sponsored links
TED2014

Nancy Kanwisher: A neural portrait of the human mind

March 14, 2014

Brain imaging pioneer Nancy Kanwisher, who uses fMRI scans to see activity in brain regions (often her own), shares what she and her colleagues have learned: The brain is made up of both highly specialized components and general-purpose "machinery." Another surprise: There's so much left to learn.

Nancy Kanwisher - Brain researcher
Using fMRI imaging to watch the human brain at work, Nancy Kanwisher’s team has discovered cortical regions responsible for some surprisingly specific elements of cognition. Full bio

sponsored links
Double-click the English subtitles below to play the video.
Today I want to tell you
00:12
about a project being carried out
00:13
by scientists all over the world
00:15
to paint a neural portrait of the human mind.
00:18
And the central idea of this work
00:21
is that the human mind and brain
00:23
is not a single, general-purpose processor,
00:25
but a collection of highly specialized components,
00:28
each solving a different specific problem,
00:31
and yet collectively making up
00:34
who we are as human beings and thinkers.
00:37
To give you a feel for this idea,
00:41
imagine the following scenario:
00:43
You walk into your child's day care center.
00:45
As usual, there's a dozen kids there
00:47
waiting to get picked up,
00:50
but this time,
00:51
the children's faces look weirdly similar,
00:53
and you can't figure out which child is yours.
00:56
Do you need new glasses?
00:59
Are you losing your mind?
01:00
You run through a quick mental checklist.
01:02
No, you seem to be thinking clearly,
01:05
and your vision is perfectly sharp.
01:07
And everything looks normal
01:09
except the children's faces.
01:11
You can see the faces,
01:13
but they don't look distinctive,
01:15
and none of them looks familiar,
01:17
and it's only by spotting an orange hair ribbon
01:18
that you find your daughter.
01:21
This sudden loss of the ability to recognize faces
01:23
actually happens to people.
01:26
It's called prosopagnosia,
01:28
and it results from damage
01:30
to a particular part of the brain.
01:31
The striking thing about it
01:33
is that only face recognition is impaired;
01:35
everything else is just fine.
01:37
Prosopagnosia is one of many surprisingly specific
01:40
mental deficits that can happen after brain damage.
01:44
These syndromes collectively
01:48
have suggested for a long time
01:49
that the mind is divvied up into distinct components,
01:52
but the effort to discover those components
01:55
has jumped to warp speed
01:58
with the invention of brain imaging technology,
01:59
especially MRI.
02:02
So MRI enables you to see internal anatomy
02:05
at high resolution,
02:08
so I'm going to show you in a second
02:10
a set of MRI cross-sectional images
02:11
through a familiar object,
02:15
and we're going to fly through them
02:16
and you're going to try to figure out what the object is.
02:17
Here we go.
02:20
It's not that easy. It's an artichoke.
02:24
Okay, let's try another one,
02:26
starting from the bottom and going through the top.
02:27
Broccoli! It's a head of broccoli.
02:32
Isn't it beautiful? I love that.
02:33
Okay, here's another one. It's a brain, of course.
02:35
In fact, it's my brain.
02:38
We're going through slices through my head like that.
02:39
That's my nose over on the right, and now
02:41
we're going over here, right there.
02:43
So this picture's nice, if I do say so myself,
02:46
but it shows only anatomy.
02:51
The really cool advance with functional imaging
02:53
happened when scientists figured out how to make
02:55
pictures that show not just anatomy but activity,
02:57
that is, where neurons are firing.
03:00
So here's how this works.
03:03
Brains are like muscles.
03:04
When they get active,
03:05
they need increased blood flow to supply that activity,
03:07
and lucky for us, blood flow
control to the brain is local,
03:10
so if a bunch of neurons, say, right there
03:14
get active and start firing,
03:16
then blood flow increases just right there.
03:17
So functional MRI picks up
on that blood flow increase,
03:20
producing a higher MRI response
03:24
where neural activity goes up.
03:26
So to give you a concrete feel
03:29
for how a functional MRI experiment goes
03:30
and what you can learn from it
03:33
and what you can't,
03:34
let me describe one of the first studies I ever did.
03:36
We wanted to know if there was a special
part of the brain for recognizing faces,
03:39
and there was already reason to
think there might be such a thing
03:43
based on this phenomenon of prosopagnosia
03:46
that I described a moment ago,
03:48
but nobody had ever seen that part of the brain
03:50
in a normal person,
03:52
so we set out to look for it.
03:54
So I was the first subject.
03:56
I went into the scanner, I lay on my back,
03:58
I held my head as still as I could
04:01
while staring at pictures of faces like these
04:03
and objects like these
04:08
and faces and objects for hours.
04:10
So as somebody who has
pretty close to the world record
04:15
of total number of hours spent inside an MRI scanner,
04:18
I can tell you that one of the skills
04:22
that's really important for MRI research
04:23
is bladder control.
04:26
(Laughter)
04:28
When I got out of the scanner,
04:29
I did a quick analysis of the data,
04:31
looking for any parts of my brain
04:33
that produced a higher response
when I was looking at faces
04:35
than when I was looking at objects,
04:38
and here's what I saw.
04:39
Now this image looks just awful by today's standards,
04:42
but at the time I thought it was beautiful.
04:45
What it shows is that region right there,
04:48
that little blob,
04:50
it's about the size of an olive
04:51
and it's on the bottom surface of my brain
04:53
about an inch straight in from right there.
04:55
And what that part of my brain is doing
04:58
is producing a higher MRI response,
05:01
that is, higher neural activity,
05:04
when I was looking at faces
05:06
than when I was looking at objects.
05:07
So that's pretty cool,
05:10
but how do we know this isn't a fluke?
05:11
Well, the easiest way
05:13
is to just do the experiment again.
05:15
So I got back in the scanner,
05:17
I looked at more faces and I looked at more objects
05:18
and I got a similar blob,
05:21
and then I did it again
05:23
and I did it again
05:25
and again and again,
05:27
and around about then
05:30
I decided to believe it was for real.
05:31
But still, maybe this is
something weird about my brain
05:34
and no one else has one of these things in there,
05:38
so to find out, we scanned a bunch of other people
05:40
and found that pretty much everyone
05:43
has that little face-processing region
05:45
in a similar neighborhood of the brain.
05:47
So the next question was,
05:50
what does this thing really do?
05:52
Is it really specialized just for face recognition?
05:53
Well, maybe not, right?
05:57
Maybe it responds not only to faces
05:58
but to any body part.
06:00
Maybe it responds to anything human
06:02
or anything alive
06:05
or anything round.
06:07
The only way to be really sure that that region
06:08
is specialized for face recognition
06:10
is to rule out all of those hypotheses.
06:13
So we spent much of the next couple of years
06:15
scanning subjects while they looked at lots
06:18
of different kinds of images,
06:20
and we showed that that part of the brain
06:21
responds strongly when you look at
06:23
any images that are faces of any kind,
06:25
and it responds much less strongly
06:29
to any image you show that isn't a face,
06:31
like some of these.
06:34
So have we finally nailed the case
06:35
that this region is necessary for face recognition?
06:37
No, we haven't.
06:41
Brain imaging can never tell you
06:42
if a region is necessary for anything.
06:44
All you can do with brain imaging
06:46
is watch regions turn on and off
06:48
as people think different thoughts.
06:50
To tell if a part of the brain is
necessary for a mental function,
06:52
you need to mess with it and see what happens,
06:55
and normally we don't get to do that.
06:58
But an amazing opportunity came about
07:00
very recently when a couple of colleagues of mine
07:03
tested this man who has epilepsy
07:05
and who is shown here in his hospital bed
07:08
where he's just had electrodes placed
07:11
on the surface of his brain
07:12
to identify the source of his seizures.
07:14
So it turned out by total chance
07:17
that two of the electrodes
07:20
happened to be right on top of his face area.
07:22
So with the patient's consent,
07:25
the doctors asked him what happened
07:27
when they electrically stimulated
that part of his brain.
07:30
Now, the patient doesn't know
07:34
where those electrodes are,
07:35
and he's never heard of the face area.
07:37
So let's watch what happens.
07:39
It's going to start with a control condition
07:41
that will say "Sham" nearly invisibly
07:43
in red in the lower left,
07:45
when no current is delivered,
07:47
and you'll hear the neurologist speaking
to the patient first. So let's watch.
07:49
(Video) Neurologist: Okay, just look at my face
07:53
and tell me what happens when I do this.
07:55
All right?
07:59
Patient: Okay.
08:00
Neurologist: One, two, three.
08:02
Patient: Nothing.
Neurologist: Nothing? Okay.
08:07
I'm going to do it one more time.
08:10
Look at my face.
08:12
One, two, three.
08:15
Patient: You just turned into somebody else.
08:20
Your face metamorphosed.
08:23
Your nose got saggy, it went to the left.
08:25
You almost looked like somebody I'd seen before,
08:28
but somebody different.
08:31
That was a trip.
08:34
(Laughter)
08:36
Nancy Kanwisher: So this experiment —
08:39
(Applause) —
08:41
this experiment finally nails the case
08:45
that this region of the brain is not only
08:48
selectively responsive to faces
08:49
but causally involved in face perception.
08:52
So I went through all of these details
08:55
about the face region to show you what it takes
08:57
to really establish that a part of the brain
08:59
is selectively involved in a specific mental process.
09:02
Next, I'll go through much more quickly
09:05
some of the other specialized regions of the brain
09:07
that we and others have found.
09:10
So to do this, I've spent a lot of time
09:12
in the scanner over the last month
09:14
so I can show you these things in my brain.
09:16
So let's get started. Here's my right hemisphere.
09:18
So we're oriented like that.
You're looking at my head this way.
09:21
Imagine taking the skull off
09:24
and looking at the surface of the brain like that.
09:25
Okay, now as you can see,
09:27
the surface of the brain is all folded up.
09:29
So that's not good. Stuff could be hidden in there.
09:30
We want to see the whole thing,
09:32
so let's inflate it so we can see the whole thing.
09:34
Next, let's find that face area I've been talking about
09:37
that responds to images like these.
09:40
To see that, let's turn the brain around
09:42
and look on the inside surface on the bottom,
09:43
and there it is, that's my face area.
09:45
Just to the right of that is another region
09:48
that is shown in purple
09:50
that responds when you process color information,
09:52
and near those regions are other regions
09:55
that are involved in perceiving places,
09:58
like right now, I'm seeing
this layout of space around me
10:00
and these regions in green right there
10:03
are really active.
10:05
There's another one out on the outside surface again
10:06
where there's a couple more face regions as well.
10:08
Also in this vicinity
10:11
is a region that's selectively involved
10:14
in processing visual motion,
10:15
like these moving dots here,
10:17
and that's in yellow at the bottom of the brain,
10:19
and near that is a region that responds
10:21
when you look at images of bodies and body parts
10:25
like these, and that region is shown in lime green
10:27
at the bottom of the brain.
10:30
Now all these regions I've shown you so far
10:32
are involved in specific aspects of visual perception.
10:35
Do we also have specialized brain regions
10:39
for other senses, like hearing?
10:41
Yes, we do. So if we turn the brain around a little bit,
10:44
here's a region in dark blue
10:47
that we reported just a couple of months ago,
10:50
and this region responds strongly
10:52
when you hear sounds with pitch, like these.
10:54
(Sirens)
10:57
(Cello music)
10:59
(Doorbell)
11:01
In contrast, that same region
does not respond strongly
11:03
when you hear perfectly familiar sounds
11:07
that don't have a clear pitch, like these.
11:08
(Chomping)
11:11
(Drum roll)
11:13
(Toilet flushing)
11:15
Okay. Next to the pitch region
11:18
is another set of regions that
are selectively responsive
11:21
when you hear the sounds of speech.
11:23
Okay, now let's look at these same regions.
11:26
In my left hemisphere, there's a similar arrangement —
11:28
not identical, but similar —
11:30
and most of the same regions are in here,
11:32
albeit sometimes different in size.
11:34
Now, everything I've shown you so far
11:36
are regions that are involved in
different aspects of perception,
11:38
vision and hearing.
11:41
Do we also have specialized brain regions
11:43
for really fancy, complicated mental processes?
11:44
Yes, we do.
11:48
So here in pink are my language regions.
11:49
So it's been known for a very long time
11:53
that that general vicinity of the brain
11:54
is involved in processing language,
11:56
but we showed very recently
11:58
that these pink regions
12:00
respond extremely selectively.
12:02
They respond when you understand
the meaning of a sentence,
12:04
but not when you do other complex mental things,
12:07
like mental arithmetic
12:10
or holding information in memory
12:12
or appreciating the complex structure
12:14
in a piece of music.
12:17
The most amazing region that's been found yet
12:21
is this one right here in turquoise.
12:24
This region responds
12:27
when you think about what another person is thinking.
12:30
So that may seem crazy,
12:34
but actually, we humans do this all the time.
12:35
You're doing this when you realize
12:39
that your partner is going to be worried
12:42
if you don't call home to say you're running late.
12:43
I'm doing this with that region of my brain right now
12:46
when I realize that you guys
12:49
are probably now wondering about
12:51
all that gray, uncharted territory in the brain,
12:53
and what's up with that?
12:56
Well, I'm wondering about that too,
12:58
and we're running a bunch of
experiments in my lab right now
12:59
to try to find a number of other
13:02
possible specializations in the brain
13:04
for other very specific mental functions.
13:06
But importantly, I don't think we have
13:09
specializations in the brain
13:12
for every important mental function,
13:13
even mental functions that may be critical for survival.
13:16
In fact, a few years ago,
13:19
there was a scientist in my lab
13:21
who became quite convinced
13:23
that he'd found a brain region
13:24
for detecting food,
13:26
and it responded really strongly in the scanner
13:28
when people looked at images like this.
13:30
And further, he found a similar response
13:32
in more or less the same location
13:35
in 10 out of 12 subjects.
13:37
So he was pretty stoked,
13:39
and he was running around the lab
13:41
telling everyone that he was going to go on "Oprah"
13:43
with his big discovery.
13:45
But then he devised the critical test:
13:47
He showed subjects images of food like this
13:50
and compared them to images with very similar
13:53
color and shape, but that weren't food, like these.
13:56
And his region responded the same
13:59
to both sets of images.
14:02
So it wasn't a food area,
14:04
it was just a region that liked colors and shapes.
14:05
So much for "Oprah."
14:08
But then the question, of course, is,
14:12
how do we process all this other stuff
14:14
that we don't have specialized brain regions for?
14:16
Well, I think the answer is that in addition
14:19
to these highly specialized components
that I've been describing,
14:21
we also have a lot of very general-
purpose machinery in our heads
14:25
that enables us to tackle
14:28
whatever problem comes along.
14:30
In fact, we've shown recently that
14:32
these regions here in white
14:34
respond whenever you do any difficult mental task
14:36
at all —
14:39
well, of the seven that we've tested.
14:41
So each of the brain regions that I've described
14:44
to you today
14:46
is present in approximately the same location
14:48
in every normal subject.
14:50
I could take any of you,
14:52
pop you in the scanner,
14:54
and find each of those regions in your brain,
14:55
and it would look a lot like my brain,
14:57
although the regions would be slightly different
14:59
in their exact location and in their size.
15:01
What's important to me about this work
15:05
is not the particular locations of these brain regions,
15:07
but the simple fact that we have
15:10
selective, specific components of mind and brain
15:13
in the first place.
15:15
I mean, it could have been otherwise.
15:17
The brain could have been a single,
15:19
general-purpose processor,
15:21
more like a kitchen knife
15:23
than a Swiss Army knife.
15:24
Instead, what brain imaging has delivered
15:26
is this rich and interesting picture of the human mind.
15:29
So we have this picture of very general-purpose
15:33
machinery in our heads
15:35
in addition to this surprising array
15:37
of very specialized components.
15:39
It's early days in this enterprise.
15:43
We've painted only the first brushstrokes
15:45
in our neural portrait of the human mind.
15:48
The most fundamental questions remain unanswered.
15:51
So for example, what does each
of these regions do exactly?
15:54
Why do we need three face areas
15:58
and three place areas,
16:00
and what's the division of labor between them?
16:02
Second, how are all these things
16:04
connected in the brain?
16:07
With diffusion imaging,
16:09
you can trace bundles of neurons
16:10
that connect to different parts of the brain,
16:13
and with this method shown here,
16:15
you can trace the connections of
individual neurons in the brain,
16:17
potentially someday giving us a wiring diagram
16:20
of the entire human brain.
16:23
Third, how does all of this
16:25
very systematic structure get built,
16:27
both over development in childhood
16:30
and over the evolution of our species?
16:33
To address questions like that,
16:36
scientists are now scanning
16:38
other species of animals,
16:40
and they're also scanning human infants.
16:42
Many people justify the high
cost of neuroscience research
16:48
by pointing out that it may help us someday
16:52
to treat brain disorders like Alzheimer's and autism.
16:55
That's a hugely important goal,
16:58
and I'd be thrilled if any of my work contributed to it,
17:00
but fixing things that are broken in the world
17:03
is not the only thing that's worth doing.
17:06
The effort to understand the human mind and brain
17:09
is worthwhile even if it never led to the treatment
17:12
of a single disease.
17:15
What could be more thrilling
17:17
than to understand the fundamental mechanisms
17:19
that underlie human experience,
17:22
to understand, in essence, who we are?
17:24
This is, I think, the greatest scientific quest
17:27
of all time.
17:31
(Applause)
17:34
Translator:Joseph Geni
Reviewer:Mad Aronson

sponsored links

Nancy Kanwisher - Brain researcher
Using fMRI imaging to watch the human brain at work, Nancy Kanwisher’s team has discovered cortical regions responsible for some surprisingly specific elements of cognition.

Why you should listen

Does the brain use specialized processors to solve complex problems, or does it rely instead on more general-purpose systems?

This question has been at the crux of brain research for centuries. MIT researcher Nancy Kanwisher seeks to answer this question by discovering a “parts list” for the human mind and brain. "Understanding the nature of the human mind," she says, "is arguably the greatest intellectual quest of all time."

Kanwisher and her colleagues have used fMRI to identify distinct sites in the brain for face recognition, knowing where you are, and thinking about other people’s thoughts. Yet these discoveries are a prelude to bigger questions: How do these brain regions develop and function? What are the actual computations that go on in each region, and how are these computations implemented in circuits of neurons? And how do these work together to produce human intelligence?

To learn more, see Kanwisher's collection of short talks on how scientists actually study the human mind and brain and what they have learned so far.

The original video is available on TED.com
sponsored links

If you need translations, you can install "Google Translate" extension into your Chrome Browser.
Furthermore, you can change playback rate by installing "Video Speed Controller" extension.

Data provided by TED.

This website is owned and operated by Tokyo English Network.
The developer's blog is here.