'Brain Dictionary' Maps Where Specific Words Are Represented In The Mind

It's like Google Maps for your cerebral cortex: A new interactive atlas, developed with the help of such unlikely tools as public radio podcasts and Wikipedia, purports to show which bits of your brain help you understand which types of concepts.
This post was published on the now-closed HuffPost Contributor platform. Contributors control their own work and posted freely to our site. If you need to flag this entry as abusive, send us an email.
a business man reaching towards a head shape figure
a business man reaching towards a head shape figure

By Eric Boodman

It's like Google Maps for your cerebral cortex: A new interactive atlas, developed with the help of such unlikely tools as public radio podcasts and Wikipedia, purports to show which bits of your brain help you understand which types of concepts.

Hear a word relating to family, loss, or the passing of time -- such as "wife," "month," or "remarried"-- and a ridge called the right angular gyrus may be working overtime. Listening to your contractor talking about the design of your new front porch? Thank a pea-sized spot of brain behind your left ear.

The research on the "brain dictionary" has the hallmarks of a big scientific splash: Published on Wednesday in Nature, it's accompanied by both a video and an interactive website where you can click your way from brain region to brain region, seeing what kinds of words are processed in each.

Yet neuroscientists aren't uniformly impressed.

"This is technically very savvy," said David Poeppel, a neuroscientist who studies language at New York University and who was not involved in the study. But he invoked an old metaphor to explain why he isn't convinced by the analysis: He compared it to establishing a theory of how weather works by pointing a video camera out the window for 7 hours.

Indeed, among neuroscientists, the new "comprehensive atlas" of the cerebral cortex is almost as controversial as a historical atlas of the Middle East.

That's because every word has a constellation of meanings and associations -- and it's hard for scientists to agree about how best to study them in the lab.

For this study, neuroscientist Jack Gallant and his team at the University of California, Berkeley played more than two hours' worth of stories from the Moth Radio Hour for seven grad students and postdocs while measuring their cerebral blood flow using functional magnetic resonance imaging.

Then, they linked the activity in some 50,000 pea-sized regions of the cortex to the "meaning" of the words being heard at that moment.

How, you might ask, did they establish the meaning of words? The neuroscientists pulled all the nouns and verbs from the podcasts. With a computer program, they then looked across millions of pages of text to see how often the words from the podcasts are used near 985 common words taken from Wikipedia's List of 1,000 Basic Words.

"Wolf," for instance, would presumably be used more often in proximity to "dog" than to, say, "eggplant."

Using that data, the program assigned numbers that approximated the meaning of each individual word from the podcasts -- and, with some fancy number crunching, they figured out what areas of the brain were activated when their research subjects heard words with certain meanings.

Everyone agrees that the research is innovative in its method. After all, linking up the meanings of thousands of words to the second-by-second brain activity in thousands of tiny brain regions is no mean feat. "That's way more data than any human being can possibly think about," said Gallant.

What they can't agree on is what it means.

"In this study, our goal was not to ask a specific question. Our goal was to map everything so that we can ask questions after that," said Gallant. "One of the most frequent questions we get is, 'What does it mean?' If I gave you a globe, you wouldn't ask what it means, you'd start using it for stuff. You can look for the smallest ocean or how long it will take to get to San Francisco."

This "data-driven approach" still involves assumptions about how to break up language into different categories of meaning. To MIT neuroscientist Evelina Fedorenko, though, it's exciting that Gallant's team has been able to record what's going on in the brain during the real-life situation of listening to stories. "Of course it's a very simplified version of how meaning is captured in our minds, but it seems to be a pretty good proxy," she said.

Gallant agrees that there are still hordes of unanswered questions: "We can map where your brain represents the meaning of a narrative text that is associated with family, but we don't know why the brain is responding to family at that location. Is it the word 'father' itself? Is it your memories of your own father? Is it your own thinking about being a parent yourself?" He hopes that it's just those types of questions that researchers will ask, using his brain map as a guide.

Follow STAT on Facebook: www.facebook.com/statnews

Popular in the Community

Close

What's Hot