1
00:00:00,080 --> 00:00:01,670
The following
content is provided

2
00:00:01,670 --> 00:00:03,820
under a Creative
Commons license.

3
00:00:03,820 --> 00:00:06,550
Your support will help MIT
OpenCourseWare continue

4
00:00:06,550 --> 00:00:10,160
to offer high quality
educational resources for free.

5
00:00:10,160 --> 00:00:12,700
To make a donation or to
view additional materials

6
00:00:12,700 --> 00:00:16,620
from hundreds of MIT courses,
visit MIT OpenCourseWare

7
00:00:16,620 --> 00:00:17,267
at ocw.mit.edu.

8
00:00:25,940 --> 00:00:28,810
PROFESSOR: All right,
so today our topic

9
00:00:28,810 --> 00:00:32,500
is going to be depth
perception, which,

10
00:00:32,500 --> 00:00:34,690
as I have mentioned
to you before,

11
00:00:34,690 --> 00:00:38,620
is certainly one of the
most intriguing achievements

12
00:00:38,620 --> 00:00:39,160
in vision.

13
00:00:40,900 --> 00:00:43,870
Because the impressions
onto the retinal surface

14
00:00:43,870 --> 00:00:46,560
essentially two dimensional.

15
00:00:46,560 --> 00:00:49,000
And from that, somehow
the brain needs

16
00:00:49,000 --> 00:00:52,470
to reconstruct the
third dimension.

17
00:00:52,470 --> 00:00:54,850
And what is interesting
about this also,

18
00:00:54,850 --> 00:00:59,300
that even in the most primitive
animals, this is a must.

19
00:00:59,300 --> 00:01:02,280
And so annals with
tiny brains also

20
00:01:02,280 --> 00:01:08,370
have mechanisms to be
able to calculate depth

21
00:01:08,370 --> 00:01:11,330
from the information that
comes in through their eyes.

22
00:01:11,330 --> 00:01:15,150
And to demonstrate
that, I have here

23
00:01:15,150 --> 00:01:21,020
a frog, which has a tiny little
brain like that, has big eyes,

24
00:01:21,020 --> 00:01:24,150
and this frog,
for its existence,

25
00:01:24,150 --> 00:01:27,080
needs to know exactly
where things are in depth.

26
00:01:27,080 --> 00:01:29,820
Because if he doesn't,
he would starve to death.

27
00:01:29,820 --> 00:01:32,830
And so what a frog does
looks something like this,

28
00:01:32,830 --> 00:01:33,405
very crudely.

29
00:01:35,170 --> 00:01:40,870
He will stick out his
tongue, grab a flying insect

30
00:01:40,870 --> 00:01:42,940
and consume it.

31
00:01:42,940 --> 00:01:46,440
And because of this
incredible capability,

32
00:01:46,440 --> 00:01:49,930
it is a well-adjusted,
healthy animal

33
00:01:49,930 --> 00:01:52,500
in most parts of the world.

34
00:01:52,500 --> 00:01:55,340
Now the big question
then comes up,

35
00:01:55,340 --> 00:01:57,495
how do we carry out
these computations.

36
00:01:58,710 --> 00:02:01,360
What kind of
mechanisms are involved

37
00:02:01,360 --> 00:02:06,200
in being able to compute
where things are in space,

38
00:02:06,200 --> 00:02:08,830
either in absolute sense
where it is from you

39
00:02:08,830 --> 00:02:11,870
and in a relative sense
where one object is

40
00:02:11,870 --> 00:02:13,560
relative to another one.

41
00:02:13,560 --> 00:02:20,000
Now it turns out that this
became such a serious problem,

42
00:02:20,000 --> 00:02:22,010
in the course of
evolution, that actually

43
00:02:22,010 --> 00:02:26,120
several different
mechanisms have evolved

44
00:02:26,120 --> 00:02:30,090
to make possible our ability
to see things in depth.

45
00:02:30,090 --> 00:02:34,800
And so when one looks at this as
a list, as a fairly brief list,

46
00:02:34,800 --> 00:02:37,890
we can make a distinction
between so called ocular motor

47
00:02:37,890 --> 00:02:40,090
cues and visual cues.

48
00:02:40,090 --> 00:02:44,130
The ocular motor cues are
accommodation and vergence.

49
00:02:44,130 --> 00:02:47,575
So if various objects are
at a very distance from you,

50
00:02:47,575 --> 00:02:49,920
your eyes converge or diverge.

51
00:02:49,920 --> 00:02:53,550
And you your lens gets
thicker and thinner.

52
00:02:53,550 --> 00:02:58,530
And that information can be
utilized in a rather crude way

53
00:02:58,530 --> 00:03:03,670
to tell you about where things
are in that relative to you.

54
00:03:03,670 --> 00:03:07,166
Now as far as visual
cues are concerned,

55
00:03:07,166 --> 00:03:09,310
the very significant
one we are going

56
00:03:09,310 --> 00:03:12,160
to talk about quite a
bit, is a binocular cue,

57
00:03:12,160 --> 00:03:14,490
which is called stereopsis
as you all know.

58
00:03:14,490 --> 00:03:17,300
And then we have a whole
bunch of monocular cues motion

59
00:03:17,300 --> 00:03:20,340
parallax, shading,
interposition, size,

60
00:03:20,340 --> 00:03:21,470
and perspective.

61
00:03:21,470 --> 00:03:24,240
And so we will talk
about many of these

62
00:03:24,240 --> 00:03:28,110
to give you a sense
of what it is like

63
00:03:28,110 --> 00:03:32,040
and to give you a sense of what
various brain structures do

64
00:03:32,040 --> 00:03:36,470
with this as a result
of extensive research

65
00:03:36,470 --> 00:03:38,550
that had been done in this area.

66
00:03:38,550 --> 00:03:42,220
So now, first of all, let's
talk about stereopsis.

67
00:03:42,220 --> 00:03:44,710
And we talk about
stereopsis, we're

68
00:03:44,710 --> 00:03:46,970
going to talk about
the basic facts of it,

69
00:03:46,970 --> 00:03:50,330
and then we are going to
have some demonstrations.

70
00:03:50,330 --> 00:03:55,462
First of all, the so
called stereoscope,

71
00:03:55,462 --> 00:03:59,690
of whichever modern version
that has been handed out to you,

72
00:03:59,690 --> 00:04:04,570
the stereoscope was invented
in the late 19th century.

73
00:04:04,570 --> 00:04:10,170
And when that was done, the
initial approach to this

74
00:04:10,170 --> 00:04:14,280
was to be able to present
to each eye separately

75
00:04:14,280 --> 00:04:18,420
an image that was
taken by a camera that

76
00:04:18,420 --> 00:04:21,829
has two lenses, which are apart
about as much as your two eyes

77
00:04:21,829 --> 00:04:22,840
are apart.

78
00:04:22,840 --> 00:04:24,930
And each of those
created a separate image

79
00:04:24,930 --> 00:04:26,429
of what's out there.

80
00:04:26,429 --> 00:04:28,720
And, of course, each eye gets
a very slightly different

81
00:04:28,720 --> 00:04:30,550
perspective of what's there.

82
00:04:30,550 --> 00:04:33,260
And then when you
present these two images

83
00:04:33,260 --> 00:04:36,030
that you had collected
separately to each eye,

84
00:04:36,030 --> 00:04:38,830
you get a very strong
sense of real depth,

85
00:04:38,830 --> 00:04:40,670
as you will see
in just a minute.

86
00:04:40,670 --> 00:04:44,420
Now another way to do it,
which nowadays is easier

87
00:04:44,420 --> 00:04:48,090
because you can
barely ever find even

88
00:04:48,090 --> 00:04:55,550
one of these two-lens
cameras, even in stores that

89
00:04:55,550 --> 00:04:58,300
sell ancient materials,
antique stores.

90
00:04:59,860 --> 00:05:01,950
So sometimes what
you do instead,

91
00:05:01,950 --> 00:05:05,550
if you only want to take a
picture of a static image,

92
00:05:05,550 --> 00:05:08,330
that you can take a
camera, put in a track,

93
00:05:08,330 --> 00:05:11,774
and have it take two
pictures in succession.

94
00:05:11,774 --> 00:05:13,190
And then you can
do the same thing

95
00:05:13,190 --> 00:05:15,960
as you do with a serial camera.

96
00:05:15,960 --> 00:05:18,390
You can present one to each eye.

97
00:05:18,390 --> 00:05:21,210
OK, so, what we are
going to do now,

98
00:05:21,210 --> 00:05:23,670
we are going to have
a series of demos.

99
00:05:23,670 --> 00:05:28,040
And so we have a handout
for each of you, the paper.

100
00:05:28,040 --> 00:05:30,650
And that you can
keep and take home.

101
00:05:30,650 --> 00:05:34,370
But the stereoscope that
I have for each of you,

102
00:05:34,370 --> 00:05:36,430
that you're going to
have to leave behind,

103
00:05:36,430 --> 00:05:39,650
because I need to use
that in other classes.

104
00:05:39,650 --> 00:05:42,830
So what I want you
to do then, there

105
00:05:42,830 --> 00:05:45,280
are two pictures
on the first page

106
00:05:45,280 --> 00:05:48,880
that you put the stereoscope
down onto the page

107
00:05:48,880 --> 00:05:51,550
so that the vertical
line cuts it in half

108
00:05:51,550 --> 00:05:54,290
so that one goes into each eye.

109
00:05:54,290 --> 00:05:56,250
And then you put your
head right down to it

110
00:05:56,250 --> 00:05:58,320
to look into it, all right?

111
00:05:58,320 --> 00:06:00,950
And if you do that, if you
have it properly sectioned,

112
00:06:00,950 --> 00:06:04,075
you're going to have a sense
that that image is actually

113
00:06:04,075 --> 00:06:04,825
three dimensional.

114
00:06:06,600 --> 00:06:08,990
It's an ancient, ancient
old picture on purpose.

115
00:06:10,200 --> 00:06:15,080
But you should still be
able to see it in depth.

116
00:06:15,080 --> 00:06:17,430
So that's the initial thing.

117
00:06:17,430 --> 00:06:21,310
This became quite a parlor
game and for that case,

118
00:06:21,310 --> 00:06:24,570
for many, many decades,
whenever you went to a party,

119
00:06:24,570 --> 00:06:27,960
they would hand out to you a
stereoscope, a handheld one,

120
00:06:27,960 --> 00:06:30,370
and they would show you
all kinds of images.

121
00:06:30,370 --> 00:06:32,530
And you can even
do this today when

122
00:06:32,530 --> 00:06:36,690
you get on the internet
to find such displays.

123
00:06:36,690 --> 00:06:41,670
Now, then a very important
discovery was made.

124
00:06:41,670 --> 00:06:43,220
I shouldn't say
discovery, really,

125
00:06:43,220 --> 00:06:48,660
I should say an invention was
made by Bela Julesz who came up

126
00:06:48,660 --> 00:06:50,855
with a so-called
random dot stereograms.

127
00:06:52,620 --> 00:06:55,520
By the way, don't look at the
bottom one, that just tells you

128
00:06:55,520 --> 00:06:56,728
what it's going to look like.

129
00:06:56,728 --> 00:06:59,750
There's nothing to look at
the bottom, the bottom set.

130
00:06:59,750 --> 00:07:01,820
Now if you look
in the middle set,

131
00:07:01,820 --> 00:07:04,250
that looks like a
random dot stereogram.

132
00:07:04,250 --> 00:07:09,130
And the idea here was that
the only cue that you provide

133
00:07:09,130 --> 00:07:11,300
is stereo cue, nothing else.

134
00:07:11,300 --> 00:07:12,160
It's pure.

135
00:07:12,160 --> 00:07:14,150
And so what can
be done here, you

136
00:07:14,150 --> 00:07:18,540
can take a section here,
or the same on each side,

137
00:07:18,540 --> 00:07:23,814
and simply move a few pixels,
those images as a unit, over.

138
00:07:23,814 --> 00:07:26,230
And when you do that, they're
going to stick out in depth.

139
00:07:27,720 --> 00:07:33,430
So now take you stereoscope
and look in the middle display,

140
00:07:33,430 --> 00:07:35,770
and you look through it,
you should see something

141
00:07:35,770 --> 00:07:37,130
sticking out in depth.

142
00:07:38,700 --> 00:07:43,000
And the first question I'm going
to ask you is how many of you

143
00:07:43,000 --> 00:07:45,020
can see something
stick out in depth?

144
00:07:46,220 --> 00:07:46,920
What do you see?

145
00:07:48,220 --> 00:07:48,730
You see

146
00:07:48,730 --> 00:07:49,680
AUDIENCE: A square.

147
00:07:49,680 --> 00:07:51,680
PROFESSOR: A little
square sticking out?

148
00:07:51,680 --> 00:07:56,820
All right, so now, don't try
to look at the bottom one.

149
00:07:56,820 --> 00:08:00,140
That simply tells you what the
procedure was in that center

150
00:08:00,140 --> 00:08:02,630
section where you see
the square sticking out.

151
00:08:02,630 --> 00:08:09,130
The pixels were moved a few
steps inward from both the left

152
00:08:09,130 --> 00:08:11,623
and to the right creating
what is called a disparity.

153
00:08:12,680 --> 00:08:16,480
And that's what the brain
then can calculate for depth.

154
00:08:16,480 --> 00:08:20,190
So now to provide you
with the acid test,

155
00:08:20,190 --> 00:08:21,360
go to the second page.

156
00:08:24,510 --> 00:08:27,330
Now you look at the
second page, everybody

157
00:08:27,330 --> 00:08:28,530
see the letter on top?

158
00:08:29,780 --> 00:08:32,840
You don't even have to look
through the stereoscope,

159
00:08:32,840 --> 00:08:34,534
obviously you see
the letter E, right?

160
00:08:35,830 --> 00:08:39,890
That's because that
section is made darker.

161
00:08:39,890 --> 00:08:43,690
But now if you do the same thing
at the bottom, the only cue

162
00:08:43,690 --> 00:08:45,860
you have is the disparity cue.

163
00:08:45,860 --> 00:08:50,065
And the question comes up,
what letter do you see there?

164
00:08:53,230 --> 00:08:57,390
And let me just add, this can be
used as a quick, general test.

165
00:08:57,390 --> 00:09:01,246
You can present
these two subjects,

166
00:09:01,246 --> 00:09:03,620
and if it can present a whole
bunch of different letters,

167
00:09:03,620 --> 00:09:07,050
and if they can see the letters,
that means they can see stereo.

168
00:09:07,050 --> 00:09:10,150
If they cannot see the letters,
then it looks like they may not

169
00:09:10,150 --> 00:09:11,350
see stereo.

170
00:09:11,350 --> 00:09:13,710
Now let me add one
other fact here.

171
00:09:15,570 --> 00:09:20,420
As you move these progressively
closer to each other like this,

172
00:09:20,420 --> 00:09:22,050
you increase the disparity.

173
00:09:22,050 --> 00:09:25,645
And that causes the image to
be seen at increasing depths.

174
00:09:27,390 --> 00:09:29,570
Our sensitivity is
so great that when

175
00:09:29,570 --> 00:09:33,410
you look at this on a
computer, a standard computer,

176
00:09:33,410 --> 00:09:36,750
if you move those
images just one

177
00:09:36,750 --> 00:09:39,190
pixel from the left
to the right, then

178
00:09:39,190 --> 00:09:42,240
from right to the left,
you will see it in depth.

179
00:09:42,240 --> 00:09:46,165
And even monkeys can see
as small step as one pixel.

180
00:09:47,170 --> 00:09:48,690
So now how many
of you can tell me

181
00:09:48,690 --> 00:09:51,076
what was the letter
at the bottom there?

182
00:09:51,076 --> 00:09:51,576
AUDIENCE: H.

183
00:09:51,576 --> 00:09:52,730
PROFESSOR: H, H, good.

184
00:09:52,730 --> 00:09:55,660
Anybody not being able
to see the letter?

185
00:09:55,660 --> 00:09:57,210
Everybody sees it.

186
00:09:57,210 --> 00:09:58,300
Well, you guys are lucky.

187
00:09:58,300 --> 00:10:01,080
Because there are a
significant number

188
00:10:01,080 --> 00:10:04,780
of people in the world
who lack stereopsis,

189
00:10:04,780 --> 00:10:08,330
something like 5% to
10% of the population

190
00:10:08,330 --> 00:10:14,260
lacks stereopsis for
a variety of reasons.

191
00:10:14,260 --> 00:10:16,200
We'll talk about that
a bit more later.

192
00:10:16,200 --> 00:10:20,450
But one is sometimes
you're born,

193
00:10:20,450 --> 00:10:22,200
and you're amblyopic in one eye.

194
00:10:23,220 --> 00:10:26,100
Sometimes you are
strabismic, which

195
00:10:26,100 --> 00:10:29,340
means that your two
eyes are not aligned,

196
00:10:29,340 --> 00:10:32,080
which in commonplace
language is often

197
00:10:32,080 --> 00:10:35,410
called as being cross
eyed or wall eyed.

198
00:10:35,410 --> 00:10:38,220
Those types of
people very seldom

199
00:10:38,220 --> 00:10:42,880
will have stereoscopic
depth perception,

200
00:10:42,880 --> 00:10:46,035
even after it's
corrected, especially

201
00:10:46,035 --> 00:10:51,470
if the correction is made by the
time you're 8 or 10 years old,

202
00:10:51,470 --> 00:10:52,660
the correction won't help.

203
00:10:52,660 --> 00:10:55,400
It has to be done
much, much earlier.

204
00:10:55,400 --> 00:10:59,800
All right, so that then
is the very, very basics

205
00:10:59,800 --> 00:11:04,520
of the stereo procedures.

206
00:11:04,520 --> 00:11:07,720
And now another
procedure that had

207
00:11:07,720 --> 00:11:11,040
been developed more
recently is one

208
00:11:11,040 --> 00:11:13,100
which is called the
auto stereogram.

209
00:11:13,100 --> 00:11:16,610
So then if you go to the next
page, and what you want to do

210
00:11:16,610 --> 00:11:19,480
is you want to look at
this horizontally like that

211
00:11:19,480 --> 00:11:20,530
with the T on top.

212
00:11:23,240 --> 00:11:26,280
And then just look at it at
sort of, I don't know, maybe

213
00:11:26,280 --> 00:11:30,540
about 20 inches from you,
normal reading length.

214
00:11:30,540 --> 00:11:34,460
And what you want to do
is to look beyond it.

215
00:11:34,460 --> 00:11:36,040
So stare beyond it.

216
00:11:36,040 --> 00:11:38,600
And if you keep doing
that for a while,

217
00:11:38,600 --> 00:11:42,750
you will suddenly see an image,
a three-dimensional image,

218
00:11:42,750 --> 00:11:46,550
as this comes actually from
a book called The Magic Eye.

219
00:11:46,550 --> 00:11:48,120
There are several
magic eye books

220
00:11:48,120 --> 00:11:49,990
in which all kinds
of displays are

221
00:11:49,990 --> 00:11:52,820
done using these
auto stereograms.

222
00:11:53,930 --> 00:11:57,820
Does everybody see-- who
can see what's sticking out?

223
00:11:58,900 --> 00:12:00,009
OK, what do you see?

224
00:12:00,009 --> 00:12:01,050
AUDIENCE: Was it a shark?

225
00:12:01,050 --> 00:12:02,450
PROFESSOR: You see a shark?

226
00:12:02,450 --> 00:12:03,110
OK.

227
00:12:03,110 --> 00:12:07,590
Now let me see if any
of you don't see it.

228
00:12:07,590 --> 00:12:08,640
Keep staring at it.

229
00:12:08,640 --> 00:12:10,170
Look beyond it.

230
00:12:10,170 --> 00:12:12,190
Another thing that
helps, if you look at it,

231
00:12:12,190 --> 00:12:14,200
bring it a little
closer to you so you

232
00:12:14,200 --> 00:12:16,590
can look beyond it
easier, gradually

233
00:12:16,590 --> 00:12:17,910
move it back and forth.

234
00:12:17,910 --> 00:12:19,900
And if you're patient,
eventually you

235
00:12:19,900 --> 00:12:21,970
may be able to do this.

236
00:12:21,970 --> 00:12:24,410
The reason this is
difficult is because you

237
00:12:24,410 --> 00:12:27,960
have to uncouple the
vergence in your two eyes.

238
00:12:27,960 --> 00:12:29,565
You have to look
beyond it slightly.

239
00:12:30,630 --> 00:12:34,850
And, in fact, that is
one of the reasons why

240
00:12:34,850 --> 00:12:37,980
testing people for
stereopsis, an auto stereogram

241
00:12:37,980 --> 00:12:40,210
is not a very good procedure.

242
00:12:40,210 --> 00:12:44,160
Whereas virtually everybody
can use a stereoscope

243
00:12:44,160 --> 00:12:45,020
without any trouble.

244
00:12:49,199 --> 00:12:50,240
AUDIENCE: That's so cool.

245
00:12:50,240 --> 00:12:51,685
PROFESSOR: Did you
get it finally?

246
00:12:51,685 --> 00:12:53,170
AUDIENCE: Yeah, that's so cool.

247
00:12:53,170 --> 00:12:55,540
PROFESSOR: Yeah, all
right so, if anybody

248
00:12:55,540 --> 00:12:58,180
is really interested in
this auto stereograms

249
00:12:58,180 --> 00:13:01,440
is I say go to the
store, the bookstore,

250
00:13:01,440 --> 00:13:04,090
and get one of those
magic eye books.

251
00:13:04,090 --> 00:13:05,320
They're just a lot of fun.

252
00:13:05,320 --> 00:13:07,050
And you can just
leaf through it.

253
00:13:07,050 --> 00:13:09,670
You don't even have to buy
the book, just look through it

254
00:13:09,670 --> 00:13:10,720
at the store.

255
00:13:10,720 --> 00:13:14,010
And you'll see one interesting,
clever image after the next.

256
00:13:15,550 --> 00:13:17,490
All right, so that's
the stereoscope.

257
00:13:17,490 --> 00:13:20,660
And now let me
explain to, I think

258
00:13:20,660 --> 00:13:25,430
I've mentioned this briefly
before, the principles involved

259
00:13:25,430 --> 00:13:30,530
behind being able to see
stereoscopic depth perception.

260
00:13:31,550 --> 00:13:34,010
And what I've mentioned
to you before was

261
00:13:34,010 --> 00:13:36,370
that if you have the
two eyes fixating

262
00:13:36,370 --> 00:13:40,114
at a particular distance, if you
then draw a circle around that,

263
00:13:40,114 --> 00:13:42,030
that's sometimes called
a Veith-Muller circle,

264
00:13:42,030 --> 00:13:46,620
or the sometimes called a
horopter, then any target,

265
00:13:46,620 --> 00:13:49,400
like this one here, will
hit equivalent points

266
00:13:49,400 --> 00:13:52,840
on the retinal surface of
the left and right eyes.

267
00:13:52,840 --> 00:13:55,770
However, if you
do the same thing,

268
00:13:55,770 --> 00:13:58,620
and you put a target
either beyond or closer

269
00:13:58,620 --> 00:14:01,430
than the Vieth-Muller
circle, then they're

270
00:14:01,430 --> 00:14:05,390
going to hit nonequivalent
points on the retinal surface.

271
00:14:05,390 --> 00:14:09,920
So then by nonequivalency, we
can do this and calculate this

272
00:14:09,920 --> 00:14:12,940
as to where the
image falls relative

273
00:14:12,940 --> 00:14:15,960
to the central fixation
spot in the foveola.

274
00:14:18,070 --> 00:14:21,780
And so then, when these
nonequivalent points are hit,

275
00:14:21,780 --> 00:14:24,410
somehow the brain can
measure this nonequivalence.

276
00:14:25,900 --> 00:14:29,690
And that is then
converted into an estimate

277
00:14:29,690 --> 00:14:31,650
of where things are in depth.

278
00:14:31,650 --> 00:14:34,490
Now the idea behind
this was that

279
00:14:34,490 --> 00:14:38,150
these nonequivalent
points, that you

280
00:14:38,150 --> 00:14:42,230
have on the retinal surface,
can connect in the cortex

281
00:14:42,230 --> 00:14:43,980
to single cells.

282
00:14:43,980 --> 00:14:47,910
So they have a cell in the
cortex that is binocular,

283
00:14:47,910 --> 00:14:49,310
by virtue of the
fact that inputs

284
00:14:49,310 --> 00:14:50,590
from the left and right eyes.

285
00:14:50,590 --> 00:14:56,765
But they don't necessarily have
to come from equivalent points.

286
00:14:56,765 --> 00:14:59,080
They can come from
nonequivalent points.

287
00:14:59,080 --> 00:15:03,030
And that may be then
the mechanism whereby

288
00:15:03,030 --> 00:15:06,640
it can tell you the
degree of nonequivalence,

289
00:15:06,640 --> 00:15:08,720
and, therefore, convert
that into depth.

290
00:15:08,720 --> 00:15:10,710
And, therefore, there
could be single neurons

291
00:15:10,710 --> 00:15:14,210
in the brain that are
selective to certain depths.

292
00:15:14,210 --> 00:15:17,630
And so people began to do all
kinds of experiments with this.

293
00:15:17,630 --> 00:15:20,340
And the way these
experiments were done

294
00:15:20,340 --> 00:15:22,080
is you presented
images separately

295
00:15:22,080 --> 00:15:23,850
to the left eyes and right eyes.

296
00:15:25,310 --> 00:15:32,080
And you could then present them
to both eyes at the same time

297
00:15:32,080 --> 00:15:35,940
and vary the amount of
disparity systematically

298
00:15:35,940 --> 00:15:38,390
to see what kind
of tuning function

299
00:15:38,390 --> 00:15:40,130
you would get in the cortex.

300
00:15:40,130 --> 00:15:42,960
So this, some of the most
beautiful work of this

301
00:15:42,960 --> 00:15:46,650
was done by a person
called John [? Porgio ?]

302
00:15:46,650 --> 00:15:49,970
And I will tell you briefly
about some of his experiments.

303
00:15:49,970 --> 00:15:51,450
So here we go with then.

304
00:15:51,450 --> 00:15:53,670
We're going to look at
the neural responses,

305
00:15:53,670 --> 00:16:00,235
neural responses in V1 as
initially done in the monkey.

306
00:16:01,410 --> 00:16:03,410
So here's an example of a cell.

307
00:16:03,410 --> 00:16:07,703
We have here different
degrees of disparity.

308
00:16:09,690 --> 00:16:13,840
And we have the
neuron responding

309
00:16:13,840 --> 00:16:16,480
each time they're
four repeated trials.

310
00:16:16,480 --> 00:16:20,730
And you can see the action
potentials by these dark lines

311
00:16:20,730 --> 00:16:21,580
here.

312
00:16:21,580 --> 00:16:24,700
And what you do is you move
the stimuli back and forth

313
00:16:24,700 --> 00:16:28,320
across the two eyes, the
way it's actually done,

314
00:16:28,320 --> 00:16:29,130
you have a mirror.

315
00:16:29,130 --> 00:16:31,410
And then you have two
in this experiment.

316
00:16:31,410 --> 00:16:35,480
You have two monitors, one to
the left and one to the right.

317
00:16:35,480 --> 00:16:37,264
And then you can set
it up almost exactly

318
00:16:37,264 --> 00:16:39,305
the same as what you would
do with a stereoscope.

319
00:16:40,370 --> 00:16:43,810
So this particular cell,
as you can readily see,

320
00:16:43,810 --> 00:16:46,130
responds best when
there's zero disparity.

321
00:16:48,180 --> 00:16:53,330
Now by contrast, here is a
cell that responds vigorously

322
00:16:53,330 --> 00:16:57,800
at the far disparity and
not to the close one.

323
00:16:59,560 --> 00:17:03,010
So then, when you do this, you
can study hundreds of cells

324
00:17:03,010 --> 00:17:05,730
to see what kinds
of distributions

325
00:17:05,730 --> 00:17:09,471
you have in the cortex for
different degrees of disparity,

326
00:17:09,471 --> 00:17:10,054
cell activity.

327
00:17:11,130 --> 00:17:14,510
Now again, this
hearkens back to what

328
00:17:14,510 --> 00:17:18,010
I talked about with
respect to color vision

329
00:17:18,010 --> 00:17:20,680
where the question came up,
if you want to see color,

330
00:17:20,680 --> 00:17:23,670
how many receptors
would you need that peak

331
00:17:23,670 --> 00:17:26,220
at different wavelengths, right?

332
00:17:26,220 --> 00:17:31,300
And one idea was that maybe you
need as many photo receptors

333
00:17:31,300 --> 00:17:32,290
as there are colors.

334
00:17:32,290 --> 00:17:35,410
But in the end, it turned
out that we have only three

335
00:17:35,410 --> 00:17:36,440
of them.

336
00:17:36,440 --> 00:17:39,430
And on the basis of that, we
can recreate all the colors out

337
00:17:39,430 --> 00:17:40,280
there.

338
00:17:40,280 --> 00:17:43,590
Now the same thing
applies to stereopsis.

339
00:17:43,590 --> 00:17:45,450
So when this was
done systematically,

340
00:17:45,450 --> 00:17:50,170
here's an example of
the tuning functions.

341
00:17:50,170 --> 00:17:52,990
This one here is the
same, very much the same,

342
00:17:52,990 --> 00:17:55,260
as the first figure
I showed you.

343
00:17:57,310 --> 00:17:59,420
And so we have a bunch
of different ones.

344
00:17:59,420 --> 00:18:02,430
And if you then study, as
I've said hundreds of cells,

345
00:18:02,430 --> 00:18:05,640
you can come up with a
distribution of this.

346
00:18:05,640 --> 00:18:08,060
And what John
[? Porgio ?] came up with,

347
00:18:08,060 --> 00:18:10,405
he thought that there
were four major classes.

348
00:18:11,510 --> 00:18:13,680
And the relative amount
of activity of these four

349
00:18:13,680 --> 00:18:17,570
major classes then is used
to compute all the very fine

350
00:18:17,570 --> 00:18:19,020
differences in depth.

351
00:18:19,020 --> 00:18:20,370
So there's a right on one.

352
00:18:22,460 --> 00:18:27,730
This cell is right
on the fixation spot.

353
00:18:27,730 --> 00:18:29,880
And then you have
near and far cells.

354
00:18:29,880 --> 00:18:31,920
And you have some
in between cells.

355
00:18:31,920 --> 00:18:34,720
Initially he thought
there were four classes.

356
00:18:34,720 --> 00:18:37,670
Some people argue that
there may be as many as six.

357
00:18:37,670 --> 00:18:40,030
But at any rate,
there's a limited number

358
00:18:40,030 --> 00:18:42,190
on the basis of which
you can calculate

359
00:18:42,190 --> 00:18:44,960
almost an unlimited
number of depths,

360
00:18:44,960 --> 00:18:46,590
which is quite remarkable.

361
00:18:46,590 --> 00:18:51,550
All right, so now, what
you can next turn to

362
00:18:51,550 --> 00:18:56,740
is to ask the question to what
degree do various extrastriate

363
00:18:56,740 --> 00:19:02,140
areas contribute to
stereoscopic depth perception.

364
00:19:02,140 --> 00:19:05,390
And some people
thought that this

365
00:19:05,390 --> 00:19:08,590
is a unique function
for area MT,

366
00:19:08,590 --> 00:19:10,620
some people argue
that maybe it's

367
00:19:10,620 --> 00:19:16,940
area V4, and so experiments were
done in which it was examined

368
00:19:16,940 --> 00:19:23,920
to what degree stereoscopic
depth deception is altered when

369
00:19:23,920 --> 00:19:31,180
you eliminate, say, area MT
or you eliminate area V4.

370
00:19:31,180 --> 00:19:34,725
So that is what's been done.

371
00:19:34,725 --> 00:19:36,350
And you can think
about it for a minute

372
00:19:36,350 --> 00:19:38,070
and say, well,
what do you think?

373
00:19:38,070 --> 00:19:42,160
What do you think would
happen in a monkey

374
00:19:42,160 --> 00:19:44,300
once you no longer had area V4?

375
00:19:44,300 --> 00:19:46,950
What you think would happen
if the monkey no longer had

376
00:19:46,950 --> 00:19:47,535
area MT?

377
00:19:48,740 --> 00:19:51,990
Well, the results were
actually quite surprising.

378
00:19:51,990 --> 00:19:55,720
And they're shown here,
same experiment as before.

379
00:19:55,720 --> 00:19:57,645
The Bela Julesz random
dot stereograms.

380
00:19:57,645 --> 00:20:00,060
And then you're presenting
one of four locations

381
00:20:00,060 --> 00:20:05,090
of however many
little area where,

382
00:20:05,090 --> 00:20:07,640
like little square,
that's sticks out in depth

383
00:20:07,640 --> 00:20:10,470
and you vary the amount
of depth that sticks out

384
00:20:10,470 --> 00:20:12,830
by varying the
number of pixels you

385
00:20:12,830 --> 00:20:17,700
moved the images
into this place.

386
00:20:17,700 --> 00:20:20,520
And when that was
done systematically,

387
00:20:20,520 --> 00:20:21,680
this is what was found.

388
00:20:21,680 --> 00:20:26,830
It was found that neither
V4 lesion nor an MT lesion

389
00:20:26,830 --> 00:20:29,070
cause a significant deficit.

390
00:20:29,070 --> 00:20:32,060
The only deficit
that was significant

391
00:20:32,060 --> 00:20:34,660
had to do with a
response latency.

392
00:20:34,660 --> 00:20:37,660
And as I should have
mentioned earlier,

393
00:20:37,660 --> 00:20:40,620
like when we talked
about the frog, one

394
00:20:40,620 --> 00:20:43,090
of the very important things
about processing depth,

395
00:20:43,090 --> 00:20:45,440
again, is to be able
to do it quickly.

396
00:20:45,440 --> 00:20:48,380
So when you have that's frog
and the fly is flying along,

397
00:20:48,380 --> 00:20:50,640
he has to be very
quick to compute it

398
00:20:50,640 --> 00:20:52,860
so that he can catch it,
right, as you had seen.

399
00:20:54,220 --> 00:20:57,090
So in this case,
what you see here,

400
00:20:57,090 --> 00:20:59,960
that there is about a 20
millisecond difference

401
00:20:59,960 --> 00:21:04,700
after V4 lesion, increase in
latency, and quite a bit more,

402
00:21:04,700 --> 00:21:09,090
almost 40, 30, 40
after and MT lesion.

403
00:21:09,090 --> 00:21:13,510
So that contributes to some
aspect of depth processing

404
00:21:13,510 --> 00:21:16,100
in terms of being
able to do it quickly.

405
00:21:16,100 --> 00:21:21,840
But neither the MT or V4
are unique in processing

406
00:21:21,840 --> 00:21:22,820
stereopsis.

407
00:21:22,820 --> 00:21:25,190
It looks like that
it is processed

408
00:21:25,190 --> 00:21:28,180
in several different
areas in the brain

409
00:21:28,180 --> 00:21:32,560
and inspire conjoint
computation that you

410
00:21:32,560 --> 00:21:36,290
can arrive at the actual depth.

411
00:21:36,290 --> 00:21:38,890
And it's by virtue of
that joint computation

412
00:21:38,890 --> 00:21:40,345
that you can do
this very quickly.

413
00:21:41,790 --> 00:21:49,970
So now you come to the
next important depth

414
00:21:49,970 --> 00:21:52,380
cue, which is called
motion parallax.

415
00:21:53,790 --> 00:22:01,440
This one is a capacity
that we had acquired

416
00:22:01,440 --> 00:22:03,160
in the course of
evolution, which

417
00:22:03,160 --> 00:22:05,800
is extremely potent
and powerful.

418
00:22:05,800 --> 00:22:09,170
And it's based on a very
simple physical fact.

419
00:22:09,170 --> 00:22:16,274
And the physical fact it that
either when you are in motion,

420
00:22:16,274 --> 00:22:17,690
or something in
the environment is

421
00:22:17,690 --> 00:22:23,100
in motion, the rate at
which these images travel

422
00:22:23,100 --> 00:22:29,330
across the retinal surface is
heavily distance dependent.

423
00:22:29,330 --> 00:22:32,120
And so let me demonstrate
this concretely for you.

424
00:22:33,830 --> 00:22:36,010
Here we have an
eye that's fixed.

425
00:22:36,010 --> 00:22:37,780
It's always looking
straight ahead.

426
00:22:37,780 --> 00:22:43,680
And we have two rods
here, sorry, just one rod,

427
00:22:43,680 --> 00:22:46,460
that we are going to move into
position gradually from here

428
00:22:46,460 --> 00:22:49,420
to here, and then back up
as shown by the arrows here.

429
00:22:49,420 --> 00:22:53,300
And we examine the
range over which

430
00:22:53,300 --> 00:22:56,790
these near and far and
middle objects move

431
00:22:56,790 --> 00:23:00,720
across the retinal surface
when you engage in this motion

432
00:23:00,720 --> 00:23:02,160
and the eye is stable.

433
00:23:02,160 --> 00:23:04,020
And you can see that
the far object moves

434
00:23:04,020 --> 00:23:06,685
over a much shorter distance
than the near object.

435
00:23:08,990 --> 00:23:11,720
This you can readily do it
yourself in an experiment.

436
00:23:11,720 --> 00:23:14,660
You can stick out your thumb and
move your head back and forth.

437
00:23:14,660 --> 00:23:20,330
And you see that your thumb will
move a lot more than the object

438
00:23:20,330 --> 00:23:21,330
that you're looking at.

439
00:23:22,660 --> 00:23:27,070
So now the same thing also
applies when you actually

440
00:23:27,070 --> 00:23:29,310
are engaged in the eye
movements, which of course you

441
00:23:29,310 --> 00:23:30,420
do all the time.

442
00:23:30,420 --> 00:23:35,790
In this case, they eye is set up
so that it's fixated initially

443
00:23:35,790 --> 00:23:38,020
on this object and
then tracks it to here

444
00:23:38,020 --> 00:23:39,560
and then tracks it back.

445
00:23:39,560 --> 00:23:42,830
And when you do that, you
get the same kind of effect,

446
00:23:42,830 --> 00:23:46,460
namely that the distance over
which a far and a near object

447
00:23:46,460 --> 00:23:50,080
move is quite different.

448
00:23:50,080 --> 00:23:53,280
The near object moves over
much, much greater distance

449
00:23:53,280 --> 00:23:56,290
than the far object, even
though the eyes are tracking.

450
00:23:57,460 --> 00:24:03,060
So this, then, being
a basic physical fact,

451
00:24:03,060 --> 00:24:08,780
was then used in the
course of evolution

452
00:24:08,780 --> 00:24:12,350
to create mechanisms that are
sensitive to this differential

453
00:24:12,350 --> 00:24:12,850
motion.

454
00:24:14,080 --> 00:24:16,790
And, of course, because
of the rate of motion

455
00:24:16,790 --> 00:24:20,320
also varies a little
bit, it became

456
00:24:20,320 --> 00:24:24,000
possible to create mechanisms
to make that computation

457
00:24:24,000 --> 00:24:26,180
to tell you where
things are in depth.

458
00:24:26,180 --> 00:24:30,420
So here I'm going to show
you an actual demo of this

459
00:24:30,420 --> 00:24:32,620
to make it clear to you.

460
00:24:32,620 --> 00:24:37,100
In this case, again, we have
a bunch of random dots, much

461
00:24:37,100 --> 00:24:40,280
like in the Bela Julesz
random dot stereograms

462
00:24:40,280 --> 00:24:41,890
but just a single one.

463
00:24:41,890 --> 00:24:43,900
And everybody agrees
there's no depth here.

464
00:24:43,900 --> 00:24:44,345
Is there any depth?

465
00:24:44,345 --> 00:24:45,300
Do you see any depth?

466
00:24:47,890 --> 00:24:50,130
So now what I'm
going to do is I'm

467
00:24:50,130 --> 00:24:54,230
going to set this image
into rocking motion.

468
00:24:54,230 --> 00:24:57,090
And when I do this,
almost instantly, you're

469
00:24:57,090 --> 00:24:58,560
going to see something in depth.

470
00:24:58,560 --> 00:24:59,230
Are you ready?

471
00:25:02,670 --> 00:25:06,210
So what you see here is
are three levels, right,

472
00:25:06,210 --> 00:25:07,320
very clearly.

473
00:25:07,320 --> 00:25:11,350
In milliseconds, in 20
milliseconds, you can see this.

474
00:25:11,350 --> 00:25:14,260
And let me explain to
you why you see this.

475
00:25:14,260 --> 00:25:17,550
If, OK, let me go
back and do it again.

476
00:25:17,550 --> 00:25:21,810
If I keep this stable, you can
see that the dots move over

477
00:25:21,810 --> 00:25:24,250
a great distance here,
a lesser distance

478
00:25:24,250 --> 00:25:26,770
here, and practically
not at all here.

479
00:25:27,930 --> 00:25:30,470
So there is a
differential motion.

480
00:25:30,470 --> 00:25:33,700
And the greater the motion,
the closer the image

481
00:25:33,700 --> 00:25:36,760
is in your analysis.

482
00:25:36,760 --> 00:25:39,300
So that's called
motion parallax.

483
00:25:40,730 --> 00:25:43,040
And then what you
can do actually,

484
00:25:43,040 --> 00:25:45,346
you can play all kinds
of games, do experiments

485
00:25:45,346 --> 00:25:48,990
in which you can present
this kind of image.

486
00:25:48,990 --> 00:25:51,700
You can put this into each
of the eyes separately.

487
00:25:51,700 --> 00:25:54,790
And you can present
this image alone

488
00:25:54,790 --> 00:26:01,240
or you can present it paired
with disparity for stereopsis.

489
00:26:04,460 --> 00:26:07,970
And you can do each separately
or you can do the two together.

490
00:26:07,970 --> 00:26:16,060
So let's now first summarize
the essence of motion parallax.

491
00:26:16,060 --> 00:26:18,710
To derive depth information
from motion parallax,

492
00:26:18,710 --> 00:26:21,480
neurons are needed that
provide information

493
00:26:21,480 --> 00:26:24,380
about velocity and
direction of motion

494
00:26:24,380 --> 00:26:27,190
and perhaps also about
differential motion.

495
00:26:27,190 --> 00:26:30,710
Secondly, the
majority of V1 cells

496
00:26:30,710 --> 00:26:33,240
are direction and
velocity selective,

497
00:26:33,240 --> 00:26:36,410
as we had discussed before,
and some appear also

498
00:26:36,410 --> 00:26:40,200
to be selective for
differential motion, which

499
00:26:40,200 --> 00:26:42,100
I did not mention before.

500
00:26:42,100 --> 00:26:44,630
But, indeed, there are such
cells in the visual cortex.

501
00:26:46,250 --> 00:26:48,410
Now, the third
important point is

502
00:26:48,410 --> 00:26:54,170
that such cells that are
motion selective and direction

503
00:26:54,170 --> 00:26:57,870
selective and selective
for differential motion

504
00:26:57,870 --> 00:26:59,960
are very, very common area MT.

505
00:27:01,660 --> 00:27:04,580
So those are some of the
very, very basic facts.

506
00:27:04,580 --> 00:27:10,780
And now we can move on and ask
what kind of brain activation

507
00:27:10,780 --> 00:27:12,730
occurs by stereopsis
and motion parallax

508
00:27:12,730 --> 00:27:15,180
in normal and serial
blind subjects

509
00:27:15,180 --> 00:27:17,840
using a recently
developed technique, which

510
00:27:17,840 --> 00:27:22,940
is magnetic resonance imaging,
functional, functional

511
00:27:22,940 --> 00:27:24,520
magnetic resonance imaging.

512
00:27:24,520 --> 00:27:26,530
So how do you do
this kind of stuff?

513
00:27:26,530 --> 00:27:30,000
Well, what you do here,
here's an example,

514
00:27:30,000 --> 00:27:34,180
you have a very
large stereoscope

515
00:27:34,180 --> 00:27:35,680
with a mirror at the end.

516
00:27:35,680 --> 00:27:37,880
And you have a subject
who is lying down.

517
00:27:37,880 --> 00:27:41,340
And this whole unit, except
not of course that part,

518
00:27:41,340 --> 00:27:44,000
is put into the magnet.

519
00:27:44,000 --> 00:27:46,200
And we have a magnet
down here at MIT.

520
00:27:46,200 --> 00:27:47,790
Most of you probably
have seen that.

521
00:27:47,790 --> 00:27:49,870
It's on the ground floor.

522
00:27:49,870 --> 00:27:53,990
So you can do this, and then you
can present those images here.

523
00:27:53,990 --> 00:27:57,200
And so the stereoscope
will present two images

524
00:27:57,200 --> 00:28:01,740
and then you can vary this
by rocking him back and forth

525
00:28:01,740 --> 00:28:06,140
either to present
only motion parallax

526
00:28:06,140 --> 00:28:09,860
or present only stereopsis
and to present both.

527
00:28:09,860 --> 00:28:12,515
And so now the question is, this
is a very primitive question

528
00:28:12,515 --> 00:28:17,070
at this stage, where
in the brain are

529
00:28:17,070 --> 00:28:21,630
these processes analyzed?

530
00:28:21,630 --> 00:28:25,110
And so you can find out
what brain areas are

531
00:28:25,110 --> 00:28:29,770
active by doing this repeatedly
collecting the fMRI data

532
00:28:29,770 --> 00:28:32,140
and then printing
them out and looking

533
00:28:32,140 --> 00:28:34,190
at them to see what happens.

534
00:28:34,190 --> 00:28:36,810
So I'm going to show you
a couple examples of that.

535
00:28:36,810 --> 00:28:40,710
Here is the basic
figure that the person

536
00:28:40,710 --> 00:28:43,340
sees but done in such a
way that you can see it.

537
00:28:43,340 --> 00:28:45,280
Of course, he doesn't
see anything like this.

538
00:28:45,280 --> 00:28:47,440
He just sees different depths.

539
00:28:47,440 --> 00:28:49,400
There are one, two,
three, four, five, six,

540
00:28:49,400 --> 00:28:51,100
seven different depths here.

541
00:28:52,140 --> 00:28:53,610
And this rocks back and forth.

542
00:28:54,660 --> 00:28:57,110
And then you can, as I
say, present this only

543
00:28:57,110 --> 00:29:00,830
with differential motion
or you can present it only

544
00:29:00,830 --> 00:29:05,870
with disparity or you
can present it with both.

545
00:29:05,870 --> 00:29:08,107
And then finally, as a
control, what you do is

546
00:29:08,107 --> 00:29:09,190
you can do the same thing.

547
00:29:09,190 --> 00:29:11,300
But you don't have
any depth of any sort.

548
00:29:11,300 --> 00:29:14,810
You just have a flat surface
rocking back and forth.

549
00:29:14,810 --> 00:29:17,130
And then when you
do data analysis,

550
00:29:17,130 --> 00:29:21,070
you actually subtract the last
one from the rest of the data

551
00:29:21,070 --> 00:29:24,280
so that you're not looking at
the data for the activation

552
00:29:24,280 --> 00:29:26,840
just by the spots but
for the activation that's

553
00:29:26,840 --> 00:29:30,390
specific for stereopsis
or motion parallax.

554
00:29:30,390 --> 00:29:32,090
So now if you do
this experiment,

555
00:29:32,090 --> 00:29:34,770
here's an example
of a normal subject

556
00:29:34,770 --> 00:29:36,440
and a stereobind subject.

557
00:29:36,440 --> 00:29:42,100
And we have here a sagittal cut
adding up the images sideways.

558
00:29:42,100 --> 00:29:45,740
And what you see here, this is
posterior cortex, of course.

559
00:29:45,740 --> 00:29:49,430
Here in the normal subject,
when you present only motion

560
00:29:49,430 --> 00:29:52,180
parallax, you only
analyze motion parallax.

561
00:29:52,180 --> 00:29:56,890
But you have a huge amount of
activation in the visual areas.

562
00:29:56,890 --> 00:29:59,040
And then if you
do the same thing

563
00:29:59,040 --> 00:30:02,170
when you do a binocular
stereopsis only,

564
00:30:02,170 --> 00:30:04,465
you also get a great
deal of activation,

565
00:30:04,465 --> 00:30:05,915
in quite similar set of areas.

566
00:30:07,030 --> 00:30:09,770
And then the big
crucial test comes up.

567
00:30:09,770 --> 00:30:13,320
What happens if you present
the stereo under monocular

568
00:30:13,320 --> 00:30:15,860
conditions when you
don't see stereo?

569
00:30:15,860 --> 00:30:20,470
And if you do that, using this
same calculation procedures,

570
00:30:20,470 --> 00:30:23,530
there is no brain
activation here.

571
00:30:23,530 --> 00:30:27,940
And, therefore, what we see here
is due, indeed, to the analysis

572
00:30:27,940 --> 00:30:31,970
that we do for stereopsis.

573
00:30:31,970 --> 00:30:33,790
Now if you do the
same experiment

574
00:30:33,790 --> 00:30:38,275
in a stereoblind subject who
has been tested, on tests

575
00:30:38,275 --> 00:30:43,150
similar we had shown you,
when that person even

576
00:30:43,150 --> 00:30:46,190
looks at it under
binocular conditions,

577
00:30:46,190 --> 00:30:48,660
there is no brain
activation meaning

578
00:30:48,660 --> 00:30:51,540
that this person doesn't have
any mechanisms in the brain

579
00:30:51,540 --> 00:30:54,400
to analyze stereopsis.

580
00:30:54,400 --> 00:30:56,970
Now the fortunate
thing is that we

581
00:30:56,970 --> 00:30:59,380
have these several
different mechanisms

582
00:30:59,380 --> 00:31:00,680
for depth perception.

583
00:31:00,680 --> 00:31:04,190
And so people who are
stereoblind and have

584
00:31:04,190 --> 00:31:10,320
no analysis for disparity, they
can still see depth reasonably

585
00:31:10,320 --> 00:31:11,490
well.

586
00:31:11,490 --> 00:31:13,700
And, indeed, they can get
driver's license and all

587
00:31:13,700 --> 00:31:17,690
that, because we have all these
other mechanisms that include,

588
00:31:17,690 --> 00:31:21,520
that we have talked
about, motion parallax.

589
00:31:23,810 --> 00:31:26,830
So that then is one
way of looking at it.

590
00:31:26,830 --> 00:31:31,370
Now the other way to look at it,
especially when I ask as well,

591
00:31:31,370 --> 00:31:34,600
are the same brain areas
doing both or what.

592
00:31:34,600 --> 00:31:36,660
And so what you
can do is instead

593
00:31:36,660 --> 00:31:40,410
of doing a sagittal section,
you can take sections coronally

594
00:31:40,410 --> 00:31:42,240
like bang, bang,
bang, bang like that

595
00:31:42,240 --> 00:31:43,860
and see what that looks like.

596
00:31:43,860 --> 00:31:48,130
And here's an example in which
I've isolated the stereo.

597
00:31:48,130 --> 00:31:49,830
And here are a
bunch of sections.

598
00:31:49,830 --> 00:31:54,435
And this shows the
activation for stereopsis.

599
00:31:56,300 --> 00:31:58,084
You can see there are
all kinds of areas

600
00:31:58,084 --> 00:31:59,125
that are being activated.

601
00:32:00,840 --> 00:32:02,470
And then if you
do the same thing

602
00:32:02,470 --> 00:32:05,340
and just look at
the parallax alone,

603
00:32:05,340 --> 00:32:08,720
here we have the
activation for that.

604
00:32:08,720 --> 00:32:11,980
And then lastly here,
we do both of them.

605
00:32:11,980 --> 00:32:13,180
Now we can go back.

606
00:32:13,180 --> 00:32:15,135
The way to look at
the question, how

607
00:32:15,135 --> 00:32:21,210
are these two areas, these
two types of depth perception

608
00:32:21,210 --> 00:32:24,450
analyses, differently
activating in the brain.

609
00:32:24,450 --> 00:32:26,620
And so I'm going to
go back and forth

610
00:32:26,620 --> 00:32:29,740
between stereo and parallax.

611
00:32:29,740 --> 00:32:31,980
You can see, and that, you
can see the difference.

612
00:32:31,980 --> 00:32:34,600
Now some regions,
which have a perfect

613
00:32:34,600 --> 00:32:36,285
overlap, and
there's some regions

614
00:32:36,285 --> 00:32:37,620
that are quite separate.

615
00:32:37,620 --> 00:32:40,480
Notable here are
these areas here,

616
00:32:40,480 --> 00:32:44,060
which are activated by
stereo but not by parallax.

617
00:32:44,060 --> 00:32:48,190
So this then can provide
you with an initial idea

618
00:32:48,190 --> 00:32:51,490
that there are some brain
areas in which both of these

619
00:32:51,490 --> 00:32:53,090
are analyzed together.

620
00:32:53,090 --> 00:32:54,950
And there's some
brain areas in which

621
00:32:54,950 --> 00:32:58,760
are uniquely analyzed for either
stereopsis or motion parallax

622
00:32:58,760 --> 00:32:59,260
alone.

623
00:33:00,570 --> 00:33:03,460
Now this tells you where it
takes place in the brain.

624
00:33:03,460 --> 00:33:07,660
But how it takes place requires
a totally different approach.

625
00:33:07,660 --> 00:33:11,600
Namely, the most
comfortably to record

626
00:33:11,600 --> 00:33:14,650
from individual neurons
in various areas

627
00:33:14,650 --> 00:33:17,750
just like I had shown
you that nice work done

628
00:33:17,750 --> 00:33:24,500
by John [? Porgio ?] recording
from the one demonstrating

629
00:33:24,500 --> 00:33:27,980
there that there are disparity
selective neurons that

630
00:33:27,980 --> 00:33:32,940
are tuned that then provide
the hardware, if you will,

631
00:33:32,940 --> 00:33:36,470
for being able to analyze
stereoscopic depth.

632
00:33:36,470 --> 00:33:40,100
So that then summarizes
what I wanted

633
00:33:40,100 --> 00:33:44,480
to tell you about
motion parallax.

634
00:33:46,260 --> 00:33:48,990
And now we are going
to go on and talk

635
00:33:48,990 --> 00:33:51,410
about yet another
important depth

636
00:33:51,410 --> 00:33:56,900
cue that is utilized by the
brain, which is called shading.

637
00:33:56,900 --> 00:34:07,550
Now remember that our ability to
use light to illuminate things

638
00:34:07,550 --> 00:34:11,020
is something that was
practically nonexistent

639
00:34:11,020 --> 00:34:14,400
for endless millions of years.

640
00:34:14,400 --> 00:34:17,340
And so because of that,
both animals and us,

641
00:34:17,340 --> 00:34:19,920
we have to heavily
rely on information

642
00:34:19,920 --> 00:34:23,610
based on light coming from
the sun, coming from above.

643
00:34:23,610 --> 00:34:29,270
And shading is based on those
millions of years of evolution

644
00:34:29,270 --> 00:34:32,500
utilizing the fact that
most of the light that

645
00:34:32,500 --> 00:34:34,810
illuminates things
comes from above.

646
00:34:34,810 --> 00:34:38,300
So there are all kinds
of nice examples of this.

647
00:34:38,300 --> 00:34:39,550
And here is one of them.

648
00:34:41,219 --> 00:34:45,900
What you can do here is you
can take a bunch of disks

649
00:34:45,900 --> 00:34:47,330
and set them up so.

650
00:34:47,330 --> 00:34:50,139
You can do this on a computer
to make the upper part

651
00:34:50,139 --> 00:34:52,219
light and the lower part
dark or the other way

652
00:34:52,219 --> 00:34:55,100
around, the upper part
dark and lower part light.

653
00:34:55,100 --> 00:34:58,990
And all of you readily
can see that these images,

654
00:34:58,990 --> 00:35:02,130
the first and third row, seem
to be protruding towards you.

655
00:35:02,130 --> 00:35:05,780
And the images in the second and
fourth row seem to be receding.

656
00:35:05,780 --> 00:35:10,680
Now that is because the
brain is interpreting

657
00:35:10,680 --> 00:35:17,456
that on the basis of the fact
that the light at least used

658
00:35:17,456 --> 00:35:18,830
to come predominantly
from above.

659
00:35:20,220 --> 00:35:26,830
So that is the basic
arrangement for seeing depth.

660
00:35:26,830 --> 00:35:29,470
And now I'm going to give
you some demonstrations

661
00:35:29,470 --> 00:35:35,370
to indicate that this cue
is actually quite powerful,

662
00:35:35,370 --> 00:35:37,580
even when you would not
necessarily expect it to be.

663
00:35:41,670 --> 00:35:47,930
So these shading cues have
also been extensively used

664
00:35:47,930 --> 00:35:51,460
in art work to provide
an impression of depth.

665
00:35:52,630 --> 00:35:54,200
And I will show
you some examples

666
00:35:54,200 --> 00:35:57,340
that will give you a
sense of how that is done.

667
00:35:57,340 --> 00:36:02,060
So let me make one more point
before I proceed that namely it

668
00:36:02,060 --> 00:36:05,320
is, indeed, the
degree of illumination

669
00:36:05,320 --> 00:36:06,460
that's crucial here.

670
00:36:06,460 --> 00:36:10,330
We have the same change
from red to, sorry,

671
00:36:10,330 --> 00:36:14,090
from some greenish to yellowish.

672
00:36:14,090 --> 00:36:16,680
And you have no sense of
depth here whatsoever.

673
00:36:16,680 --> 00:36:21,444
In other words, you do need
the shading information,

674
00:36:21,444 --> 00:36:22,860
meaning the amount
of light that's

675
00:36:22,860 --> 00:36:25,550
being reflected from
various surfaces, that

676
00:36:25,550 --> 00:36:29,950
is crucial for perceiving depth.

677
00:36:29,950 --> 00:36:33,240
So now what we
are going to do is

678
00:36:33,240 --> 00:36:37,570
we are going to present a series
of slides that will highlight

679
00:36:37,570 --> 00:36:41,850
the power that shading has
for the perception of depth.

680
00:36:42,870 --> 00:36:46,460
So here is an example
of how we do this.

681
00:36:46,460 --> 00:36:48,240
And the reason I'm
showing this in some

682
00:36:48,240 --> 00:36:53,630
detail because if you really are
interested in stuff like this,

683
00:36:53,630 --> 00:36:55,800
you can do all this
on your own computer.

684
00:36:55,800 --> 00:36:58,200
You can play games,
endless games with it.

685
00:36:58,200 --> 00:37:01,150
You can spend hours and
hours having a lot of fun

686
00:37:01,150 --> 00:37:05,940
thinking about how depth
works on the basis of shading.

687
00:37:05,940 --> 00:37:09,050
So what you have here are
a whole bunch of disks.

688
00:37:09,050 --> 00:37:12,410
And each of these can
be shaded differently

689
00:37:12,410 --> 00:37:14,920
by many different
computer programs.

690
00:37:16,900 --> 00:37:18,070
So that's what you can do.

691
00:37:18,070 --> 00:37:19,430
That's the basics.

692
00:37:19,430 --> 00:37:22,770
So now what we can do
is we can play a game.

693
00:37:22,770 --> 00:37:25,870
And we can say, present just
two different objects here.

694
00:37:26,900 --> 00:37:30,570
But we're going to present them
repeatedly on a big display.

695
00:37:30,570 --> 00:37:35,090
And then we can shade these
differently as we please.

696
00:37:35,090 --> 00:37:37,530
So here is a whole
bunch of them.

697
00:37:37,530 --> 00:37:42,090
And all the rows, this, the
first, third, and so on rows

698
00:37:42,090 --> 00:37:44,210
are the same shape
and the second, fourth

699
00:37:44,210 --> 00:37:47,150
and so on are the other
shape, these two shapes.

700
00:37:48,950 --> 00:37:51,563
So we only have two shapes
here that are juxtapositioned.

701
00:37:52,950 --> 00:37:57,595
Now what we can do is say, well,
this is a peculiar sensation.

702
00:37:59,650 --> 00:38:02,230
I have a vague
sense that there's

703
00:38:02,230 --> 00:38:04,520
something maybe in
the third dimension.

704
00:38:04,520 --> 00:38:07,900
But it's not too well
defined, because this

705
00:38:07,900 --> 00:38:11,990
is not in accordance with
the rules and laws of shading

706
00:38:11,990 --> 00:38:13,850
of light coming from above.

707
00:38:13,850 --> 00:38:16,440
So now what we can do
instead, we can selectively

708
00:38:16,440 --> 00:38:22,280
shade these to be in accordance
with the rules of light coming

709
00:38:22,280 --> 00:38:24,460
from above to create
shading and depth.

710
00:38:24,460 --> 00:38:27,620
And when you do that,
here's an example of that.

711
00:38:27,620 --> 00:38:31,070
What you can see here is
a very compelling image

712
00:38:31,070 --> 00:38:36,000
of these protruding elements,
sort of protruding to the left,

713
00:38:36,000 --> 00:38:36,500
right?

714
00:38:37,870 --> 00:38:40,850
Everybody see, have a
strong sense of depth here?

715
00:38:42,330 --> 00:38:45,500
So now what you can do
is you can play with it

716
00:38:45,500 --> 00:38:47,510
and decide, well,
can we do something

717
00:38:47,510 --> 00:38:51,250
that, keeping the
very, very same shapes,

718
00:38:51,250 --> 00:38:53,220
shade them differently
and see what

719
00:38:53,220 --> 00:38:55,050
it does to our
perception of depth.

720
00:38:55,050 --> 00:38:56,596
And so what we are
going to do next

721
00:38:56,596 --> 00:38:58,470
is we're going to take
each of these elements

722
00:38:58,470 --> 00:39:00,380
here, the same ones
here, and we're

723
00:39:00,380 --> 00:39:02,000
going to reverse the contrast.

724
00:39:02,000 --> 00:39:04,710
You see the contrast
here on top is white

725
00:39:04,710 --> 00:39:06,250
and the bottom is black.

726
00:39:06,250 --> 00:39:08,780
So we're going to
reverse that contrast.

727
00:39:08,780 --> 00:39:12,520
And when we do so, the question
is what are you going to see.

728
00:39:12,520 --> 00:39:14,600
And if you do that,
low and behold,

729
00:39:14,600 --> 00:39:16,540
you still have a
strong sense of depth.

730
00:39:16,540 --> 00:39:18,570
But it's a very confusing sense.

731
00:39:18,570 --> 00:39:23,430
You may see sometimes these
objects pointing to the left

732
00:39:23,430 --> 00:39:24,720
and sometimes to the right.

733
00:39:24,720 --> 00:39:28,820
It's unstable, because you're
confusing those computations

734
00:39:28,820 --> 00:39:31,200
that have evolved
over millions of years

735
00:39:31,200 --> 00:39:35,140
for interpreting depth
in terms of shading.

736
00:39:35,140 --> 00:39:37,910
Now you can play also
some additional games.

737
00:39:37,910 --> 00:39:41,080
You can make this even more
complicated, make more changes,

738
00:39:41,080 --> 00:39:42,324
and here is another one.

739
00:39:42,324 --> 00:39:43,740
You still have a
feeling of depth.

740
00:39:43,740 --> 00:39:45,540
But it's totally confusing.

741
00:39:45,540 --> 00:39:46,360
It's very hard.

742
00:39:46,360 --> 00:39:48,610
You can't organize it
any way, because it

743
00:39:48,610 --> 00:39:52,310
is not in accordance
with the law of light

744
00:39:52,310 --> 00:39:55,300
coming from above
to a real object.

745
00:39:55,300 --> 00:39:57,010
And lastly, you
can also make this

746
00:39:57,010 --> 00:39:59,500
so that it would be in
accordance with the laws.

747
00:39:59,500 --> 00:40:01,350
But you can change
it around so that you

748
00:40:01,350 --> 00:40:03,040
get a completely
different perception,

749
00:40:03,040 --> 00:40:04,430
a strong sense of depth.

750
00:40:04,430 --> 00:40:07,030
It's still the very,
very same elements

751
00:40:07,030 --> 00:40:08,280
that you had seen before.

752
00:40:08,280 --> 00:40:10,590
But now the shading is
done, again, differently.

753
00:40:10,590 --> 00:40:13,260
And then now gives you,
again, a unified sense

754
00:40:13,260 --> 00:40:20,050
of a display that
is not conflicting.

755
00:40:20,050 --> 00:40:23,430
Because in this case, it's
in accordance with that

756
00:40:23,430 --> 00:40:25,430
some of the basic
principles of shading.

757
00:40:26,930 --> 00:40:32,310
So now, what we can do next,
having talked about stereo

758
00:40:32,310 --> 00:40:34,620
and we have talked
about shading,

759
00:40:34,620 --> 00:40:36,550
is to look at some
more of the demos.

760
00:40:36,550 --> 00:40:39,100
So let's go back
to the stereoscope.

761
00:40:39,100 --> 00:40:42,250
And let's go back
to the handouts.

762
00:40:42,250 --> 00:40:50,380
And so if you now come to the
next page that has a heading

763
00:40:50,380 --> 00:40:52,210
called stereo and shading.

764
00:40:53,430 --> 00:40:55,790
So, again, take the
stereoscope and we

765
00:40:55,790 --> 00:40:58,430
are going to look
at these in steps.

766
00:40:58,430 --> 00:41:01,360
So let's start by looking
at the top display

767
00:41:01,360 --> 00:41:04,840
first, which is
called stereo only.

768
00:41:06,380 --> 00:41:08,330
So if you look at
that, first of all,

769
00:41:08,330 --> 00:41:12,170
if you just look at it
without the stereoscope,

770
00:41:12,170 --> 00:41:16,630
you see pretty much a
sort of flat display

771
00:41:16,630 --> 00:41:18,250
of a truncated pyramid.

772
00:41:18,250 --> 00:41:23,240
Then if you put the stereoscope
there and look through it,

773
00:41:23,240 --> 00:41:26,230
you should see, if you look
at it for a little while,

774
00:41:26,230 --> 00:41:29,730
that one of those
sticks out towards you

775
00:41:29,730 --> 00:41:32,280
and the other one
seems to recede.

776
00:41:33,940 --> 00:41:35,145
Does everybody see that?

777
00:41:37,010 --> 00:41:38,830
So let's stop there
for a minute, because I

778
00:41:38,830 --> 00:41:41,360
want to add one more
fact here, which

779
00:41:41,360 --> 00:41:43,590
I should have mentioned earlier.

780
00:41:43,590 --> 00:41:47,620
So what you do here
is-- so you have

781
00:41:47,620 --> 00:41:49,500
these two displays like that.

782
00:41:50,880 --> 00:41:55,430
And you have one image here
and another image here.

783
00:41:55,430 --> 00:41:57,980
When you-- this is
greatly exaggerated,

784
00:41:57,980 --> 00:41:59,990
these coming together like this.

785
00:41:59,990 --> 00:42:01,650
That means it's
going to stick out

786
00:42:01,650 --> 00:42:03,200
towards you as you look at it.

787
00:42:03,200 --> 00:42:10,350
But if you do the
opposite like that,

788
00:42:10,350 --> 00:42:12,460
they're further apart
than the rest of them.

789
00:42:12,460 --> 00:42:13,935
Then actually you
see it receding.

790
00:42:15,430 --> 00:42:18,800
And that's why, if you now
take away the stereoscope

791
00:42:18,800 --> 00:42:24,640
and look at it, you can see
that the top left image is

792
00:42:24,640 --> 00:42:28,800
in each, or facing
towards each other,

793
00:42:28,800 --> 00:42:32,780
whereas the other ones are
facing away from each other.

794
00:42:32,780 --> 00:42:35,800
And that's what creates, that's
what the brain interprets,

795
00:42:35,800 --> 00:42:39,060
as protruding versus receding
using the stereoscope.

796
00:42:40,740 --> 00:42:42,570
So that is very
similar, in a way,

797
00:42:42,570 --> 00:42:44,470
to what happens with shading.

798
00:42:44,470 --> 00:42:48,300
So now if you look
at the second image

799
00:42:48,300 --> 00:42:52,060
there, first without the
stereoscope, what you see here,

800
00:42:52,060 --> 00:42:54,130
again, is one that
sticks out just

801
00:42:54,130 --> 00:42:56,920
like in the original display.

802
00:42:56,920 --> 00:42:59,100
And the rest of
them are receding.

803
00:42:59,100 --> 00:43:03,090
That's because the shading,
the one that sticks out

804
00:43:03,090 --> 00:43:05,360
is light on top and
dark on the bottom.

805
00:43:05,360 --> 00:43:07,380
And it's the obverse
for the other ones.

806
00:43:07,380 --> 00:43:09,410
Now if you do the
same thing looking

807
00:43:09,410 --> 00:43:12,420
through the stereoscope,
what you will see

808
00:43:12,420 --> 00:43:16,300
is still some degree of depth,
but it's not very pronounced.

809
00:43:16,300 --> 00:43:22,111
Because there is no
corresponding disparity

810
00:43:22,111 --> 00:43:22,610
information.

811
00:43:24,250 --> 00:43:28,960
But now if you look
at the third display,

812
00:43:28,960 --> 00:43:33,500
where stereo and shading are
in harmony, then what you see

813
00:43:33,500 --> 00:43:36,880
is an extremely
compelling dramatic sense

814
00:43:36,880 --> 00:43:39,810
of depth with the top
left one sticking out

815
00:43:39,810 --> 00:43:42,660
towards you and the
other three receding.

816
00:43:42,660 --> 00:43:47,582
So shading appeared to have
added to the compelling nature

817
00:43:47,582 --> 00:43:49,665
of the depth that you see
through the stereoscope.

818
00:43:51,240 --> 00:43:57,200
Now the last image in here is
that we put stereo and shading

819
00:43:57,200 --> 00:43:59,040
and conflict with each other.

820
00:43:59,040 --> 00:44:02,932
And when you do that, you can
look at it, first with just one

821
00:44:02,932 --> 00:44:04,140
eye, then with the other eye.

822
00:44:04,140 --> 00:44:06,695
When you look at it with both
of them, you, for a while,

823
00:44:06,695 --> 00:44:08,540
you see something unstable.

824
00:44:08,540 --> 00:44:11,800
And when you see it
well, eventually, you

825
00:44:11,800 --> 00:44:14,330
realize that there
is a conflict there

826
00:44:14,330 --> 00:44:17,880
because of the
shading and the stereo

827
00:44:17,880 --> 00:44:19,720
being in opposition
to each other.

828
00:44:26,640 --> 00:44:31,990
Now then, this kind of effect,
if you go to the next page,

829
00:44:31,990 --> 00:44:35,250
we're going to go now to
page five, six, and seven.

830
00:44:35,250 --> 00:44:37,380
Again, what you
need to do here is

831
00:44:37,380 --> 00:44:41,850
to look at it sideways
with the F's on top.

832
00:44:41,850 --> 00:44:50,260
And when you look at this, those
of you who can use your eyes so

833
00:44:50,260 --> 00:44:53,930
that they are divergent,
and you look beyond it,

834
00:44:53,930 --> 00:44:57,240
then this is very much like,
or it is the same actually,

835
00:44:57,240 --> 00:44:58,770
as an auto stereogram.

836
00:44:58,770 --> 00:45:01,680
So if you look at this for a
while, and you look beyond it,

837
00:45:01,680 --> 00:45:03,570
eventually it's going to gel.

838
00:45:03,570 --> 00:45:05,500
And when it gels,
what you should

839
00:45:05,500 --> 00:45:08,180
see is where the
F's are, the images

840
00:45:08,180 --> 00:45:12,475
are protruding towards you,
and the others are receding.

841
00:45:12,475 --> 00:45:14,610
Now it may take you awhile.

842
00:45:14,610 --> 00:45:16,810
It's much more difficult
than what we just

843
00:45:16,810 --> 00:45:18,430
did with the stereoscope.

844
00:45:18,430 --> 00:45:21,860
But you should be
able to see that.

845
00:45:23,420 --> 00:45:27,580
How many of you are able to
actually see those images?

846
00:45:27,580 --> 00:45:31,360
A little bit di-- move it back
and forth a little bit slowly.

847
00:45:31,360 --> 00:45:33,940
And maybe eventually
you manage it.

848
00:45:38,550 --> 00:45:40,410
So as I say, where
the F's are you

849
00:45:40,410 --> 00:45:44,020
see these images-- these
truncated pyramids protruding

850
00:45:44,020 --> 00:45:44,980
towards you.

851
00:45:44,980 --> 00:45:47,070
And the rest of
them are receding.

852
00:45:50,989 --> 00:45:52,900
If you have difficulty
seeing this,

853
00:45:52,900 --> 00:45:55,670
I'm not surprised, because
it takes a lot of practice.

854
00:45:55,670 --> 00:45:58,530
But once you get
a sense of it, I

855
00:45:58,530 --> 00:46:03,647
think that you will enjoy
doing this and actually

856
00:46:03,647 --> 00:46:05,105
showing it to some
of your friends.

857
00:46:07,210 --> 00:46:10,760
So then if you go
to the next page,

858
00:46:10,760 --> 00:46:15,570
there we have added shading.

859
00:46:15,570 --> 00:46:18,270
And the shearing is
the same everywhere.

860
00:46:18,270 --> 00:46:22,220
But the stereo cues are not.

861
00:46:22,220 --> 00:46:25,400
Once again, what
happens is that they

862
00:46:25,400 --> 00:46:30,120
are stereo cues where
the F's are stick out

863
00:46:30,120 --> 00:46:32,695
towards you much
greater than the others.

864
00:46:35,740 --> 00:46:38,530
They stick out a lot less
because of the added stereo.

865
00:46:45,300 --> 00:46:51,990
And then, in the last demo
there, the last page, we

866
00:46:51,990 --> 00:46:56,860
are putting, just like in that
figure with a stereoscope,

867
00:46:56,860 --> 00:47:00,490
we are putting them into
opposition with each other.

868
00:47:00,490 --> 00:47:02,350
And so when you
look at these, this

869
00:47:02,350 --> 00:47:04,950
would be very difficult
to see for a while.

870
00:47:04,950 --> 00:47:10,740
Because there's a tendency
to see it differently

871
00:47:10,740 --> 00:47:15,692
for stereo and for
motion parallax.

872
00:47:15,692 --> 00:47:19,170
And so it's going to
be an unstable percept.

873
00:47:20,310 --> 00:47:23,310
So what you can do then
is you can play around

874
00:47:23,310 --> 00:47:29,220
with this at your leisure and
especially once you become more

875
00:47:29,220 --> 00:47:32,130
proficient looking at
all the stereograms,

876
00:47:32,130 --> 00:47:36,420
if you go and get one of these
magic eye books to look at,

877
00:47:36,420 --> 00:47:40,160
you will be able to see
these displays as well.

878
00:47:44,790 --> 00:47:49,370
So now, this is the one was the
first one that I showed you.

879
00:47:49,370 --> 00:47:52,330
As I said, this one, with
the F's, are the ones

880
00:47:52,330 --> 00:47:54,360
that should stick
out closest to you.

881
00:47:54,360 --> 00:47:57,190
Once you see that, then
you can go on to the next

882
00:47:57,190 --> 00:48:01,636
to add the shading or
subtract the shading from it.

883
00:48:05,040 --> 00:48:10,310
So now, an interesting
question that arises

884
00:48:10,310 --> 00:48:20,400
is to what degree are we
able, or are animals able,

885
00:48:20,400 --> 00:48:26,140
to integrate these different
kinds of depth cues.

886
00:48:26,140 --> 00:48:28,270
And in particular,
in this case, you're

887
00:48:28,270 --> 00:48:33,185
going to ask the question, what
about integrating stereopsis,

888
00:48:33,185 --> 00:48:34,425
parallax, and shading.

889
00:48:35,590 --> 00:48:40,160
So the experiment is
one done on monkeys

890
00:48:40,160 --> 00:48:44,120
in which you can present
these cues either singly

891
00:48:44,120 --> 00:48:46,850
or in combination.

892
00:48:46,850 --> 00:48:48,380
And we can ask the
question, well,

893
00:48:48,380 --> 00:48:52,350
does the monkey do better
with one or the other,

894
00:48:52,350 --> 00:48:56,300
or does he integrate really
and does really much better

895
00:48:56,300 --> 00:48:58,290
when you provide all three cues.

896
00:48:58,290 --> 00:49:00,640
And so here is a procedure.

897
00:49:00,640 --> 00:49:03,700
What you do here, again,
you have a rocking display

898
00:49:03,700 --> 00:49:06,230
like this, and you
can present this

899
00:49:06,230 --> 00:49:08,700
either with shading
as shown here

900
00:49:08,700 --> 00:49:11,440
or with motion parallax where
it rocks back and forth,

901
00:49:11,440 --> 00:49:15,240
and lastly also with stereopsis.

902
00:49:15,240 --> 00:49:19,330
So if you do that, the results
you get are quite dramatic.

903
00:49:19,330 --> 00:49:23,910
What happens is shown here as
a percent correct performance

904
00:49:23,910 --> 00:49:26,980
and here is the latency
in milliseconds.

905
00:49:26,980 --> 00:49:32,510
And it shows that the monkey
does extremely well when

906
00:49:32,510 --> 00:49:34,400
you present-- this
is percent correct.

907
00:49:34,400 --> 00:49:36,600
This is degrees of disparity.

908
00:49:36,600 --> 00:49:40,390
The monkey does extremely well
when you present all three cues

909
00:49:40,390 --> 00:49:42,890
and does worse when you present
each of those cues alone.

910
00:49:44,080 --> 00:49:46,050
Even more dramatic
is the fact, and this

911
00:49:46,050 --> 00:49:49,220
is-- I keep coming back to
this, that the ability for us

912
00:49:49,220 --> 00:49:52,395
to respond quickly to things
is very important for survival.

913
00:49:54,220 --> 00:49:57,500
And here what we can see is
that when you present all three

914
00:49:57,500 --> 00:50:00,950
cues, performance is much,
much, much, much faster than

915
00:50:00,950 --> 00:50:04,010
when you present
each of those alone.

916
00:50:04,010 --> 00:50:06,080
And, of course, as
you might expect,

917
00:50:06,080 --> 00:50:09,150
when you present parallax
only, because that's

918
00:50:09,150 --> 00:50:13,130
motion over time, that
takes the longest to do.

919
00:50:15,120 --> 00:50:18,680
So even though motion
parallax cues are great,

920
00:50:18,680 --> 00:50:21,330
it became important in
the course of evolution

921
00:50:21,330 --> 00:50:25,180
to create mechanisms that
can detect these things more

922
00:50:25,180 --> 00:50:27,195
quickly and more efficiently.

923
00:50:28,930 --> 00:50:32,120
So now we come to
yet another cue

924
00:50:32,120 --> 00:50:36,300
that we know very little about
at the level of the brain

925
00:50:36,300 --> 00:50:40,450
or single units, because
it's so complicated, which

926
00:50:40,450 --> 00:50:41,930
is called perspective.

927
00:50:41,930 --> 00:50:45,570
But I want you to just be aware
of it and have a sense of it.

928
00:50:45,570 --> 00:50:48,770
And here is one of
those cartoon examples

929
00:50:48,770 --> 00:50:51,220
that gives you a very
strong sense of depth.

930
00:50:51,220 --> 00:50:52,510
And you almost cringe.

931
00:50:52,510 --> 00:50:54,330
If you were there,
you would worry

932
00:50:54,330 --> 00:50:56,690
that you would be falling down.

933
00:50:56,690 --> 00:50:59,285
This is done strictly by
virtue of perspective.

934
00:51:01,400 --> 00:51:04,910
It's very similar to what you
encounter all the time when

935
00:51:04,910 --> 00:51:08,972
you're driving down a road and
the road seems to converge,

936
00:51:08,972 --> 00:51:10,430
even though you're
not aware of it.

937
00:51:10,430 --> 00:51:12,586
But that's what's happening
on the rental surface.

938
00:51:12,586 --> 00:51:14,710
Because things further away
are smaller than things

939
00:51:14,710 --> 00:51:16,082
that are close by.

940
00:51:16,082 --> 00:51:18,040
And that's when you look
down a railroad track,

941
00:51:18,040 --> 00:51:20,170
the same thing happens,
even though you

942
00:51:20,170 --> 00:51:22,920
know that the railroad
track is not converging,

943
00:51:22,920 --> 00:51:24,450
it's going parallel.

944
00:51:24,450 --> 00:51:27,820
But because of the
distances involved,

945
00:51:27,820 --> 00:51:29,590
that's what falls on the retina.

946
00:51:29,590 --> 00:51:33,220
And you're smart enough to
know that even though that's

947
00:51:33,220 --> 00:51:34,900
what falls on the
retina, you can

948
00:51:34,900 --> 00:51:37,070
make the right kind
of interpretation.

949
00:51:37,070 --> 00:51:40,090
Conversely, you can
also compute the depth

950
00:51:40,090 --> 00:51:42,270
on the basis of that
kind of convergence.

951
00:51:42,270 --> 00:51:44,930
Now here's another example
of that, a much simpler way

952
00:51:44,930 --> 00:51:47,200
that people can do
with experiments.

953
00:51:47,200 --> 00:51:48,890
Here we have a bunch of dots.

954
00:51:48,890 --> 00:51:53,000
And we have two basic cues that
have to do with perspective.

955
00:51:53,000 --> 00:52:00,715
One of them is this gradually
decreasing size of these dots.

956
00:52:02,920 --> 00:52:05,220
I should say elongated
disks, if you will,

957
00:52:05,220 --> 00:52:06,780
and also that they
are converging

958
00:52:06,780 --> 00:52:08,800
much like a railroad
track converges.

959
00:52:08,800 --> 00:52:10,830
And so we have a
very strong sense

960
00:52:10,830 --> 00:52:13,160
of having a third
dimension here.

961
00:52:13,160 --> 00:52:15,130
Now the fact that
this is so strong

962
00:52:15,130 --> 00:52:21,030
can be mitigated by
adding a few things here.

963
00:52:21,030 --> 00:52:25,150
If you add some more dots,
it's not question as dramatic.

964
00:52:25,150 --> 00:52:27,460
And then if you start
mixing up the sizes,

965
00:52:27,460 --> 00:52:29,160
you are beginning to lose it.

966
00:52:29,160 --> 00:52:30,910
And then if you
totally mix it up,

967
00:52:30,910 --> 00:52:33,590
then you have no sense
of that left at all.

968
00:52:33,590 --> 00:52:37,200
So it is that progression
of steps and sizes

969
00:52:37,200 --> 00:52:40,380
and whatnot that
gives you the sense

970
00:52:40,380 --> 00:52:45,310
of the depth of the images
that you're looking at.

971
00:52:46,710 --> 00:52:49,870
Now here's another
converse example

972
00:52:49,870 --> 00:52:55,370
of this that is an illusory
effect that what you see here

973
00:52:55,370 --> 00:52:58,770
is three barrels, if you will.

974
00:53:00,050 --> 00:53:04,380
And this barrel is a lot
bigger than this barrel, right,

975
00:53:04,380 --> 00:53:05,730
or is it?

976
00:53:05,730 --> 00:53:07,880
Well, so what we're
going to do here,

977
00:53:07,880 --> 00:53:11,820
we have an inducing element here
by this hallway, if you will,

978
00:53:11,820 --> 00:53:12,950
with a door at the end.

979
00:53:12,950 --> 00:53:15,200
And we're going to
remove this hallway

980
00:53:15,200 --> 00:53:19,120
keeping the barrels
exactly as they are.

981
00:53:19,120 --> 00:53:20,900
And if you do that,
low and behold,

982
00:53:20,900 --> 00:53:22,725
those barrels are
all the same size.

983
00:53:24,340 --> 00:53:27,230
It's induced by virtue
of the surround that

984
00:53:27,230 --> 00:53:29,100
gives you a false
sense of depth.

985
00:53:33,500 --> 00:53:39,700
So now let me show you
another picture because

986
00:53:39,700 --> 00:53:42,380
of the purpose behind this.

987
00:53:42,380 --> 00:53:45,190
This a picture
that's in a museum

988
00:53:45,190 --> 00:53:47,250
in Worcester, Massachusetts.

989
00:53:47,250 --> 00:53:50,375
And it was created by a
fellow called Edward Savage.

990
00:53:50,375 --> 00:53:53,930
And it's a pretty
unpleasant picture.

991
00:53:53,930 --> 00:53:56,300
But the main reason
I'm showing this to you

992
00:53:56,300 --> 00:53:59,400
is that there seems to be
very poor sense of depth

993
00:53:59,400 --> 00:54:01,080
in this picture.

994
00:54:01,080 --> 00:54:06,670
Now the reason
this is interesting

995
00:54:06,670 --> 00:54:12,290
is because when artists
began, centuries

996
00:54:12,290 --> 00:54:20,140
ago in the 13th, 12th
centuries, draw things,

997
00:54:20,140 --> 00:54:24,240
they did not have a concept
of an understanding of how

998
00:54:24,240 --> 00:54:29,895
to create depth, a third
dimension, in their drawings.

999
00:54:31,130 --> 00:54:34,930
So what they did
eventually, they came up

1000
00:54:34,930 --> 00:54:37,060
with a so-called
vanishing point,

1001
00:54:37,060 --> 00:54:42,160
and they drew very much
like what we had here.

1002
00:54:42,160 --> 00:54:45,620
Lines that converged
at a point and then

1003
00:54:45,620 --> 00:54:49,080
scaled the images
accordingly rather

1004
00:54:49,080 --> 00:54:50,740
than keeping it the same size.

1005
00:54:50,740 --> 00:54:54,180
And that way you got
a good sense of depth.

1006
00:54:54,180 --> 00:54:58,290
So now that has a number
of interesting stories

1007
00:54:58,290 --> 00:55:03,410
about it that we
are going to discuss

1008
00:55:03,410 --> 00:55:07,909
next time you talk about
the perception of shapes,

1009
00:55:07,909 --> 00:55:08,450
OK, patterns.

1010
00:55:10,270 --> 00:55:13,350
But I will leave that
discussion until that.

1011
00:55:13,350 --> 00:55:17,120
What I'm going to do
next, however, I'm

1012
00:55:17,120 --> 00:55:26,720
going to try to give you a sense
of how important stereopsis can

1013
00:55:26,720 --> 00:55:28,995
be for the perception
of fine depths.

1014
00:55:30,190 --> 00:55:33,860
And so to do that, what I'm
going--I'm going to show you

1015
00:55:33,860 --> 00:55:35,500
actually a film.

1016
00:55:35,500 --> 00:55:40,600
And here what we have is
a so-called needle test.

1017
00:55:40,600 --> 00:55:42,940
What you have here is a
fine needle protruding.

1018
00:55:42,940 --> 00:55:48,870
And here we have a
bunch of different size

1019
00:55:48,870 --> 00:55:54,300
circular openings, a little bit
like a needle, but it's round.

1020
00:55:55,540 --> 00:55:59,590
And the task is to take these
one at a time and hang them up.

1021
00:55:59,590 --> 00:56:01,900
And one can time how
quickly you can do that,

1022
00:56:01,900 --> 00:56:05,670
or you can make a film to
see how well you can do it.

1023
00:56:05,670 --> 00:56:07,940
And then what we
can do is we can

1024
00:56:07,940 --> 00:56:11,020
test the subject under
binocular conditions,

1025
00:56:11,020 --> 00:56:13,460
and test them under
monocular conditions.

1026
00:56:13,460 --> 00:56:18,060
So I'm going to show you a film
of this, actually two films.

1027
00:56:20,580 --> 00:56:22,786
It will just take just
few seconds to do it.

1028
00:56:28,850 --> 00:56:31,900
OK, be ready, it's going
to come up in a second.

1029
00:56:35,050 --> 00:56:37,820
OK, here's the subject under
binocular viewing conditions.

1030
00:56:48,400 --> 00:56:51,870
So that's the condition
under binocular viewing.

1031
00:56:51,870 --> 00:56:56,510
And now I'm going to show it to
you, same subject, same time,

1032
00:56:56,510 --> 00:56:58,090
but with one eye closed off.

1033
00:57:26,680 --> 00:57:29,800
So that then just
even looking at it

1034
00:57:29,800 --> 00:57:32,350
without taking any
careful measurements.

1035
00:57:32,350 --> 00:57:34,370
It's obvious that
it's much, much

1036
00:57:34,370 --> 00:57:38,050
more difficult to thread
a needle under monocular

1037
00:57:38,050 --> 00:57:40,700
than under binocular
viewing conditions.

1038
00:57:40,700 --> 00:57:43,530
And so what you can do
is when you go home,

1039
00:57:43,530 --> 00:57:47,400
and next time you want
to sew something up,

1040
00:57:47,400 --> 00:57:50,920
try threading the needle with
one eye closed and with the two

1041
00:57:50,920 --> 00:57:51,460
eyes open.

1042
00:57:51,460 --> 00:57:54,490
And you will see immediately
what a huge difference it is.

1043
00:57:54,490 --> 00:57:58,350
And that difference,
therefore, is due

1044
00:57:58,350 --> 00:58:02,260
to you're having the
mechanism of steropsis.

1045
00:58:17,750 --> 00:58:19,620
Just a few seconds here.

1046
00:59:09,920 --> 00:59:15,270
Another test that has been
used in a similar fashion which

1047
00:59:15,270 --> 00:59:20,960
allows you to actually calculate
exactly what your error is

1048
00:59:20,960 --> 00:59:24,620
in reaching, you
can have a subject

1049
00:59:24,620 --> 00:59:27,940
sit in front of one
of these touch panels,

1050
00:59:27,940 --> 00:59:30,400
and then do this experiment
either binocularly or

1051
00:59:30,400 --> 00:59:31,245
monocularly.

1052
00:59:31,245 --> 00:59:33,590
And after he presses
this a dot comes up,

1053
00:59:33,590 --> 00:59:35,540
and then the person
has to touch it.

1054
00:59:35,540 --> 00:59:38,020
And you have about 30
or 40 trials like that.

1055
00:59:38,020 --> 00:59:42,070
And then you have recorded
where the person touched.

1056
00:59:42,070 --> 00:59:44,480
And, therefore, you
can calculate the error

1057
00:59:44,480 --> 00:59:47,350
between where we touched
and where the dot is.

1058
00:59:47,350 --> 00:59:49,790
And then, again, you
get a huge effect

1059
00:59:49,790 --> 00:59:53,040
between monocular and
binocular viewing conditions.

1060
00:59:54,410 --> 00:59:56,850
Now when you come to
monocular and binocular

1061
00:59:56,850 --> 01:00:00,950
viewing conditions, another
thing important to test

1062
01:00:00,950 --> 01:00:04,580
is to what degree a person
who does or does not

1063
01:00:04,580 --> 01:00:08,300
have stereopsis is capable
of integrating information

1064
01:00:08,300 --> 01:00:10,130
between the two eyes.

1065
01:00:10,130 --> 01:00:13,530
So to do that, we
have here examples

1066
01:00:13,530 --> 01:00:16,100
of what is called
binocular integration.

1067
01:00:16,100 --> 01:00:19,910
So what we do here, again,
you can use a stereoscope.

1068
01:00:19,910 --> 01:00:21,370
You look at a monitor.

1069
01:00:21,370 --> 01:00:23,940
And this represent the left
eye, this is the right eye.

1070
01:00:23,940 --> 01:00:25,500
And you flash these on.

1071
01:00:25,500 --> 01:00:28,249
If you integrate this,
this is what you see.

1072
01:00:28,249 --> 01:00:29,790
This would be actually
what you would

1073
01:00:29,790 --> 01:00:32,205
show in the control
part of the experiment.

1074
01:00:33,690 --> 01:00:36,440
So you see the Star of David.

1075
01:00:36,440 --> 01:00:39,390
And if a subject is
shown this, and they

1076
01:00:39,390 --> 01:00:42,700
don't see the Star
of David, you worry

1077
01:00:42,700 --> 01:00:46,520
that their ability to integrate
the information between the two

1078
01:00:46,520 --> 01:00:48,530
eyes is deficient.

1079
01:00:48,530 --> 01:00:51,940
And I would say 90% of the
cases, those people who

1080
01:00:51,940 --> 01:00:57,290
are deficient on this
also show major deficiency

1081
01:00:57,290 --> 01:00:59,525
in stereoscopic viewing.

1082
01:00:59,525 --> 01:01:02,780
Now another way to do
this is an experiment

1083
01:01:02,780 --> 01:01:06,400
in which you can present
two words here, sud and try,

1084
01:01:06,400 --> 01:01:08,690
so the two are separate.

1085
01:01:08,690 --> 01:01:11,920
And when you present
them simultaneously,

1086
01:01:11,920 --> 01:01:14,200
you actually see
the word sturdy.

1087
01:01:14,200 --> 01:01:17,070
So you ask the subject, please
tell us what do you see.

1088
01:01:17,070 --> 01:01:18,160
What is it word you see.

1089
01:01:18,160 --> 01:01:19,920
And the subject says sud.

1090
01:01:19,920 --> 01:01:22,746
And the subject
says try, then you

1091
01:01:22,746 --> 01:01:26,010
know that that subject
sees, if he says try,

1092
01:01:26,010 --> 01:01:28,390
he sees mostly with his
right and prefers it,

1093
01:01:28,390 --> 01:01:30,360
doesn't see too well
with his left eye.

1094
01:01:30,360 --> 01:01:33,790
If he says sturdy, then
he integrates the two.

1095
01:01:33,790 --> 01:01:35,900
And, therefore,
you can safely say

1096
01:01:35,900 --> 01:01:39,661
that this guy has very good
integration between the two

1097
01:01:39,661 --> 01:01:40,160
eyes.

1098
01:01:41,900 --> 01:01:48,360
So what I would
like to do next then

1099
01:01:48,360 --> 01:01:53,330
is to provide you with any
questions that you have.

1100
01:01:53,330 --> 01:01:56,172
This was a complicated
topic, that you have,

1101
01:01:56,172 --> 01:01:57,630
and then we are
going to summarize.

1102
01:01:58,970 --> 01:02:02,184
Does anybody have a question
about motion parallax,

1103
01:02:02,184 --> 01:02:03,100
stereopsis, and so on?

1104
01:02:06,700 --> 01:02:09,540
Let me maybe add one
more important factor.

1105
01:02:11,080 --> 01:02:15,655
Your eyes are separated
only by so many centimeters.

1106
01:02:16,880 --> 01:02:18,425
Now, can you think
of an animal where

1107
01:02:18,425 --> 01:02:19,800
there's a much
larger separation?

1108
01:02:21,064 --> 01:02:22,122
AUDIENCE: Hammerheads.

1109
01:02:22,122 --> 01:02:23,455
PROFESSOR: The hammerhead shark.

1110
01:02:25,210 --> 01:02:29,170
Yeah, that has a separation
of over a foot between the two

1111
01:02:29,170 --> 01:02:29,870
eyes.

1112
01:02:29,870 --> 01:02:33,360
And so you could ask the
question, why on earth did

1113
01:02:33,360 --> 01:02:36,460
that animal evolve to such a
huge separation between the two

1114
01:02:36,460 --> 01:02:36,960
eyes?

1115
01:02:40,260 --> 01:02:43,565
Well, that brings one to yet
another interesting point.

1116
01:02:48,760 --> 01:02:52,080
This I think may have started
during the Second World War.

1117
01:02:55,440 --> 01:02:59,460
It was realized that
when you're flying

1118
01:02:59,460 --> 01:03:03,210
over some territory, where
there are all kinds of weapons

1119
01:03:03,210 --> 01:03:07,910
and whatnot, which
are well camouflaged,

1120
01:03:07,910 --> 01:03:11,630
that just looking down at
them, you can't see them.

1121
01:03:12,800 --> 01:03:14,730
But obviously if you're
going to have a tank

1122
01:03:14,730 --> 01:03:23,800
or you're going to have a
gun or other that may be more

1123
01:03:23,800 --> 01:03:28,620
like a cannon, it sticks
out of the ground.

1124
01:03:28,620 --> 01:03:36,180
So it was discovered that
if you had in your airplane

1125
01:03:36,180 --> 01:03:39,140
two lenses which
are far apart, that

1126
01:03:39,140 --> 01:03:41,510
would greatly magnify the depth.

1127
01:03:41,510 --> 01:03:43,890
You could defeat
that camouflage.

1128
01:03:43,890 --> 01:03:45,785
And you could find
those weapons down there

1129
01:03:45,785 --> 01:03:48,160
by virtue of the fact they're
sticking out of the ground.

1130
01:03:49,750 --> 01:03:56,710
So the fact then is that the
more you separate the images

1131
01:03:56,710 --> 01:04:00,970
from the two eyes, if you
will, or your two cameras,

1132
01:04:00,970 --> 01:04:04,800
the more likely it is
that you can calculate

1133
01:04:04,800 --> 01:04:08,840
the disparity of information
between the two images.

1134
01:04:08,840 --> 01:04:16,330
So that then is probably
one of the reasons, not

1135
01:04:16,330 --> 01:04:19,050
the sole reason, but
maybe one of the reasons

1136
01:04:19,050 --> 01:04:23,400
why, in some animals, is
an excessive separation

1137
01:04:23,400 --> 01:04:24,530
between the two eyes.

1138
01:04:25,770 --> 01:04:28,780
And that brings me to get me
to yet another point, which

1139
01:04:28,780 --> 01:04:33,380
is that stereopsis actually
works best at relatively

1140
01:04:33,380 --> 01:04:36,070
short distances, like
threading a needle.

1141
01:04:37,280 --> 01:04:42,760
It doesn't work too well beyond,
I don't know, 10 feet or so.

1142
01:04:43,875 --> 01:04:46,404
It becomes progressively
less effective.

1143
01:04:46,404 --> 01:04:48,195
But at short distance,
it's very effective.

1144
01:04:49,380 --> 01:04:51,590
And so I presume
also many animals

1145
01:04:51,590 --> 01:04:58,400
that have to hunt for
food are able to utilize

1146
01:04:58,400 --> 01:05:00,550
the mechanism of stereopsis,
because everything

1147
01:05:00,550 --> 01:05:04,830
is at a close distance when they
hunt for food on the ground.

1148
01:05:06,515 --> 01:05:11,920
And by contrast, when you
talk about motion parallax,

1149
01:05:11,920 --> 01:05:15,535
that works extremely well
over very long distances.

1150
01:05:18,290 --> 01:05:21,400
So does anybody have any
questions about motion parallax

1151
01:05:21,400 --> 01:05:23,640
or about steropsis?

1152
01:05:26,580 --> 01:05:28,410
Oh, once again, I'm
crystal clear, huh?

1153
01:05:29,460 --> 01:05:31,610
So, therefore, I
think it's time for us

1154
01:05:31,610 --> 01:05:33,900
to summarize what we
had covered today.

1155
01:05:33,900 --> 01:05:37,213
First of all, there
are numerous mechanisms

1156
01:05:37,213 --> 01:05:41,550
that have emerged
for analyzing depth.

1157
01:05:41,550 --> 01:05:47,190
And they include the ocular
motor cues, which are vergence

1158
01:05:47,190 --> 01:05:51,220
an accommodation and then the
binocular cue of stereopsis

1159
01:05:51,220 --> 01:05:53,969
and then the monocular
cues of parallax shading

1160
01:05:53,969 --> 01:05:54,635
and perspective.

1161
01:05:56,040 --> 01:05:58,205
Then you have several
cortical structures

1162
01:05:58,205 --> 01:06:00,180
that process stereopsis.

1163
01:06:00,180 --> 01:06:03,570
You don't have one
specific brain area

1164
01:06:03,570 --> 01:06:05,020
that uniquely does this.

1165
01:06:06,140 --> 01:06:10,240
The number of disparities that
are represented in the brain,

1166
01:06:10,240 --> 01:06:13,270
as studies in the area of
the one by John [? Porgio, ?]

1167
01:06:13,270 --> 01:06:14,360
is limited.

1168
01:06:14,360 --> 01:06:17,840
And maybe, maybe four,
but there may be six,

1169
01:06:17,840 --> 01:06:21,650
but certainly there are
not a large number of them.

1170
01:06:21,650 --> 01:06:24,720
And so it's analogous
to the way things

1171
01:06:24,720 --> 01:06:28,430
had been resolved for us to
be able to process color.

1172
01:06:29,680 --> 01:06:32,370
Utilizing motion parallax
for depth processing

1173
01:06:32,370 --> 01:06:36,250
necessitates neuron specific
for direction, velocity,

1174
01:06:36,250 --> 01:06:38,150
and differential velocity.

1175
01:06:38,150 --> 01:06:45,090
Several areas getting V1 and
MT process motion parallax,

1176
01:06:45,090 --> 01:06:46,190
which I did not say.

1177
01:06:46,190 --> 01:06:51,110
But indeed, if you make
a lesion in area MT,

1178
01:06:51,110 --> 01:06:54,506
you go get a deficit in
motion parallax, even

1179
01:06:54,506 --> 01:06:57,700
thought you don't get a
major deficit in steropsis.

1180
01:06:59,130 --> 01:07:01,990
Now, area MT
combines the analysis

1181
01:07:01,990 --> 01:07:05,700
of motion parallax,
depth, and flicker.

1182
01:07:05,700 --> 01:07:07,960
However, these analyses
are also carried out

1183
01:07:07,960 --> 01:07:10,850
by several other structures
as I've already said.

1184
01:07:10,850 --> 01:07:13,100
And lastly, little
is know at present

1185
01:07:13,100 --> 01:07:16,810
about the manner in which
information about shading

1186
01:07:16,810 --> 01:07:19,490
and prospective are
analyzed in the brain.

1187
01:07:19,490 --> 01:07:23,640
And hopefully, that will
be one of the future tasks

1188
01:07:23,640 --> 01:07:25,170
by neuroscientists.

1189
01:07:25,170 --> 01:07:30,900
And so if any of you ever
get involved in neuroscience,

1190
01:07:30,900 --> 01:07:33,170
this certainly is
a big open area

1191
01:07:33,170 --> 01:07:37,396
that we hope people
will start to analyze.

1192
01:07:41,260 --> 01:07:46,800
So that then is the essence of
what I wanted to cover today.

1193
01:07:46,800 --> 01:07:50,230
And once again, if any of
you has a question, please,

1194
01:07:50,230 --> 01:07:51,560
please don't hesitate to ask.

1195
01:07:51,560 --> 01:07:53,490
I'll be very happy
to answer them.

1196
01:07:58,890 --> 01:08:03,750
OK, lastly then, did everybody
sign the attendance sheet?

1197
01:08:03,750 --> 01:08:07,470
If not, please come up after the
class and sign your name to it.

1198
01:08:08,731 --> 01:08:09,230
Very good.

1199
01:08:09,230 --> 01:08:11,100
So next time then,
we are going to talk

1200
01:08:11,100 --> 01:08:13,530
about pattern perception.

1201
01:08:13,530 --> 01:08:18,279
And hopefully you will
find that also interesting.