Brain Development from Stem Cell to Organoid | Lex Fridman Podcast #32