Tag Archives: language

does functional specialization exist in the language system?

One of the central questions in cognitive neuroscience–according to some people, at least–is how selective different chunks of cortex are for specific cognitive functions. The paradigmatic examples of functional selectivity are pretty much all located in sensory cortical regions or adjacent association cortices. For instance, the fusiform face area (FFA), is so named because it (allegedly) responds selectively to faces but not to other stimuli. Other regions with varying selectivity profiles are similarly named: the visual word form area (VWFA), parahippocampal place area (PPA), extrastriate body area (EBA), and so on.

In a recent review paper, Fedorenko and Kanwisher (2009) sought to apply insights from the study of functionally selective visual regions to the study of language. They posed the following question with respect to the neuroimaging of language in the title of their paper: Why hasn’t a clearer picture emerged? And they gave the following answer: it’s because brains differ from one another, stupid.

Admittedly, I’m paraphrasing; they don’t use exactly those words. But the basic point they make is that it’s difficult to identify functionally selective regions when you’re averaging over a bunch of very different brains. And the solution they propose–again, imported from the study of visual areas–is to identify potentially selective language regions-of-interest (ROIs) on a subject-specific basis rather than relying on group-level analyses.

The Fedorenko and Kanwisher paper apparently didn’t please Greg Hickok of Talking Brains, who’s done a lot of very elegant work on the neurobiology of language.  A summary of Hickok’s take:

What I found a bit on the irritating side though was the extremely dim and distressingly myopic view of progress in the field of the neural basis of language.

He objects to Fedorenko and Kanwisher on several grounds, and the post is well worth reading. But since I’m very lazy tired, I’ll just summarize his points as follows:

  • There’s more functional specialization in the language system than F&K give the field credit for
  • The use of subject-specific analyses in the domain of language isn’t new, and many researchers (including Hickok) have used procedures similar to those F&K recommend in the past
  • Functional selectivity is not necessarily a criterion we should care about all that much anyway

As you might expect, F&K disagree with Hickok on these points, and Hickok was kind enough to post their response. He then responded to their response in the comments (which are also worth reading), which in turn spawned a back-and-forth with F&K, a cameo by Brad Buchsbaum (who posted his own excellent thoughts on the matter here), and eventually, an intervention by a team of professional arbitrators. Okay, I made that last bit up; it was a very civil disagreement, and is exactly what scientific debates on the internet should look like, in my opinion.

Anyway, rather than revisit the entire thread, which you can read for yourself, I’ll just summarize my thoughts:

  • On the whole, I think my view lines up pretty closely with Hickok’s and Buchsbaum’s. Although I’m very far from an expert on the neurobiology of language (is there a word in English for someone’s who’s the diametric opposite of an expert–i.e., someone who consistently and confidently asserts exactly the wrong thing? Cause that’s what I am), I agree with Hickok’s argument that the temporal poles show a response profile that looks suspiciously like sentence- or narrative-specific processing (I have a paper on the neural mechanisms of narrative comprehension that supports that claim to some extent), and think F&K’s review of the literature is probably not as balanced as it could have been.
  • More generally, I agree with Hickok that demonstrating functional specialization isn’t necessarily that important to the study of language (or most other domains). This seems to be a major point of contention for F&K, but I don’t think they make a very strong case for their view. They suggest that they “are not sure what other goals (besides understanding a region’s computations) could drive studies aimed at understanding how functionally specialized a region is,” which I think is reasonable, but affirms the consequent. Hickok isn’t saying there’s no reason to search for functional specialization in the F&K sense; as I read him, he’s simply saying that you can study the nature of neural computation in lots of interesting ways that don’t require you to demonstrate functional specialization to the degree F&K seem to require. Seems hard to disagree with that.
  • Buchsbaum points out that it’s questionable whether there are any brain regions that meet the criteria F&K set out for functional specialization–namely that “A brain region R is specialized for cognitive function x if this region (i) is engaged in tasks that rely on cognitive function x, and (ii) is not engaged in tasks that do not rely on cognitive function x.Buchsbaum and Hickok both point out that the two examples F&K give of putatively specialized regions (the FFA and the temporo-parietal junction, which some people believe is selectively involved in theory of mind) are hardly uncontroversial. Plenty of people have argued that the FFA isn’t really selective to faces, and even more people have argued that the TPJ isn’t selective to theory of mind. As far as I can tell, F&K don’t really address this issue in the comments. They do refer to a recent paper of Kanwisher’s that discusses the evidence for functional specificity in the FFA, but I’m not sure the argument made in that paper is itself uncontroversial, and in any case, Kanwisher does concede that there’s good evidence for at least some representation of non-preferred stimuli (i.e., non-faces in the FFA). In any case, the central question here is whether or not F&K really unequivocally believe that FFA and TPJ aren’t engaged by any tasks that don’t involve face or theory of mind processing. If not, then it’s unfair to demand or expect the same of regions implicated in language.
  • Although I think there’s a good deal to be said for subject-specific analyses, I’m not as sanguine as F&K that a subject-specific approach offers a remedy to the problems that they perceive afflict the study of the neural mechanisms of language. While there’s no denying that group analyses suffer from a number of limitations, subject-specific analyses have their own weaknesses, which F&K don’t really mention in their paper. One is that such analyses typically require the assumption that two clusters located in slightly different places for different subjects must be carrying out the same cognitive operations if they respond similarly to a localizer task. That’s a very strong assumption for which there’s very little evidence (at least in the language domain)–especially because the localizer task F&K promote in this paper involves a rather strong manipulation that may confound several different aspects of language processing.
    Another problem is that it’s not at all obvious how you determine which regions are the “same” (in their 2010 paper, F&K argue for an algorithmic parcellation approach, but the fact that you get sensible-looking results is no guarantee that your parcellation actually reflects meaningful functional divisions in individual subjects). And yet another is that serious statistical problems can arise in cases where one or more subjects fail to show activation in a putative region (which is generally the norm rather than the exception). Say you have 25 subjects in your sample, and 7 don’t show activation anywhere in a region that can broadly be called Broca’s area. What do you do? You can’t just throw those subjects out of the analysis, because that would grossly and misleadingly inflate your effect sizes. Conversely, you can’t just identify any old region that does activate and lump it in with the regions identified in all the other subjects. This is a very serious problem, but it’s one that group analyses, for all their weaknesses, don’t have to contend with.

Disagreements aside, I think it’s really great to see serious scientific discussion taking place in this type of forum. In principle, this is the kind of debate that should be resolved (or not) in the peer-reviewed literature; in practice, peer review is slow, writing full-blown articles takes time, and journal space is limited. So I think blogs have a really important role to play in scientific communication, and frankly, I envy Hickok and Poeppel for the excellent discussion they consistently manage to stimulate over at Talking Brains!

estimating bias in text with Ruby

Over the past couple of months, I’ve been working on and off on a collaboration with my good friend Nick Holtzman and some other folks that focuses on ways to automatically extract bias from text using a vector space model. The paper is still in progress, so I won’t give much away here, except to say that Nick’s figured out what I think is a pretty clever way to show that, yes, Fox likes Republicans more than Democrats, and MSNBC likes Democrats more than Republicans. It’s not meant to be a surprising result, but simply a nice validation of the underlying method, which can be flexibly applied to all sorts of interesting questions.

The model we’re using is a simplified variant of Jones and Mewhort’s (2007) BEAGLE model. Essentially, similarity between words is quantified by looking at the degree to which words have similar co-occurrence patterns with other words. This basic idea is actually common to pretty much all vector space models, so in that sense, there’s not much new here (there’s plenty that’s new in Jones and Mewhort (2007), but we’re mostly leaving those features out for the sake of simplicity and computational speed). The novel aspect is the contrast coding of similarity terms in order to produce bias estimates. But you’ll have to wait for the paper to read more about that.

In the meantime, one thing we’ve tried to do is develop software that can be used to easily implement the kind of analyses we describe in the paper. With plenty of input from Nick and Mike Jones, I’ve written a set of tools in Ruby that’s now freely available for download here. The tools are actually bundled as a Ruby gem, so installation should be a snap on most platforms. We’re still working on documentation, so there’s no full-blown manual yet, but the quick-start guide should be sufficient to get many users up and running. And for people who share my love of Ruby and are interested in using the tools programmatically, there’s a fairly well-commented RDoc.

The code should really be considered an alpha release at the moment; I’m sure there are plenty of bugs (if you find any, email me!), and the feature set is currently pretty limited. Hopefully it’ll grow over time. I also plan to throw the code up on GitHub at some point in the near future so that anyone who’s interested can help out with the development. In the meantime, if you’re interested in semantic space models and want to play around with a crude (but relatively fast) implementation of one, there’s a (very) small chance you might find these tools useful.

why do we sing up?

While singing loudly to myself in the car the other day (windows safely rolled up, of course–I don’t want no funny looks from pedestrians), I noticed that the first few notes of the vocal melody of most songs seem to go up rather than down. That’s to say, the first pitch change in most songs seems to be from a lower note to a higher note; there don’t seem to be very many that show the opposite pattern. It actually took me a while to find a song that goes down at the beginning (Elliott Smith‘s Angeles–incidentally also my favorite song—which hangs on B for the first few bars before dropping to A); the first eight or nine I tried all went up. After carefully inspecting millions thousands hundreds several more songs on my drive home, I established that only around 10% of vocal melodies dip down initially (95% confidence interval = 0 – 80%); the rest all go up.

When I got home I did a slightly more systematic but still totally ascientific analysis. I write songs occasionally, so I went through them all, and found that only three or four go down; the rest all go up. But that could just be me. So then I went through a few of my favorite albums (not a random sample–I picked ones I knew well enough to rehearse mentally) and found the same pattern. I don’t know if this is a function of the genre of music I listen to (which I’d charitably describe as wuss-rock) or a general feature of most music, but it seems odd. Not having any musical training or talent, I’m not really sure why that would be, but I’d like to know. Does it have anything to do with the fact that most common chord progressions go up the scale initially? Is there some biological reason we find ascending notes more pleasant at the beginning of a melody? Is it a function of our speech production system? Does Broca’s Area just like going up more than going down? Is it just an arbitrary matter of convention, instilled in songwriters everywhere by all the upward-bound music that came before? And is the initial rise specific to English, or does it happen when people sing in other languages as well? Could it be something about the emotional connotation of rises versus drops? Do drops seem too depressing to kick off a song with? Are evil clowns behind it all? Or am I just imagining the whole thing, and there isn’t actually any bias toward initial upness?

Can we get an update on this? Are there any musicians/musicologists in the house?