1
00:00:01,024 --> 00:00:03,707
I want you to take a look at this baby.

2
00:00:04,179 --> 00:00:08,975
What you're drawn to are her eyes
and the skin you love to touch.

3
00:00:09,386 --> 00:00:12,526
But today I'm going to talk to you
about something you can't see.

4
00:00:12,550 --> 00:00:15,302
What's going on
up in that little brain of hers.

5
00:00:16,135 --> 00:00:20,056
The modern tools of neuroscience
are demonstrating to us

6
00:00:20,080 --> 00:00:23,975
that what's going on up there
is nothing short of rocket science.

7
00:00:24,736 --> 00:00:28,173
And what we're learning
is going to shed some light

8
00:00:28,197 --> 00:00:34,095
on what the romantic writers and poets
described as the "celestial openness"

9
00:00:34,119 --> 00:00:35,976
of the child's mind.

10
00:00:37,013 --> 00:00:40,471
What we see here is a mother in India,

11
00:00:40,495 --> 00:00:44,399
and she's speaking Koro,
which is a newly discovered language.

12
00:00:44,797 --> 00:00:46,604
And she's talking to her baby.

13
00:00:47,089 --> 00:00:48,245
What this mother --

14
00:00:48,269 --> 00:00:51,488
and the 800 people who speak
Koro in the world --

15
00:00:51,512 --> 00:00:54,722
understands is that,
to preserve this language,

16
00:00:54,746 --> 00:00:57,332
they need to speak it to the babies.

17
00:00:57,356 --> 00:01:00,231
And therein lies a critical puzzle.

18
00:01:00,255 --> 00:01:02,505
Why is it that you can't
preserve a language

19
00:01:02,529 --> 00:01:05,764
by speaking to you and I, to the adults?

20
00:01:05,788 --> 00:01:07,976
Well, it's got to do with your brain.

21
00:01:08,538 --> 00:01:12,976
What we see here is that language
has a critical period for learning.

22
00:01:13,000 --> 00:01:16,800
The way to read this slide is to look
at your age on the horizontal axis.

23
00:01:16,824 --> 00:01:19,507
(Laughter)

24
00:01:19,531 --> 00:01:23,170
And you'll see on the vertical
your skill at acquiring a second language.

25
00:01:23,789 --> 00:01:27,648
The babies and children are geniuses
until they turn seven,

26
00:01:27,672 --> 00:01:29,976
and then there's a systematic decline.

27
00:01:30,000 --> 00:01:32,497
After puberty, we fall off the map.

28
00:01:32,934 --> 00:01:35,466
No scientists dispute this curve,

29
00:01:35,490 --> 00:01:37,445
but laboratories all over the world

30
00:01:37,469 --> 00:01:39,976
are trying to figure out
why it works this way.

31
00:01:40,518 --> 00:01:44,552
Work in my lab is focused on the first
critical period in development,

32
00:01:44,576 --> 00:01:46,421
and that is the period in which babies

33
00:01:46,445 --> 00:01:49,810
try to master which sounds
are used in their language.

34
00:01:49,834 --> 00:01:52,324
We think, by studying
how the sounds are learned,

35
00:01:52,348 --> 00:01:54,613
we'll have a model
for the rest of language,

36
00:01:54,637 --> 00:01:57,575
and perhaps for critical periods
that may exist in childhood

37
00:01:57,599 --> 00:02:00,462
for social, emotional
and cognitive development.

38
00:02:01,035 --> 00:02:02,976
So we've been studying the babies

39
00:02:03,000 --> 00:02:05,503
using a technique
that we're using all over the world

40
00:02:05,527 --> 00:02:07,383
and the sounds of all languages.

41
00:02:07,407 --> 00:02:09,198
The baby sits on a parent's lap,

42
00:02:09,222 --> 00:02:12,143
and we train them to turn
their heads when a sound changes --

43
00:02:12,167 --> 00:02:13,625
like from "ah" to "ee."

44
00:02:13,649 --> 00:02:17,082
If they do so at the appropriate time,
the black box lights up

45
00:02:17,106 --> 00:02:19,090
and a panda bear pounds a drum.

46
00:02:19,114 --> 00:02:21,532
A six-monther adores the task.

47
00:02:22,160 --> 00:02:23,321
What have we learned?

48
00:02:23,345 --> 00:02:25,505
Well, babies all over the world

49
00:02:25,529 --> 00:02:29,187
are what I like to describe
as "citizens of the world."

50
00:02:29,646 --> 00:02:32,574
They can discriminate
all the sounds of all languages,

51
00:02:32,598 --> 00:02:35,830
no matter what country we're testing
and what language we're using,

52
00:02:35,854 --> 00:02:38,849
and that's remarkable
because you and I can't do that.

53
00:02:38,873 --> 00:02:40,844
We're culture-bound listeners.

54
00:02:40,868 --> 00:02:43,263
We can discriminate the sounds
of our own language,

55
00:02:43,287 --> 00:02:45,134
but not those of foreign languages.

56
00:02:45,158 --> 00:02:48,325
So the question arises:
When do those citizens of the world

57
00:02:48,349 --> 00:02:51,233
turn into the language-bound
listeners that we are?

58
00:02:51,257 --> 00:02:54,394
And the answer:
before their first birthdays.

59
00:02:54,418 --> 00:02:57,550
What you see here is performance
on that head-turn task

60
00:02:57,574 --> 00:03:00,037
for babies tested in Tokyo
and the United States,

61
00:03:00,061 --> 00:03:01,499
here in Seattle,

62
00:03:01,523 --> 00:03:03,272
as they listened to "ra" and "la" --

63
00:03:03,296 --> 00:03:06,102
sounds important to English,
but not to Japanese.

64
00:03:06,126 --> 00:03:09,504
So at six to eight months,
the babies are totally equivalent.

65
00:03:09,528 --> 00:03:12,082
Two months later,
something incredible occurs.

66
00:03:12,106 --> 00:03:14,786
The babies in the United States
are getting a lot better,

67
00:03:14,810 --> 00:03:16,734
babies in Japan are getting a lot worse,

68
00:03:16,758 --> 00:03:20,208
but both of those groups of babies
are preparing for exactly the language

69
00:03:20,232 --> 00:03:21,724
that they are going to learn.

70
00:03:21,748 --> 00:03:26,372
So the question is: What's happening
during this critical two-month period?

71
00:03:26,396 --> 00:03:28,739
This is the critical period
for sound development,

72
00:03:28,763 --> 00:03:30,178
but what's going on up there?

73
00:03:30,202 --> 00:03:32,181
So there are two things going on.

74
00:03:32,205 --> 00:03:35,081
The first is that the babies
are listening intently to us,

75
00:03:35,105 --> 00:03:38,838
and they're taking statistics
as they listen to us talk --

76
00:03:38,862 --> 00:03:40,843
they're taking statistics.

77
00:03:40,867 --> 00:03:43,406
So listen to two mothers
speaking motherese --

78
00:03:43,430 --> 00:03:46,472
the universal language
we use when we talk to kids --

79
00:03:46,496 --> 00:03:48,707
first in English and then in Japanese.

80
00:03:48,731 --> 00:03:52,170
(Video) Ah, I love your big blue eyes --

81
00:03:52,194 --> 00:03:55,000
so pretty and nice.

82
00:03:56,049 --> 00:04:01,976
(Japanese)

83
00:04:02,452 --> 00:04:05,819
Patricia Kuhl: During the production
of speech, when babies listen,

84
00:04:05,843 --> 00:04:10,612
what they're doing is taking statistics
on the language that they hear.

85
00:04:11,353 --> 00:04:14,263
And those distributions grow.

86
00:04:14,287 --> 00:04:18,660
And what we've learned is that babies
are sensitive to the statistics,

87
00:04:18,684 --> 00:04:22,512
and the statistics of Japanese
and English are very, very different.

88
00:04:22,536 --> 00:04:25,193
English has a lot of Rs and Ls.

89
00:04:25,217 --> 00:04:26,976
The distribution shows.

90
00:04:27,000 --> 00:04:29,840
And the distribution of Japanese
is totally different,

91
00:04:29,864 --> 00:04:32,865
where we see a group
of intermediate sounds,

92
00:04:32,889 --> 00:04:35,304
which is known as the Japanese "R."

93
00:04:35,328 --> 00:04:39,436
So babies absorb
the statistics of the language

94
00:04:39,460 --> 00:04:41,138
and it changes their brains;

95
00:04:41,162 --> 00:04:43,313
it changes them
from the citizens of the world

96
00:04:43,337 --> 00:04:46,283
to the culture-bound
listeners that we are.

97
00:04:46,307 --> 00:04:51,020
But we as adults are no longer
absorbing those statistics.

98
00:04:51,044 --> 00:04:53,601
We are governed
by the representations in memory

99
00:04:53,625 --> 00:04:56,174
that were formed early in development.

100
00:04:56,198 --> 00:04:57,976
So what we're seeing here

101
00:04:58,000 --> 00:05:01,233
is changing our models
of what the critical period is about.

102
00:05:01,257 --> 00:05:04,026
We're arguing from
a mathematical standpoint

103
00:05:04,050 --> 00:05:07,190
that the learning of language
material may slow down

104
00:05:07,214 --> 00:05:09,594
when our distributions stabilize.

105
00:05:09,618 --> 00:05:12,480
It's raising lots of questions
about bilingual people.

106
00:05:12,918 --> 00:05:16,949
Bilinguals must keep two sets
of statistics in mind at once

107
00:05:16,973 --> 00:05:19,856
and flip between them,
one after the other,

108
00:05:19,880 --> 00:05:21,652
depending on who they're speaking to.

109
00:05:21,676 --> 00:05:22,830
So we asked ourselves,

110
00:05:22,854 --> 00:05:26,261
can the babies take statistics
on a brand new language?

111
00:05:26,285 --> 00:05:29,563
And we tested this
by exposing American babies

112
00:05:29,587 --> 00:05:31,276
who'd never heard a second language

113
00:05:31,300 --> 00:05:34,016
to Mandarin for the first time
during the critical period.

114
00:05:34,040 --> 00:05:37,577
We knew that, when monolinguals
were tested in Taipei and Seattle

115
00:05:37,601 --> 00:05:40,263
on the Mandarin sounds,
they showed the same pattern.

116
00:05:40,287 --> 00:05:42,758
Six to eight months,
they're totally equivalent.

117
00:05:42,782 --> 00:05:45,318
Two months later,
something incredible happens.

118
00:05:45,342 --> 00:05:48,894
But the Taiwanese babies are getting
better, not the American babies.

119
00:05:48,918 --> 00:05:53,800
What we did was expose American babies,
during this period, to Mandarin.

120
00:05:53,824 --> 00:05:56,832
It was like having Mandarin relatives
come and visit for a month

121
00:05:56,856 --> 00:06:00,653
and move into your house
and talk to the babies for 12 sessions.

122
00:06:00,677 --> 00:06:02,883
Here's what it looked like
in the laboratory.

123
00:06:02,907 --> 00:06:08,463
(Mandarin)

124
00:06:24,662 --> 00:06:26,908
PK: So what have we done
to their little brains?

125
00:06:26,932 --> 00:06:28,976
(Laughter)

126
00:06:29,000 --> 00:06:31,671
We had to run a control group to make sure

127
00:06:31,695 --> 00:06:34,907
that coming into the laboratory
didn't improve your Mandarin skills.

128
00:06:34,931 --> 00:06:37,584
So a group of babies came in
and listened to English.

129
00:06:37,608 --> 00:06:39,010
And we can see from the graph

130
00:06:39,034 --> 00:06:41,617
that exposure to English
didn't improve their Mandarin.

131
00:06:41,641 --> 00:06:45,236
But look at what happened to the babies
exposed to Mandarin for 12 sessions.

132
00:06:45,260 --> 00:06:47,660
They were as good as the babies in Taiwan

133
00:06:47,684 --> 00:06:50,660
who'd been listening
for 10 and a half months.

134
00:06:50,684 --> 00:06:54,508
What it demonstrated is that babies
take statistics on a new language.

135
00:06:54,532 --> 00:06:57,976
Whatever you put in front of them,
they'll take statistics on.

136
00:06:58,000 --> 00:06:59,590
But we wondered what role

137
00:06:59,614 --> 00:07:03,851
the human being played
in this learning exercise.

138
00:07:04,168 --> 00:07:08,484
So we ran another group of babies
in which the kids got the same dosage,

139
00:07:08,508 --> 00:07:10,976
the same 12 sessions,
but over a television set.

140
00:07:11,000 --> 00:07:14,661
And another group of babies
who had just audio exposure

141
00:07:14,685 --> 00:07:16,773
and looked at a teddy bear on the screen.

142
00:07:16,797 --> 00:07:18,976
What did we do to their brains?

143
00:07:19,000 --> 00:07:22,375
What you see here is the audio result --

144
00:07:22,923 --> 00:07:24,661
no learning whatsoever --

145
00:07:24,685 --> 00:07:26,443
and the video result --

146
00:07:27,236 --> 00:07:29,081
no learning whatsoever.

147
00:07:29,105 --> 00:07:33,246
It takes a human being
for babies to take their statistics.

148
00:07:33,622 --> 00:07:35,788
The social brain is controlling

149
00:07:35,812 --> 00:07:37,917
when the babies
are taking their statistics.

150
00:07:37,941 --> 00:07:41,434
We want to get inside the brain
and see this thing happening

151
00:07:41,458 --> 00:07:45,396
as babies are in front of televisions,
as opposed to in front of human beings.

152
00:07:45,420 --> 00:07:50,117
Thankfully, we have a new machine,
magnetoencephalography,

153
00:07:50,141 --> 00:07:51,434
that allows us to do this.

154
00:07:51,458 --> 00:07:54,043
It looks like a hair dryer from Mars.

155
00:07:54,067 --> 00:07:58,550
But it's completely safe,
completely noninvasive and silent.

156
00:07:58,574 --> 00:08:01,017
We're looking at millimeter accuracy

157
00:08:01,041 --> 00:08:04,357
with regard to spatial
and millisecond accuracy

158
00:08:04,381 --> 00:08:07,003
using 306 SQUIDs --

159
00:08:07,027 --> 00:08:10,127
these are superconducting
quantum interference devices --

160
00:08:10,151 --> 00:08:13,782
to pick up the magnetic fields
that change as we do our thinking.

161
00:08:14,249 --> 00:08:19,858
We're the first in the world
to record babies in an MEG machine

162
00:08:19,882 --> 00:08:22,273
while they are learning.

163
00:08:22,297 --> 00:08:23,976
So this is little Emma.

164
00:08:24,000 --> 00:08:25,908
She's a six-monther.

165
00:08:25,932 --> 00:08:31,050
And she's listening to various languages
in the earphones that are in her ears.

166
00:08:31,074 --> 00:08:33,298
You can see, she can move around.

167
00:08:33,322 --> 00:08:37,092
We're tracking her head
with little pellets in a cap,

168
00:08:37,116 --> 00:08:40,091
so she's free to move
completely unconstrained.

169
00:08:40,115 --> 00:08:42,328
It's a technical tour de force.

170
00:08:42,352 --> 00:08:43,936
What are we seeing?

171
00:08:43,960 --> 00:08:45,854
We're seeing the baby brain.

172
00:08:45,878 --> 00:08:51,164
As the baby hears a word in her language,
the auditory areas light up,

173
00:08:51,188 --> 00:08:55,976
and then subsequently areas surrounding it
that we think are related to coherence,

174
00:08:56,000 --> 00:08:59,976
getting the brain coordinated
with its different areas, and causality,

175
00:09:00,000 --> 00:09:02,976
one brain area
causing another to activate.

176
00:09:03,601 --> 00:09:08,985
We are embarking on a grand
and golden age of knowledge

177
00:09:09,009 --> 00:09:11,245
about child's brain development.

178
00:09:11,269 --> 00:09:13,690
We're going to be able
to see a child's brain

179
00:09:13,714 --> 00:09:17,955
as they experience an emotion,
as they learn to speak and read,

180
00:09:17,979 --> 00:09:21,181
as they solve a math problem,
as they have an idea.

181
00:09:21,515 --> 00:09:24,864
And we're going to be able to invent
brain-based interventions

182
00:09:24,888 --> 00:09:27,158
for children who have difficulty learning.

183
00:09:27,571 --> 00:09:30,637
Just as the poets and writers described,

184
00:09:30,661 --> 00:09:34,815
we're going to be able to see, I think,
that wondrous openness,

185
00:09:34,839 --> 00:09:38,263
utter and complete openness,
of the mind of a child.

186
00:09:39,024 --> 00:09:41,700
In investigating the child's brain,

187
00:09:41,724 --> 00:09:45,720
we're going to uncover deep truths
about what it means to be human,

188
00:09:45,744 --> 00:09:46,899
and in the process,

189
00:09:46,923 --> 00:09:49,994
we may be able to help keep
our own minds open to learning

190
00:09:50,018 --> 00:09:51,709
for our entire lives.

191
00:09:51,733 --> 00:09:52,976
Thank you.

192
00:09:53,000 --> 00:09:56,000
(Applause)