Stephen Wilson 0:06 Welcome to Episode 10 of the Language Neuroscience Podcast. I'm Stephen Wilson. Today my guest is Pascale Tremblay, who is Professor of Rehabilitation Sciences at Universite Laval in Quebec City, Canada. Pascale has done outstanding work on the neuroscience of speech and language, especially speech production. And one of her major lines of work looks at the effects of aging on these systems. She has a number of insightful theoretical papers on the architecture of the language network, in which she emphasizes and advocates for the importance of white matter tracks. Papers with titles such as "Broca and Wernicke are dead, or moving past the classic model of language neurobiology". This work is going to be the focus of our discussion today. Pascal has also performed an incredible service for our field by co-founding the Society for the Neurobiology of Language. We'll talk about that, too. Okay, let's get to it. Hi, Pascal. How are you? Pascale Tremblay 0:55 Hi, Stephen. I'm good, and you? Stephen Wilson 0:57 Yeah, I'm good. And so you're in Quebec City? Pascale Tremblay 1:01 Yes. Stephen Wilson 1:02 And how are things in Quebec City today? Pascale Tremblay 1:05 It's getting much better. There's no snow anymore. And, you know, we're starting to get over the third wave. Stephen Wilson 1:14 So you're like, having the end of snow? I mean, like here in Nashville, it's like, sort of almost the beginning of summer. Snow is a distant memory. Pascale Tremblay 1:22 I'm just kidding. Because, you know, Americans tend to think that it's always cold in Canada. It's actually going to be around 30 today. 30 Celsius. Stephen Wilson 1:33 Oh, okay. That's like actually quite warm. Yeah, I mean, I really enjoyed visiting Quebec City for the Neurobiology of Language conference a couple of years ago. I don't know that I would have had a chance to go there otherwise, but such a cool city. Pascale Tremblay 1:47 It's very pretty, yeah, it's a very old. Old in our American standards. Stephen Wilson 1:52 Right. So you know, on the podcast, I'd like to start by kind of getting to know how people got interested in the field of brain and language. So did you have any interests as a kid in either brain or language that would kind of like point to your future career? Pascale Tremblay 2:09 Yeah, I was always interested in language. From as far as I remember, I wanted to study language, but I didn't know exactly what aspect of language I was going to be most interested in. So I was writing a lot as a kid. I thought I was going to be a writer. You know, my parents had these comic strips and even novels that I wrote. And then I went on to an undergraduate degree in linguistics, as I think many language researchers have. And that's where I discovered that what I was really interested was mostly in the mechanistics aspect and phonetics and speech, more so than we typically refer to as language. And then I sort of hesitated, I thought maybe I was interested in speech pathology. But in the end, I decided to go on to study the basics of speech at the PhD level, at McGill University with Vince Gracco. And from there, I continued studying the neural basis of language, later on at University of Chicago with Steve Small and in Italy with Uri Hasson. Stephen Wilson 3:27 Right, okay. Yeah. I was wondering who you had worked with over there. I saw that on your CV that you'd been to Trento? Pascale Tremblay 3:33 Yeah. Such a wonderful place. Stephen Wilson 3:37 And then and then you've ended up back at the university that you did your training at, right? Pascale Tremblay 3:41 Yes. And that was part of my sales pitch when I interviewed. It worked so. Stephen Wilson 3:47 What's it like being a professor at your undergrad institution? Pascale Tremblay 3:51 Well, it's not in the same department at all. I only did my undergrad here in linguistics and now I'm a professor in the School of Rehabilitation Sciences. So it's in the Faculty of Medicine, so it doesn't really feel like home. It had been so many years. Like, I think it's great, but I felt like I had spent so many years away. I spent six years at McGill and three other years away. So it had been a long time. Stephen Wilson 4:21 Right. So it didn't really feel like a homecoming like academically. But how about like family wise, are your family from Quebec City? Pascale Tremblay 4:28 Actually, my family is from up north. In the region that's called Saguenay where there's a fjord in the province of Quebec. It's a very beautiful region. Stephen Wilson 4:39 I bet. I'd love to go up there. I like the idea of being where there is no other people. Pascale Tremblay 4:46 And at the junction of the fjord and the St. Lawrence River and the ocean and there's whales, it's beautiful. Stephen Wilson 4:57 Wow, how far is that from where you are now? Pascale Tremblay 5:01 Well, not from the junction, but from here to the region where my parents are, it's about two hours drive. So it's close. Stephen Wilson 5:10 Okay, maybe that's like an optimal distance to be away from your parents. Pascale Tremblay 5:16 Actually I"m very close to my family. So, to me, that's a little far. Stephen Wilson 5:20 It's a little far. Yeah. So I haven't seen my family for going on two years. And like, you know, the Australian border is closed and doesn't appear that it's gonna get open anytime soon. Pascale Tremblay 5:30 So where is your family? Stephen Wilson 5:32 I mean, they're in different parts of Australia. Pascale Tremblay 5:35 Wow, yeah, yeah. That's far. Stephen Wilson 5:39 Yeah, two hours would be nice. Two hours seems really good right now. Pascale Tremblay 5:42 Yeah, two hours is perfect. Stephen Wilson 5:44 Yeah. And so obviously, you're a native French speaker, bilingual. Did you grow up fully bilingual in both languages? Like many people in Canada? Pascale Tremblay 5:55 No, I mean, I think that's a misconception. I think in Quebec, most people grow up French speaking. But they learn English at school. And in the rest of Canada, you know, they grew up English speaking and they never learn French. Although there are French communities in the rest of Canada too, like in Ontario, in Manitoba in eastern Canada. . Stephen Wilson 6:16 Okay. So when did you learn English? Pascale Tremblay 6:19 At school. But I think it really did learn it at McGill when I did my PhD. I did learn it as a kid. And you know, American culture is very much present throughout Canada. But using it on a daily basis, that was really at the PhD level. Stephen Wilson 6:35 Right, interesting. Were you like a natural language learner? Do you think? When you encountered it like, in school? Did English come easily to you? I guess is question. Pascale Tremblay 6:45 If it came easy? It's a hard question to answer. Relatively easy, because as I said, there's the influence of American culture, but also Canadian English culture onto Quebec. So for sure, we grew up and we sing English songs, and we sometimes watch English TV. And so I think, yeah, probably came relatively easy. Stephen Wilson 7:11 Cool. Okay, so one of the many roles that you've performed in your career is being one of the founders of the Society for Neurobiology of Language with Steve Small. Can you tell me and our listeners about how that came to be? I think that that's something that many of us value greatly for what you have done for the field. And I'd love to hear more about how that came to be. Pascale Tremblay 7:35 And I remember distinctly that, at a labn meeting in Steve Small's lab, we were sort of complaining that there was no venue that really fit our needs. A place where, you know, linguist, and neuroscientists, psychologists, everybody interested in language could meet up and talk about mostly -- I say language in a very broad sense, obviously, including speech -- and then we said, why not do it? And it's basically, I think from that conversation we just followed through with it. I was always involved throughout my academic career in organizing things. At all the levels I've been, I've organized something. So I was very much happy or willing to do a lot of the heavy lifting. And of course, Steve knew everybody. So you combine those two types of people, and you've got a great committee with great connections. And, you know, people to do the manpower. And also, Steve had a relatively large lab at the time. That was very helpful to be able to do a lot of the work because in the first years, we we did most of the work, because the society wasn't established. We didn't have a company running the show, it was really us that did everything. Stephen Wilson 9:07 Right. It started with a meeting and I remember it, it was like a satellite to Society for Neuroscience in Chicago. That was the first thing... Did you already think that there was going to be a society? Or did it kind of just start with that meeting? Pascale Tremblay 9:21 We just wanted to start with that meeting and see how it would go. I don't know if you remember this, but we we ran out of space in that first meeting. We thought that there was going to be 50 people that would register but there ended up being over 300. Stephen Wilson 9:37 I remember a thing about 300. I didn't realize that there was a space problem, though. Pascale Tremblay 9:41 Well, there wasn't. We had an overflow room. I don't know if you remember that. And a lot of people complained. But we were like, look we really didn't expect that and we're so happy. That's a very good problem. And then we we made it a satellite of SfN. It was sort of a philosophical stand that we wanted to tie it closely to neuroscience, not just to sort of to be a little distinct from other societies that have a more of a cognitive underpinning. So ours was going to be more of a biological underpinning. And then the year after we did it again, since there was such a good reception. And then people asked to another one, and we were willing to do another one. The second one was in California, in San Diego. Stephen Wilson 10:35 Yeah, I think I went to that one, too. Pascale Tremblay 10:39 It was really fun. It was bigger. And there's even more people. And that's when we started thinking about a society. Because we thought that, you know, the reaction was so good. It was because people, just like us, they felt like they were missing a place for exchange. Stephen Wilson 10:59 So what are the functions did you want the society to perform besides arranging the conference? Pascale Tremblay 11:04 Well, to be a place where discussion can occur more broad sense are not just at the meeting, but outside the meeting. So now the society has gone into, you know, editing this open access journal. So I think that's a great move. That's something we had in mind at the beginning, just didn't have the means, of course to do this. So that's great. Stephen Wilson 11:34 Yep. So you're talking about Neurobiology of Language, the journal? Pascale Tremblay 11:38 Yes, exactly. And now, you know, the society has moved in recent years to having these career awards for young researcher, older researcher. And I think that also helps shape a community. So, there's a little scholarship for traveling that the society has always offered. And so that's also something that is a role that the society can play. Stephen Wilson 12:09 Well, I appreciate what you've done. And I think it's made a big difference to the field. Pascale Tremblay 12:14 I think it has. I'm so happy that, you know, I'm not involved at all anymore. But I'm just so happy that it continues on. I mean, I did organize the the 10th anniversary in Quebec City. Stephen Wilson 12:28 Oh right. Yeah, of course, you would have been involved in that. Pascale Tremblay 12:31 There is so much fun. But yeah, I mean, it continues, and it's nice to see that it grows, and it changes and diversifies. And everybody we know is getting involved somehow and taking turns and you know, there's never a shortage of interested people, it seems. Stephen Wilson 12:50 Right. And do you think we're going to be having an in-person meeting in 2022? Pascale Tremblay 12:58 That's a toughy. Isn't that the one in Australia? Stephen Wilson 13:00 No. What did they do? I think that they push that back further? Pascale Tremblay 13:05 I think Pittsburgh, isn't it? Some place in the US is the next one. Stephen Wilson 13:11 Yeah, there was going to be Philadelphia this year. But then they made it virtual. Yeah, I don't remember what's coming up. But I hope we get to be in person again. Pascale Tremblay 13:21 Yeah, it would be nice. I mean, it is all great. Because I mean, it facilitates access to information for sure when it's online, but we're missing the human factor, I think. Stephen Wilson 13:34 Yeah, there's just this intangible aspect of it that can't really be replaced. Pascale Tremblay 13:39 No, not at all. And plus, I mean, I don't know if you've experienced the same thing. But it feels like there's so much now that everybody double books, and then you think that you have this day dedicated to this scientific day. But in the end, you end up working on your computer, either at the same time, or you miss the conference entirely. Stephen Wilson 14:00 Exactly. Yeah. I mean, I think it is really hard to carve out the space unless you're physically removed from your responsibilities. Pascale Tremblay 14:08 Exactly. So, I mean, we think we have so much work. We need to incentive to just turn off our computer and listen. Stephen Wilson 14:19 Yeah. Yeah, we deserve to like, you know, go and spend five days in some really interesting city and not have to take care of anything or anyone. I think we do. Pascale Tremblay 14:32 Yeah, I think everybody was looking forward to going to Australia. Stephen Wilson 14:37 Yeah, it'll happen. We'll get there. We'll get there. Pascale Tremblay 14:40 They've been waiting for a long time for this meeting so I'm sure they won't let go. Stephen Wilson 14:45 Yeah. So let's talk about your paper, entitled, "Broca and Wernicke are dead, or moving past the classic model of language neurobiology". And with your co-author, Anthony Dick. Published in Brain Language, 2016. It's quite a striking title. It's become a highly cited and well-known paper. Can you tell me how you came up with this title? Pascale Tremblay 15:13 Trying to remember how we came up with the title. I know that Anthony had this idea to link it to the title of this play, that we quote recurrently in the in the paper. But I don't remember if it came afterwards, or if it initiated the title. But we were basically thinking it. And so we wanted to have a very clear title that, you know, was going to position the paper. Stephen Wilson 15:49 Yeah. I mean, it makes your point very clear. It's like, move over old dudes. So what are your big-picture goals in this paper? Pascale Tremblay 16:03 So it was to basically question the use of certain concepts and certain models, that still is in use, actually, today. And to try to have the field, think about it. I mean, of course, a lot of people have actually said very similar things to what we have said in this paper, right, but it didn't stick. And I'm not sure that our papers will do better in the science. I mean, people are citing it. But I mean, I keep on reading other articles that still use the same terminology and concepts. So I don't know that it's had the impact that we wanted it to have. But basically just reflect upon a few notions that I think that our basic scientific notion of precision in defining concepts and agreeing on concepts in order to reach precision, for example. Stephen Wilson 17:02 Okay, so the concepts in question relate to what you guys call the classic model, what, you know, other people call the Wernicke-Lichtheim model, or the Broca-Wernicke-Lichtheim model or various other names? Sometimes Geschwind gets a call out, too. So you know, what do you sort of see as being the central tenants of this classic model? Pascale Tremblay 17:24 Well, the central tenants of this models are that language is a basically specialized system, very simple specialized system composed of a superior temporal region and an inferior frontal region. So basically known as Wernicke's area and Broca's area connected by a single pathway, the arcuate fasciculus. And that this is the core of the language architecture. Stephen Wilson 17:54 I guess like, I feel like Wernicke had a somewhat more nuanced and fleshed out vision than that. No? Pascale Tremblay 18:04 Oh, yeah. I mean, people have been arguing over this model, even from the very first time that it was called a model. Right? That is true. But some of the ideas of Wernkicke never made it through and Lichtheim also argued against certain notions that we're presented by Wernicke and obviously people have been going back and forth. I think the model that's most often cited is Geschwind's version of it, which is actually simplified. Stephen Wilson 18:33 Yeah, I guess I don't know if I entirely agree. Because, Geschwind has his famous figure, right, where it's like Broca's and Wernicke's, and they're linked by the arcuate fasciculus. But that's not all that there is to his model, right. I mean, he has, like, you know, the angular gyrus playing this kind of role as a semantic hub. And he's got the kind of following on from Wernicke and Lichtheim, he's got this whole sort of superstructure beyond the language system whereby the language system has to kind of connect with conceptual representations and the rest of the brain. So it's not like, that's all there is to it. Right? Pascale Tremblay 19:07 No, that's true. I mean, now when he said the core, I think I thought you said, the core tenants. Stephen Wilson 19:13 Oh, okay. Pascale Tremblay 19:14 Yes, of course. I don't think that. I don't think anyone believes that this, or ever did believe that that was all there was to language, right? Neither Geschwind, certainly not Wernicke, and nobody else between the two of them. I think that one of the notions that, for example, in Geschwind's model, yes, of course, he talks about other regions. One thing that we wanted to clarify was also the terminology. I think that even just the terminology, especially for Wernicke's area. Stephen Wilson 19:52 Yeah, I find that piece of it really fascinating, right. So, you guys did this survey in 2015. I remember being a participant. I can say that without being a human subjects violation. And yeah, so you you've surveyed people in the field of language neuroscience about, you know what they thought about the classic model as well as how they define those terms, Broca's area and Wernicke's area. I mean, one thing that I found really, very striking, is that you say in the paper, only 2% of the respondents endorsed the idea that the classic model is the best available theory of language neurobiology. That's kind of striking, isn't it? Given that, as you say, it is still such a cornerstone of many people's thinking? Pascale Tremblay 20:41 Well, I think we weren't surprised by this. Because I think that we know that most people, if not everyone in the field, knows that even the most complex representation of this model still doesn't scratch the surface of the complexity of, let's say, the language in a broad sense architecture. So, of course, that's what they said. It's just it's a model that has had a huge impact for developing the field. Of course, historically, and even in this paper, we actually we never said that it had no use. But I think what the paper shows, and it's something that we really wanted to get at, was that the use of this terminology, since nobody agrees on it is not helpful. It's confounding. Stephen Wilson 21:31 You want to kind of go through like what, you know, you ask people to basically pick, you know, you should pictures of what Broca's area and Wernicke's area could look like, six or seven different versions of each and you ask people, which one is the the correct... Pascale Tremblay 21:45 Which one do you think is the correct version? Stephen Wilson 21:46 Yeah. And so what did you find there? And were you surprised? or What did it kind of confirm what you're expecting, Pascale Tremblay 21:52 I wasn't surprised because Wernicke's area has never been properly defined. So for Wernicke's area that really did match what was thinking, although maybe it was lower even than we thought. So basically, the definition that was the most popular among the respondents was one that we basically made up, and it was only 26% of the respondents. And the other definitions were distributed, you know, one of the other ways. The Geschwind definition at 23%, and then the Penfield one at 12%, and the Dejerine one at 8%, and so on. There were seven different versions. And, and so this really does show that we are not in agreement with this. Nobody is. Stephen Wilson 22:38 So it's a very imprecise term, I guess it's like your point. And just to kind of spell things out a bit in case, you know, obviously, our listeners are going to be like doing the dishes or whatever, and they're not going to have a figure in front of them. But we're talking about for instance, you got Wernicke's original model, where it's like the posterior superior temporal gyrus. You've got other versions, where it's like closer to primary auditory cortex, other versions where it starts to get into angular gyrus and middle temporal gyrus. Other versions, where it's more anterior. This is the kind of variability that we're talking about here. So you just think it's just basically a kind of problematic term because it doesn't have a shared definition? Pascale Tremblay 23:20 Well, yes. And I think a lot of people have said that and propose different solutions. So for example, Jeff Binder, had a paper where he actually talked about the exact same problem. He instead proposed a new definition. We went around it in the other direction, that is, if we can't agree on it, why not stop using it instead? Obviously, I think none of those solutions have worked out, as far as I can tell from the literature, people are still using those terms without defining them properly. Or, sometimes they do define them properly. It's like, okay, I'm going to use this label. And this is what I'm going to refer to. This is at least helpful. But then one may might wonder why then if you have to define it, why do you not use the other labels? So for example, you read a paper and they would say, Okay, I'm going to use Wernicke's area. I'm going to refer to it as the posterior superior temporal gyrus. Well, then why not use posterior superior temporal gyrus? Why use this ill-defined functional term. Stephen Wilson 24:27 Well, maybe it's useful sometimes to have a functional term instead of an anatomical term. If you're, especially if it's kind of still up for debate, like, what the relationship is between the two. I mean, so I guess like that would be another possibility. Like I would say, one of the big challenges with Wernicke's area in particular is that even though Wernicke was an absolute genius, he was a little bit wrong about exactly where. I mean, to the extent that you buy into the model, which I do, you know, up to a point. Pascale Tremblay 24:56 The model isn't wrong. We're saying it's more like incomplete. Stephen Wilson 25:02 The part of it that I think is true, is that there's a region in somewhere in the back of the brain where there's some kind of representation of phonological forms. He doesn't use those words exactly. And I just think the problem is it's not in the STG it's in the STS. It's in the superior temporal sulcus, not in the gyrus. That's my view. I mean, it's not the view of everybody. But you know, that immediately for me creates a problem, right? Because if I use the word Wernicke's area, which he clearly described as the STG, that's a problem, because that's not that the STG to me doesn't have the function that he ascribed to it. So generally speaking, I don't use the word Wernicke's area. I actually do accept your argument. It's too problematic. I very rarely use it. But I would, if I did use that, I would probably just redefine that as the STS. And I would say Wernicke's area, which is actually in the STS not quite where Wernicke's thought it was. Because I think the function of it is more important than the location. Would that have been an alternative for you, or you just don't want to go there? Pascale Tremblay 26:05 I go back and forth with those functional labels. I don't know, I think sometimes they're very helpful to try to sort of integrate physiology into anatomy. But Wernicke is not a functional name even, right? It's just a person's name. Like what you're saying, for example, if you talked about phonological processing, for example, what that means, to me, that's actually at least a better label. It's more specific than if you defined it as being the posterior STS, for example, at least you would be clear. So I would think that that's a perfectly acceptable way to doing it. But but then mapping Wernicke onto all of this, I don't know how it would have been effect because Wernicke never even talked about phonological processing, for example. Stephen Wilson 26:50 No. I mean, he called them the sound images of words or something in German. I don't know, the original German, but it's translated into English as the sound images of words. And I kind of like, ascribe him to mean phonology by that. He wasn't, you know, a linguist or anything. I think he thought that these representations were somewhat abstract. And so I kind of gave him credit for thinking that they were phonological representations, not just auditory ones. But maybe I'm giving him too much credit. Pascale Tremblay 27:20 He was clearly a visionary. And it's not at all his fault that I mean, of course, I think he did way better than expected at this time. Right. I mean, he created the field, in a sense, right? He did create the field of neurobiology of language. This is a humongous accomplishment. And I think we should never forget about this. I mean, we're not criticizing his work. We're just saying that, perhaps, right now, in 2021, when we have such better tools, and it's been a long time, right? My take on it is that it's been detrimental to some level to have this public image or idea that language is such a simple system. It doesn't help promote the field, it doesn't help understand the complexity of the disease that people have, for most of which we don't have effective treatments. Or even we don't understand the exact sources have their, you know, their language or speech issues. Right. And so you have this so simplistic model, and then people are asking, so why you're still studying this? Well, that's because the model is probably correct in large part. It's just incomplete. So incomplete, right. So it's also in the sense that we're sort of advocating. You read any neuroscience textbook and there's no other system presented as simplistically as language. Stephen Wilson 28:46 Right? Yeah, I do agree with that. And certainly in the in sort of the neurology world, you definitely get that minimal, very classic and very minimal kind of depiction. Pascale Tremblay 28:58 Exactly. So I think the people that work in the field that know that, you know, it's a lot more complex. But I don't know that it's crossed the boundary. It's still in psychology textbooks, and I mean, psychology being taught by people not working on language. Stephen Wilson 29:17 All right. So yeah, so you're making that argument you're not downplaying the role of the classic model or its sort of formative accomplishment, but you think we just basically need to ditch the terminology of it entirely. Because it's just kind of too limiting, to be having these broad terms that sort of have functions associated with them. But they are very loosely associated with anatomy in ways that apparently nobody agrees on. Pascale Tremblay 29:47 If we were to come up with an alternative to put in the basic textbook, at least we should make it more representative of what we all know is happening. This a very complex system fully integrated with multiple other systems in the brain. It helps nobody to present languages as having such a simplistic architecture. Stephen Wilson 30:10 Yeah. Cool. Well, I liked the paper a lot. We read it in our lab meeting a couple of years ago and had a good time. Yeah. And I reread it just now to talk to you. So it was enjoyable again. Okay, so in place of the the classic model, to some extent in this paper, but also in some of the other papers that you've written, lay out a more complex model of language in the brain, where one of the major features is white matter tracts. And you really place a lot of importance on these tracts. Which go far beyond the arcuate fasciculus. And they have different anatomy and roles, functional roles. So I was hoping we could kind of talk through your view of the major tracts in the brain that contribute to language. And I was gonna mostly go back to your 2012 Brain paper, again, with Anthony Dick. And just kind of talk through the tracts and where they are and what you think they do. And again, you know, this is going to be challenging, right? Because this is a very visual thing. And we're going to be we're auditory only here. So we're going to try and talk about these tracts in a way that people can like, think about them while they're listening. So let's talk first about the superior longitudinal fasciculus and arcuate fasciculus. What is the relationship between those two concepts? Pascale Tremblay 31:41 I think it still isn't entirely clear for for most people. Because the architecture of the arcuate fasciculus, even though that's the one track, you know, connecting originally Wernicke to Broca's area. That's the classic definition of the arcuate fasciculus. I think that the anatomy of this tract actually is still very much debated, and that's sort of what we were trying to get a little bit at this paper on. So there's been a lot of different suggestions. So, one of them you know, that was based on the macaque mostly was that the superior longitudinal fasciculus which you know, is proposed to have four different components connecting different parts. Going in the same sort of general area as the arcuate fasciculus, but connecting superior parietal regions to frontal region, different parts of frontal region. Also also inferior parietal lobule to frontal regions, inferior frontal region, but sometimes more dorsal to like, not just inferior frontal gyrus. And so for example, in the view of these researcher working on with macaques and doing tracer injection, like Michael Petrides and Pandya and Schmahmann. There's these four different components with three of them SLF superior longitudinal fasciculus, plus the arcuate fasciculus. What they would say, though, is that there is no direct connection between the posterior STG and the IFG, which is the very definition of the arcuate fasciculus as it was classically defined. Stephen Wilson 33:28 Yeah, so that work, as you mentioned, is based on tracer experiments and macaques, which can't be done in humans, because it involves the tracer being injected into a live animal that's then sacrificed. So the SLF, basically, it has these three components, they're basically all fronto-parietal. It goes SLF one, two, and three, from dorsal to ventral. And so SLF three starts to look very language-y because it's kind of like, going from inferior frontal regions to inferior parietal regions, both of which we certainly think are involved in language. And then in their view, the arcuate like you said, doesn't go actually to the IFG. Where does the arcuate go in their view? Pascale Tremblay 34:10 To BA 6, 8, 9, 46. So a little more dorsal, than the IFG as it's defined in the macque. Of course, one has to wonder because the anatomy of the IFG is different in the macaque than in the human. It's a region that's extended in humans. And I mean, we can argue probably all day about this, but, you know, macaques, they communicate for sure, but they don't speak in the same way as we do. So it's very possible that the architecture of the macaque tractography would be different from human one. Stephen Wilson 34:46 Yeah, I think that's kind of like the elephant in the room. Right. I mean, like, we're talking about the function, which maybe most distinguishes us from macaques. Pascale Tremblay 34:55 Exactly. Yeah. So I mean, I find this research fascinating and I wish there was a way to do something this precise with humans. That would not, of course involve sacrifice. Stephen Wilson 35:10 I thought you Canadians were supposed to be nice. And now you're advocating for like... Pascale Tremblay 35:16 Obviously something that we could achieve this kind of precision, obviously, without having to sacrifice. Something entirely different, obviously, but that doesn't exist. Stephen Wilson 35:28 Why doesn't DTI or DSI, diffusion tensor imaging or diffusion spectrum imaging, why doesn't that gives us the same precision as these autoradiography studies? Pascale Tremblay 35:42 I think that in humans, I'll just get back to it. But I think in humans, you can do dissections and where you get a very good precision too right. You get specimen, cadaver brains and a trained anatomist can actually dissect the fibers and identify their origin and termination very precisely. That has a limitation that usually very few specimens are available. So the studies are conducted on obviously people that have died. And also that are an unlimited number. Though, with the macaque work. It's also on very limited number of macaques. One has to point out I think that's the second elephant in the room. The first is, it's a macaque. And the second is there's few of the macaques. So in humans, I think the best alternative we have is what you said is the diffusion MRI technique. The diffusion MRI technique, I think, is a very powerful and very interesting method. But it involves a lot of data processing, and, you know, statistical analysis in a way that those analyses don't. So I think that the choice of the metrics that the researcher will use, and the pipeline of analysis that they will choose can actually have a huge impact on the results. I think that's why we are having this problem. You know, in preparing for this interview, I thought, Oh, yeah, let me, you know, see if we have solved some of the question that we raise in those papers. And in fact, if you ask me later that my answer will be actually, no. Stephen Wilson 37:22 Can ask you right now, retroactively to what you just said. So, you know, yeah, you know, you wrote this paper in 2012. But yeah, we're thinking about also 2021. And have we learned anything further. I kind of agree with you to the best of my knowledge. Pascale Tremblay 37:37 We have learned a lot, but that has not come into being a unified vision. I have to admit that, as being like a researcher focused on speech, mostly, I have focused much more on the dorsal tracts in on the FAT track, the frontal aslant tract. And as far as I'm concerned with these tracks, we know more, but at the same time, we're further from a model that will resolve all the questions that we have. The questions are, yeah, where are the terminations? Do we have really a connectivity between posterior superior temporal, and inferior frontal gyrus in human? And so there's many different models nowadays. And people are continuing even in my lab, we've done some work with the arcuate fasciculus. The question remains entirely what model is correct? So nowadays, it furthers this question of whether there's this connection. But then the tract is so big. I mean, you know, most people have started to segment it into multiple subtracts. We talked about the SLF, as you talked about. And so some of the distinction of the arcuate. So in the human literature, people talk a lot less about dSLF. And they oftentimes will segment the arcuate into multiple different tracts that have similar connectivity to the SLF. Some kind of subtracts, but, you know, people like Marco Catani, he has his three component model. Others have two component models. And some of the models have a connection to premotor cortex, some don't. Stephen Wilson 39:23 Right. Yeah, you're talking about the Catani paper where they kind of have like, what they call the direct segment of the arcuate, which goes between inferior frontal and posterior temporal. And then they have two indirect segments that basically connect those same two regions, but via the inferior parietal lobe. And you know, you could view that differently and say, well, the one that goes from frontal to inferior parietal, that's kind of a lot like SLF 3. And I think this really, you guys in your in your 2012 paper, make this really great point that I think a lot of people could do to like, pay attention to. Which is that these tracts are not really natural kinds, right. It's like it's really just like a giant ball of spaghetti. And you can divide it up in different ways. Depending on your theory or depending on your method, it's not like the arcuate comes labeled in pink. And the SLF is like a nice shade of purple. Pascale Tremblay 40:11 No, no, exactly. And it's difficult when you do diffusion MRI research. Most of the time you basically testing your model. It's not the same as letting it emerge from the data, right? It's not data driven. It's model-driven. But if you start with model A, and you'll get something. And if you start with model B, you'll also get something. It's difficult to see what value to attribute to the different results with those all the different models. Stephen Wilson 40:42 I mean, is it even that much of a problem if the arcuate doesn't go to the IFG? Like maybe the critical speech motor regions are really more posterior than that the IFG. That's kind of something that's been emerging out of a lot of different recent literature, I think. Pascale Tremblay 40:55 Well, that's true. But then this connection can still be useful. I don't think it's necessarily exactly for motor speech, because I don't think that the IFG even the posterior part really is motor at all. That, for me, is premotor cortex. But, you know, Petrides and his colleagues in the macaque, they did find connection to premotor. In some of the models like Friederici, she does have this connection with premotor cortex. So there you have your sensory-motor sort of tract, if you will, for speech. Stephen Wilson 41:30 Yes. And is that the most important role for this tract in speech and language? Do you think? Pascale Tremblay 41:38 No, I think there's probably multiple roles. One thing that's interesting is to note that this track, for example, is is modified in people that, for example, we've been working on musicians and other kinds of experts. Musicians don't necessarily engage the motor speech system, it's just that they engage this system. So maybe this is something about sensory motor, more generally, multimodal integration, to other kinds of phonological processing is also very possible if there's a connection to IFG posterior, for example. So I think these are some of the possible roles, but there's probably a lot of other roles, given the connectivity so extended. Stephen Wilson 42:30 Yeah, it's a lot of brain regions connected to a lot of other brain regions. It's not gonna have a simple story. I mean, I think if we try and do a simple story, we're going to be back with that kind of simple model that you're arguing against earlier. Pascale Tremblay 42:47 No, in a recent paper we did argue for maybe a bit more general role in auditory motor transformation. Not just for speech, but also for music, singing. Because, again, these tracts seem to evolve as a function of different types of activities that are not necessarily speech at all. So yeah, I think with those kinds of drawings, like in the paper, you don't necessarily get the sense of how massive the tract is. Massive. Stephen Wilson 43:21 Yeah. Okay, so we've kind of talked about the dorsal tracks, the arcuate and SLF components. Let's talk now about the ventral tracts. So you started in your paper with the uncinate. Can you tell us like, where does that go? And what do you think that that's involved in. Pascale Tremblay 43:41 So the uncinate tract is one of the ventral tracks. So basically it tries to connect inferior frontal as well in more like anterior temporal regions to define it broadly. So it's ought to have a role in the semantic processing. But multiple other functions, I think, are possible with this tract as well. The inferior temporal cortex and a temporal pole, in particular, being seen as a semantic hub. You know, that would make sense that there will be a function that would be related to semantic processing. But, like for every other tract, I don't think that any tract as having just this one purpose. Stephen Wilson 44:25 Yeah, no, no. We kind of like want to move past that. But I think there was still sort of differential importance of different brain regions and connections for different functions. And I agree, I mean, I think that the evidence for semantic role for the uncinate fasciculus is quite strong. Pascale Tremblay 44:42 I think so too. And I think that's a tract that's, although I think it's more of a difficult tract to study. I mean, any tract that goes sort of more like medially into the brain is a bit more difficult to assess. With some of the tools that we have now. But it seems to be relatively stable in the function and people have attributed to it. Stephen Wilson 45:07 Right. Okay, so apart from the uncinate, what other ventral tracts do you guys lay out in your model? Pascale Tremblay 45:16 We talked about the extreme capsule fiber system, the inferior longitudinal fasciculus, or ILF, and the inferior frontal-occipital fasciculus, or IFOF. So those are tracts that connect either more posterior region to lower posterior region, but very posterior like occipital lobe on to anterior temporal cortex. Or, in the case of the extreme capsule, and ILF, excuse me, more inferior frontal and even regions around the insula into the temporal cortex. So this is the connectivity that connects the ventral part of the brain to the temporal lobe or the occipital lobe to both the inferior frontal or frontal lobe and the temporal cortex. Stephen Wilson 46:12 So there's like quite some controversy about whether the whether there really is a connection between the occipital lobe and the frontal lobe through the IFOF. Pascale Tremblay 46:21 Right. I think this is a tract whose anatomy will probably become, as it is more and more studied, the model will be refined. And actually, this connection may or may not be validated. Stephen Wilson 46:40 Right, yeah. Because I think that the guys that do the monkey stuff, they don't see direct connections between occipital and frontal. And they kind of argue that the IFOF that emerges from DTI studies is really conflating other tracts, like maybe parts of the ILF, inferior longitudinal fasciculus, many parts of the extreme capsule tract. In DTI, you know, it never provides clear evidence that you're actually tracking single fiber. Pascale Tremblay 47:09 It's true, it's true. There's also the case of the middle longitudinal fasciculus. That's actually been very little studied in the past in humans. Stephen Wilson 47:22 So where's that? Where does that go? Pascale Tremblay 47:24 The middle longitudinal fasciculus? Well, that's another one for which I won't have such a clear answer. But the basic idea is that it would connect, at least it goes through the entire temporal cortex from anterior to posterior that nobody argues against. It's the posterior connectivity that's unclear. Does it go to the inferior parietal lobe, superior parietal lobe, or occipital cortex? Stephen Wilson 47:50 It runs mostly under the STG, right? It's like pretty dorsal in the temporal lobe. Pascale Tremblay 47:56 Yeah, it's a very deep tract. And that this could be why it escaped attention for a while. It was first talked about in the 80s in the macaque. But currently, I would say there's at least two models that are conflicting about the anatomy of the MdLF. And the MdLF, the middle longitudinal fasciculus tract. I mean, it's very interesting, because no matter the definition, it could play a part in broadly defined language, but maybe more speech than language because of the the temporal cortex connectivity. But so the two models are the ones that the Makris model that propose that it really does go to IPL, the inferior parietal lobe. And because, we talked about it before, with the arcuate fasciculus being, you know, mainly composed of the angular and the supramarginal gyrus and those regions being involved for language and speech processing. So, of course, the connectivity between the rest of the temporal cortex including some of the primary auditory regions and those regions is interesting in the sense of language processing. But more recent models by the group of Kalyvas, they suggest three components instead of two. I don't know that I said what the two were but basically one in the Makris model goes to IPL, so temporal cortex-IPL, and the other one is temporal cortex-SPL. So superior parietal lobe. But we place a lot of emphasis on the IPL and then that's why they interpreted that that connectivity made it probably a region that played a role in language processing. But then in further studies, in all of the diffusion studies, the connectivity to the IPL is contested. Stephen Wilson 50:01 Okay, what is the alternative? Pascale Tremblay 50:03 It's that it runs sort of deep in it, but that it's not a termination. And that the termination really is superior parietal lobe. And even further occipital lobe, precuneus. Stephen Wilson 50:17 Which would make it not as obvious as a language tract. Pascale Tremblay 50:20 Exactly. I think it would remain important, because if we have this tract, you know, voice and sound processing travels through the temporal lobe, right, somehow. It could still be important in the preliminary steps. And I think this is also the point that we made in all of those papers, or we tried to, is that language does not occur independently or suddenly in the brain. It follows from, you know, auditory processing, and, you know, voice processing and multiple steps. And so maybe the MdLF could be one of those paths that is involved in some of the pre-processing of sounds before they're actually processed at the linguistic level. Stephen Wilson 51:04 Right. Yeah. I mean, that would make sense with it's sort of being underlying the superior temporal gyrus, where things are still pretty auditory. Pascale Tremblay 51:12 Yeah, no, exactly. I think it would probably be more auditory than language. And we haven't really defined it. I'm sure, probably your listeners, they do know this. But you know, I keep on going back and forth between speech and language. But you know, it's not like there's a clear break between speech and language. And I think we don't all agree about what speech and what language is. Stephen Wilson 51:37 Yeah, no, I never try to draw a hard distinction there. And I always like to talk to people that would consider themselves speech people. I mean, I just say I'm on the motor end, too. Whether it's perception or production, I don't think that there's anything to be gained by drawing a distinction there. Pascale Tremblay 51:51 No, but it's just that the distinction exists, mostly because of the field of speech pathology. Yeah, they would distinguish a speech disorder from a language disorder. Stephen Wilson 52:03 Right. I think that there's validity to that. It's useful clinically. Although, apraxia of speech, I think is something that very much straddles that border and makes the border a little bit not useful. Pascale Tremblay 52:16 Yeah, mostly, I struggle with phonology. Because for me, phonology would be more speechy than language-y. But then, you know, for people with this strict distinction, then phonology is language. Stephen Wilson 52:29 For me, phonology is language. Pascale Tremblay 52:31 But it depends on how you define language. It's just like this is the point where sounds become like, phonemes. To me, there's still like, below linguistic meaning and maybe for me language is when there starts to be meaning. Stephen Wilson 52:47 Oh no language definitely includes phonology. I mean, like, you go into a linguistics department, like there's gonna be a syntacticians, semanticists, and phonologists. Right. I mean, they're allowed in there. Pascale Tremblay 52:58 You'll be a phonetician too, right? Stephen Wilson 52:59 That's different. I mean, yeah, phoneticians is different. Like, they're in linguistics departments, too. But like, phonologists, you know... Pascale Tremblay 53:07 Well, I know. But I feel like for me, it's actually the interface. If I put it exactly how I see it. To me, that's the interface between speech and language. Stephen Wilson 53:17 Interesting. Pascale Tremblay 53:18 But for, for the MdLF, we had some results recently. But again, they're diffusion results, and we sort of basically looked at how this tract aged and how, you know, also the arcuate fasciculus aged. And you know, whether there were correlations between this and speech processing. And we did find that it did correlate with speech processing skills. Stephen Wilson 53:44 So this is an aging people, you said? Pascale Tremblay 53:47 Yeah, actually, if I recall correctly, in the MdLF, it was pretty much age unrelated that you had a connection between how the metrics for this tract or how the quality of the white matter in this tract and speech processing performance. It was speech in noise, so we can argue whether that's a speech or an auditory function all you want, I don't have an answer. But it was correlated. There was some relation also with the arcuate fasciculus. So I think it's still possible that this tract is even if it doesn't go to IPL. I mean in the model we use though, we use a model that did go to the IPL. Stephen Wilson 54:36 Cool. Is this paper published? Pascale Tremblay 54:37 Yes. Maybe two years ago. Stephen Wilson 54:40 Okay. I will link it in the podcast notes. Pascale Tremblay 54:43 Yeah. And more recently we look at the arcuate in aging people that sing and those who don't sing. We did find that the arcuate fasciculus was different in singers and non-singers. But we couldn't find a relationship between their performance in the speech perception in the sense that the tract was more developed in the singers, but that didn't make them better at the task. Yeah, there's a relationship for everyone with speech, but just it wasn't special for the singers. So yeah, so I think MdLF, the way that I see it and the arcuate, are two of the tracks that we we need a lot more information about because the connectivity is really not clear. Some of the recent studies are convincing that because of these have combined diffusion, I refrain from saying DTI because it's not always DTI, and dissection. So I think that provides information. I think what's missing that I don't know if it's possible is, you know, like, when you sort of test different models, how they fit the data, right? That would be nice to be able to do that and see, okay, if I use this model, how well does it account for the data versus that other model of the tract? And I think maybe that would help us sort of resolve some of these ambiguities. So, yeah, the arcuate fasciculus and MdLF for me are the two very interesting tracts, because every time you read a paper, you learn something new. Stephen Wilson 56:34 Yeah, for sure. And the arcaute certainly got plenty of attention. But the MdLF not as much as you'd expect, for a tract that basically runs anterior-posterior underneath the superior temporal gyrus. I mean, that seems like a track that we should be very interested in. Pascale Tremblay 56:49 But there's more study now. Like I think it's picking up slowly. But I think it's going to be from the auditory side that people will be interested in it. But it probably will spill over to us even though it may not be done for the purpose of understanding the language connectome. We will still learn from it. I think the auditory processing people will probably jump on it. Stephen Wilson 57:16 Yeah, cool. Okay, so the last tract that we haven't really talked about yet, that you have a recent paper about is called the frontal aslant tract, or the FAT, for those in the know. Can you tell us about the FAT? Pascale Tremblay 57:32 Yeah, the FAT is such an interesting tract, for me, at least because I've done most of my thesis work on the supplementary motor area and the pre-supplementary motor area. The medial frontal regions that are involved in sort of like conflict monitoring, response selection for speech and non-speech? These seem to be domain-general regions. And so the frontal aslant tract, or FAT, does connect this region to the inferior frontal gyrus. When I discovered that tract, it was something to me that made so much sense. Because even though the SMA and the pre-SMA are usually not in models, certainly they're not in the classical models by any means. But I think they're getting acknowledgment a lot more that they play an important part. Stephen Wilson 58:25 Yeah, no, Wernicke and Lichtheim didn't know anything about them. It was I think, Penfield that was the first to really point out their, like, role in language. Pascale Tremblay 58:33 Right, exactly. And so this tract, basically, that's what it does. It connects those regions involved in both some something like I would say, higher order motor planning, and also executive function to this region that we've been talking about a lot, because it's also the potential termination of the arcuate fasciculus, and one of the core regions for language processing. So that sort of connects a region that used to be thought as motor region for speech, the IFG. Which I would argue is not at all a motor region, since it has no motor connection. But it does connect this region to centers that are involved in some high level motor function. And that does make sense if you think about, you know, going from say, a motor plan to a representation of a word for example, like during lexical access for speech production. LIke in fluency tasks, for example. And so that these are some of the roles that have been proposed for this tract, being involved in fluency, for example, on the left side. And on the right side, maybe more an executive control role. But like with many of the other tracts that we talked about, the anatomy of this tract is debated. I mean, it's interesting, because that's a tract that was discovered using diffusion MRI and not labeling or dissection, though it was validated afterwards with labeling and dissection. So I think it does exist, although I think some people still have doubts about this. Stephen Wilson 1:00:12 And that you're saying that the terminations are in the IFG? Is it also in sort of ventral premotor cortex on the other side of the precentral sulcus? Or is it really going into the IFG, or is that what's up for debate? Pascale Tremblay 1:00:25 I haven't seen anything pointing to the premotor cortex. I think that would make so much sense, of course, because these regions are supposed to be connected. But I haven't seen this. What I've seen is that there's a Spanish team, that's the sound talking about a big FAT, what they call the extended FAT. I think the title of their paper was, "When the FAT goes wide" or something like this. So a really cool title. But basically they argue the FAT actually comprises the entire medial frontal region. So SMA, pre-SMA and anterior regions to this. They argue that it is related on the left to language. So really consistent what we have proposed, but that on the right that it will be involved in working memory. But so I think that, yeah, I think it remains to be seen whether it's actually such a massive tract because if it does involve the entire medial SFG then it's humongous. Stephen Wilson 1:01:33 Right. Yeah, there's an enormous amount of cortex in there. Pascale Tremblay 1:01:38 So then they probably will need to be segmented just like we're trying to do with the arcuate, to try to piece it out and understand what different pieces might be doing. But I think this was a very important discovery for language neurobiology, as well as, of course, connectomics in general. Because it's not every day that you discover a new tract. Stephen Wilson 1:01:58 It definitely makes sense that there has to be a connection. You know, the role of the medial wall of the frontal lobe in speech land anguage is pretty clear. We don't often see strokes there. But when we do see them, from ACA strokes, anterior cerebral artery, you definitely have this sort of rather unique aphasia that involves basically a failure to initiate, and relatively spared repetition. It's very distinctive. Pascale Tremblay 1:02:29 It's called SMA syndrome, right? Stephen Wilson 1:02:32 Yeah, I guess so. I mean, in the classic model, it would be called like a trans-cortical motor aphasia. But I mean, there's probably other ways of thinking about it, too. Pascale Tremblay 1:02:45 Are you an SLP? Stephen Wilson 1:02:46 No. Pascale Tremblay 1:02:55 I've always wondered if you had a master's degree in SLP. Stephen Wilson 1:02:58 No, no. Oh, hang on there. My dog has like decided to like knock the door open. Hang on. Let me get the dog out of here. Okay, so what were we saying? Pascale Tremblay 1:03:08 SMA syndrome or what you were calling transcritical motor aphasia. But basically, I've read that it can start with complete mutism. But it does recover quite quickly. Stephen Wilson 1:03:22 Exactly. That's what we see. Yeah, it can be really deceptive. It can look deceptively severe, but it recovers quite well. Which is great, thank God. I mean, but you know, a lot of aphasias recover really well, for reasons that are very mysterious. Okay, cool. Well, we talked about the FAT. So thank you for going through all these tracts with us. With with me and our listeners. I think we made a real good shot at trying to talk about something that's inherently very visual. But without having any visuals. And yeah, I appreciate it. Pascale Tremblay 1:04:00 Thank you so much for having me. I mean, I think you had this very great idea to do these podcasts. I've been enjoying myself listening to them. Stephen Wilson 1:04:08 Oh, I'm glad you listen to it. Yeah, it's been a lot of fun for me to make them and to kind of meet new people and catch up with people that I have gotten to know over the years. So thanks a lot. And yeah, have a good rest of your day. Pascale Tremblay 1:04:20 You too. Thank you so much. Stephen Wilson 1:04:22 All right. Take care. Okay, well, that's it for Episode 10. If you'd like to learn more about Pascale's work, I've linked her lab website, the papers we talked about, and the society website in the show notes at www.langneurosci.org/podcast. I'd like to thank Latane Bullock for editing the transcript of this episode. Thank you all for listening, and I hope to see you next time.