While brain-scanning techniques have enabled researchers to explore which regions are active when using language, most subjects in these studies have spoken English or one of just a handful of other languages—and it’s been unclear whether the findings also applied to other languages. Recently, researchers evaluated the brain activity in native speakers of 45 different languages to determine whether their language networks—specific brain regions that specialize in processing linguistic information—behaved similarly. The analysis, published Monday (July 18) in Nature Neuroscience, finds that these distinctive languages do indeed involve similar patterns of brain activity.
The authors write in their study that out of the more than 7,000 languages that humans across the globe use to communicate, research has largely focused on a single language family—the Indo-European family, and in particular, English. The new work covered 45 languages from twelve language families and evaluated the brain activity of two native speakers (one male and one female) from each language included in the study, which represented a more comprehensive survey than previously examined. During testing, fMRI data were collected while each person performed specific linguistic or nonlinguistic tasks.
For more about the study and its results, The Scientist spoke to coauthor Saima Malik-Moraleda, a cognitive neuroscientist at MIT.
The Scientist: What is the language network?
Saima Malik-Moraleda: The language network is a set of regions that cover frontal and parietal areas of your brain. We have seen that there are six main regions that tend to respond to language and [that] are selective to language, which means they do not respond to other stimuli . . . such as math or spatial working memory tasks.
TS: Can you can you explain the properties of the language network that you specifically studied?
SM-M: We looked at selectivity, which means “Does [a region] respond to language but not respond to other tasks that are not linguistic?” We looked at lateralization: The language network tends to be left-lateralized. We looked at functional integration: The language network tends to be highly integrated within itself, and highly dissociated [from] other networks in the brain, such as the multiple demand network, which is a network that takes care of executive functioning, [for example, performing] a spatial working memory task.
TS: You and your coauthors write that these [properties] haven’t been studied extensively beyond English speakers. Why is it important to include people who speak a variety of languages?
SM-M: Well, English is not a prototypical language. [Certain] properties of English are not found in other languages. For instance, English has a very strict word order, [while] there’s a lot of languages that have a different word order or have a free word order.
English is not a prototypical language.
I would also like to add that one of the points of this study is . . . not just to be able to study the language network across different languages, but also so other labs around the world can use what we call functional localizers—so any researcher [can] track which areas of the brain respond to language. The other reason to study different languages is to be able to make localizers more accessible to the language community if possible.
TS: Can you explain what you mean by localizers?
SM-M: Language localizers are a way to find the chunks of your brain that respond to . . . language. And broadly, these areas tend to be similar across participants. As I said, there [are] six areas in the brain that tend to be responsive to language, and they’re in the left hemisphere. However, the exact boundaries of these areas are different between you and I. When you’re trying to see [different responses] . . . you want to make sure that you are localizing first those language areas in the brain and then looking within those areas [to see] what is happening. Otherwise the [adjacent] networks in the brain that are not responsive to language [can blur] the signal.
TS: How did you evaluate these parameters in the participants from your study?
SM-M: We had participants come in and listen to a story. In this case, it was Alice in Wonderland, one of the most translated books in the world besides the Bible. And we had them come in for a functional MRI scan [while] they were listening to parts of the story [in their native language] and then acoustically degraded versions. The way you can visualize these areas in the brain is by creating a contrast between the intact version and the degraded version. [P]articipants . . . also performed [nonlanguage tasks such as a] spatial working memory task and a math task that would allow us to assess the selectivity of the area.
TS: And then you compared [results] across different participants who spoke different native languages. What was the main conclusion from that analysis?
SM-M: The main conclusion is that the properties that we were looking at—whether [language] was left-lateralized, whether language was selective, whether it was functionally integrated within [the] network and dissociated with other networks—those properties hold across all the languages that we looked at. And there’s always interindividual variability, which is the reason why we want to have localizers to account for variability across individuals. But the variability that we saw across languages was lower than the variability that we see across participants, meaning that the language network seems to be incredibly stable and similar across languages.
If we did a study with more participants per language, maybe we’d be able to find more nuanced distinctions within particular areas of the language network.
Now, I would want to point out that . . . because we’re looking at very broad properties of the brain, we only scanned two people per language. If we did a study with more participants per language, maybe we’d be able to find more nuanced distinctions within particular areas of the language network.
TS: What might more extensive sampling in future studies reveal in terms of those nuances? In other words, why are those nuances important?
SM-M: One of the questions that cognitive neuroscientists who particularly study language wonder is: “Why do we have six areas? What does each area do?” We haven’t quite figured out the function of each separate area. We know these areas are very functionally integrated and work closely together, but there’s a lot of debate as to what the function of each different area is. So, potentially, if we can leverage the variability within languages, maybe that would be a way where scanning more participants would allow us to understand even more about how the language network functions.
Editor’s note: This interview has been edited for brevity.