The power of learning a second language: look to the caudate

Feb 16 2011 Published by under Behavioral Neuro, Neuroscience, Uncategorized

Sci's terrible at languages. TERRIBLE. In my time, I've successfully mastered English, and attempted to master four other languages (five if you count a brief foray into Elvish when I was 15, but that doesn't really count) in my time. I have failed at ALL of them. Every once in a while I would achieve some semblance of competency, but there's no doubt that either I didn't start early enough, didn't try hard enough (two semesters of immersion definitely weren't enough), or there's only one language for this geek.

But what about people who DO learn more than one language? For those who learn as adults, it's often an extremely hard won skill. For those who learn from early childhood to mid adolescence, it's often a little easier. But no matter when it happens, it's an impressive skill. Those learning a second language will master additional tens of THOUSANDS of words along with the ones from their first language.

Two languages, but you've only got one brain. And that brain only have one major language circuit. Right now, scientists think that people who are bilingual use processing in their major language circuit to monitor and to control their FIRST language, so it doesn't butt in on their second (or third, or fourth). We know that some people are better at learning a second language than others. But what makes them different? Is it a different in your anatomy (which could depend on several factors, including genes, environment, development, etc), or is it just training? And then, can training influence your anatomy?

Tan et al. "Activity levels in the left hemisphere caudate–fusiform circuit predict how well a second
language will be learned" PNAS, 2011.

Reading in your second language is a complex task that requires more than you might think. If you are not fluent, often the word has to pass through your first language and get translated in your brain. You begin to have voluntary control over the second language. But as you become better, you keep your voluntary control over speaking your second language, and ADD the inhibitory control over your FIRST language, so that you aren't producing words in two languages at once. But where and how does this happen? And what makes some people better at it than others?

To look at the anatomy behind the learning of a second language, the scientists took a group of 10 year old kids. The kids were all Chinese, and in the same school system, being taught English as a second language at the same level of difficulty. They had all started learning English at age 6, and weren't yet proficient. They looked at them once, and then looked at them again one year later. Each time, the kids were given a reading test in English, and then they were placed in an MRI. While they were getting scanned, the kids were given English real words (like bucket) or English nonsense words (like guge), and had to identify what was a real English word and what was not. One year later, they were given the same test, at the reading level they should have reached one year later.

Obviously they found activity in a bunch of brain regions implicated in language processing. Areas like the inferior frontal gyrus (containing Broca's area), lingual gyrus (which is actually called that because it's apparently shaped like a tongue. I don't see it, but whatever. Also it helps in recognition of words), and parahippocampal gyrus.

But what they wanted to know what whether any of these regions were correlated with the reading ability of the kids, and more importantly, with how much they had progressed after a year. And two of them were: the caudate and the fusiform gyrus.

You can see here the two areas that showed higher activity correlated with better second language reading skills. This was at time 1. One year later, they brought them back, and the correlation persisted:

You can see there that the activation in the caudate was actually stronger at the year mark. Here, the activity level also reflects how much the kids had improved in their second language knowledge. This activation difference (higher activation and better reading skills) wasn't there when the students were working in their primary language.

What this shows is that people who have high activity in these brain areas MAY be better able to learn a second language than those who have less activity. And this ALSO means that the activity in the brain when reading in a second language is different from that of reading in your primary language.

Caveat emptor, obviously. We don't know if maybe some of those kids were getting outside help on their language skills before the test began. It's possible that the activation of the fusiform and caudate correlates with skills in a second language, but can't PREDICT how well you do in a second language. In order to really see that, they would have to scan the kids before they learned a second language at all, and see what the activation pattern looked like then, and look again over the time they learned the language (and I bet you someone is on that study already). It would also be interesting to see if this is something that some people have, or can it be the result of training?

But what I really like about this study is that it shows that learning a second language really does require an alternate path through the brain, correlating only with that second language, not with the primary language. It would be REALLY interesting to see how brain activation changed as people become FLUENT in their second language. Does it always take this different processing path through the brain? Or as you become fluent, do the two paths begin to appear more similar?

If this does turn out to be predictive of the ability to learn a second language, it might be useful in knowing who might need extra help when they need to become bilingual. I can tell you I'm probably lacking some extra activity in my fusiform and caudate. My continuing attempts to learn French may need more training, or I may just be stuck monolingual. 🙁

Tan LH, Chen L, Yip V, Chan AH, Yang J, Gao JH, & Siok WT (2011). Activity levels in the left hemisphere caudate-fusiform circuit predict how well a second language will be learned. Proceedings of the National Academy of Sciences of the United States of America, 108 (6), 2540-4 PMID: 21262807

27 responses so far

  • Janne says:

    What about more languages? I speak four languages, two that I learned at the same time as an infant, one during adolescence and adulthood and one only as an adult. Do those first two languages share the same pathways or does my non-primary language use the above second-language pathways? Do the third and fourth languages share the second-language pathways or do all languages carve out their own bits and pieces in my brain?

  • Not having read the paper, it seems like there's a confound here since Chinese is logographic and English is alphabetic. (But knowing the research group, I assume they must account for this?) The primary language circuits are somewhat different if you're raised speaking a logographic language than an alphabetic language, making the role of the fusiform and caudate unclear: do they have to do with learning to read in a second language, per se, or do they have to do with learning to read in an alphabetic language in the first place?

    • scicurious says:

      Yeah, I thought they would control for that...but it looks like they may NOT have. Or at least, it's not mentioned in the body of the paper. Unfortunately, my knowledge of this subject isn't good enough to speculate on HOW they might have controlled for something like this, do you know?

      OTOH, we know that the caudate has been shown before to be involved in French and German, and in making the "switch" from one language to another. So maybe they just took it as given, but what if Chinese involves the caudate LESS? It's a really good question.

      • Well, its certainly possible that while the Chinese and English left-hemisphere reading systems are different, they still could share the fusiform and/or the caudate. I know for certain there are anterior differences, but its possible that the posterior portion of the reading network is preserved across the two languages. So that could - at least partially - allow for a more direct comparison.

        • Sonia Wu says:

          I started learning English at age 6 in Taiwan as a native Taiwanese and Mandarin speakers. I am perfectly proficient in three, but English is obviously my weakest out of the three. I've been in the US for 12 yrs, and I sound really good compared to many who have lived here for decades. I have a very subtle accent, but it's pretty and not unpleasant. It only comes out when I am either emotional or physically uncomfortable. I found that I have to get in the "mindset" when I speak one of the languages. They are so different than one another. Switching back and forth daily is something I'm very used to. It took me some time to do it well. I am now teaching Mandarin to American children. This area of study is fascinating. I actually am a very effective Mandarin teacher because of my experience with language learning with both Mandarin and English. So nice to read your article here.

  • HP says:

    The thing is, failing to learn a language is still a great experience, and teaches you so much about how to communicate. Like you, I've failed to learn at least four languages, but each time I fail, I learn something about humans.

    I've attempted and failed German, Spanish, French, Dutch, and Portuguese.

    If you haven't yet tried and failed to learn Portuguese, I highly recommend it. It's one of the most popular languages on the Internet (after English, Mandarin, and Persian, IIRC), so there's plenty of it to read. If you think of it as French phonology mixed with Archaic Spanish vocabulary, you can get about 40 percent right off the bat. And there are shitloads of Brazilian and Portuguese scientists out there blogging. Not to mention pop culture, skepticism, music, etc.

    Next up: Either Mandarin or Japanese. I will fail spectacularly, but I'll learn something.

  • Mountainmums says:

    I haven't taken the time to read the article, but I agree with Jason's comment. The left fusiform gyrus is considered an important area in reading and reading development (Dehaene's Visual Word Form Area). Activity in this area is routinely associated with reading skill in alphabetic languages. To me it is quite possible that this increase is solely due to their increased skills in reading English. The increase would not appear for the Mandarin reading because of a putative ceiling effect.
    Again, I haven't read the article (which I'll try to do), but I am surprised that they have associated fusiform activity more strongly to langague learning than to reading.

  • Marianna says:

    I was born in a Latin American country, and thus, raised with Spanish as my native language. However, I went to a bilingual school where I was taught English as well, from a very young age, until I graduated. I consider myself fluent in both languages, and, in the first years of achieving fluency, I've had a strange thing happen to me while reading anything in English or Spanish. When I am, often, it takes me a few minutes to determine whether what I am reading is either English or Spanish, though it may be obvious to anyone else. Maybe this is because, since I became fluent in English, I don't translate words any longer and just acquire knowledge of them as concepts? It's a little weird, but mostly pleasant. Maybe someone should place me in an MRI, I wouldn't mind!

    • HP says:

      Marianna, how do you do with Portuguese? I have a friend from Buenos Aires who's bilingual (Argentine Spanish and American English), but totally baffled by Brazilians. Whereas I find my Romance language skills infinitely transferable, and my tiny comprehension extends to not only to Portuguese but Romansch, Neopolitan, Catalan, etc.

      • Marianna says:

        I can understand some, but I pretty much suck at Portuguese! I actually have a better time understanding Italian and French, but all of these are pretty easy to understand if I am reading them. The problem would be listening.

        Oh, Catalan is very cool! I am envious.

  • Connie says:

    I grew up with one language (German), and learned English and French at School (first English, than French). Very quickly, I was fluent in English (on student exchange, the teachers in England did think I was English). In my twenties, I moved to South Africa, where I started to learn a fourth language - Afrikaans, and become even more proficient at English (I consider myself to be bi-lingual now). Funnily enough, when I started to learn Afrikaans, it displaced my French - I tried to speak to a French person once, and all that I could think of were Afrikaans words. Now, nearly fifteen years later, I have moved to France, and try to resurrect a very dormant French (which was kind of fluent 20 years ago). The first thing I lost is my knowledge of Afrikaans - composing Afrikaans letters is getting very hard, as I now remember the French words.

    I always think that German and English have their "own" areas in my brain, and that the third and any other consecutive language are being stored in the same place, leading to those suppositions. So what would the MRI show in my case?

    My daughter (who is now 2 1/2) was bi-lingual (German/English) until a year ago. Now she is growing up with German and French as her two main languages - I wonder how this would show up?

  • SamW says:

    This is really cool.
    I agree with what Marianna said - I am bilingual and it also sometimes takes me a while to figure out what language I'm currently acting in.
    Also, if someone suddenly speaks to me in the other language (most commonly, an English person trying to show of their German), it takes me several seconds to comprehend. I cannot instantly switch. (Unless, I'm expecting it...) It would be interesting to see if that means that the two languages actually occupy the same area of the brain (and thus only one can be efficiently used at a time).

  • Katharine says:

    I'm not Romanian at all, and yet it's the second language I have the most fluency in.

    You sort of don't expect how it's descended from Latin, but it is, and it has a heavy dose of Slavic influence. It is almost nothing like the other Latin languages. It tells you how many empires have run through Romania. (They've been battered.)

    Italian is not hard for me to decipher.

    I am totally useless at reading French.

    • Katharine says:

      What's a major bitch, of course, is the morphological changes: you get this wacky-ass change from



      esti (pronounced something similar to 'yesht')
      este (pronounced 'yeste' or in its short form, 'e', 'ye'. SLAVIC INFLUENCE LOL.)
      sunteti (pronounced something similar to 'suntets')


    • Katharine says:

      Also, have some adorable Romanian children singing about communism:

  • tbell1 says:

    The literature on late learners of a second language differentiates pretty clearly between mastery of vocabulary and grammatical fluency. Vocabulary, even foreign, doesn't not appear to have any critical period, and indeed we add to our native vocabulary constantly. It's grammatical fluency that late learners have a problem with.

    This makes me very suspicious about interpreting the above study. Sure, there may be several reasons that people have trouble learning a second language late in life, but a vocab based assessment (lexical decision) seems like it misses the mark pretty badly here in terms of what predicts foreign language success.

    • Scicurious says:

      Actually, the lexical assessment was just the test IN the MRI, they got a full reading fluency test which would presumably involve grammar prior, and it was this fluency test which they correlated with the brain activation during the lexical task. But the point you make is still valid.

  • ryandake says:

    hey, elvish totally counts! besides, you speak geek.

    and HP, beware Japanese... i totally failed at learning it but got sucked irrevocably into Japanese culture. permanent case of japanophile, now.

  • This is really cool. Do they know how these brain areas change over time in kids who are not learning a second language?

    Also, does this apply learning non-language languages, like a computer language? I'm trying to learn Perl right now, and I could really do with some more activity in my caudate-fusiform circuit...

  • Beth says:

    I started learning a second language when I was 15. (I've no idea when "mid-adolescence" is.) I wouldn't call myself fluent, but I stopped having to mentally translate after the third semester. I'm not at all surprised a second language uses different paths in the brain, because I think I've experienced it. I have lesions which occasionally leave me unable to speak (and even think in) my primary language while my second language remains unimpeded. I figure that surely means there are different pathways for the first and second language. (I can't tell you where the lesion or lesions are that cause the trouble because I have too many.) I'm very glad that I didn't have much trouble learning my second language as well as I did and that my brain gets to the languages differently: aphasia is terribly frustrating!

  • anne says:

    Very interesting! My experience is similar to Connie's in that I feel like I have two separate spaces in my brain for my first and second languages (English and German), and any attempts to learn a third get mixed into, or displace, the second. When I try to speak Spanish, for instance, and I can't think of the word for something, I revert to German, rather than English. So it's like I have a place for my mother tongue, and a place for the "other".

  • Renata says:

    Ive been reading an interesting book by Norman Doige MD, who states that the more we use our native language the more it dominates our linguistic area, and therefore its hard to learn a new language as we age, but as a child learning there is less competition and the languages share a large area in the brain.
    So it seems that as you age the dominant language would be the one you use more easily, as the brain has devoted more area to this language than any other with stronger pathways .

  • [...] The power of learning a second language: look to the caudate: “Right now, scientists think that people who are bilingual use processing in their major language circuit to monitor and to control their FIRST language, so it doesn’t butt in on their second (or third, or fourth). We know that some people are better at learning a second language than others. But what makes them different?” [...]

  • Qin says:

    I found that most interesting thing is in which language you count numbers. Even though I'm fluent in both Mandarin and English, I most of the time count numbers in Chinese as it's my first language but sometimes I also do count and calculate in English but it won't get far. I'd really wanna do a MRI test and see how my brain works...

  • Randy says:

    Similar experience with Norwegian ( mother tongue) and English

  • :Lynn Porter says:

    I never mastered english grammar. I falied the language test at Penn State in the late 50
    s . I took spanish in High Scdhool, Latin in college and was hopless. Then French and struggled with a d. I went to language lab and thought I was trying. Primcipally because of my failure in Latin and French i dropped out of Penn State. I tried to qualify for the Army Language school but did not make the grade. The test was in Esperanto and I thought I did ok. I did not and ended up as a clerk typist.

    I was stationed in Japan and worked very hard at learning the language. I spoke japan-es at every opportunity and was able to get along OK. I gainsay that I was more fluent than any of my cohorts in Japan. I worked very hard for the two years I was there. I had many Japanese friends and a couple of nonenglish speaking girlfriend.

    It has been like beating my head against a wall. I feel this is one of my major life failures.

Leave a Reply