Welcome to Episode Five of the Language Neuroscience Podcast. I'm Stephen Wilson and I'm a neuroscientist at Vanderbilt University Medical Center in Nashville, Tennessee. Today, I'm gonna try something a bit different. Instead of the wide ranging chats that I've been having with leading scientists in our field, we're going to focus in on a single paper that came out recently that I thought it looked interesting and
important. The paper is called "Damage to Broca's area does not contribute to long term speech production outcome after stroke". And it came out in Brain a bit over a month ago. I'm really thrilled to be joined by the co-first authors of this paper, Andrea Gajardo-Vidal and Diego Lorca-Puls. They're a married couple, originally from Chile, and they're fantastic early career researchers. They did this work at University College London in the lab of Cathy Price, who is the senior
author. They've recently started new postdoctoral positions at the University of California, San Francisco. Hi, Andrea and Diego, how are you guys doing today?
Fine, and you?
Very well.
I'm doing good, thanks. And where are you guys?
So we are in San Francisco, in the US. We arrived here like two months ago at the beginning of January to start to do a new postdoc at the lab led by Marilu Gorno-Tempini at UCSF.
Yeah, I'm very familiar with her lab, because that's where I did my postdoc training as well.
Yeah, we know about that. Yeah. How was your experience here?
Oh, it was great. Everything that I know about clinical research. I learned from Marilu and the team at the Memory and Aging Center.
It is a great place.
Well, almost everything that I know.
Surely.
I don't want to like downplay the contributions of some of my other mentors, either. But yeah, it was great. And what kind of projects are you guys going to be working on at the Memory and Aging Center?
So in my case, I'm interested in looking at how executive functions influence language processing at behavioral and neural levels in patients with primary progressive aphasia. So that will be my focus during my time here.
Okay.
And in my case, I would like to gain a better understanding of the brain regions that support specific stages of speech production, by integrating behavioral and neuroimaging data from patients with primary progressive aphasia, and post-stroke aphasia. So it's an international collaboration between Cathy Price in London and Marilu Gorno-Tempini, here in San Francisco.
Oh, that's really cool. Because I know that Cathy, of course, trained Marilu, back in the day, so you're kind of like bring it all full circle.
Yeah.
Yeah, exactly.
And kind of being the bridge to start new collaborations between the two of them.
Yeah.
Exactly, exactly. Yeah.
So one of the things I like just before we get into the, so what we're doing today is we're going to kind of do a deep dive on your new paper, which I found really interesting. As soon as I saw it, I thought, "Okay, this is an important paper", It's called "Damage to Broca's area does not contribute to long term speech production outcome after stroke". So what we're doing today, as you guys know, is we're going to kind of talk in detail about this paper, focus on it, get into the science.
Before we get into that, I kind of like to get to know people a bit and just kind of understand who you are, and how you got to be the kind of scientists that you are. So should we go first with Andrea?
Yeah, sure.
So what kind of scientist would you say that you are? And how did you become that kind of scientist?
Well, I define myself now, I define myself as a cognitive neuroscientist with a strong focus on clinically relevant research. So my main, my long term plan, I will say, is to understand how the language system is organized in the brain and how it adapts after brain damage. So ideally, I would like to use this knowledge to boost the recovery of language abilities in brain damaged
patients. And I got interested in this relationship, the relationship between language and the brain, while studying my undergraduate in speech and language pathology. So yep, so this interest only increased when I started treating patients with acquired speech/language deficit due to stroke. So that's kind of my background.
So where did you do that undergrad degree in speech-language pathology?
So yes, I grew up in Concepcin, which is a city located in the south of Chile. I studied there at the University of Concepcin.
Yeah.
So probably you don't know that.
I'm not familiar with Chile. I know that it has some of the driest desert in the world but not in the south, I think. That's further north.
The south is beautiful and very green. I 's completely opposit
It's wet as well. It rains a lot.
You're probably at a similar latitude to where I grew up in Australia.
Oh, you grew up in Australia!
Yeah.
Oh, yeah I didn't know that, okay. So originally you're from Australia, not the US.
Oh, yeah. No, I have an accent.
I can see your accent, yes.
It will be apparent to native speakers.
Yeah, I noticed that your accent has some kind of British something there.
Yeah, some people find the Australian accent to sound British.
It's kind of a mixture, isn't it?
It's ti own thing. You know? I mean, obviously, the Australians, many of our ancestors came from Britain, somewhat later than the US. So that's why the accent resembles the British accent more than the American. So were you always interested in language as a kid? Or did you just kind of like find your way into speech pathology?
I will say yes. I will say that I found my way during my undergraduate in speech language pathology.
Yeah. When we started to treat patients, I think that's when we realized the importance of gaining a better knowledge about how brain damage to specific parts of your brain relates to very specific behavioral deficits.
Right. Yeah.
Yeah. And then, I think it's also important to mention that after our undergrad, then we traveled to London to do our PhD with Cathy Price to University College London. And we also completed a postdoc with her. So we stayed there for quite a bit.
So a master first.
Yeah a master first.
A master program in cognitive neuroscience. Yeah.
And Diego, did you have a similar kind of path into this field? Did you
Well, pr tty much the same. So Andrea an I were classmates, at un versity. So we studied the sa e degree. And we had similar ex eriences. And so I will only ad that we had a class focus on ne ropsychology and there we learn d about the work by Luria.
Yeah.
So the Russian neuropsychologist. That really got me interested in the relationship between brain and behavior. And then, you know, after that when we started seeing patients at the hospital, that also kind of motivated me to pursue a career in research, neuroscience research.
Right. And does an undergraduate speech-language pathology degree in Chile, does that make you into a clinician or is there further training beyond that to become a clinician?
It's actually a relatively long degree. So it lasts for five years, and during the last year, it's an internship. So you basically work as a speech and language pathologist with a supervisor who is, you know, guiding you throughout the experience. And then when you graduate, you can basically work as a speech and language pathologist.
Okay, that's great. So it's almost like the combination of an undergrad and master's degree in the US in terms of time commitment and the amount of time you spend on that.
Yeah, I would say that.
That's cool. I think it's neat that you kind of got some Luria exposure in your undergrad degree. I think he's given not enough attention in, you know, Western language neuroscience.
Yeah. I mean, his work was amazing. Yeah. So ahead of his time, I think.
Yeah. And he has a different kind of way of thinking about aphasia than the tradition that I'm more familiar with. And I'm definitely hoping to learn more about his work as I, you know, kind of continue to learn. Okay, so let's get to the paper that we're going to be talking about. So this paper kind of comes out of the PLORAS study, led by your mentor in London, Cathy Price. Can you guys tell me about the PLORAS study?
Of course. So PLORAS stands for 'Predicting Language Outcome and Recovery After Stroke.' So this is a project that Cathy price started about 10 years ago in the UK, with the aim of identifying lesion and non-lesion factors that explain interpatient variability in language outcome and recovery after a stroke. So far, that data from more than 1300 stroke patients have been collected, including structural MRI, along with demographic
data, and behavioral scores. So this project also has an fMRI component, in which patients are invited who, despite heavy lesions in areas associated with language processing, perform relatively well on language tasks.
That's a good summary. I will only add that this is a huge project that has been going on for a while. I think it's, you know, first of its kind, I would say, actually. There's a very rich dataset, it's amazing.
Yeah. I mean, I've been very impressed by it ever since I heard about it. And I've kind of been eagerly following each paper that's come out of that project. Because I think, you know, Cathy probably sees the importance of having very large numbers of patients to really capture the patterns of variability and to obtain generalizable findings. I mean, I think many of us in this field kind of agree that like, a lot of the studies to date have been
underpowered. And, you know, that's not gonna be a problem with 1300 patients. You're gonna be able to do a lot of things that were not possible with smaller sample sizes.
Yeah, so that's a particular focus of Cathy, you know, on interpatient variability. She really wants to understand that. And I totally agree with her, I think that's so important. We shouldn't just dismiss variability.
No, not at all. So your paper involves a group of patients with chronic post-stroke aphasia. They all had their lesions mapped, and they have language evaluations, like you said. What you're going to do is look for relationships between damage to specific regions and tracts, and language outcomes. So let's kind of go through and talk about the different pieces of the puzzle, starting with the patients. So can you guys tell me how many patients are included in the
study? What kind of inclusion criteria and lesion locations did you have for your patient sample?
Sure. So we studied 134 stroke patients with left frontal lesions that didn't extent into posterior speech production areas in lateral inferior parietal and superior temporal association cortices. All patients were right handed, native speakers of English, between three months and ten years after stroke. And importantly, the presence of aphasia wasn't part of the inclusion criteria. So some of the patients did have aphasia,
while others didn't. And this is because patients without aphasia allow us to investigate how consistently damage to a specific part of the brain is associated with a deficit of interest.
Yes. So you're defining inclusion based on lesion location, not on outcome. And I think that makes a whole lot of sense. And so just to kind of recap for our listeners, you're excluding patients whose lesions extend significantly into the parietal or temporal lobes. And that's really important, because now you've got a very well-defined sample. And this is the thing, when I saw your paper that really kind of blew my mind. 134 frontal
patients. Like most people that have even large datasets of patients with aphasia don't even have 134 total. But you guys have got 134 with lesions restricted to the frontal lobe. And that's going to give you like a lot of potential power to discriminate between the different contributions of different regions, right.
Yeah, definitely. I mean, that's the beauty of the PLORAS project.
Of PLORAS?
Yeah.
Okay, cool. So what you did is you quantified damage to a set of brain regions within the frontal lobe. Can you tell me which brain regions you included and why? And which ones you didn't include and why? How did you decide on your set of brain regions to quantify?
We exclusively select brain regions that had been previously associated with the speech production by a large body of evidence. These were BA 44, BA 45, which are together known as Broca's area. The ventral lateral premotor cortex, the parts of the primary motor cortex that are implicated in the motor control of articulators. Both parts of the superior part of the central sulcus of the insula, the putamen, the frontal aslant tract, the anterior arcuate fasciculus. And finally, the
uncinate fasciculus. Yeah, so there were other regions such as the supplementary motor cortex, posterior parts of the arcuate fasciculus that we didn't include because they didn't overlap with the patient's issues. So in the case of the inferior frontal occipital fasciculus, we excluded because a previous study led by Fridriksson and colleagues from 2013, did not find a significant relationship between persistent speech production impairments and damage to this particular
white matter tract. So that's kind of, anything you want to add Diego?
That's pretty much everything. Yeah.
Yeah, great. So I'm trying to stay on top of defining abbreviations for our listeners. I think most people know that BA stands for Brodmann area. So Andrea, you mentioned Brodmann area, 44 and 45. They would also be called the pars opercularis and pars triangularis of the inferior frontal gyrus. And on those tracts, you know, you've kind of got the uncinate fasciculus included, which is one of the
frontotemporal tracks. But you're not including what you guys are calling the IFOF, inferior frontal occipital fasciculus, which is the other major frontotemporal tract. And your justification for that is that Fridriksson et al. didn't find a relationship between that tract and speech production deficits. I guess including it would probably result in, it probably would be kind of correlated with uncinate damage, I'm guessing? I don't know if you looked at it at all?
Yeah, we didn't look at that. But, yeah, probably. That would be the case. We haven't looked at that. But, yes, probbaly.
Yeah, that was kind of the one that I wondered about. I think everything else was very justifiable. That was the one where I was like, oh, I wonder what happened if you had included that one. Because there are some documented relationships between damage of that tract and certain language deficits. Not production deficits, though. So I get it where you, you know, went that way.
Yeah.
Okay, so let's now talk about the language assessment. How does aphasia get assessed in the PLORAS project, and what outcome measure that you guys decide to focus on?
All patients recruited to the PLORAS database are assessed with the Comprehensive Aphasia Test, which is a standardized language battery widely used in the UK. It comprises 27 different tasks that evaluate a wide range of language abilities, including naming, repetition, reading and
writing. Our behavioral index of speech production abilities, was the spoken picture description task from the comprehensive aphasia test, where the patient is prompted to describe verbally what is happening in a scene for one minute. We decided to use connected speech production rather than single word production as our outcome measure because it is more likely to capture the difficulties that a patient might experience in everyday communication.
Okay, so it's more ecologically valid than picture naming or something like that.
Exactly.
It's like a proxy for natural speech production, I would say.
Yep, I think it's a good measure. And I definitely prefer it to the fluency measure from the Western Aphasia Battery, which is what a lot of other related researchers used, and everybody hates, but everybody uses anyway.
And one of the advantages that this task is particularly sensitive to is speech production difficulties.
Yeah. So now we've kind of described the regions that you quantify damage in and the behavioral outcome measure. You then constructed statistical models to see in which regions would damage predict speech production outcomes. So let's kind of talk about some of those models. So you kind of fit a sequence of models. I think in the paper, you describe them all really nicely and clearly. So I appreciated that as a reader. In the first model, you had only BA
44 and BA 45. What did you find in that model?
So, let me start by saying that the aim of this first model was to test whether our decision to investigate BA 45 and BA 44 individually rather than a single area was supported by the data. So that's was the aim of model one, our first model. Here, we found that damage to BA 44, but not BA 45, significantly predicted worse speech production scores, which is not entirely surprising, because previous work has shown that BA 44 probably plays a more active role in speech production than
BA 45. So, since damaged to BA45, pars triangularis, did not make a unique contribution to the prediction of speech production scores above and beyond that of BA 44, we excluded it from other analyses. So that is a summary of our first analysis.
Okay, so the first analysis is primarily aimed at figuring out which part of Broca's area you should be focused on. And you've basically found that if any part of Broca's area was going to be predictive of the speech production deficits, and it's going to be pars opercularis, BA 44, rather than the pars triangularis, BA 45. And then in your subsequent models, you basically don't really look at BA45 anymore, you're focusing on 44. Okay, so to my mind, model 2 in your paper was probably the
most central one. Can you tell me about model 2 and what you found in that model?
Yeah. So in model 2, we accounted for the fact of damage to regions neighboring Broca's area, by entering the seven remaining regions of interest into the regression. We also covaried out variance unrelated to lesion site or speech production by adding lesion volume, time post stroke, age at the stroke, and semantic memory scores. And we found that across all eight regions of interest included the model two, only damage to the anterior arcuate fasciculus, significantly predicted worse
speech production scores. In other words, damage to BA44, no longer predicted speech production scores when damage to surrounding regions was controlled for.
Okay, so only the anterior arcuate fasciculus made an independent contribution to predicting speech production outcomes, when everything else was included in the model, and not the part of Broca's area that was your candidate.
Yeah, exactly. Yeah.
So I think one of the great strengths of this paper, like we talked about earlier, is the sample size, right. You have this unique sample size of patients with damage to different adjacent frontal regions,that kind of gives you the potential to parcellate out the contributions of these neighboring regions and tracts. Few of us have have ever
had that opportunity before. But even though you have this very large sample size with 134 patients, you still have pretty strong patterns of adjacent regions tending to be damaged together. So if something's next to you know, if region A is next to region B, most lesions that affect region A are going to affect region B. And that creates multicollinearity in the model, which you guys address in the paper. So can you tell me about how you go about addressing the multicollinearity challenge?
Yeah. So high multicollinearity is probably
Yeah. one of the biggest challenges we face when relating behavioral deficits to specific lesion sites after stroke. And this is because when damage one region can almost perfectly be predicted from damage to another region, disassociating their effects becomes virtually
impossible. So for example, if damage to region A, which is associated with function X, very frequently co-occurs with damage to region B, which is is not associated with function X, we won't be able to know which of these regions is driving the
observed behavioral effects. Or even worse, we might wrongly conclude that both regions contribute to function X. In our case, we were fortunate enough to have access to a large number of stroke patients who differed in the degree of damage or regions of interest, making it possible for us to dissociate their effects except for the ventral lateral premotor cortex, and the frontal aslant tract, which were affected by high
multicollinearity. In these cases, we dealt with high multicollinearity by rerunning the regression after removing the regions of interest with which the ventral lateral premotor cortex or frontal aslan tract were highly correlated. Importantly, this didn't change our results because damage to he anterior arcuate fasciculu continued to be the on y significant predictor of spee h production scores.
Right. Okay. So I think it's really neat that the arcuate fasciculus is the only region that can explain additional variance once everything else is included in the model. But I'm interested in knowing how well the model would do without the arcuate. So in your final model, you basically explained 51.5% of the variance, which is quite a lot. Half of the variance being explained by damage to these frontal regions,
in combination. Did you ever run a model that didn't include the arcuate and get a sense of how much you could explain without the arcuate?
Um, yeah. So the unique effect of damaging anterior arcuate fasciculus accounts for around 6% of the variance. This is after adjusting for everything else. However, the total effect of damage to the anterior arcuate fasciculus is much larger, accounting for around 37% of the variance, which is very close to the 40% of variance explained by damage to this white matter tract reported in the
Fridriksson et al study. Yeah, and I think it's also important to mention that adding the arcuate fasciculus to a covariate-only model comprising lesion volume, time post-stroke, age of stroke, and semantic memory, of course, leads to a significant r-squared change. This isn't the case for any of the other regions of interest examined in our study. Taken together, our results suggest that damage to the anterior arcuate fasciculus is a particularly robust predictor of long term speech production.
Right, that's clear, yes. Okay, so you next address the possibility that arcuate fasciculus damage might impact language by disconnecting Broca's area. You argue against this interpretation. You've got several analyses agreeing against that interpretation. So just for the sake of keeping our conversation intelligible to our listeners, can we focus on just
one of those analyses? Can you tell us about the analysis where you derive two groups of patients that had lesions to either one or the other of these regions, either to the arcuate alone or to Broca's area alone.
Sure. So it could be argued that long lasting effects of damage to the anterior arcuate fasciculus might be explained by a disconnection of Broca's area. So this basically means that Broca's area might disconnect as a result of damage to the anterior arcuate fasciculus. And this what really drives the behavioural effect. So to investigate this possibility, we selected three groups of patients. So with different lesion sites, who were for lesion volume, time post-stroke,
and age at the stroke. So the first group comprises seven patients with damage to BA 44, and relative sparing of the anterior arcuate fasciculus. The second group comprised seven patients with damage to the anterior arcuate fasciculus, and relatively sparing of BA 44. And the third group comprised of patients with damage to both BA 44 and the anterior arcuate
fasciculus. And we found that the speech production ability of patients with damage to BA 44 (group one), were on average, within the normal range and significantly better than those patients with damage to the anterior arcuate fasciculus, those from group two. So these results cannot be explained by Broca's area, this connection hypothesis, from our point of view.
Right, yeah. Because if it was just a disconnection, then the patients with Broca's damage ought to have been at least as bad as the patients with arcuate damage. Yeah.
Yeah. Yeah.
Yeah. So you argue against that disconnection hypothesis? I think it's pretty compelling argument. So many of your analyses kind of play off the arcuate fasciculus versus Broca's area. I wonder if like a stronger candidate for a cortical predictor of speech production deficits might have been ventral premotor cortex. As you guys know, is the area with where stimulation actually leads to speech arrest as opposed to
Broca's area. Did you do any analyses where you looked at whether ventral premotor cortex might contribute to predicting long term deficits?
Indeed. Damage to the ventral lateral premotor cortex is a significant predictor of speech production scores when analyzed in isolation. However, as soon as the effect of damage to anterior arcuate fasciculus is controlled for, this relationship stops being statistically significant.
In fact, in the paper, we showed that more than 80% of the contribution of damage to the ventral lateral premotor cortex to long term speech production outcome can be explained by co-occurring damage to the anterior arcuate fasciculus.
This is consistent with the findings from van Geeman et al. reported in 2014, where the partial resection, the partial or complete resection, actually, of tumors infiltrating the ventral lateral premotor cortex only resulted in transient speech production deficits.
Right, yeah, I think so. So really, that's no more plausible as a single cortical region than Broca's area?
Not really.
Right. Okay, cool. So we've discussed your main results. I mean, there's a lot of other analyses and results in the paper that our listeners can read if they want to kind of delve more into the details. I mean, I'm kind of skipping over the mediation models that are super complicated and hard to talk about in a podcast. But it's, you know, I think we kind of covered the main ground and the main structure of your
arguments. So you conclude, in your discussion, you say, I'm going to quote, "we are not refuting the role that Broca's area has been shown to play in speech production in the undamaged brain. But we are challenging the long held assumption that damage to Broca's area contributes to long term speech production impairments after stroke". So if you think that Broca's area does play a role in normal speech production, then how can it be that damaging it doesn't have a long term effect?
Our findings indicate that Broca's area is not as critical for speech production, as commonly thought, because we have shown that damage to Broca's area doesn't contribute to long term speech production outcome after a stroke. However, many studies of the undamaged brain have sho n that Broca's area does pla a role in a speech pro uction. This apparent dis repancy might be explained by
he effect of recovery. In oth r words, it might be the cas that the recovery potential is uch greater after damage to Bro a's area than the anterior arc ate fasciculus, as an inc easing number of studies hav started to show. For exa ple, the results reported by Her ert et al. in Brain in 2016, b sed on intraoperative mappin in patients with brain tum rs, suggests that the degre of functional compensatio is generally high after cort cal damage, and low after w
ite matter damage. However, the strokes and tumors are not the same thing. So fu ure longitudinal studies are ne ded to investigate the sp ech production recovery in the stroke patients with relati ely focal damage to Broca's area or anterior arcuate fascicu
Well, conveniently, that's the kind of data that my lab is acquiring. Maybe in a year or two. I mean, you know, our primary project is a longitudinal study of the first year of recovery from aphasia.
Oh, that's fantastic.
Although we don't, yeah, we don't have the kind of numbers that you guys have in the PLORAS project, because we're recruiting from a single site. We have been working on this four or five years, and we've got a pretty good cohort. So hopefully, we'll be able to kind of shed some light on those recovery questions.
Looking forward to those papers.
Yeah, me too. Maybe if I don't spend so much time working on this podcast, we can write some stuff up. Okay, so you guys, you've kind of argued against the effect of the arcuate fasciculus as being a disconnection of Broca's area. How do you think the damage to the anterior arcuate fasciculus exerts its long-term effect on speech production?
Yeah, so we have already stated why its disconnection of Broca's area cannot logically explain our results. It is possible, however, that the long lasting effect of damage to the anterior arcuate fasciculus on speech production might be the consequence of this rapid functional integration among the multiple regions in inferior frontal, inferior parietal, and inferior temporal cortices that are involved in speech
production. So in other words, I would say that damage to anterior arcuate fasciculus might disrupt a substantial part of the speech production network, thereby resulting in a speech production impairments. So that's my view about it.
Yeah. Yeah, I think that connects well with several other observations in the literature. I'm thinking about And Turken and Nina Dronkers' 2011 paper on the white matter underlying the posterior middle temporal gyrus, kind of making this case that damage to that white matter basically kind of impacts a much wider network than you might think. And that's maybe the reason why. You know, that's the area that's most predictive of long term aphasia, severe long term aphasia. I mean, that's not
about production, right. It's about aphasia more generally. But this might be a similar kind of argument about production. Yeah.
Yeah, exactly.
So let's talk about your paper in the context of other studies that have also kind of expressed this skepticism of the central role of Broca's area. And you guys cite all of these studies. So I know that you know this literature well, so it kind of goes back to Pierre Marie, 1906, you know, kind of trash talking his fellow Frenchman. And writing this paper that basically, I'm not going to embarrass myself by trying to state the name of the paper in
French. But the translation is, "the third left frontal convolution plays no special role in language". So you know, as you guys know, Pierre Marie and his student, Moutier, had a lot of patients and did post mortem and careful anatomy, and also documented similar findings to you all, which is that isolated damage to Broca's area didn't seem to cause Broca's aphasia. They also argued that it was a white matter damage, in what they call the lenticular zone, that was predictive of
speech production deficits. Then you've got JP Mohr. Again, who you guys cite. I had the pleasure of seeing him give a talk a few years ago, where he declared, and I quote, "Broca's area is bullshit".
Oh wow.
Really?
That's what he said. "It's all bullshit." I'll never forget it. So he kind of made a somewhat different argument. He looked at patients that had damage restricted to Broca's area, as you guys have done, and found that they had really good long term language outcomes. Not just with respect to production, but lack of aphasia in general. Whereas, in his view, you needed to have kind of the whole upper division of the MCA territory, middle cerebral artery territory, needed to be infarcted to get a
persistent Broca's aphasia. And then, most recently, which you guys already mentioned earlier, is Fridriksson et al. 2013 in Brain, where they actually argued, similar to you all, that it was the anterior part of the accurate fasciculus that is most predictive of speech production deficits after stroke measured with the fluency subscale of the Western Aphasia Battery, which is not my favorite. But, you know, probably more or less gets the job done. So your paper goes beyond this prior work in
several ways. Do you want to talk about how you're findings sort of converge with this prior work as well as how you extend it?
Yeah, sure. So, our study was certainly informed and motivated by previous work. For example, the importance of the anterior arcuate fasciculus, as you have said, for a speech production had already been highlighted by Fridriksson et al. who showed the damage to this white matter tract was associated with nonfluent speech production in chronic stroke patients. We were able able to replicate and extend these findings in a much larger sample with stroke
patients. On the other hand, some unique aspects of our study is that we focus on the effect of damage to Broca's area in the context of relatively focal left frontal strokes. We tested the Broca's area, this connection hypothesis, by conducting post-hoc group comparisons on a small group of patients with
different lesion sites. We also provided evidence for the absence of a long lasting effect of damage to Broca's area on speech production, which is much more informative than just showing a lack of significant
association. And finally, we showed that more than 70% of the influence of damage to different parts of the posterior inferior frontal cortex on speech production can be explained by conquering damage to the anterior arcuate fasciculus, which is consistent with the lesion sites of Paul Broca's two historic cases, both of whom had co-ocurring damage to the anterior arcuate fasciculus.
Yeah, that's really neat. So I would it be fair to summarize and say that the unique contributions of your paper are mainly related to your ability to really discriminate between these adjacent structures and tracts, which is made possible by your extraordinary sample size of frontal patients?
Yeah, totally. Yeah, I would say that's one of the main strengths of the paper,
Right. Why do you think the role of Broca's area has remained so compelling in our field, despite the mounting evidence against its central importance?
That's a great question. And I'm not sure actually, but it might be due to its simplicity, in some way. And the fact that many traditional textbooks from the 70s, 80s, and 90s, teaching the Broca's area is critical for speech production, are still used to train a speech language pathology, psychology, and medical students. So that might be one reason, but I'm not sure why. The evidence is so compelling.
Yeah, it's interesting. I sort of understand why clinicians hold on to that old idea. But we can't use that as an excuse for our field. You know, the language neuroscientists, you know, I mean, I know that I've been guilty of it. I just think it comes up in imaging. Especially with imaging, you know, it's so easy to find effects in the pars opercularis, especially. And it's hard to get away from even though you know deep down that it's not real. You know, it's hard to get away from.
So what's your opinion about Broca's area, Stephen? What is the role for Broca's area in speech production and language?
Hey, you guys are not allowed to ask me questions.
I would like to know your opinion about it.
I don't know. Your findings comport with what we've seen in my lab, you know, unpublished data that is longitudinal and related to the first year after stroke. We see even very early on, we don't see significant speech production deficits in patients who have isolated damage to Broca's area.
That's interesting.
I think it's an area that, certainly its role in language has been seriously overstated. And I still think it's somewhat mysterious as to why it's so easy to activate in millions of different language paradigms. I mean, some of it is confusion with probably multiple-demand regions in the
left inferior frontal cortex. So yeah, as you know, Ev Fedorenko who I talked to on the show a couple months ago now, has done this work showing that, you know, language regions and MD regions are side by side in Broca's territory. So I think some of it is confusion between those regions. But I don't think that's the whole thing. I think there there are genuine language
regions. Well, they're language regions in the sense that they're activated by language paradigms in the inferior frontal gyrus, and yet damage to them does not cause significant lasting aphasia. And that's, I mean, honestly, it's a mystery. It's one of the things that keeps me up at night. I don't think it's fully explained.
Yeah. Well, it might be because the brain is so plastic, you know. We haven't been able to actually investigate it in real detail.
I think the adult brain is a little bit less plastic than people assume. Just because we don't see, in our unpublished data, dramatic shifts in the localization of the language network. But I think it might be more redundant than we assume. So it's not so much that it's plastic, and that a whole other region is gonna
take over. But it might be that there's a much wider range of regions that can perform those functions and that when some of them are damaged, others are very much ready to step in very quickly, maybe even immediately.
I mean, that's what I meant. Yeah.
Oh, okay. Well, yeah. Different definitions of plasticity. But yeah, there might be more redundancy than we assume. I think that's probably part of the explanation. But I don't think it fully explains why Broca's area is so amenable to activating.
I've been looking at some fMRI data from patients who have damaged Broca's area, but you know, not focal damage. But they do overlap in Broca's area. And what we see is that they activate a substantial part of the intact speech production system, without showing effects in regions that are not seen in in controls. Although, we do see increased activation. So some indication of functional compensation. It's particularly strong in right
You see that where? Crus I in the right posterolat Well, that would be a result out of, I won't say out of left field, but out of right field, something like that. Yes, the right cerebellum is certainly, that's the appropriate cerebellar hemisphere to be doing something in language.
Yeah, totally. So I think it's a very interesting result. Because, you know, there are an increasing number of studies showing that if you stimulate the right posterolateral cerebellum, it might help to boost the recovery of speech production and language abilities in general, in stroke patients. I think it it kind of aligns well, with current evidence. Yeah. Working on that paper at the moment.
All right. Well yeah, I'm looking forward to seeing that. That's an interesting substrate for reorganization that I haven't heard talked about too much.
Yeah. So before we finish, I just wanted to make clear that I do think that Paul Broca made an important contribution to your understanding of functional specialization in the brain. Although, the term Broca's area might not be as useful as we used to think.
Yeah. Well, I really appreciate you guys joining me for the podcast.
Thank you very much Stephen for having us. It was a pleasure to discuss our paper with you.
Yeah. It's great to talk to you. And I really enjoyed the paper. Congratulations on it. And I hope that your time in the Memory and Aging Center will be as productive and enjoyable as my time there was.
Yeah, certainly will do, will be.
Say "hi" to Marilu for me.
Oh, yeah, sure.
All right. Well, take care you guys.
Thank you very much.
You too. Thank you very much, Stephen.
Bye.
Bye.
Bye.
Okay, well, that's it for Episode Five. Please subscribe to the podcast on your favorite podcast app. And if you can, rate and review the show on Apple podcasts. If you'd like to learn more about Andrea and Diego's work, I've provided some relevant links on the podcast website, which is langneurosci.org/podcast. I'd be grateful for any feedback. You can reach me at
[email protected]. I'd like to thank Sam Harvey for assistance with audio engineering and Latan Bullock for editing the transcript of this episode. All right, well, thanks everybody. And I will see you next time. Bye for now.