CARTA: The Idea Organ - Questions, Answers & Closing Remarks

0:25 Thank you, everyone.

0:28 That was a wonderful series of thought provoking talks about the human brain.

0:32 On behalf of CARTA and my co chair, I'd like to start the Q&A session.

0:37 Speakers, if you haven't already,

0:39 please turn on your cameras, unmute your microphones.

0:43 We've collected audience questions throughout the day

0:46 and we will address them now.

0:48 With that, I'm going to present the first question,

0:51 and this is addressed to Dean Falk.

0:55 Does the theory you posed linking auditory and trainment and language

1:01 to a mother's footsteps rely on the assumption

1:05 that language first evolved from vocalizations?

1:09 Some researchers argue that our linguistic capabilities evolved

1:14 first in gestures and later in spoken language.

1:18 I'm wondering if that view could also be consistent with your hypothesis.

1:23 We do address the vocal channel,

1:27 mainly we're well aware that gestures are part of the package.

1:32 In fact, when you talk about motherese, for instance,

1:36 we talk about baby talk and the vocalizations and the affect and the prosody,

1:40 but also there's a huge component that is gestural.

1:45 We acknowledge that, but we did not speak to that or incorporate

1:50 a theory to do with the evolution of gestures,

1:55 per se within the ideas about entrainment of movement to sounds.

2:02 Thank you for the question.

2:05 Do you know the time range for novel one selection sweep?

2:08 An environmental disturbance relating to lead makes me think of volcanic trends.

2:14 As volcano plums deposit large amounts of lead

2:17 and other heavy metals like cadimu and contemporany people

2:20 who live through them exhibit elevated urinary and blood

2:26 concentration of these half metals for long periods afterwards.

2:30 That's a great question.

2:32 It's hard to estimate when we see the Nova one swap,

2:36 but we estimate about 200,000 years ago.

2:40 It's quite recent.

2:42 But definitely the environmental exposure was there before that.

2:47 The genetic comes later.

2:50 Great.

2:50 I have the next question for Nenad Sestan.

2:54 Nenad, the person ass.

2:57 I'm curious about the role of retinoc acid in shaping the neural development.

3:02 In immunology, retinoc acid guides T cells to go

3:07 to the gut and also regulate immune cell development.

3:10 I'm wondering if some of retinoc acid's effect on brain

3:14 development could actually be indirect working through the immune cells,

3:19 it shapes rather than acting on neurons themselves.

3:23 Thank you.

3:24 That's a very good question that I don't have a direct experimental answer.

3:31 Just to try to little speculate,

3:35 I would not be surprised that that brain does not develop in a vacuum.

3:41 I would not be surprised that immune cells don't have effect.

3:46 We can see microphages that come from immune cells

3:50 and microglia that is a part of immune cells,

3:52 so that's not native to the central nervous system as early

3:56 as even early fetal to early migestation in humans and other species.

4:01 That would not be surprising.

4:03 We don't have any evidence that these cells do release,

4:06 the source of retinoic acid seems to be multiple cell types, not just neurons.

4:11 Also we see glial cells, we see cells outside the cortex too.

4:16 Also it's a very promiscuous signaling pathway that affects

4:20 many cells and systems not just within the brain,

4:24 so I would not be surprised at that effect,

4:27 but we don't have any evidence for it, but we cannot exclude that.

4:31 Thank you.

4:32 I have one that is for me as well from Adrian.

4:39 In the Nova one organoid experiment

4:42 where those homozygotes genomes introduced or heterozygotes.

4:47 If the later case which gene is dominant,

4:50 curious how the neanderthal sapiens hybrid phenotypes

4:53 would have fair with the potential lead exposure.

4:56 For the lead experiment, we only work with homozygotes.

5:00 These are clones that have either the two copies

5:04 of the modern version or the two cops of their cake version.

5:07 We haven't done this experiment that you are suggesting, which is pretty cool.

5:12 I have one for Joe Paradiso.

5:16 The question asks, in your project, Jam Underscore boot,

5:22 you described the AI model as acting like a musical brain.

5:27 Based on your research,

5:29 how does this musical brain behave differently than the human composer's brain?

5:34 How do you see the collaboration between the two evolving?

5:38 We've at least working with living composers like Jordan Rudus,

5:42 the person we've worked with most closely.

5:44 He it provided hours of data of him playing in different styles and basically,

5:52 it took a transformer model that was developed

5:55 for generic music and honed it into his playing style.

5:59 He could anticipate what it was going to do because it's what he would do.

6:05 In the beginning, especially, it would take some wild jumps once in a while.

6:09 You see Jordan smile and laugh and play another line to bring it back.

6:12 He called it Baby Jordan.

6:14 Once in a while, it would get frantic and take a tantrum,

6:17 so to speak, and he'd bring it back.

6:19 As we honed it a little more with reinforcement learning,

6:22 it got a little better behaved.

6:24 It's more minor, and it still would give him ideas,

6:28 but they'd be within the bounds of what he would expect.

6:31 It is a partner where you can trade lines back.

6:34 For me playing with Jordan's model,

6:36 I would never play those quarts because I don't know them.

6:39 He does.

6:40 I have somebody else's expertise at my disposal that's looking at what

6:43 I'm doing to come in and bring me to a different place.

6:47 Certainly, as you train the model, it evolves in its capability.

6:52 They're just going to keep getting better.

6:54 But the same token, as a player working with it, it can put you into new places.

6:59 It works on both ends.

7:01 I have one for Bruce Miller.

7:05 Question is or the statement, you explain in your talk that as patients

7:11 begin to experience the symptoms of SVPPA,

7:14 they often develop increased visual attention

7:17 and artistic output which you argue

7:20 reflects something evolutionarily fundamental about

7:23 the artistic nature of the human brain.

7:25 Do you think there could be any shared neural mechanisms between

7:30 the visually heightened experience associated

7:33 with SVPPA and those seen in synesthesia?

7:38 With Ann Adams, in particular, we wondered whether she had been synesthetic

7:44 because she was really presenting individual notes visually.

7:51 Looking at her notebooks,

7:54 I think this may have been a component of her creativity,

7:57 that transmodal association of sound with visual output.

8:04 I think it's a really smart question,

8:07 and I think in some cases, that really does occur, yes.

8:11 There was a question to me asking I

8:16 mentioned evidence that OPCs can participate in synaptic pruning.

8:21 Do you think this contributes to the increased

8:24 cortical plasticity seen compared other primates.

8:28 Yeah, that's a great question.

8:29 The evidence for OPCs contributing to synaptic

8:34 pruning is actually found in rodents,

8:36 so it's not something specific to primates or specific to humans.

8:40 But I tried to lay the stage or propose that if we have more OPCs,

8:45 perhaps there's more around to do this in a more efficient

8:50 way or to participate more broadly in this compared to other species.

9:01 It's something that we're trying to test

9:03 in a dish by doing these mix and match experiments

9:06 of human and chimpanzee OPCs with neurons to test

9:10 at least between those two genetically closely related species,

9:13 whether there's a difference in activity or function.

9:17 But yeah, it's really speculative at this point.

9:22 Great question.

9:25 Sorry, that I'm doing all the talking.

9:28 But I'll just continue asking the questions.

9:32 The next one's for Miles Wilkinson.

9:36 The HR 123 knockout mice showed a decrease in cognitive flexibility.

9:43 But as you pointed out,

9:44 mice lack many of the human specific mutations in HR 123 to begin with.

9:50 How confident are you that the knockout phenotype reflects

9:54 what the human specific changes to HR 123 actually do?

10:00 I guess the short answer would be that we're not confident at all.

10:04 Basically, one of our questions was, what does HR 123 do?

10:10 I spent a lot of my talk talking about that.

10:13 That was one of our main focuses and in Bivo,

10:18 as you said, evidence that it's involved in cognitive flexibility.

10:22 But of course, we only know that in a mouse.

10:24 And so we don't know if that can relate to humans,

10:27 but we're obviously excited about the possibility that it might.

10:30 By the way, of course,

10:32 our big interest is what's the difference between humans and chimps.

10:37 By the way, there's actually evidence that chimps

10:40 have more cognitive flexibility than humans actually.

10:45 I would say the evidence is limited, but there's some evidence for that.

10:48 I think it's an intriguing question for the future to try

10:52 to my guess is it depends on the kind of cognitive flexibility,

10:55 whether humans are better than chimps.

10:58 We'll be fascinated to try to deal with that in the future.

11:01 But obviously, it's a little bit difficult

11:03 to answer that question in humans in Bivo.

11:06 Genevieve, I might take on this question

11:09 at the top for speakers studying speech areas.

11:13 How does asymmetry differences dyslexic humans with no other

11:18 learning issues affect your understanding of the necessity of asymmetry?

11:23 I love the question.

11:27 I think there has not been enough

11:30 research and study of individuals with dyslexia emphasizing

11:35 the enormous strengths that they have and things

11:38 that make them unique and different in, you know, very positive ways to others.

11:44 One paper on this from Virginia Sturm,

11:48 looked at a group of dyslexic children and found better spatial navigation.

11:57 Visual memory and I think this is good,

12:03 but it's complex increased emotional sensitivity in her neurophysiology

12:09 lab to stimuli that we're potentially emotion evoking.

12:18 We know often that individuals with dyslexia,

12:23 there smaller white matter tracts in the left hemisphere.

12:28 What happens to the other hemisphere?

12:31 Is it possible and I'm sure it is that many people

12:36 with this asymmetry develop extraordinary capabilities

12:41 using the other side of the brain?

12:43 There's a follow up question about HR 123 that was in the Q&A.

12:49 Is there anything known regarding

12:51 homo-sapiens differences versus neanderthal and Deniseven?

12:56 No sequence differences that we've seen.

13:01 The alteration seemed to be prior to that.

13:05 Then on sent to me for Allison, how does the Nova 1 protein sequence in non

13:12 human primates compare to modern and archaic humans?

13:15 Have you tested organoids derived from non

13:18 human primate IPSCs and investigated their maturation patterns?

13:21 Yes, we've done this before with chimps and other non human primates.

13:27 They all cluster together.

13:29 Humans is really the outlier.

13:31 We are the only ones with this modern version of Nova 1

13:36 and probably part of that is causing

13:40 our developmental delay compared to other species.

13:44 All other primates would have the archaic version of Nova 1.

13:49 Right next questions for Alex Pollen.

13:51 Thanks for a great talk.

13:53 I'm wondering if you view the differential in brain evolution

13:58 across species in reward areas as an evolutionary trade off.

14:02 For example, we seem to have

14:04 significantly enhanced connectivity within dopaminergic systems

14:08 associated with our capacities for sustained

14:11 attention and sophisticated cognition surrounding reward.

14:15 However, the same systems closely associated with pathological

14:19 forms of cognition such as ADHD and addiction.

14:23 Yeah, thank you for that question.

14:24 I think it's really important to consider

14:27 evolution in terms of trade offs and many

14:30 of us are interested rightly in the genetic

14:33 and cellar changes that enhance our cognitive functions.

14:37 But these come at a cost and can change a fitness optimum,

14:41 creating new challenges for the brain.

14:43 Certainly new functional requirements for working

14:48 memory or reward or other things that increase demand on the diplomurgic system

14:54 are likely to also potentially confer vulnerabilities.

14:59 I'm not sure how much of the vulnerabilities

15:02 of our dopamergic system reflect the new demands.

15:05 There's some evidence for increased connectivity,

15:08 innervation density to certain areas of medial

15:12 caudate and nucleus succumbens versus how

15:15 much of the vulnerability comes from the unequal

15:17 scaling that occurred during brain expansion,

15:20 the sheer expansion of the territories to which these project.

15:24 I think both of those models are worth

15:27 exploring through a combination of comparative and experimental methods.

15:31 It's tough to distinguish.

15:32 Then, of course, the third force would be the increased lifespan.

15:35 Maybe not for those neuropsychiatric disorders mentioned in the question,

15:38 but for the neurodegenerative disorders.

15:40 These neurons must persist for decades longer in humans,

15:44 coupled with these increased demands.

15:46 I think for all of these traits, it's worth considering the trade offs.

15:50 What we're interested in is the secondary idea that the cell types

15:55 themselves have evolved to our large brain cellar ecology and these new

16:00 demands and that we might be

16:01 able to discover compensatory adaptations that might

16:05 shed light on protective factors that we could harness for therapeutic purposes.

16:09 Yeah, thank you for that question.

16:12 The next question from the audience is for James Rilling.

16:16 What do you think drove the evolution of these neuroanatomical differences?

16:21 Did a need for language come first?

16:23 Was it the chicken or the egg?

16:26 Yeah, well, hi, everybody.

16:27 It's great to be here.

16:29 Yeah, it's interesting to think about why language

16:33 evolved in only humans among all the primates.

16:38 There's some different ideas about that.

16:42 One idea that I think is compelling is that language

16:47 may have evolved to allow us to transmit culture,

16:51 especially across generations that we accumulate

16:57 this body of knowledge and somehow we need to transmit that to the next

17:04 generation and language facilitates that effort.

17:07 Language, the other really cool thing about language,

17:12 very powerful thing about language is that it allows us

17:17 to have knowledge about things that we never directly experience ourselves.

17:23 Think about how powerful that is something

17:26 can happen somewhere else in my environment, and I will not have witnessed it,

17:32 but someone else who did can tell me about it,

17:35 and it expands my knowledge of my environment dramatically.

17:41 That's something else that I think we should

17:44 think about when thinking about the evolution of language.

17:46 Then also the ability to just have this shared

17:51 intentionality with others so that we can collaborate on things.

17:57 I think all of those are potential selection pressures.

18:02 Anyway, those are some of my thoughts on that interesting question.

18:07 I'm not sure that I fully answered it, but yeah.

18:11 These are hard questions to answer, I think, so that was a great job.

18:15 The next question is open to everyone,

18:18 so we'll see who wants to answer first and everyone

18:22 can take a shot at it if they want.

18:24 How do we escape the accusation of human exceptionalism after

18:29 hearing all these amazing examples of uniquely human brain specializations?

18:35 Go ahead, Alex.

18:37 I think I could take a stab at this from a few different perspectives,

18:41 and that is a great question.

18:44 It's one of the reasons I always like

18:46 to use the word specialization instead of unique adaptation.

18:50 Different lineages across the phylogeny are specialized for their own

18:54 environments and optimized in a way that we would not be.

18:57 I think one way we can tackle

18:59 this and take the caution suggested by the questioner,

19:02 at least from a genomic perspective is

19:05 to look equally for signals on the human lineage,

19:08 as well as the chimpanzee lineage or other lineages.

19:11 We heard the beautiful talk about HRS,

19:13 and if there were more time folks who have identified

19:18 human accelerated reasons have also

19:19 reciprocally identified chimpanzee accelerated regions.

19:22 Now with the Zoonomia consortium, we can do this across any branch.

19:26 That allows us to ask whether certain functional ontologies of genes

19:30 might be uniquely changed in human versus recurrently changed in other lineages,

19:35 and it may give us perspective on what is truly distinct on our branch.

19:40 I'll open up the discussion there,

19:42 but I think this is absolutely worth considering.

19:45 Thank you.

19:46 Anyone else?

19:48 I think Alex has done an extremely eloquent job explaining this.

19:55 But the reason we are in this situation that we

19:57 can do this and we can think about it.

19:59 It's not that we are not recognizing other species.

20:03 The second one, it's part of our curiosity

20:06 and our ability to do this as a species.

20:09 That's why there is a bias.

20:10 But there is a second bias which Alex and other have already mentioned.

20:14 We are interesting in human biology and diseases,

20:17 and there is evidence as my colleagues have done extremely well,

20:22 is that some of the susceptibility may come

20:27 from unique features and changes across the human lineage.

20:32 We need to also explore that.

20:34 To what extent that explains, that's a separate question.

20:37 Those are two reasons why we have

20:39 this extreme anthropocentric bias in explaining this just

20:43 because we are the only one that can do this, or at least study this.

20:47 Secondly, we think that this is important

20:49 for human biology and disease and conditions.

20:53 There's an argument that we could

20:56 learn from convergent changes in other lineages.

21:00 I agree with the point.

21:01 There is maybe rightly some human aceptualism because we are

21:04 the species that's able to study and ask these questions.

21:07 That said, our brains only reached the size

21:11 they did in the last few million years,

21:14 and other lineages like the cetacean clade invented

21:17 large brains 30 or 40 million years earlier.

21:20 They may not have the same degree of association cortex as many talked about,

21:24 so they may not have some of our unique capacities for abstraction.

21:29 But from a cells perspective at dealing

21:32 with the challenges of connectivity in a large brain,

21:34 they may have had an order of magnitude, more time to adapt to the challenges

21:40 of connectivity in a large brain environment.

21:42 We may be able to learn cellular

21:44 adaptations by moving beyond the human exceptionalism perspective.

21:49 I might give a slightly broad answer to the question.

21:55 Thank you, Alex.

21:56 Really well said.

21:57 I think we have to understand our deficiencies as a species as well,

22:03 and we have wreaked unbelievable havoc on this planet,

22:07 destroyed species after species.

22:11 I think understanding why that happens,

22:15 what are the deficiencies in the human brain that allows us to continue and how

22:21 to remedy it is I think a pretty

22:24 important question for the survival of the planet.

22:28 The next question is specifically directed to Alex.

22:32 Is there evidence that the connectivity issues

22:36 you highlighted were indeed the selective pressure

22:39 for the dampened response to oxidative stress we see in the modern human brain,

22:45 or is it possible that exposures to pollutants such

22:49 as fire smoke or the lead exposures discussed by Allison,

22:53 could have driven that change?

22:55 Thank you.

22:55 That's a great question.

22:56 Sometimes it's easier to approach the how the the why.

23:03 When we want to explain the why, we can come up with potential stories,

23:07 but these could be considered hand waving.

23:10 We started with a hypothesis that there might

23:13 be compensatory adaptations that buffer increased metabolic stress,

23:18 and we found some evidence consistent with that.

23:21 I think to defend whether these are

23:23 adaptations would take much stronger evidence.

23:26 It would require both functional evidence as well

23:29 as evidence for signatures of selection on the genomic level.

23:34 But then your question is even more

23:38 challenging is what drove these adaptive signatures?

23:42 We have a plausible hypothesis,

23:44 but it would be very difficult to fully address that here,

23:50 we could think about what is most likely based

23:53 on when these are manifest and which cell types,

23:56 do we see it broadly across tissues in the body where you might be managing,

24:00 say if the hypothesis is pollutants, do we see things in the liver,

24:04 the kidney or other changes that have kind of co-evolved or adapted with this?

24:08 Or is it very specific to a vulnerable

24:11 population of neurons that would be more affected

24:14 by unequal scaling or new functional requirements that might

24:17 be more consistent with this connectivity vulnerability hypothesis?

24:23 Answering the why is tough,

24:25 but it is something that I think drove all of our interests.

24:28 Thank you for that.

24:30 Why is hard.

24:32 The next question is for Joe Paradiso, in your talk,

24:38 you mentioned that we now have true AI.

24:41 Were you referring to artificial general intelligence or to what

24:46 Dario Amode calls powerful AI in his essay, machines of loving grace?

24:52 Do you think it is important to maintain distinctions in AI terminology?

24:57 I wouldn't say it's a slippery slope because eventually one leads to the other.

25:02 This is powerful AI compared to what we had before.

25:06 Is it a transition or is it a rapid evolution?

25:09 But between 2020 and 2023, suddenly, the natural language problem is solved.

25:14 This is something we worked on forever.

25:16 LLMs just nailed it and we were all shocked that they did it so early.

25:22 Again, I'm speaking to a group of experts,

25:24 but the Snapta density of the equivalent

25:26 neurons was 10% of the brain and suddenly

25:29 you had to think speak multiple languages

25:30 and know all the stuff would make mistakes.

25:33 But it was incredible what we had achieved when no

25:36 one expected it to wake up like that, so to speak.

25:40 We crossed the threshold, and that's what I was referring to.

25:43 Compared to before, everything was fragile.

25:45 We'd make these models and they had to do some things,

25:48 but they wouldn't deal with anything that was out of context.

25:51 It got to a point where a lot of these things were solved.

25:55 Of course, who knows where it's going to end?

25:57 There are so many scenarios I gave a few.

26:00 We have to see.

26:02 Things will happen fast, I suspect.

26:04 But I'd say, right now, powerful, general AI.

26:07 What will that be?

26:08 What does intelligence really mean?

26:09 These are deep questions that people in the room

26:11 have been dealing with biologically for years.

26:13 Now we have artificial brains, so to speak, we can begin to study and maybe

26:17 answer some of these philosophical questions as well.

26:20 Hey, the next question is for Miles,

26:24 is there any indication or speculation on here's another why,

26:29 why HRs are so small.

26:32 Interesting question.

26:33 I suspect that'd be a better question for Katie Pollard,

26:38 who is the first person to start this field

26:41 and others that have used bioinformatics to define them.

26:44 I think it mostly comes down

26:45 to bioinformatics and that short sequences are easier

26:48 to identify that have undergone these statistically

26:52 unique changes in human, but not elsewhere.

26:56 I don't think that they're necessarily the functional units.

27:01 The functional units could be smaller than HRs,

27:03 and they could also be bigger than HRs.

27:05 I think most likely it's a biostatistics reason for their size.

27:11 Great.

27:11 Thank you.

27:13 The next question is for Dean.

27:16 The questioner asks, I'd love to hear more about

27:20 the hypothesis that motherese came before full language.

27:24 What are the differences between the two that makes

27:27 you think this motherese infant directed speech is common,

27:31 but not a cultural universal in humans?

27:34 Does this pose a problem for the theory?

27:37 To begin with, I've addressed the allegation

27:40 or the assertion that motherese is not

27:43 universal by looking at those specific societies

27:47 that were claimed not to have it.

27:49 It turned out they do have motherese.

27:52 It's just different in different cultures.

27:55 For instance, if you have a culture in which eye contact is taboo,

28:00 you don't have eye contact in the motherese.

28:05 What was the beginning of the question?

28:08 Yes, sorry.

28:10 The hypothesis that motherese came before full language.

28:15 Only humans have motherese to the extent they do,

28:20 lots of animals have contact calls and a little

28:24 bit of interaction between mothers and infants,

28:26 but only humans do it's an extremely special register in which

28:31 parents can be fathers too or other individuals speak to babies.

28:36 It is shown.

28:38 There are lots of articles that shown absolutely bootstraps,

28:43 acquisition of language and young infants.

28:46 They're born without language.

28:48 They're exposed incessantly to motherese.

28:52 By around about the end of the first year, they're starting to get language.

28:57 Part of the reason is because one of the functions of motherese is

29:02 that it highlights certain things infants need

29:04 to know in order to acquire language, like how to parse a sentence.

29:09 People speaking motherese, most are completely unconscious of the fact

29:14 that they are hyper articulating vowels, for instance.

29:18 There's a large literature which shows

29:21 that the more motherese you speak to your infant,

29:25 the better in terms of its acquisition of language.

29:29 Don't worry about, are you speaking down to your infant?

29:33 They like it.

29:35 It's unique to humans.

29:37 In humans, it precedes and bootstraps the acquisition of language.

29:44 It seems logical then to ask, when did it rise?

29:49 Other animals don't have it.

29:51 When in home and and evolution, did it rise and how?

29:55 That's what the putting the baby down hypothesis address.

29:59 I've now corrected it in light of David Lindsey's work,

30:03 I would now call it the picking the baby

30:05 up hypothesis because of some very important caveats he raised.

30:11 Yeah, I think motherese it evolved.

30:16 It evolved into our lineage.

30:19 We know it's function besides bonding,

30:22 expressing love and becoming educational when the infant is

30:26 old enough because the motherese expands as the infant develops,

30:31 motherese changes to accommodate the infant.

30:35 I hope that addressed most of the question.

30:37 Yeah, that was great.

30:39 Thanks.

30:40 Thank you.

30:41 The next questions for Allison.

30:44 Do you have any thoughts on why?

30:47 Again, the modern Nova one version only evolved

30:51 in humans as a response to lead if,

30:54 as you mentioned, lead has been contaminating humans

30:57 and other primates for about two million years.

30:59 Why was this change specifically selected for in humans but not our relatives?

31:04 Yeah, that's a great question and I don't have the answer.

31:08 I can only imagine that this is all about genetics,

31:12 it turns out that the mutation arises in modern humans and then was selected.

31:21 I would say that this could have happened in other species as well.

31:27 It just happened to us and we have the right environment,

31:33 the right brain, the right cell, the right genome for these to be selected.

31:37 Especially on this cluster with FOXG1.

31:41 I think this was an important observation that we have those as well.

31:45 We know that the regulation of those genes might be different in other species.

31:50 I think we were just the ones.

31:52 Can I jump in here?

31:56 Perhaps it's a bad news,

31:59 good news situation where this is obviously highly speculative,

32:03 but what if the human precursors were in a particularly lead rich area?

32:11 Of course, at first that was bad news,

32:14 but the good news is then there was selection for the changes in Nova,

32:18 which then led to good news, which is good things for the brain.

32:22 We became smarter.

32:24 I'm obviously simplifying it.

32:25 But you know what I'm saying that sometimes

32:27 a bad situation can lead to a good situation.

32:31 Yes, we hope all bad situations lead to good situations.

32:36 Great.

32:37 The next question is for Nenad.

32:40 The question is, the effects on neuronal development trajectories

32:44 and knockout mice brains in these studies is intriguing.

32:48 Have any behavioral studies been conducted in SAP B2,

32:53 SIP 26 B1 or mice to knockout mice,

32:56 and what do these animals look like behaviorally?

33:00 Very good question and the answer is, yes,

33:05 the behavioral studies have been done and they did poorly,

33:11 at least in case of mice to in tests

33:14 that require working memory and prefrontal cortex.

33:18 But you have to again,

33:19 I think Mile said that really clearly nicely and eloquently,

33:23 it's very hard to study behavior in mice.

33:27 We are separated but hundreds of millions of years from last common ancestor.

33:32 I think whatever you get there,

33:35 you have to be very careful with what you are trying to interpret

33:39 in the context of human evolution and human cognition and or behavior.

33:43 Just to let you, also one important thing to also understand these are

33:47 extremely important genes and so we shared more than we are different,

33:56 even with other non human primates.

33:59 We have to keep that in mind.

34:01 Also what Miles showed us was very beautiful example

34:04 of her that should and has human specific properties.

34:09 But yet it is affecting a gene that is probably highly conserved.

34:14 That gene is also affected by lineage specific events in other species,

34:20 including let's say in chimpanzee lineage.

34:24 Again, we have to, we are really peeling an onion

34:27 here and we have to, maybe a couple of layers, maybe even one layer.

34:31 We have to be very careful.

34:34 Yes, there are clear behavioral deficits in all these mice.

34:38 The phenotypes are very profound, which also tells us that probably the genes

34:42 that are under some a human or primary specific selections,

34:46 it's good to have a gene that is being changed in its

34:52 expression levels or spatio temporal that are

34:55 important developmental genes that are important.

34:58 That's all great.

34:59 But I have to just say that yes,

35:01 there are behavioral deficits interpretation of them has

35:04 to be done with a lot of caution.

35:06 Yes, totally agree.

35:08 Thank you.

35:09 The next question for Bruce,

35:13 in your patients with degenerative neurological conditions,

35:17 did you observe upregulation in positive emotion

35:21 moods associated with proclivity for artistic endeavors,

35:25 or was there more or less a neutral affect present?

35:30 Yeah, it's a great question,

35:32 I don't think we've answered that in the subset with artistic creativity.

35:39 The person who studies this in most detail

35:42 is Virginia Sturm and she's contrasted frontoltemporal dementia,

35:48 which hits frontal circuits, salience Network.

35:52 Sometimes with upregulation of posterior brain versus Alzheimer's,

35:57 where there's a loss of default mode network,

36:00 posterior brain circuits, and sometimes upregulation of the salience network.

36:06 In the Alzheimer patients that she studied,

36:12 who are amyloid positive but not symptomatic, there are some emotional signals.

36:20 I think also in one study, of a cohort Baltimore Longitudinal aging study,

36:30 the personality changes that seem to precede any memory cognitive issues.

36:36 I think they are associated things like increased neuroticism,

36:42 increased anxiety, and more introversion.

36:50 I think that these diseases affect emotions very differently.

36:56 Then there's left versus right,

36:59 which I think also influences exactly how we see these emotional disorders.

37:04 But I think it's an astoundingly good question.

37:09 What has happened to the emotional system in these artistic,

37:13 and we just haven't even touched that.

37:16 The next questions for me,

37:19 I feel like almost Alex could answer this a little too.

37:22 What is the potential downside in having

37:25 a reduced number of mature oligodendrocytes in humans?

37:28 What's the trade off there?

37:30 I alluded to this a bit at the end.

37:33 I mean, of course, we don't know, but we speculate that this may be linked

37:38 to human risk for demolinating disorders like multiple sclerosis,

37:44 but we don't really know how this change in ratio affects

37:48 differentiation or if there's increased apoptosis of one versus the other.

37:53 This is something that we're trying to study in the dish.

37:57 But right now we're going after really the disease aspect of it.

38:01 Alex, do you want to comment?

38:03 Yeah, I think it's a great question.

38:06 I always love to think about trade offs.

38:07 I think we could extend this from the oligo lineage alone

38:12 to thinking about trade-offs reflecting

38:15 the protracted maturation time in humans.

38:18 We have this potentially unique period of early childhood.

38:23 We have extended adolescence window.

38:27 Our development is slower in many capacities,

38:30 and one of the trade-offs related to protracted development,

38:33 including delayed maturation of oligodendrocytes,

38:37 relates to these increased windows of vulnerability to environmental factors.

38:43 We have more plasticity in our circuits, but that may also in a general

38:47 way confer vulnerability to neuropsychiatric disorders.

38:51 This additional plasticity and wiring could

38:54 lead to aberrant reinforcement of pathological circuits,

39:00 I think those may be some of the most

39:03 dramatic trade-offs related to this slow maturation that we face,

39:07 Nenad has done extensive work on the molecular

39:10 staging of these periods across species.

39:13 Maybe I'll turn it over to you.

39:14 Thank you, Alex.

39:19 There is very little to add to this.

39:22 Except that I completely agree and Alex has done

39:26 a very good job trying to describe all of this.

39:32 There is a reaction to every action and so whichever cell type you change,

39:37 whatever you change in brain development, there's something else,

39:41 you have to pay a price for it.

39:43 This is true for all species.

39:46 It's like if you look at marsupials, they are born prematurely,

39:51 and so they develop a specializations to really deal with that.

39:55 They also have some advantages and many disadvantages to their life history.

40:03 We have to also be careful to what extent changes in human life history,

40:09 not just but including aging and stuff like

40:11 that, are due to our environment and lifestyles.

40:15 We all treat each other differently than we

40:18 are caged animals or living in a wild, as Bruce says, in some cases, badly.

40:24 I think those are all things that we have to take in consideration.

40:28 We're looking at the changes in life history and pros and cons of every change.

40:35 Nothing is free.

40:37 Or without consequence.

40:40 Excellent.

40:41 Thanks, team.

40:43 The next questions for James.

40:46 There are a number of species such as dogs that seem

40:51 to outperform great apes on cooperative

40:54 communication and sometimes theory of mind tasks.

40:57 Based on your research on the default mode network,

41:00 would you expect convergent evolution of a novel PCC to MPFC

41:06 pathway or the emergence of a distinct

41:10 neural mechanism underlying these abilities?

41:13 Yeah, what an interesting question.

41:16 I guess because, humans have been living with dogs for such a long

41:21 period of time and working together

41:24 with them and selecting them in different ways,

41:29 it makes sense that dogs have come

41:33 to understand humans better than maybe other species,

41:37 and that there and that's part of why they're great

41:40 companions is because it seems as though they have some,

41:44 human-like social cognitive abilities.

41:48 What an intriguing question.

41:51 Do we think they have a default mode network or they have,

41:58 specializations related to theory of mind?

42:02 I'm not sure, there are different levels of theory of mind.

42:08 Sometimes when we're talking about theory of mind,

42:11 we're talking about really high level, do you know what I know?

42:17 Can we both share the same goals and ideas.

42:21 Sometimes it's lower level,

42:24 do I understand your intentions and that sort of thing.

42:29 It's hard to know exactly.

42:31 I don't know enough about dog cognition to know

42:34 what level of the theory of mind hierarchy they're at.

42:37 But yeah, that's an intriguing question.

42:41 There are some people who are doing a little bit of dog fMRI.

42:46 One of my colleagues here at Emory, Greg Berns has done some,

42:51 and I think it's a really exciting area for exactly this reason

42:55 that they have so many unique or special social cognitive abilities.

43:01 But yeah, thanks for a great question.

43:05 That's really fascinating dog fMRI studies.

43:08 That's great.

43:09 Yeah.

43:10 The next question is open to everyone.

43:14 What do the speakers think newer

43:17 methods such as spatial transcriptomics might add

43:21 to comparative brain studies beyond what we've

43:23 already learned from bulk and single cell approaches?

43:27 I mean, I think it's the future of genomics.

43:31 I'm a single cell genomics person because we can look for rare cells.

43:35 We can look developmentally where cells are migrating in the tissue.

43:40 We can get a larger area when we dissociate the tissue,

43:43 we lose so much information.

43:44 But I've already drank the kolate,

43:47 so I don't know if anyone else on the panel has other thoughts about this.

43:52 I'll just say I totally agree.

43:54 Yeah.

43:55 Alex?

43:57 Yeah.

43:58 Yeah, so we know so much about the genome sequence differences between species.

44:04 Of course, there's this gap between genotype and phenotype.

44:07 Allison talked about how we can

44:10 take a strong effect potentially candidate mutation

44:13 from signatures of selection and then try

44:16 to link it to phenotypes by reconstructing it.

44:19 We can tell some genomic regions that are under adaptive selection.

44:24 But it's very hard to tell

44:25 with these neural cell type or neighborhood phenotypes,

44:30 what is selection and what is constrained.

44:32 I imagine by really extending our comparative

44:37 phenotyping to the level of cellular organization,

44:40 we can get a sense for what neural

44:43 circuit and cell type motifs are constrained within species,

44:47 but divergent across species, and we might be able to get more of a signature

44:52 consistent with adaptive selection for some

44:54 aspects of nervous system organization.

44:57 The big challenge, of course,

44:58 is the limited amount of tissue that we can work with.

45:04 There are resources out there that are ethically collected,

45:08 but it's difficult to use these.

45:10 It's also a challenge to think about when

45:12 is the right time to use these scarce resources?

45:15 When is the technology ready, given that we might not be able

45:20 to continue to expand these collections of postmortem samples?

45:25 I think there's things to think about.

45:26 I'm certainly enthusiastic for more extensive comparative phenotyping,

45:31 including spatial transcriptomics, as well as some of these and diffusion

45:36 tensor imaging and MRI approaches that we heard about.

45:39 If I may add to this is, I think, excellent question,

45:43 but also it's a future as you said, Gene,

45:46 and Alex has also really pointed out what is important.

45:50 But one thing that we need to do better

45:52 is to integrate across sub disciplines and scales and modalities.

45:59 There is a lot of classical literature that is really important in neuronatomy.

46:05 But we have a single cell where you are just breaking up things.

46:09 We have a lot of literature in imaging.

46:13 I think the spatial technologies,

46:15 whether it's transcriptomics or something else,

46:17 will I think, help us bridge that.

46:20 And I think this is particularly also

46:22 important for development where things are extremely dynamic.

46:27 There are transient cell types.

46:29 There are transient circuits that we know from working experimental animals.

46:35 We know some a little bit about it in human brain development,

46:41 but we actually have very little knowledge.

46:43 I mean, almost none in what happening in closely related species,

46:49 particularly chimpanzee.

46:51 For that reason, I think it's absolutely something

46:54 that I think will be critical for moon dog old.

46:57 The next question is for Allison.

46:59 This is a long one.

47:01 What do you think are the best and worst

47:05 case scenarios as actualization technology continues to develop?

47:10 Do you have any thoughts about a recent

47:12 paper calling for regulation in this space,

47:15 pointing out that wealthy donors could theoretically

47:18 fund private efforts in the extinct human lineage

47:21 in the same way colossal Biosciences has been pursuing

47:25 the Direol woolly mammoth and other so called resurrections?

47:30 That's a great question.

47:33 Yeah.

47:36 There are thousands of differences between humans and the Neanderthals.

47:41 When I say a 61 variant, these are protein-coding genes.

47:47 Even if we change the 61, there's other thousands of regulatory regions

47:52 that would be very hard to manipulate genetically,

47:56 at least with the tools that we

47:59 currently have to reconstruct entirely a Neanderthal genome.

48:03 Someone might even think about reconstructing using hypotypes from humans.

48:09 But even though, as far as I can tell, will never be a complete Neanderthal.

48:14 I'm not too worried about,

48:16 but there is a cool movie called "William" from Tim Disney,

48:23 who discussed this possibility, of course, not through genome editing,

48:27 but supposingly that you can find an intact cell on a Neanderthal that is frozen

48:32 and you revive the nucleus and do a nuclear cell transfer into a human egg,

48:40 and you might raise a Neanderthal.

48:42 But even though it's on the fiction level.

48:45 Yes, thank goodness; it's still fiction.

48:48 Anyway, we should touch on that there

48:50 are legal prohibitions against germline editing.

48:54 I don't know how they would apply

48:54 to somatic cell nuclear transfer from your frozen tissue,

48:57 but there's many legal and ethical restrictions

49:01 on editing the human germline that govern this.

49:05 I think what Allison does in vitro in cell lines,

49:09 testing a small number of these mutations at once for experimental studies,

49:12 is very distinct from this concept of trying

49:17 to create a human organism carrying many Neanderthal variants.

49:24 But I think an important addition is

49:27 that many of these variants still exist today.

49:31 30-50% of the Neanderthal genome is preserved among humans,

49:35 just at low frequency in any individual human.

49:38 For many of the variants, but not the 61 that Allison mentioned,

49:42 we actually have the opportunity to study and learn

49:44 from their function by standard

49:46 genetic association mapping studies among humans.

49:49 The next question actually is for you, Alex.

49:53 In the common garden environments where

49:55 you grow cells from different species together,

49:58 how do you rule out potentially unwanted

50:01 or unnatural signaling that could change cellular behavior?

50:06 This is a great question.

50:08 It's a pretty new approach to mix

50:10 cells from different species in the same environment.

50:13 Also, anytime you mix human and non-human material,

50:16 there's ethical considerations,

50:18 which we've considered extensively before these studies.

50:21 But in terms of the experimental points alluded to by the question.

50:26 In our studies, we both do within

50:29 species pools as well as between species pools.

50:32 One of the first things we analyzed was,

50:34 to what extent is gene expression divergence governed by extrinsic mechanisms?

50:41 The question asks versus cell intrinsic mechanisms.

50:45 Some of the most exciting candidate genes

50:47 from recent studies could point to either mechanism,

50:51 like Debra Silver's lab has this example of HARE, near the gene Frizzled-8,

50:56 that could increase wind signaling in a cell intrinsic way.

50:59 But then Sofie Salama, David Haussler, Pierre Vanderhaeghen,

51:04 and others have found a human-specific gene,

51:06 NOTCH2NL, that could increase notch signaling in a cell extrinsically.

51:11 What does it look like when

51:12 we do the within-species versus between-species organoids?

51:17 What we find in our hands, at least,

51:19 is that nearly all of the cell-type-specific

51:22 gene expression divergence is governed by intrinsic mechanisms,

51:27 and that supports the use of these mixed-species

51:30 organoids for discovery of intrinsic mechanisms.

51:33 That's not to say that there aren't

51:36 extrinsic mechanisms in the right cellular environment,

51:38 and we're very interested in investigating this as well.

51:41 But there's a lot of evolutionary

51:43 theory that modular cis-regulatory changes driving

51:47 gene regulatory networks will guide adaptive

51:49 changes because they're favored under selection,

51:51 and those would excel intrinsically.

51:54 We actually think it's an advantage of the system

51:56 that we can dissect what's intrinsic and extrinsic.

52:00 Then the other part is that there's a practical utility to mixing these cells,

52:05 and that controls for batch effects allows us to scale

52:10 too many individuals to identify

52:11 species differences rather than individual differences.

52:14 Then, for our experiments where we

52:16 want to deliver some external environmental stimulus,

52:20 in our study, we delivered an oxidative stress stimulation.

52:23 We can really control for these being in a similar

52:26 cellular stress oxidative environment going into the experiment.

52:30 That's why we made these experimental choices.

52:34 I would love to discover what the small number of extrinsic

52:38 changes are because that could point to potential driver changes,

52:42 like this NOTCH2NL example.

52:44 But we can also use this system to discover extrinsic changes,

52:48 and that's something that we're interested in working on moving forward.

52:51 Thank you.

52:53 The next question's for James.

52:56 The gradation of size of the lateral temporal cortex was intriguing,

53:01 as well as the finding of pronounced difference

53:04 between this area and humans and its chimpanzee homologue.

53:08 I wonder about cell types there and if

53:11 there are mirror neurons or connections from mirror neurons,

53:15 for example, from the IPL present in the area.

53:21 The pathway that we were focusing on the arcuate fasciculus,

53:26 is really connecting the temporal lobe,

53:31 so the posterior superior temporal gyrus,

53:35 the superior temporal sulcus, and the middle temporal gyrus,

53:39 as well as the inferior temporal gyrus,

53:41 so lateral temporal cortex, with the frontal lobe, with the prefrontal cortex,

53:46 and especially the ventral prefrontal areas like Broca's area.

53:51 That pathway doesn't include the parietal lobe,

53:56 and so where there are mirror neurons.

53:59 But of course, we know there are mirror neurons

54:02 in Broca's area which is part of the pathway.

54:07 I think it's interesting to think about

54:12 differences in those mirror neurons across species.

54:17 I know there's been some work done on that.

54:20 Of course, macaque monkeys have mirror neurons in the homologue of Broca's area,

54:27 and I know human mirror neurons have

54:30 some properties that you don't find in macaques,

54:35 for example, they respond to intransitive grasping movements,

54:40 whereas macaque mirror neurons don't.

54:48 We're working at more of a systems macro level,

54:53 but it's important to think about underlying cellular differences as well.

55:00 The next one's for Bruce.

55:02 Was there any indication of correlation

55:06 with experiences of mystical-type encounters,

55:11 like notions of the divine or supernatural that accompanied the conditions?

55:16 Yeah.

55:16 The one subgroup of patients that we see this in is a group,

55:23 Teresa Wu wrote the first paper on this, our patients

55:27 who generate the right anterior temporal lobe.

55:31 We see increased religiosity in this group.

55:36 We tend to see also hypergraphia in this group.

55:42 But the religious intensification,

55:47 philosophical intensification is a feature of this subgroup

55:52 of patients with frontal temporal dementia.

55:56 There's something about the balance between the right ventral temporal

56:00 lobe and the left that I think releases this phenomenon.

56:06 Interesting.

56:07 Thank you.

56:09 The next question is for Nenad.

56:11 When we talk about information flowing through cortical circuits,

56:16 signals going up from sensory areas versus

56:19 predictions coming back down from association cortex,

56:23 how precisely can we actually track or decode that right now?

56:30 Thank you.

56:31 It's a wonderful question that I've been recently reading more extensively,

56:37 but I cannot answer that question.

56:39 I just don't have the knowledge and expertise to answer that question.

56:42 Maybe somebody else on the panel.

56:45 Those are very technical details,

56:48 just I don't have that knowledge to understand in the imaging.

56:52 I apologize.

56:52 It's a fabulous question.

56:54 I just don't have the expertise.

56:56 Hard question to answer.

56:59 There's one more for me.

57:03 Do I believe that the between-species difference in immature

57:07 oligodendrocyte markers persist throughout the lifespan of humans,

57:10 or expected to follow a degenerative course across all three species you tested?

57:15 That's a great question.

57:17 That's something we're trying to answer by sampling across

57:21 the lifespan with a different project from birth through adult.

57:27 For degeneration, that's a separate question on top of that.

57:32 One is developmental, and then one is degeneration.

57:37 We were middle-aged when we did that study,

57:39 and we need to go before, we need to go after.

57:41 But yeah, I really like that question.

57:43 I'm going to do a wrap-up, I think,

57:46 unless anyone has any follow-up answers to any

57:50 of the questions that were asked already.

57:56 To our speakers and everyone still in the audience,

57:59 this is the end of the question-and-answer period.

58:03 I'm just going to take a few minutes now to do my best to summarize the day.

58:06 There's a lot of information.

58:08 But first, I really want to thank all of the speakers.

58:11 These were amazing talks.

58:13 I'd also like to thank all of the symposium participants who I can't

58:17 see for listening and asking all of these really excellent questions that have,

58:23 I think, really got all of us thinking

58:26 further about our own research in the future.

58:29 Today, we've explored many different aspects of the human brain,

58:33 what we call the idea organ,

58:35 and the role of evolution in its function at many different levels, really.

58:40 We've discussed how the brain has biologically evolved,

58:44 how it's culturally embedded.

58:47 What I think is really cool,

58:49 how technology has allowed us to extend that beyond our own intrinsic biology.

58:56 At the biological level, we've heard many different things.

59:00 We've heard about developmental mechanisms that impact

59:03 circuits and how they shape neural architecture.

59:08 We can measure that neural architecture in a number

59:11 of ways through brain imaging and through model systems.

59:15 Then how ultimately does this neural architecture underlie human behaviors,

59:20 like tool use or picking up or dropping babies

59:24 to the emergence of language and theory of mind?

59:29 Then how did this tool use language and social cognition enable culture?

59:35 Beyond behaviors, we heard about cell types and alterations in cell types,

59:41 whether it's a specific cell type,

59:42 such as the von Economo neuron, or proportions of cell types,

59:46 like oligodendrocytes, how do genes influence these features?

59:51 For example, how do HAREs influence cell differentiation or proportions,

59:56 or the use of genes from understanding ancient genomes,

1:00:00 such as those from Neanderthals, to identify modern human-specific alterations?

1:00:05 Then taking those genetic alterations

1:00:08 and circling back to behaviors like language.

1:00:11 We saw the case of NOVA1 intersecting with FOXP2,

1:00:15 a so-called speech and language gene,

1:00:17 or how our understanding of cortical development that's important

1:00:21 for brain evolution in the case of mice too.

1:00:25 Beyond then, cell types and genes,

1:00:28 we also talked a lot about modeling these features to dissect

1:00:32 them and to understand them

1:00:34 and really get some experimentally tractable information.

1:00:39 We can do this by doing evolution in a dish or in a model system.

1:00:43 We heard a lot about cells organoids, and model systems like rodents.

1:00:49 What was refreshing to me and something that I

1:00:52 don't often get to hear about in a similar symposia,

1:00:55 the intersection of the human brain with art, both visual and musical.

1:01:00 Music came up in the talks in several different ways.

1:01:05 We also talked a little bit about pathology,

1:01:07 what happens when the brain goes bad.

1:01:09 But then, refreshingly, how that can reveal human-specific functions,

1:01:14 like in the case of FTD, revealing new creative abilities.

1:01:19 Then finally, how do computers intersect with our biological

1:01:23 systems to allow our brains to reach full potential?

1:01:26 Again, hearing music and computers in the brain intersect,

1:01:30 and the future of augmentation, whether that's to play music or to even

1:01:35 change one's surroundings, is very remarkable.

1:01:38 I'm just going to seal one of the quotes from the last talk from Joe's talk,

1:01:41 where he said, "What will the human become as everything

1:01:45 merges?" I think the future directions are really exciting.

1:01:49 We have some limitations.

1:01:50 We touched upon them a little bit in the Q&A with respect to ethics and what

1:01:56 we can actually manipulate either in the model

1:01:59 systems or even in the human brain.

1:02:01 We can't really change our genetics, our circuits,

1:02:04 or cells ethically, or even with the tools at hand.

1:02:08 But we can clearly model some of these features,

1:02:11 and now we've seen that we can augment them in living humans,

1:02:15 and then we can study behaviors,

1:02:16 and then how those new behaviors with augmentation may impact culture.

1:02:20 We're affecting our own evolution in a sense.

1:02:23 There's so many more discoveries to make about the idea organ.

1:02:28 This has been a remarkable day.

1:02:31 Again, thanks to all the speakers,

1:02:33 and I'm going to pass the baton to Pascal for final closing remarks.

1:02:38 Thank you so much, Genevieve and Allison,

1:02:40 our two co-chairs for this amazing symposium, and all the speakers.

1:02:45 I will thank all of you for attending online.

1:02:48 This was virtual.

1:02:50 If you want to catch some of the talks you

1:02:52 missed or any of the hundreds of past CARTA talk,

1:02:55 you can do so on our website or on YouTube.

1:02:58 We also want to let you know that we will have an in-person symposium in fall,

1:03:03 that will be about We Are What We Ate, The Diets That Fueled Human Evolution.

1:03:09 There will be food involved.

1:03:11 We look forward to seeing you then and, of course,

1:03:14 count on your continued support to make these free symposia possible.

1:03:18 Thank you all very much.

1:03:20 [MUSIC].