0:00:12.640,0:00:15.459
I'm from a fairly traditional family.

0:00:15.459,0:00:19.148
So like a good Asian girl,[br]I grew up studying violin and piano,

0:00:19.148,0:00:21.701
but I was also expected[br]to take all the premed courses

0:00:21.701,0:00:25.216
and go to med school and become a doctor.

0:00:25.576,0:00:27.330
But then I went to college,

0:00:27.330,0:00:28.362
and in college,

0:00:28.362,0:00:30.157
I got really seduced by this idea

0:00:30.157,0:00:34.685
of how these seemingly abstract,[br]elusive concepts -

0:00:34.685,0:00:39.972
like beauty and truth and love and art[br]and, in particular, music -

0:00:39.972,0:00:44.996
could actually be understood[br]using the objective principles of science.

0:00:45.276,0:00:46.967
So then I signed up for grad school

0:00:46.967,0:00:49.072
to study the cognitive[br]neuroscience of music.

0:00:49.072,0:00:51.546
In grad school, I got obsessed[br]with this question of,

0:00:51.546,0:00:53.307
Where does music come from?

0:00:53.717,0:00:55.915
Music is a multibillion-dollar industry,

0:00:55.915,0:00:57.990
and that's because people love music.

0:00:57.990,0:00:59.952
People love to rock out at concerts,

0:00:59.952,0:01:02.960
and I'd like to think there's something[br]about the musical signal

0:01:02.960,0:01:05.436
that appeals to what[br]is uniquely human in each of us.

0:01:05.436,0:01:07.964
Of course, this is not just true[br]of the Western world.

0:01:07.964,0:01:10.259
This is a picture[br]taken in Mali, in West Africa,

0:01:10.259,0:01:12.951
and he's holding[br]this instrument called the ngoni.

0:01:12.951,0:01:15.587
And although this seems very foreign,

0:01:15.587,0:01:17.936
what his brain does to listen to music

0:01:17.936,0:01:21.055
is probably very similar[br]to what our brains do to listen to music.

0:01:21.055,0:01:24.327
Furthermore, the physical principles[br]that make his strings vibrate

0:01:24.327,0:01:28.209
probably are the same[br]as what makes our instruments vibrate,

0:01:28.209,0:01:29.814
like the violin.

0:01:30.054,0:01:34.525
Now, we don't just love music;[br]we also know lots of things about music.

0:01:34.525,0:01:37.334
Consider, for instance,[br]this musical example.

0:01:37.334,0:01:41.874
(Simple chords on a keyboard)

0:01:41.874,0:01:44.356
All right. You might say[br]that sounds nice and normal,

0:01:44.356,0:01:46.726
kind of like saying,[br]"I took the T here today."

0:01:47.386,0:01:48.465
What about this?

0:01:48.465,0:01:51.454
(Same chords, the final one discordant)

0:01:52.394,0:01:55.669
Right - if you think that sounded normal,[br]come talk to me afterwards;

0:01:55.669,0:01:58.620
we might sign you up[br]for that tone-deafness study we're doing.

0:01:58.620,0:02:00.066
(Laughter)

0:02:00.066,0:02:03.316
When you heard that last chord,[br]your brain does a double take, right?

0:02:03.316,0:02:04.579
There's something about it

0:02:04.579,0:02:06.831
that's like saying[br]"I took the T here octopus."

0:02:06.831,0:02:08.135
Nothing wrong with octopus,

0:02:08.135,0:02:11.104
but it just doesn't fit the context[br]of what happened before it;

0:02:11.104,0:02:12.573
it doesn't fit the grammar.

0:02:12.573,0:02:14.599
Now, this double take that your brain does

0:02:14.599,0:02:17.974
can be measured using electrical[br]potentials on the surface of the scalp.

0:02:17.974,0:02:21.075
This is a picture of my mom[br]getting her brain potentials recorded,

0:02:21.075,0:02:24.473
and she's got 64 electrodes[br]on her cap there,

0:02:24.473,0:02:27.502
and what those do[br]is make recordings like this.

0:02:27.502,0:02:29.236
And so on the left, I'm showing you

0:02:29.236,0:02:32.600
brain responses to expected[br]and unexpected musical chords.

0:02:32.600,0:02:34.880
And on the right,[br]I'm showing you the difference

0:02:34.880,0:02:37.794
between expected and unexpected[br]on the surface of the scalp -

0:02:37.794,0:02:39.793
so this is a bird's-eye view of the scalp.

0:02:39.793,0:02:40.796
So right away,

0:02:40.796,0:02:44.275
you can see that 200 milliseconds[br]after the onset of the unexpected chord,

0:02:44.275,0:02:47.130
your brain does this double take:[br]"Oh, that was unexpected."

0:02:47.130,0:02:49.806
In 500 milliseconds,[br]you get the brain saying,

0:02:49.806,0:02:53.128
"Oh, how do I integrate that[br]into what happened before?"

0:02:53.368,0:02:56.825
So this is telling us,[br]with millisecond accuracy,

0:02:56.825,0:02:58.685
that we know about about music;

0:02:58.685,0:03:01.374
there's something about our brains[br]that is very sensitive

0:03:01.374,0:03:03.835
to what's grammatical in Western music.

0:03:03.835,0:03:04.849
So the question is,

0:03:04.849,0:03:06.584
Where does this knowledge come from?

0:03:06.584,0:03:08.312
How do we come to know what we know?

0:03:08.312,0:03:09.545
To answer that question,

0:03:09.545,0:03:12.306
we again have to go all the way back[br]to the ancient Greeks.

0:03:12.306,0:03:15.306
Pythagoras found that if two strings[br]are being played together

0:03:15.306,0:03:18.356
where one string[br]is twice the length of the other

0:03:18.356,0:03:21.471
those two sound good together;[br]they sound consonant.

0:03:21.471,0:03:23.773
So this two-to-one frequency ratio

0:03:24.013,0:03:27.274
is what, supposedly,[br]brought us closer to the Greek gods.

0:03:27.444,0:03:29.294
In fact, the word "symphony"

0:03:29.294,0:03:31.912
originally means [br]"vibrating in perfect harmony"

0:03:31.912,0:03:35.109
using these mathematical integer ratios.

0:03:35.109,0:03:40.020
So this two-to-one frequency ratio[br]is true of music all around the world.

0:03:40.020,0:03:42.942
Now, different cultures divide[br]that two-to-one frequency ratio

0:03:42.942,0:03:43.945
in different ways.

0:03:43.945,0:03:46.695
In our culture, the equal-tempered[br]Western chromatic scale

0:03:46.695,0:03:48.084
divides them in 12 steps.

0:03:48.084,0:03:49.345
So this is how it sounds.

0:03:49.345,0:03:52.470
(13 tones covering a 12-note scale)

0:03:56.400,0:03:57.408
Okay.

0:03:57.408,0:04:00.261
Then two guys came along,[br]said, "Does it have to be this way?

0:04:00.261,0:04:02.273
Why two-to-one? Why not three-to-one?"

0:04:02.273,0:04:05.737
So the Bohlen-Pierce scale is based[br]on a three-to-one frequency ratio,

0:04:05.737,0:04:09.336
and within that, we've got[br]13 logarithmic divisions of that scale.

0:04:09.336,0:04:12.632
So you still get some[br]mathematical integer ratios -

0:04:12.632,0:04:14.915
so the Greek gods are not offended here.

0:04:14.915,0:04:16.293
But what this sounds like

0:04:16.293,0:04:19.851
is completely different[br]from Western or other types of music.

0:04:19.851,0:04:22.843
(14 tones covering an alternate scale)

0:04:27.584,0:04:33.161
So this is a really powerful approach[br]to find out what people know about music

0:04:33.161,0:04:34.196
in the laboratory.

0:04:34.196,0:04:37.366
So we can be pretty sure people[br]have never heard this music before,

0:04:37.366,0:04:40.206
but they come in, they can listen[br]to this music for a while,

0:04:40.206,0:04:42.965
then we can measure[br]how they come to know what they know.

0:04:42.965,0:04:45.103
So I'm going to play you,[br]for about a minute,

0:04:45.103,0:04:46.865
a snippet of a piece by Stephen Yi.

0:04:46.865,0:04:48.207
It's called "Reminiscences,"

0:04:48.207,0:04:50.278
and it's written[br]in the Bohlen-Pierce scale,

0:04:50.278,0:04:51.446
just so you get an idea.

0:04:51.446,0:04:54.443
(Ethereal music)

0:05:42.624,0:05:45.861
So this really is kind of an otherworldly[br]new musical experience

0:05:45.861,0:05:47.243
that we're entering here,

0:05:47.243,0:05:48.962
and in our lab, what we wanted to do

0:05:48.962,0:05:51.587
was figure out how people learn[br]this new musical system.

0:05:51.587,0:05:53.592
So we have these well-controlled melodies

0:05:53.592,0:05:55.736
that people listen to[br]for about half an hour.

0:05:55.736,0:05:58.466
(Atonal note progression)

0:05:59.450,0:06:01.646
So you listen to these things[br]for half an hour,

0:06:01.646,0:06:04.020
and they're defined[br]using rules and principles,

0:06:04.020,0:06:05.281
or grammatical structures,

0:06:05.281,0:06:06.661
that we've defined ourselves.

0:06:06.661,0:06:07.857
And then the question is,

0:06:07.857,0:06:10.431
What can people learn[br]from this new musical experience?

0:06:10.431,0:06:13.381
First thing we found was that memory[br]increases with repetition.

0:06:13.381,0:06:16.215
Turns out, also, that preference[br]increases with repetition.

0:06:16.215,0:06:19.162
So what we're seeing is the beginning[br]of musical taste, right?

0:06:19.162,0:06:22.232
The more you listen to something,[br]the more you begin to like it.

0:06:22.232,0:06:24.246
But I'm interested in how learning occurs.

0:06:24.246,0:06:26.913
It turns out that learning[br]does not occur with repetition

0:06:26.913,0:06:27.940
but with variability.

0:06:27.940,0:06:30.594
In other words, the more ways[br]you tell people something,

0:06:30.594,0:06:32.260
the more people are able to infer

0:06:32.260,0:06:34.404
the underlying structure[br]of what you tell them

0:06:34.404,0:06:38.401
and then to generalize those[br]to new instances of the same grammar.

0:06:38.741,0:06:40.444
Our question becomes, now,

0:06:40.444,0:06:43.157
We've got 100 trillion [br]neural connections to the brain;

0:06:43.157,0:06:45.661
how did those 100 trillion[br]neural connections

0:06:45.661,0:06:48.288
give rise to what we know[br]and love in music?

0:06:49.198,0:06:53.031
Right now, these neural connections[br]are in the order of nanometers,

0:06:53.031,0:06:55.735
but what we can image[br]using the living human brain,

0:06:55.735,0:06:59.826
using this technology called[br]"diffusion tensor imaging,"

0:06:59.826,0:07:03.147
is large bundles[br]of these neural connections -

0:07:03.147,0:07:04.714
so highways, if you will.

0:07:04.714,0:07:08.390
And the highway we're most interested in[br]is called the arcuate fasciculus

0:07:08.390,0:07:10.402
and it's known[br]to be important in language.

0:07:10.402,0:07:13.416
But what we saw is the larger[br]of an arcuate fasciculus you have,

0:07:13.416,0:07:16.099
the better you are at learning[br]this new musical system.

0:07:16.099,0:07:18.099
So there's something[br]structurally different

0:07:18.099,0:07:20.315
about a good and a not-so-good[br]learner's brain.

0:07:20.315,0:07:22.351
But what's important[br]is that these pathways

0:07:22.351,0:07:24.845
that are previously known[br]to be important in language

0:07:24.845,0:07:26.771
are actually important in music as well.

0:07:26.771,0:07:29.433
So this tells us that there[br]is no single center for music

0:07:29.433,0:07:31.667
or there's no one center[br]for music in the brain.

0:07:31.667,0:07:34.526
But what we do have[br]are these shared neural networks

0:07:34.526,0:07:37.941
that are important in language[br]and in grammar and in expectation

0:07:37.941,0:07:40.228
and all these things[br]that actually make us human.

0:07:40.228,0:07:41.311
So I think music -

0:07:41.311,0:07:43.171
that's actually why people like music.

0:07:43.171,0:07:46.483
It's not because it's this individualized[br]stereotyped activity,

0:07:46.483,0:07:49.965
but it's something that tickles[br]all the different cognitive components

0:07:49.965,0:07:52.755
and neural mechanisms[br]that we already have.

0:07:52.755,0:07:53.866
Now, that sounds good,

0:07:53.866,0:07:57.982
but can we actually observe the brain[br]as it is learning in real time?

0:07:57.982,0:08:03.303
So we go back to the millisecond-accuracy[br]kind of brain-potential recording,

0:08:03.303,0:08:06.081
and it turns out that our brains[br]respond to new music

0:08:06.081,0:08:08.747
in very much the same way[br]as it does to Western music.

0:08:08.747,0:08:11.221
So we get the same[br]expected-unexpected pattern

0:08:11.221,0:08:14.182
200 milliseconds and 500 milliseconds

0:08:14.182,0:08:17.480
after the onset of anything[br]that sounds unexpected.

0:08:17.480,0:08:22.453
And furthermore, our brains respond[br]more and more towards these expectations

0:08:22.453,0:08:24.067
throughout the course of an hour,

0:08:24.067,0:08:26.580
so as if within an hour,

0:08:26.580,0:08:30.801
we're getting more and more experts[br]in the Bohlen-Pierce scale.

0:08:30.801,0:08:34.231
So there's no rules of music[br]that are written in our brains,

0:08:34.231,0:08:38.572
but what we do have that are in our brains[br]is the immense ability to learn.

0:08:38.572,0:08:41.291
So we are fundamentally[br]open-minded creatures.

0:08:41.291,0:08:44.431
So what does that mean[br]if we want to go back to West Africa?

0:08:44.431,0:08:46.626
I invite you to take off your headphones

0:08:46.626,0:08:49.920
and actually experience[br]the new musical world.

0:08:49.920,0:08:53.696
Try to come up with the grammar[br]of this seemingly foreign country.

0:08:53.696,0:08:56.351
You know, so, and what about[br]even just being here today?

0:08:56.351,0:09:00.925
How about change your radio channel[br]or listen to a new musical artist today?

0:09:00.925,0:09:03.653
There's something[br]about experiencing new things

0:09:03.653,0:09:06.421
that, to me, is what it means to thrive

0:09:06.421,0:09:10.366
because to thrive is to maximize[br]our potentials as human beings, right?

0:09:10.366,0:09:12.892
It's not to do the same thing[br]over and over every day,

0:09:12.892,0:09:14.669
but it's to seek out new experiences,

0:09:14.669,0:09:16.138
and what I've shown you today

0:09:16.138,0:09:20.241
is that the brain is fundamentally capable[br]of learning new things.

0:09:20.241,0:09:22.061
We can take - even within an hour,

0:09:22.061,0:09:25.181
we can have this flexible,[br]adaptive ability

0:09:25.181,0:09:27.126
to make sense of new sounds.

0:09:27.126,0:09:30.225
So I invite you to listen to new sounds,

0:09:30.225,0:09:31.355
see new sights

0:09:31.355,0:09:34.655
and come up with the grammar[br]of the world that's around us

0:09:34.655,0:09:36.206
so that we can learn to love it.

0:09:36.206,0:09:37.740
Thank you very much.

0:09:37.740,0:09:39.405
(Applause)