How Do We Know What Other People Are Thinking?

Feb 20, 2015
Originally published on December 8, 2017 7:57 am

Part 4 of the TED Radio Hour episode The Unknown Brain

About Rebecca Saxe's TED Talk

Sensing the motives and feelings of others is a natural talent for humans. But how do we do it? Neuroscientist Rebecca Saxe explains how one region in the brain focuses on other people's thoughts.

About Rebecca Saxe

Rebecca Saxe is a cognitive neuroscientist at MIT. While a graduate student, Saxe made a discovery: There's a specific region in our brain that becomes active when we contemplate the workings of other minds. Now, the head of MIT's Saxelab, she and her team are delving deeper into that brain region, hoping it might reveal clues about conditions such as autism.

Copyright 2017 NPR. To see more, visit


So we know the brain is this complex collection of 86 billion neurons. But until pretty recently, scientists weren't really sure how those neurons worked together. Was the brain like a utility knife, one tool that did many different things? Or was it more like a Swiss Army knife with specific mechanisms for specific tasks? So take one specific task, for example, one that your brain performs pretty much every single day - and that is recognizing someone else's face.

NANCY KANWISHER: Well, there's a bunch of different things that happen when you see a face.

RAZ: This is Nancy Kanwisher.

KANWISHER: I'm a cognitive neuroscientist at MIT.

RAZ: And just think, she says, about all the different variables your brain has to grapple with when you look at another person.

KANWISHER: Right, like, every time you see a person, they look different, from different viewpoints - a profile and a front view are totally different - from the way their hair falls on their face, the expression on their face, the lighting.

RAZ: What about, you know, like when you see a face and you know it but you can't place it?

KANWISHER: Right, you get a signal like, I know that person.

RAZ: Yeah.

KANWISHER: But that's a different thing than which particular person is it in terms of, what else do I know about them; where did I meet them? And that in turn is a different thing than what is their name.

RAZ: And the thing is, it's possible for your brain to get hung up anywhere along that chain.


UNIDENTIFIED MAN #1: Just look at my face and tell me what happens when I do this, all right?

RAZ: So this is a video that Nancy pointed us to...

KANWISHER: An incredible video.

RAZ: ...That was made by some colleagues of hers.

KANWISHER: Josef Parvizi and Kalanit Grill-Spector.

RAZ: Two neurologists who were treating a man with epilepsy. And they wanted to find the source of his seizures, so they placed electrodes on the surface of his brain. But by chance, what they found was a pulse from those electrodes...




RAZ: ...In just the right place...



RAZ: ...Had an unintended effect on their patient.


UNIDENTIFIED MAN #1: I'm going to do it one more time.

KANWISHER: And the guy was looking at the surgeon when they stimulated that region.


UNIDENTIFIED MAN #1: Look at my face. One, two, three.

KANWISHER: And he said...


UNIDENTIFIED MAN #2: You just turned in to somebody else.


RAZ: The doctor's face instantly became unrecognizable.


UNIDENTIFIED MAN #2: Your nose got saggy, went to the left. You almost like somebody I had seen before, but somebody different. That was a trip.

RAZ: The effect of that pulse had accidentally recreated a condition that actually exists in real life, a condition where you can't remember faces. It's called face blindness...

KANWISHER: That's right.

RAZ: ...Or prosopagnosia.

KANWISHER: So prosopagnosia, which has been known for a long time, is that people can selectively lose their ability to recognize faces.

RAZ: How does that happen?

KANWISHER: They have brain damage usually from a stroke or from a - you know, a physical injury to the brain. And if it's very, very focal in just a tiny part of the brain, then you are perfectly normal at absolutely everything else, but you can't recognize faces.

RAZ: Which raised a big question for Nancy and her colleagues at MIT. Here's her TED Talk.


KANWISHER: We wanted to know if there was a special part of the brain for recognizing faces. So I was the first subject. I went into the scanner. I lay on my back. I held my head as still as I could while staring at pictures of faces and objects and faces and objects for hours. So as somebody who has pretty close to the world record of total number of hours spent inside an MRI scanner, I can tell you that one of the skills that's important for MRI research is bladder control.


KANWISHER: When I got out of the scanner, I did a quick analysis of the data, looking for any parts of my brain that produced a higher response when I was looking at faces than when I was looking at objects. And here's what I saw. That region right there, that little blob, it's about the size of an olive, and it's on the bottom surface of my brain about an inch straight in, OK? And what that part of my brain is doing is producing a higher MRI response, that is higher neural activity, when I was looking at faces than when I was looking at objects. So that's pretty cool. But how do we know this isn't a fluke? Well, the easiest way is to just do the experiment again.

RAZ: So you have your brain scanned, and then, like, the part that's responding to facial recognition is, like, lighting up. And you do this, like, several, several...

KANWISHER: Oh, I've done it hundreds of times. I like it in there. (Laughter).

RAZ: What's it like?

KANWISHER: Oh, it's peaceful. I go in there, and I think, oh, we're going to get this great data. I have these wonderful students. I can't wait to see what my brain is doing when I do this task, so let's find the face region. And we found it right away. I remember I was just so excited. I remember like a week later it was like, oh, let's try it again. I can remember running out to this - we have this computer right outside the scanner. I was, like, do a quick analysis. I was like, oh, it's not going to still be there. It's not going to still be there. Oh, there it is.

RAZ: Wow.

KANWISHER: We did that, like, 10 times before we kind of really believed that the thing was for real.

RAZ: Like, when we talk about regions of the brain that light up...


RAZ: ...We're still talking about an area that encompasses billions and billions of neurons and trillions of synaptic connections, right?

KANWISHER: That's right. So the smallest unit in my data is usually called a voxel. It's like a three-dimensional pixel. It's a teeny, little cube that's 2 or 3 millimeters on a side. And so it's either - you know, it has some magnitude of response, some brightness, that's our basic unit. But that teeny, little voxel has half a million neurons in it.

RAZ: Wow.

KANWISHER: So that just tells you a lot of the stuff we can't see. We see this just drastically blurred version of the actual neural code.


RAZ: Now, to try and get a better grasp of what that blurry image means, Nancy and other scientists are asking why is it that some tasks don't seem to have their own special real estate in the brain when other tasks, like recognizing faces or processing colors or even identifying body parts, do?


KANWISHER: Do we also have specialized brain regions for other senses, like hearing? Yes, we do. Here's a region that we reported just a couple of months ago, and this region responds strongly when you hear sounds with pitch, like these...


KANWISHER: OK. In contrast, that same region does not respond strongly when you hear perfectly familiar sounds that don't have a clear pitch, like these...


KANWISHER: OK. What's important to me about this work is not the particular locations of these brain regions, but the simple fact that we have selective, specific components of mind and brain in the first place. I mean, it could have been otherwise. The brain could have been a single general-purpose processor, more like a kitchen life than a Swiss Army knife. Instead, what brain imaging has delivered is this rich and interesting picture of the human mind. It's early days in this enterprise. The most fundamental questions remain unanswered. How are all these things connected in the brain? How does all of this very systematic structure get built, both over development in childhood and over the evolution of our species? This is, I think, the greatest scientific quest of all time.

RAZ: I mean, there's some process that is happening right now as I'm speaking to you and you're speaking to me and we're stringing these words together and having a conversation. And it's just...

KANWISHER: It's the damnedest thing, isn't it?

RAZ: Yeah.

KANWISHER: Yeah. I mean, that whole question of how you take a thought and a bunch of sounds come out your mouth...

RAZ: Right.

KANWISHER: Like, what?

RAZ: Yeah.

KANWISHER: How does that ever happen?

RAZ: How does it happen?

KANWISHER: Well, I think we want to know these things to know who we are. That's why I work in this field, is I want to know who we are. And I think modern human cognitive science and cognitive neuroscience are starting to give us an answer to that, at least who we are as thinkers. And who we are as thinkers is these machines that have a bunch of highly specialized components, some very general-purpose machinery. And now we're starting to know what those components are, which each one does. Like, that's the beginning of a depiction of what a human mind is. I think that's an incredibly exciting thing to have.


RAZ: Nancy Kanwisher is a cognitive neuroscientist at MIT. You can see her full talk at Our show today - The Unknown Brain. I'm Guy Raz, and you're listening to the TED Radio Hour from NPR. Transcript provided by NPR, Copyright NPR.