TED日本語 - ウリ・ハッソン: コミュニケーション中の脳の反応

Imagine that you invented a device that can record my memories, my dreams, my ideas, and transmit them to your brain. That would be a game-changing technology, right? But in fact, we already possess this device, and it's called human communication system and effective storytelling. To understand how this device works, we have to look into our brains. And we have to formulate the question in a slightly different manner.

Now we have to ask how these neuron patterns in my brain that are associated with my memories and ideas are transmitted into your brains. And we think there are two factors that enable us to communicate. First, your brain is now physically coupled to the sound wave that I'm transmitting to your brain. And second, we developed a common neural protocol that enabled us to communicate.

So how do we know that? In my lab in Princeton, we bring people to the fMRI scanner and we scan their brains while they are either telling or listening to real-life stories. And to give you a sense of the stimulus we are using, let me play 20 seconds from a story that we used, told by a very talented storyteller, Jim O'Grady.

(Audio) Jim O'Grady: So I'm banging out my story and I know it's good, and then I start to make it better --

(Laughter)

by adding an element of embellishment. Reporters call this "making shit up."

(Laughter)

And they recommend against crossing that line. But I had just seen the line crossed between a high-powered dean and assault with a pastry. And I kinda liked it."

Uri Hasson: OK, so now let's look into your brain and see what's happening when you listen to these kinds of stories. And let's start simple -- let's start with one listener and one brain area: the auditory cortex that processes the sounds that come from the ear. And as you can see, in this particular brain area, the responses are going up and down as the story is unfolding. Now we can take these responses and compare them to the responses in other listeners in the same brain area. And we can ask: How similar are the responses across all listeners?

So here you can see five listeners. And we start to scan their brains before the story starts, when they're simply lying in the dark and waiting for the story to begin. As you can see, the brain area is going up and down in each one of them, but the responses are very different, and not in sync. However, immediately as the story is starting, something amazing is happening.

(Audio) JO: So I'm banging out my story and I know it's good, and then I start to make it --

UH: Suddenly, you can see that the responses in all of the subjects lock to the story, and now they are going up and down in a very similar way across all listeners. And in fact, this is exactly what is happening now in your brains when you listen to my sound speaking. We call this effect "neural entrainment." And to explain to you what is neural entrainment, let me first explain what is physical entrainment.

So, we'll look and see five metronomes. Think of these five metronomes as five brains. And similar to the listeners before the story starts, these metronomes are going to click, but they're going to click out of phase.

(Clicking)

Now see what will happen when I connect them together by placing them on these two cylinders.

(Clicking)

Now these two cylinders start to rotate. This rotation vibration is going through the wood and is going to couple all the metronomes together. And now listen to the click.

(Synchronized clicking)

This is what you call physical entrainment. Now let's go back to the brain and ask: What's driving this neural entrainment? Is it simply the sounds that the speaker is producing? Or maybe it's the words. Or maybe it's the meaning that the speaker is trying to convey.

So to test it, we did the following experiment. First, we took the story and played it backwards. And that preserved many of the original auditory features, but removed the meaning. And it sounds something like that.

(Audio) JO: (Unintelligible)

And we flashed colors in the two brains to indicate brain areas that respond very similarly across people. And as you can see, this incoming sound induced entrainment or alignment in all of the brains in auditory cortices that process the sounds, but it didn't spread deeper into the brain.

Now we can take these sounds and build words out of it. So if we take Jim O'Grady and scramble the words, we'll get a list of words.

(Audio) JO: ... an animal ... assorted facts ... and right on ... pie man ... potentially ... my stories

UH: And you can see that these words start to induce alignment in early language areas, but not more than that. Now we can take the words and start to build sentences out of them.

(Audio) JO: And they recommend against crossing that line. He says: "Dear Jim, Good story. Nice details. Didn't she only know about him through me?"

UH: Now you can see that the responses in all the language areas that process the incoming language become aligned or similar across all listeners. However, only when we use the full, engaging, coherent story do the responses spread deeper into the brain into higher-order areas, which include the frontal cortex and the parietal cortex, and make all of them respond very similarly. And we believe that these responses in higher-order areas are induced or become similar across listeners because of the meaning conveyed by the speaker, and not by words or sound. And if we are right, there's a strong prediction over here if I tell you the exact same ideas using two very different sets of words, your brain responses will still be similar.

And to test it, we did the following experiment in my lab. We took the English story and translated it to Russian. Now you have two different sounds and linguistic systems that convey the exact same meaning. And you play the English story to the English listeners and the Russian story to the Russian listeners, and we can compare their responses across the groups. And when we did that, we didn't see responses that are similar in auditory cortices in language, because the language and sound are very different. However, you can see that the responses in high-order areas were still similar across these two groups. We believe this is because they understood the story in a very similar way, as we confirmed, using a test after the story ended.

And we think that this alignment is necessary for communication. For example, as you can tell, I am not a native English speaker. I grew up with another language, and the same might be for many of you in the audience. And still, we can communicate. How come? We think we can communicate because we have this common code that presents meaning.

So far, I've only talked about what's happening in the listener's brain, in your brain, when you're listening to talks. But what's happening in the speaker's brain, in my brain, when I'm speaking to you? To look in the speaker's brain, we asked the speaker to go into the scanner, we scan his brain and then compare his brain responses to the brain responses of the listeners listening to the story. You have to remember that producing speech and comprehending speech are very different processes. Here we're asking: How similar are they? To our surprise, we saw that all these complex patterns within the listeners actually came from the speaker brain. So production and comprehension rely on very similar processes. And we also found the stronger the similarity between the listener's brain and the speaker's brain, the better the communication. So I know that if you are completely confused now, and I do hope that this is not the case, your brain responses are very different than mine. But I also know that if you really understand me now, then your brain ... and your brain ... and your brain are really similar to mine.

Now, let's take all this information together and ask: How can we use it to transmit a memory that I have from my brain to your brains? So we did the following experiment. We let people watch, for the first time in their life, a TV episode from the BBC series "Sherlock," while we scanned their brains. And then we asked them to go back to the scanner and tell the story to another person that never watched the movie. So let's be specific. Think about this exact scene, when Sherlock is entering the cab in London driven by the murderer he is looking for.

With me, as a viewer, there is a specific brain pattern in my brain when I watch it. Now, the exact same pattern, I can reactivate in my brain again by telling the world: Sherlock, London, murderer. And when I'm transmitting these words to your brains now, you have to reconstruct it in your mind. In fact, we see that pattern emerging now in your brains. And we were really surprised to see that the pattern you have now in your brains when I'm describing to you these scenes would be very similar to the pattern I had when I watched this movie a few months ago in the scanner. This starts to tell you about the mechanism by which we can tell stories and transmit information. Because, for example, now you're listening really hard and trying to understand what I'm saying. And I know that it's not easy. But I hope that at one point in the talk we clicked, and you got me. And I think that in a few hours, a few days, a few months, you're going to meet someone at a party, and you're going to tell him about this lecture, and suddenly it will be as if he is standing now here with us. Now you can see how we can take this mechanism and try to transmit memories and knowledge across people, which is wonderful, right?

But our ability to communicate relies on our ability to have common ground. Because, for example, if I'm going to use the British synonym "hackney carriage" instead of "cab," I know that I'm going to be misaligned with most of you in the audience. This alignment depends not only on our ability to understand the basic concept; it also depends on our ability to develop common ground and understanding and shared belief systems. Because we know that in many cases, people understand the exact same story in very different ways.

So to test it in the lab, we did the following experiment. We took a story by J.D. Salinger, in which a husband lost track of his wife in the middle of a party, and he's calling his best friend, asking, "Did you see my wife?" For half of the subjects, we said that the wife was having an affair with the best friend. For the other half, we said that the wife is loyal and the husband is very jealous. This one sentence before the story started was enough to make the brain responses of all the people that believed the wife was having an affair be very similar in these high-order areas and different than the other group. And if one sentence is enough to make your brain similar to people that think like you and very different than people that think differently than you, think how this effect is going to be amplified in real life, when we are all listening to the exact same news item after being exposed day after day after day to different media channels, like Fox News or The New York Times, that give us very different perspectives on reality.

So let me summarize. If everything worked as planned tonight, I used my ability to vocalize sound to be coupled to your brains. And I used this coupling to transmit my brain patterns associated with my memories and ideas into your brains. In this, I start to reveal the hidden neural mechanism by which we communicate. And we know that in the future it will enable us to improve and facilitate communication. But these studies also reveal that communication relies on a common ground. And we have to be really worried as a society if we lose this common ground and our ability to speak with people that are slightly different than us because we let a few very strong media channels take control of the mic, and manipulate and control the way we all think. And I'm not sure how to fix it because I'm only a scientist. But maybe one way to do it is to go back to the more natural way of communication, which is a dialogue, in which it's not only me speaking to you now, but a more natural way of talking, in which I am speaking and I am listening, and together we are trying to come to a common ground and new ideas. Because after all, the people we are coupled to define who we are. And our desire to be coupled to another brain is something very basic that starts at a very early age.

So let me finish with an example from my own private life that I think is a good example of how coupling to other people is really going to define who we are.

This my son Jonathan at a very early age. See how he developed a vocal game together with my wife, only from the desire and pure joy of being coupled to another human being.

(Both vocalizing)

(Laughter)

Now, think how the ability of my son to be coupled to us and other people in his life is going to shape the man he is going to become. And think how you change on a daily basis from the interaction and coupling to other people in your life.

So keep being coupled to other people. Keep spreading your ideas, because the sum of all of us together, coupled, is greater than our parts.

Thank you.

(Applause)