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 --> 00:00:17,275 at ocw.mit.edu. 8 00:00:28,340 --> 00:00:29,840 PROFESSOR: Good afternoon, everyone. 9 00:00:32,425 --> 00:00:34,810 Let us get started. 10 00:00:34,810 --> 00:00:36,860 The first thing we're going to do today 11 00:00:36,860 --> 00:00:39,920 is I'm going to finish up what I didn't get a chance 12 00:00:39,920 --> 00:00:43,270 to finish the last time, because we ran over a little bit. 13 00:00:43,270 --> 00:00:46,960 And what we are going to cover to begin with 14 00:00:46,960 --> 00:00:50,180 is the lateral geniculate nucleus of the thalamus. 15 00:00:50,180 --> 00:00:53,560 This is a structure in the thalamus that 16 00:00:53,560 --> 00:00:58,010 receives an extensive input from the retina 17 00:00:58,010 --> 00:01:00,250 from retinal ganglion cells. 18 00:01:00,250 --> 00:01:05,580 And it is a beautiful structure that has several layers to it. 19 00:01:05,580 --> 00:01:09,570 And here is a picture of one from a monkey. 20 00:01:09,570 --> 00:01:12,740 And this is a coronal section. 21 00:01:12,740 --> 00:01:14,260 I think you all know a little bit 22 00:01:14,260 --> 00:01:16,240 about the anatomy of the brain. 23 00:01:16,240 --> 00:01:19,910 If you cut slices this way, it's coronal cuts. 24 00:01:19,910 --> 00:01:22,521 If you cut it this way, it's sagittal cuts. 25 00:01:22,521 --> 00:01:23,020 OK? 26 00:01:23,020 --> 00:01:24,780 So this is a coronal cut. 27 00:01:24,780 --> 00:01:26,090 A thin slice. 28 00:01:26,090 --> 00:01:29,740 And it shows what the lateral geniculate nucleus 29 00:01:29,740 --> 00:01:33,450 looks like when it's cut about in its middle. 30 00:01:34,660 --> 00:01:37,390 And what you can say is a bunch of layers here. 31 00:01:37,390 --> 00:01:39,460 There's six layers as labeled here. 32 00:01:40,980 --> 00:01:47,570 And the top four layers are called the parvocellular 33 00:01:47,570 --> 00:01:48,440 layers. 34 00:01:48,440 --> 00:01:51,680 And the bottom two are called the magnocellular layers. 35 00:01:51,680 --> 00:01:53,310 And the reason they have these names 36 00:01:53,310 --> 00:01:55,940 is because the cells-- this was stained. 37 00:01:55,940 --> 00:01:57,720 You can't see it that clearly. 38 00:01:57,720 --> 00:02:00,650 But the cells are much bigger in the bottom two layers 39 00:02:00,650 --> 00:02:02,317 than in the top four layers. 40 00:02:02,317 --> 00:02:03,900 And that's why these are called parvo, 41 00:02:03,900 --> 00:02:05,850 and these are called magno. 42 00:02:05,850 --> 00:02:09,310 Now, one of the important discoveries that had been made, 43 00:02:09,310 --> 00:02:11,150 and I'll belabor that in a minute, 44 00:02:11,150 --> 00:02:14,550 is that these top four layers get input 45 00:02:14,550 --> 00:02:16,420 from the midget cells of the retina, 46 00:02:16,420 --> 00:02:19,680 and the bottom two get input from the parasol cells. 47 00:02:19,680 --> 00:02:23,010 And we'll be talking about that a great deal in the next few 48 00:02:23,010 --> 00:02:23,930 lectures. 49 00:02:23,930 --> 00:02:27,740 Now, each of these layers gets input from one eye. 50 00:02:27,740 --> 00:02:28,960 And there's an alternation. 51 00:02:28,960 --> 00:02:31,040 If you go from layer six, this is 52 00:02:31,040 --> 00:02:32,570 input for a contralateral line. 53 00:02:32,570 --> 00:02:33,480 Ipsilateral. 54 00:02:33,480 --> 00:02:34,440 Contralateral. 55 00:02:34,440 --> 00:02:35,190 Ipsilateral. 56 00:02:35,190 --> 00:02:37,230 Then there is a reversal. 57 00:02:37,230 --> 00:02:39,627 And this layer is again ipsilateral. 58 00:02:39,627 --> 00:02:40,960 And this layer is contralateral. 59 00:02:42,590 --> 00:02:45,370 So that is the basic layout. 60 00:02:45,370 --> 00:02:47,900 And I might as well anticipate to tell you 61 00:02:47,900 --> 00:02:49,780 that if you study the receptive field 62 00:02:49,780 --> 00:02:53,460 organization of these cells, they are very similar to those 63 00:02:53,460 --> 00:02:57,130 that you see in the retina. 64 00:02:57,130 --> 00:02:58,160 You have midget cells. 65 00:02:58,160 --> 00:02:58,995 Parasol cells. 66 00:02:58,995 --> 00:03:00,710 You have center-surround antagonism. 67 00:03:02,630 --> 00:03:06,010 So that there is not much of a transform 68 00:03:06,010 --> 00:03:11,640 at this level in the brain, as the inputs progress 69 00:03:11,640 --> 00:03:16,800 from the retina to the lateral geniculate nucleus. 70 00:03:16,800 --> 00:03:18,530 But once we get up to the cortex, 71 00:03:18,530 --> 00:03:21,690 we will see many major transforms 72 00:03:21,690 --> 00:03:23,630 that I will come to shortly. 73 00:03:23,630 --> 00:03:27,110 Now, the layout of the lateral geniculate nucleus 74 00:03:27,110 --> 00:03:29,520 varies quite a bit from species to species. 75 00:03:31,410 --> 00:03:34,070 In the market we have this kind of arrangement. 76 00:03:34,070 --> 00:03:36,550 And this is very similar to the kind of arrangement 77 00:03:36,550 --> 00:03:38,490 we have in our own heads. 78 00:03:38,490 --> 00:03:43,370 But then if you look at-- well, let me make one more point here 79 00:03:43,370 --> 00:03:44,100 about this. 80 00:03:45,149 --> 00:03:46,690 There's a beautiful experiment that's 81 00:03:46,690 --> 00:03:51,030 been done to confirm that indeed the inputs to the parvo 82 00:03:51,030 --> 00:03:53,410 and magnocellular layers are coming 83 00:03:53,410 --> 00:03:55,240 from different cells in the retina. 84 00:03:55,240 --> 00:03:59,660 And so an experiment was done in which a labeling material was 85 00:03:59,660 --> 00:04:01,250 put into either the parvocellular 86 00:04:01,250 --> 00:04:03,420 layers or the magnocellular layers, 87 00:04:03,420 --> 00:04:06,290 as you can see here on the right side. 88 00:04:06,290 --> 00:04:13,610 Then those labeling substances were transported back 89 00:04:13,610 --> 00:04:14,630 to the retina. 90 00:04:14,630 --> 00:04:19,190 And they labeled the cells that project to the axons 91 00:04:19,190 --> 00:04:20,640 into these two regions. 92 00:04:20,640 --> 00:04:22,810 And what you can see here is that these 93 00:04:22,810 --> 00:04:26,390 are smaller cells that project to the parvocellular layers, 94 00:04:26,390 --> 00:04:29,590 and much bigger cells that project to the magnocellular 95 00:04:29,590 --> 00:04:30,800 layers. 96 00:04:30,800 --> 00:04:33,980 And these correspond, as other studies 97 00:04:33,980 --> 00:04:37,670 have shown in more detail, to the parasol cells. 98 00:04:37,670 --> 00:04:42,820 And these correspond to the midget cells. 99 00:04:42,820 --> 00:04:46,490 So that's an important distinction here of the inputs. 100 00:04:46,490 --> 00:04:48,895 And the other thing that's important to realize here 101 00:04:48,895 --> 00:04:53,390 is that we talked about how the eye is organizing, 102 00:04:53,390 --> 00:04:55,810 and the way it projects to the nasal and temple 103 00:04:55,810 --> 00:04:56,660 parts of the retina. 104 00:04:56,660 --> 00:04:59,340 The left lateral geniculate nucleus 105 00:04:59,340 --> 00:05:01,890 sees the right half of the visual field, 106 00:05:01,890 --> 00:05:05,250 and right geniculate sees the left half of the visual field. 107 00:05:05,250 --> 00:05:08,150 Now, let me say just a few words about the fact 108 00:05:08,150 --> 00:05:11,710 that there are different kinds of geniculates. 109 00:05:11,710 --> 00:05:13,750 But they're not that different. 110 00:05:13,750 --> 00:05:16,050 So here's an example of a tree shrew 111 00:05:16,050 --> 00:05:17,520 lateral geniculate nucleus. 112 00:05:17,520 --> 00:05:19,380 It's shaped differently, but it still 113 00:05:19,380 --> 00:05:21,210 has these nice six layers. 114 00:05:21,210 --> 00:05:24,590 And the bottom two layers again are magnocellular, 115 00:05:24,590 --> 00:05:28,880 and the top four layers are essentially parvocellular. 116 00:05:28,880 --> 00:05:30,820 But then when recordings were made, 117 00:05:30,820 --> 00:05:32,610 an interesting specialization was 118 00:05:32,610 --> 00:05:36,050 seen in this animal, which was not that obvious in others. 119 00:05:36,050 --> 00:05:39,340 If you look at this here, examine the question 120 00:05:39,340 --> 00:05:43,400 to what degree are the cells in these top four layers 121 00:05:43,400 --> 00:05:45,220 on or off cells. 122 00:05:45,220 --> 00:05:49,210 And what you see here, that when you look at layers three, 123 00:05:49,210 --> 00:05:52,600 four, five, and six, there's a huge distinction here. 124 00:05:52,600 --> 00:05:57,740 Virtually all the cells in three and four are off cells, 125 00:05:57,740 --> 00:06:01,730 and all the cells in five and six are on cells. 126 00:06:01,730 --> 00:06:04,330 So there's a specialization in the layers 127 00:06:04,330 --> 00:06:08,300 as to whether the input is from the on cells or off cells. 128 00:06:08,300 --> 00:06:11,330 And yet another animal has a different kind of arrangement. 129 00:06:11,330 --> 00:06:13,740 I'm not going to go into more details about this, 130 00:06:13,740 --> 00:06:16,280 but again there's six layers here. 131 00:06:17,330 --> 00:06:20,610 But here they don't seem to have a distinction between magno 132 00:06:20,610 --> 00:06:22,020 and parvocellular layers. 133 00:06:22,020 --> 00:06:24,560 Instead we have two layers, five and four, 134 00:06:24,560 --> 00:06:26,440 where the cells are very small. 135 00:06:26,440 --> 00:06:28,855 And the rest of the layers are quite large. 136 00:06:29,920 --> 00:06:31,380 This is the galago. 137 00:06:31,380 --> 00:06:32,390 All right? 138 00:06:32,390 --> 00:06:37,050 Now most thoroughly studied had been the cat, and the monkey. 139 00:06:37,050 --> 00:06:38,700 And that's the kind of stuff we are 140 00:06:38,700 --> 00:06:42,490 going to talk about more as we progress in the course. 141 00:06:42,490 --> 00:06:42,990 All right. 142 00:06:42,990 --> 00:06:46,380 So that in a very summary form is the essence 143 00:06:46,380 --> 00:06:50,460 of what that lateral geniculate nucleus is like. 144 00:06:50,460 --> 00:06:52,060 Now, what I want to do is I'm going 145 00:06:52,060 --> 00:06:54,310 to have a quick summary of what I've 146 00:06:54,310 --> 00:06:56,060 covered in the first session. 147 00:06:56,060 --> 00:07:00,330 And here, first of all again, is a drawing 148 00:07:00,330 --> 00:07:05,150 of what the cells look like in the retina. 149 00:07:05,150 --> 00:07:06,470 These are the photoreceptors. 150 00:07:06,470 --> 00:07:09,220 Remember the light is coming in from the bottom up. 151 00:07:09,220 --> 00:07:09,890 OK? 152 00:07:09,890 --> 00:07:11,960 And so if you look at that, we find 153 00:07:11,960 --> 00:07:15,230 that all the photoreceptors hyperpolarize the light, 154 00:07:15,230 --> 00:07:18,680 and they produce only graded potentials. 155 00:07:18,680 --> 00:07:21,490 My last important fact is that all these cells 156 00:07:21,490 --> 00:07:24,880 use glutamate as the neurotransmitter that 157 00:07:24,880 --> 00:07:28,910 is released at the bottom here of these cells, which then 158 00:07:28,910 --> 00:07:31,530 innervates the subsequent elements 159 00:07:31,530 --> 00:07:36,110 in this retinal picture here, which 160 00:07:36,110 --> 00:07:39,150 are the horizontal cells and the bipolar cells. 161 00:07:39,150 --> 00:07:41,330 So if you look at those, the horizontal cells 162 00:07:41,330 --> 00:07:44,623 are all hyperpolarized to light just to the photoreceptors. 163 00:07:45,780 --> 00:07:48,710 And they also only produce graded potentials. 164 00:07:48,710 --> 00:07:51,660 But then when we proceed, and look at the bipolar cells, 165 00:07:51,660 --> 00:07:55,270 there are two major classes of bipolar cells 166 00:07:55,270 --> 00:07:57,870 that I want to deal with, and more. 167 00:07:57,870 --> 00:08:00,340 These are the important ones from our point of view. 168 00:08:00,340 --> 00:08:02,930 The so-called on and off bipolars. 169 00:08:02,930 --> 00:08:06,210 And that arises, because the on bipolars 170 00:08:06,210 --> 00:08:09,240 have signed inverting synapses with which they 171 00:08:09,240 --> 00:08:11,990 connect to the photoreceptors. 172 00:08:11,990 --> 00:08:17,170 Whereas the off bipolars have signed conserving synapses. 173 00:08:17,170 --> 00:08:20,910 So the way to think of this, if the photoreceptor 174 00:08:20,910 --> 00:08:24,660 hyperpolarizes and depolarizes, going up and down as the light 175 00:08:24,660 --> 00:08:27,220 goes in and out, an off cell will 176 00:08:27,220 --> 00:08:29,000 mimic this, and go like this. 177 00:08:29,000 --> 00:08:31,330 But the on cell will do the opposite. 178 00:08:31,330 --> 00:08:32,380 OK? 179 00:08:32,380 --> 00:08:35,864 So we have created in the retina, I shouldn't say we, 180 00:08:35,864 --> 00:08:39,669 but evolution has, from a single-ended system 181 00:08:39,669 --> 00:08:42,390 since they were hyperpolarized, a double-ended system 182 00:08:42,390 --> 00:08:44,480 at the level of the bipolar cell. 183 00:08:44,480 --> 00:08:48,090 And that's what then creates these famous on and off cells 184 00:08:48,090 --> 00:08:50,010 that we are going to talk about in a lot 185 00:08:50,010 --> 00:08:52,620 more detail the next time to try to figure out 186 00:08:52,620 --> 00:08:53,935 why do we have these cells. 187 00:08:54,930 --> 00:08:55,430 All right. 188 00:08:55,430 --> 00:08:59,320 So that then is a summary of the upper layers. 189 00:08:59,320 --> 00:09:05,040 And then when they come down to the amacrine cells, 190 00:09:05,040 --> 00:09:08,540 some of these amacrine cells produce action potentials. 191 00:09:08,540 --> 00:09:10,000 Some don't. 192 00:09:10,000 --> 00:09:12,600 There are many different kinds, as I mentioned the last time. 193 00:09:12,600 --> 00:09:14,510 There are more than 20 different kinds. 194 00:09:14,510 --> 00:09:15,210 Some are on. 195 00:09:15,210 --> 00:09:16,010 Some are off. 196 00:09:16,010 --> 00:09:17,240 Some are on-off. 197 00:09:17,240 --> 00:09:21,100 And some don't even give you action potentials. 198 00:09:21,100 --> 00:09:23,860 Then finally when you come down to the ganglion cells, 199 00:09:23,860 --> 00:09:26,930 they all give you action potentials. 200 00:09:26,930 --> 00:09:28,617 And the major two classes that we 201 00:09:28,617 --> 00:09:30,950 are going to deal with a lot, even though there are many 202 00:09:30,950 --> 00:09:33,550 more, we'll talk about some of them later, 203 00:09:33,550 --> 00:09:37,190 are the midget cells and the parasol cells. 204 00:09:37,190 --> 00:09:37,690 OK. 205 00:09:37,690 --> 00:09:41,470 So then to have an overall summary of what we covered 206 00:09:41,470 --> 00:09:44,540 the last time, first of all, we had the right brain 207 00:09:44,540 --> 00:09:47,200 receives input from the left visual hemifield, 208 00:09:47,200 --> 00:09:49,790 and the left from the right hemifield. 209 00:09:49,790 --> 00:09:52,295 That, you understood when I explained to you 210 00:09:52,295 --> 00:09:55,810 what the wiring is like in the retina. 211 00:09:55,810 --> 00:09:58,290 And then I pointed out, just seeing 212 00:09:58,290 --> 00:10:01,660 that, we have these five major classes of retinal cells. 213 00:10:01,660 --> 00:10:03,570 We have the photoreceptors themselves, 214 00:10:03,570 --> 00:10:05,200 which are the rods and the cords. 215 00:10:05,200 --> 00:10:07,790 And then we have the horizontal cells, the bipolar cells, 216 00:10:07,790 --> 00:10:12,080 the amacrine cells, and the retinal ganglion cells. 217 00:10:12,080 --> 00:10:13,780 Hopefully, by repeating this stuff, 218 00:10:13,780 --> 00:10:17,001 eventually this will stick into your brain. 219 00:10:17,001 --> 00:10:17,500 All right. 220 00:10:17,500 --> 00:10:22,300 So then the receptive fields, or the retinal ganglion 221 00:10:22,300 --> 00:10:24,690 cells, sometimes referred to as RGCs, 222 00:10:24,690 --> 00:10:27,800 have antagonistic centers around organization. 223 00:10:27,800 --> 00:10:30,280 And then when we look at the question of adaptation, 224 00:10:30,280 --> 00:10:33,890 you can explain why on earth did this evolve. 225 00:10:33,890 --> 00:10:35,470 This complex arrangement. 226 00:10:35,470 --> 00:10:39,040 Then there are several classes of retinal ganglion cells. 227 00:10:39,040 --> 00:10:42,380 We have talked about the on and off, the midget and parasol, 228 00:10:42,380 --> 00:10:45,530 and we will later on talk about several other classes. 229 00:10:45,530 --> 00:10:48,220 Now, all photoreceptors and horizontal cells 230 00:10:48,220 --> 00:10:49,820 hyperpolarize the light. 231 00:10:49,820 --> 00:10:52,100 At that level you have a single-ended system. 232 00:10:52,100 --> 00:10:56,210 But then when you come to the bipolar cells, about half 233 00:10:56,210 --> 00:10:59,380 of them are hyperpolarizing, and half of them are depolarizing. 234 00:10:59,380 --> 00:11:06,740 And the opposition arises, because the on bipolar cells 235 00:11:06,740 --> 00:11:10,120 have an inverting synapse, which is 236 00:11:10,120 --> 00:11:12,180 due to their intersynaptic junction 237 00:11:12,180 --> 00:11:17,430 where their receptors are made of molecules, 238 00:11:17,430 --> 00:11:18,610 which is called the mGluR6. 239 00:11:19,740 --> 00:11:22,242 And we'll talk about that in more detail next time. 240 00:11:22,242 --> 00:11:23,700 The action potentials in the retina 241 00:11:23,700 --> 00:11:26,960 are generated by only amacrine cells and retinal ganglion 242 00:11:26,960 --> 00:11:27,960 cells. 243 00:11:27,960 --> 00:11:30,900 The lateral geniculate nucleus that we have talked about 244 00:11:30,900 --> 00:11:34,530 is a laminated structure, but desegregating lamina 245 00:11:34,530 --> 00:11:37,800 varies with species just as I pointed out to you. 246 00:11:37,800 --> 00:11:40,950 The parvocellular layers receive input from the midget cells, 247 00:11:40,950 --> 00:11:43,600 and the magnocellular layers from the parasol cells 248 00:11:43,600 --> 00:11:46,280 in the monkey, and in the human. 249 00:11:46,280 --> 00:11:47,380 OK? 250 00:11:47,380 --> 00:11:50,859 The inputs in the left and right are segregated into lamina, 251 00:11:50,859 --> 00:11:52,400 as I have already pointed out to you. 252 00:11:52,400 --> 00:11:54,720 Lastly, the receptive few properties 253 00:11:54,720 --> 00:11:58,960 of lateral geniculate cells are similar to those 254 00:11:58,960 --> 00:12:01,000 that you see in the retina ganglion cells. 255 00:12:01,000 --> 00:12:03,020 You don't have any major transforms. 256 00:12:03,020 --> 00:12:03,520 OK. 257 00:12:03,520 --> 00:12:07,556 So that's then what we have an initial understanding of when 258 00:12:07,556 --> 00:12:10,770 it comes to the retina, and the lateral geniculate nucleus. 259 00:12:10,770 --> 00:12:15,570 We are now going to proceed, hang on for a minute, 260 00:12:15,570 --> 00:12:22,020 to the visual cortex. 261 00:12:22,020 --> 00:12:22,540 All right. 262 00:12:22,540 --> 00:12:26,445 So we are going to talk about the visual cortex. 263 00:12:33,880 --> 00:12:34,380 OK. 264 00:12:34,380 --> 00:12:37,200 So the first thing we are going to talk about that pertains 265 00:12:37,200 --> 00:12:41,490 to the visual cortex is V1, this initial area 266 00:12:41,490 --> 00:12:43,850 that gets most of the input directly 267 00:12:43,850 --> 00:12:46,860 from the lateral geniculate nucleus. 268 00:12:46,860 --> 00:12:48,440 And we're going to look at this first 269 00:12:48,440 --> 00:12:50,740 from an anatomical point of view. 270 00:12:50,740 --> 00:12:53,560 And then we're going to look at it progressively more 271 00:12:53,560 --> 00:12:55,430 from a functional point of view. 272 00:12:55,430 --> 00:12:55,970 OK. 273 00:12:55,970 --> 00:12:58,410 So here once again is a monkey brain. 274 00:12:58,410 --> 00:13:00,830 You're going to see this monkey brain over and over again, 275 00:13:00,830 --> 00:13:03,490 and gradually will become familiar with it. 276 00:13:03,490 --> 00:13:07,430 I want to point out to you that this here is a central sulcus. 277 00:13:07,430 --> 00:13:09,670 Just like we have a central sulcus, so do monkeys. 278 00:13:10,800 --> 00:13:13,961 And then this area here is called the lunate sulcus. 279 00:13:13,961 --> 00:13:15,460 And then in front here, we are going 280 00:13:15,460 --> 00:13:16,590 to talk about this region. 281 00:13:16,590 --> 00:13:18,410 It's called the frontal eye fields. 282 00:13:18,410 --> 00:13:19,420 There are two. 283 00:13:19,420 --> 00:13:21,550 The soft side, the principalis, and the arcuate. 284 00:13:21,550 --> 00:13:23,560 So you're going to see this over and over again. 285 00:13:23,560 --> 00:13:25,580 There's not too many things to remember here. 286 00:13:25,580 --> 00:13:27,590 But it will grow, because I will have 287 00:13:27,590 --> 00:13:30,190 to talk about more and more areas 288 00:13:30,190 --> 00:13:32,030 as we progress in the course. 289 00:13:32,030 --> 00:13:34,670 Now, here is area V1. 290 00:13:34,670 --> 00:13:36,950 And as I pointed out to you before, 291 00:13:36,950 --> 00:13:40,990 this region is in the monkey lissencephalic, 292 00:13:40,990 --> 00:13:43,270 meaning it's mostly flat. 293 00:13:43,270 --> 00:13:45,770 And because of that, its spatial layout 294 00:13:45,770 --> 00:13:47,970 is fairly easy to understand. 295 00:13:47,970 --> 00:13:49,530 So the first thing you want to do 296 00:13:49,530 --> 00:13:53,680 is to understand this spatial layout of the structure. 297 00:13:53,680 --> 00:13:57,270 And to do that people have done all sorts of experiments. 298 00:13:57,270 --> 00:14:01,400 Here is one that has done the spatial layout. 299 00:14:01,400 --> 00:14:03,520 Initially, the way this was done was 300 00:14:03,520 --> 00:14:06,700 kind of a difficult undertaking. 301 00:14:06,700 --> 00:14:08,560 They used single microelectrodes, 302 00:14:08,560 --> 00:14:12,020 and they would systematically move the microelectrodes here, 303 00:14:12,020 --> 00:14:13,170 and all across. 304 00:14:13,170 --> 00:14:15,750 And then they would map out where the receptive field was. 305 00:14:15,750 --> 00:14:16,730 OK? 306 00:14:16,730 --> 00:14:19,630 So if they did that, they found that in this region 307 00:14:19,630 --> 00:14:24,480 here the fovea is represented. 308 00:14:24,480 --> 00:14:31,430 And then as you progress towards the center of the brain, what 309 00:14:31,430 --> 00:14:35,290 you find is the center of the hemisphere I should really say. 310 00:14:36,770 --> 00:14:39,600 You progress to about 80 degrees out. 311 00:14:39,600 --> 00:14:41,630 And this is the horizontal meridian. 312 00:14:41,630 --> 00:14:43,950 And this below here, and above it, 313 00:14:43,950 --> 00:14:46,660 is the vertical meridian that goes around. 314 00:14:46,660 --> 00:14:47,250 OK? 315 00:14:47,250 --> 00:14:50,750 So what is represented, as you had already seen, 316 00:14:50,750 --> 00:14:54,160 you're representing each hemisphere, 317 00:14:54,160 --> 00:14:55,780 half the visual field. 318 00:14:55,780 --> 00:14:57,790 So what you have is this. 319 00:14:57,790 --> 00:14:58,680 OK? 320 00:14:58,680 --> 00:15:00,540 This is the horizontal meridian. 321 00:15:00,540 --> 00:15:01,880 This is the upper part. 322 00:15:01,880 --> 00:15:03,560 This is the lower part. 323 00:15:03,560 --> 00:15:04,370 OK? 324 00:15:04,370 --> 00:15:07,560 But that is inverted on the retina, 325 00:15:07,560 --> 00:15:10,630 as well is inverted on the visual cortex. 326 00:15:12,320 --> 00:15:12,820 All right. 327 00:15:12,820 --> 00:15:15,030 So that's the basic layout. 328 00:15:15,030 --> 00:15:19,450 But now what happened, as it always happens in science, 329 00:15:19,450 --> 00:15:21,690 new methods were developed to study 330 00:15:21,690 --> 00:15:24,970 this in more detail, and more reliably. 331 00:15:24,970 --> 00:15:27,150 And one of the methods that was developed 332 00:15:27,150 --> 00:15:33,120 is to do a combined action, and then chemical experiment, 333 00:15:33,120 --> 00:15:36,510 in which the eyes were stimulated. 334 00:15:36,510 --> 00:15:40,555 Then they were fixed, and one of the neurons were active. 335 00:15:41,810 --> 00:15:43,800 Took up the substance that was injected 336 00:15:43,800 --> 00:15:50,010 into the bloodstream, which then could be labeled subsequently. 337 00:15:50,010 --> 00:15:52,330 So I'm going to show you what this arrangement is. 338 00:15:52,330 --> 00:15:54,700 This is what the monkey is looking at. 339 00:15:54,700 --> 00:15:57,550 This is the half of the visual field that 340 00:15:57,550 --> 00:15:59,710 is being studied, because you're looking 341 00:15:59,710 --> 00:16:02,750 at the contralateral hemisphere. 342 00:16:02,750 --> 00:16:04,915 And then what is being done in these experiments 343 00:16:04,915 --> 00:16:07,570 is that you alternate the black and white stripes 344 00:16:07,570 --> 00:16:08,205 back and forth. 345 00:16:09,260 --> 00:16:11,840 So you can see if that's sort of a flicker. 346 00:16:11,840 --> 00:16:14,300 And you keep doing this for an hour or two. 347 00:16:14,300 --> 00:16:15,960 And as a result of this, the cells 348 00:16:15,960 --> 00:16:22,070 that are activated by this take up 349 00:16:22,070 --> 00:16:24,120 this substance that is being used, 350 00:16:24,120 --> 00:16:26,170 which is called 2DG, 2-Deoxyglucose. 351 00:16:27,830 --> 00:16:30,074 And then subsequently you can label that. 352 00:16:30,074 --> 00:16:31,490 So now what we are going to do, we 353 00:16:31,490 --> 00:16:34,350 can look at what the label looks like. 354 00:16:34,350 --> 00:16:38,065 And here is this beautiful example of this work. 355 00:16:38,065 --> 00:16:40,140 Here is a foveal area. 356 00:16:40,140 --> 00:16:42,340 This is about 7 or 8 degrees out. 357 00:16:42,340 --> 00:16:42,950 OK? 358 00:16:42,950 --> 00:16:45,631 So if you go back-- oops. 359 00:16:45,631 --> 00:16:46,130 Sorry. 360 00:16:47,330 --> 00:16:56,280 This outermost half disk is one that you see here. 361 00:16:56,280 --> 00:16:57,350 OK? 362 00:16:57,350 --> 00:17:00,280 So now we have a very clear understanding, 363 00:17:00,280 --> 00:17:04,670 which we can do quantitatively, of what the spatial layout is 364 00:17:04,670 --> 00:17:07,310 of the visual field on the retinal surface. 365 00:17:07,310 --> 00:17:12,656 And one important point that I want you to remember today, 366 00:17:12,656 --> 00:17:14,030 and we will discuss it again when 367 00:17:14,030 --> 00:17:19,020 we talk about visual prosthesis, is that much more brain area 368 00:17:19,020 --> 00:17:23,530 is devoted to central vision, because of course the many more 369 00:17:23,530 --> 00:17:26,589 cells in the fovea than in the periphery. 370 00:17:26,589 --> 00:17:29,710 And the cortex itself is of constant thickness. 371 00:17:29,710 --> 00:17:32,820 It's about roughly 2 millimeters in thickness. 372 00:17:32,820 --> 00:17:34,950 And so you need to give it more space 373 00:17:34,950 --> 00:17:37,280 in this lissencephalic brain to accept 374 00:17:37,280 --> 00:17:40,850 all that input from the foveal region. 375 00:17:40,850 --> 00:17:44,060 And so much more area is devoted to foveal vision 376 00:17:44,060 --> 00:17:45,570 than to peripheral vision. 377 00:17:47,120 --> 00:17:47,620 All right. 378 00:17:47,620 --> 00:17:49,710 So that then is the spatial layout. 379 00:17:49,710 --> 00:17:53,010 And then if we look at a cross section 380 00:17:53,010 --> 00:17:57,010 of the visual cortical regions here, 381 00:17:57,010 --> 00:18:01,280 this includes several cortical areas, this one right here, 382 00:18:01,280 --> 00:18:03,710 and this region here is so-called area V1. 383 00:18:05,020 --> 00:18:06,620 Now, an interesting discovery was 384 00:18:06,620 --> 00:18:10,490 made by a person called Gennari, who 385 00:18:10,490 --> 00:18:13,420 discovered when he used this kind of Nissl stain 386 00:18:13,420 --> 00:18:16,100 that when you look at V1, there seems to be almost 387 00:18:16,100 --> 00:18:21,130 what looks like an extra stripe, which gets labeled. 388 00:18:21,130 --> 00:18:24,000 And as soon as we get here, as you can see, this point, 389 00:18:24,000 --> 00:18:25,450 it stops here. 390 00:18:25,450 --> 00:18:29,720 And that defines, this extra layer, this stria of Gennari 391 00:18:29,720 --> 00:18:33,220 defines area 17 anatomically. 392 00:18:33,220 --> 00:18:35,970 If you get here, you suddenly get to area V2, 393 00:18:35,970 --> 00:18:37,680 so V2 starts here. 394 00:18:37,680 --> 00:18:40,570 So even with a simple anatomical technique, 395 00:18:40,570 --> 00:18:44,990 you can tell what is area V1, and area V2, and so on. 396 00:18:44,990 --> 00:18:45,490 OK. 397 00:18:45,490 --> 00:18:50,930 So now, if we take a cross section of the visual cortex, 398 00:18:50,930 --> 00:18:55,530 which I've said is about 2 millimeters in thickness, OK? 399 00:18:56,620 --> 00:19:03,450 People have divided this region, this so-called gray matter. 400 00:19:03,450 --> 00:19:04,730 Why is it called gray matter? 401 00:19:04,730 --> 00:19:07,470 It's called gray matter, because most of the cells 402 00:19:07,470 --> 00:19:12,150 there are not coded. 403 00:19:12,150 --> 00:19:17,070 Whereas if you go below, you have all the coded fibers 404 00:19:17,070 --> 00:19:19,200 that are coming in. 405 00:19:19,200 --> 00:19:21,550 So this gray matter then has been 406 00:19:21,550 --> 00:19:24,010 divided into six subdivisions. 407 00:19:24,010 --> 00:19:25,900 Layers one through six. 408 00:19:25,900 --> 00:19:27,800 In this case from the top down instead of 409 00:19:27,800 --> 00:19:30,550 in the genicular point, so from the bottom up. 410 00:19:30,550 --> 00:19:33,080 And these layers, subsequently, were 411 00:19:33,080 --> 00:19:35,280 realized they're more than six layers, really. 412 00:19:35,280 --> 00:19:39,100 And that's why it became 4A, 4B, 4CF, or 4C-beta. 413 00:19:40,340 --> 00:19:44,960 Now the 4C-alpha and beta is a region that gets a lot of input 414 00:19:44,960 --> 00:19:47,350 directly from the lateral geniculate nucleus. 415 00:19:47,350 --> 00:19:49,060 And what the nature of that input is, 416 00:19:49,060 --> 00:19:50,730 I'm going to show you here. 417 00:19:50,730 --> 00:19:53,050 Once again, here's the lateral geniculate nucleus. 418 00:19:53,050 --> 00:19:54,170 The six layers. 419 00:19:54,170 --> 00:19:57,200 The four parvocellular, and two magnocellular layers. 420 00:19:57,200 --> 00:20:00,510 And then what we can do, is we can trace how that connect. 421 00:20:00,510 --> 00:20:02,660 This has been done both physiologically and 422 00:20:02,660 --> 00:20:03,790 anatomically. 423 00:20:03,790 --> 00:20:08,030 And it shows that the four top layers terminate 424 00:20:08,030 --> 00:20:10,970 for the most part, almost exclusive, but not 100%, 425 00:20:10,970 --> 00:20:14,310 but pretty closely into layers of so-called 4C-alpha. 426 00:20:16,250 --> 00:20:20,030 Whereas when you come to the two bottom layers, OK? 427 00:20:20,030 --> 00:20:22,360 The so-called magnocellular layers, 428 00:20:22,360 --> 00:20:23,920 they terminate in 4C-beta. 429 00:20:24,844 --> 00:20:26,260 So some of these things are almost 430 00:20:26,260 --> 00:20:28,510 like an inversion from here to there. 431 00:20:28,510 --> 00:20:29,760 OK? 432 00:20:29,760 --> 00:20:31,270 And then another thing we haven't 433 00:20:31,270 --> 00:20:34,420 talked about before is that also some cells, even though they're 434 00:20:34,420 --> 00:20:39,220 not numerous, that reside in between the layers 435 00:20:39,220 --> 00:20:41,690 of the lateral geniculate nucleus, 436 00:20:41,690 --> 00:20:43,345 are called the koniocellular cells. 437 00:20:44,719 --> 00:20:46,510 But you don't have to worry about that now. 438 00:20:46,510 --> 00:20:48,670 You don't have to remember that at this point. 439 00:20:48,670 --> 00:20:50,960 But we'll talk about it later. 440 00:20:50,960 --> 00:20:53,145 But when you do look at them, what you find 441 00:20:53,145 --> 00:20:56,160 is that they project into the upper part 442 00:20:56,160 --> 00:21:00,950 of the visual cortex, both from the parvocellular interlaminar 443 00:21:00,950 --> 00:21:03,880 layers, and the magnocellular interlaminar layers. 444 00:21:03,880 --> 00:21:06,380 So as you might expect, things are 445 00:21:06,380 --> 00:21:08,470 complicated when it comes to the brain. 446 00:21:08,470 --> 00:21:10,347 And there are just all kinds of connections, 447 00:21:10,347 --> 00:21:11,930 and you're trying to make sense of it, 448 00:21:11,930 --> 00:21:15,470 and understand the reasons as to why we have these connections, 449 00:21:15,470 --> 00:21:16,670 and what the functions are. 450 00:21:24,270 --> 00:21:24,770 All right. 451 00:21:24,770 --> 00:21:29,610 So now, we are going to take a big step forward, and begin 452 00:21:29,610 --> 00:21:35,030 to look at the functional aspects of area V1. 453 00:21:35,030 --> 00:21:35,590 OK? 454 00:21:35,590 --> 00:21:35,730 Oops. 455 00:21:35,730 --> 00:21:36,230 Sorry. 456 00:21:37,830 --> 00:21:38,800 Hang on. 457 00:21:38,800 --> 00:21:39,459 Here we are. 458 00:21:39,459 --> 00:21:41,250 We are going to look at the receptive field 459 00:21:41,250 --> 00:21:42,690 organization of cells. 460 00:21:42,690 --> 00:21:44,020 So how do you do that? 461 00:21:44,020 --> 00:21:47,580 What you do is you can stick either a single electrode, 462 00:21:47,580 --> 00:21:50,560 or multiple electrodes into the visual cortex. 463 00:21:50,560 --> 00:21:53,330 And then you can map out the receptive fields 464 00:21:53,330 --> 00:21:55,220 to see how they respond. 465 00:21:55,220 --> 00:21:56,250 OK? 466 00:21:56,250 --> 00:21:58,640 Now, this was an interesting story 467 00:21:58,640 --> 00:22:00,360 in the beginning when this kind of search 468 00:22:00,360 --> 00:22:04,640 went on, because there were two major groups that were trying 469 00:22:04,640 --> 00:22:07,300 to understand what the visual responses are 470 00:22:07,300 --> 00:22:08,510 like in the visual cortex. 471 00:22:09,630 --> 00:22:11,130 And this was not even that long ago. 472 00:22:11,130 --> 00:22:16,480 This was in the 1930s, '40s, and '50s. 473 00:22:16,480 --> 00:22:21,050 So what was done in two areas, and one 474 00:22:21,050 --> 00:22:24,600 was in Germany, what they did there, 475 00:22:24,600 --> 00:22:30,120 they kind of followed in the spirit of Keffer Hartline. 476 00:22:30,120 --> 00:22:33,650 And they shone light into the eye, diffused light. 477 00:22:33,650 --> 00:22:38,230 And they recorded from V1, and they couldn't derive any cells 478 00:22:38,230 --> 00:22:41,260 to the extent that some people said, 479 00:22:41,260 --> 00:22:44,670 I think people have made a major mistake. 480 00:22:44,670 --> 00:22:47,330 That stuff back here, that's not a visual area. 481 00:22:47,330 --> 00:22:48,900 It's something else. 482 00:22:48,900 --> 00:22:54,660 But then, pretty much at the same time, another group, 483 00:22:54,660 --> 00:23:01,860 initially at Johns Hopkins, and subsequently at Harvard, 484 00:23:01,860 --> 00:23:03,830 did similar experiments. 485 00:23:03,830 --> 00:23:06,650 But what happened was, it's almost an amusing story. 486 00:23:06,650 --> 00:23:09,910 The people who did that were Hubel and Wiesel. 487 00:23:09,910 --> 00:23:13,160 And when they first went to Johns Hopkins, 488 00:23:13,160 --> 00:23:14,760 they worked with a person's name I've 489 00:23:14,760 --> 00:23:16,830 mentioned before, I think, Kuffler, 490 00:23:16,830 --> 00:23:19,450 who discovered center-surround antagonism. 491 00:23:19,450 --> 00:23:22,840 And Kuffler did experiments similar in technique, 492 00:23:22,840 --> 00:23:27,280 in terms of activating the eye, that 493 00:23:27,280 --> 00:23:29,590 had been done by Keffer Hartline, which shine light 494 00:23:29,590 --> 00:23:31,080 into the eye. 495 00:23:31,080 --> 00:23:34,390 Not necessarily diffused light, but spots of light 496 00:23:34,390 --> 00:23:37,570 that he would move around on the retinal surface. 497 00:23:37,570 --> 00:23:40,270 This was a very complicated piece of equipment. 498 00:23:40,270 --> 00:23:46,590 And so when Hubel and Wiesel went to become postdocs 499 00:23:46,590 --> 00:23:50,470 with Kuffler, they said, my god, what is this? 500 00:23:50,470 --> 00:23:52,860 We don't know how to work this stuff. 501 00:23:52,860 --> 00:23:56,050 Let's do something different, so we can handle it. 502 00:23:56,050 --> 00:24:01,700 And so what they decided to do, and this may sound crazy-- 503 00:24:01,700 --> 00:24:03,430 Let me explain one more thing. 504 00:24:03,430 --> 00:24:06,530 When these experiments were done in the cat, with the way 505 00:24:06,530 --> 00:24:08,430 the experiment was done, that the cat was 506 00:24:08,430 --> 00:24:11,540 put upside down on a table like this. 507 00:24:11,540 --> 00:24:12,410 Like that. 508 00:24:12,410 --> 00:24:13,878 And the light was shone into it. 509 00:24:13,878 --> 00:24:14,378 OK? 510 00:24:16,010 --> 00:24:17,850 With a piece of equipment from above, 511 00:24:17,850 --> 00:24:21,340 and you could even look at it through a microscope 512 00:24:21,340 --> 00:24:24,294 to see how that light was shone into the eye. 513 00:24:25,656 --> 00:24:29,020 So Hubel and Wiesel said, you can't handle this. 514 00:24:29,020 --> 00:24:30,810 But I tell you what we'll do instead. 515 00:24:30,810 --> 00:24:35,120 You'll take a bed sheet, and put it up on the ceiling. 516 00:24:35,120 --> 00:24:38,400 And then we'll take a projector, and move the light around 517 00:24:38,400 --> 00:24:39,600 like this. 518 00:24:39,600 --> 00:24:40,100 OK? 519 00:24:40,100 --> 00:24:43,710 Now, the projector they used was like an old fashioned camera. 520 00:24:43,710 --> 00:24:45,430 And you could put a slide in there, 521 00:24:45,430 --> 00:24:47,170 and move the slide in and out. 522 00:24:47,170 --> 00:24:48,160 OK? 523 00:24:48,160 --> 00:24:50,070 And so they would, first of all, just 524 00:24:50,070 --> 00:24:51,777 had the whole lights come on. 525 00:24:51,777 --> 00:24:54,110 And they kept recording, and they wouldn't get anything. 526 00:24:54,110 --> 00:24:55,270 They said, oh my god. 527 00:24:55,270 --> 00:24:58,922 Those people in Germany are right. 528 00:24:58,922 --> 00:25:00,740 These cells don't respond to light. 529 00:25:00,740 --> 00:25:02,220 What's going on? 530 00:25:02,220 --> 00:25:07,410 And then one day, as they were fiddling around with this, 531 00:25:07,410 --> 00:25:11,030 Torsten Wiesel pulled out a slide 532 00:25:11,030 --> 00:25:12,610 to put a different slide in. 533 00:25:12,610 --> 00:25:15,580 And when he pulled that slide out, 534 00:25:15,580 --> 00:25:19,090 they displayed the sound of the action potentials. 535 00:25:20,320 --> 00:25:21,380 Like that. 536 00:25:21,380 --> 00:25:23,210 What on earth was that? 537 00:25:23,210 --> 00:25:24,620 And so they did it repeated. 538 00:25:24,620 --> 00:25:28,390 [HIGH-PITCHED BEEPING SOUND] Like that. 539 00:25:28,390 --> 00:25:30,950 And guess what that resulted in? 540 00:25:30,950 --> 00:25:35,000 That finding [INAUDIBLE] logic that resulted in the Nobel 541 00:25:35,000 --> 00:25:39,760 Prize he and Hubel got, because they discovered 542 00:25:39,760 --> 00:25:45,455 that cells' individual cortex are orientation selective. 543 00:25:46,470 --> 00:25:47,590 OK? 544 00:25:47,590 --> 00:25:51,980 Meaning that each cell responds to a particular orientation 545 00:25:51,980 --> 00:25:53,350 of an edge. 546 00:25:53,350 --> 00:25:55,710 And this was an incredible transformation, 547 00:25:55,710 --> 00:25:57,470 which we saw in the retina, and what 548 00:25:57,470 --> 00:26:01,060 you had seen in the lateral geniculate nucleus. 549 00:26:01,060 --> 00:26:05,360 So what they then did, they became systematic about it, 550 00:26:05,360 --> 00:26:09,560 and mapped out the receptive field organization of cells. 551 00:26:09,560 --> 00:26:12,920 And they discovered two majors classes of cells, 552 00:26:12,920 --> 00:26:15,990 they called the simple cells, and the complex cells. 553 00:26:15,990 --> 00:26:19,170 The simple cells were distinctive in the sense, 554 00:26:19,170 --> 00:26:21,470 first of all, all these types of cells 555 00:26:21,470 --> 00:26:23,230 were orientation specific. 556 00:26:23,230 --> 00:26:25,220 I want to find the appropriate orientation 557 00:26:25,220 --> 00:26:29,830 by moving the edge around until you get the best response. 558 00:26:29,830 --> 00:26:31,460 Then you could do the very small spot. 559 00:26:32,530 --> 00:26:35,290 You could activate the cell, but not as 560 00:26:35,290 --> 00:26:38,390 effectively as with the ball moving, of course, but still 561 00:26:38,390 --> 00:26:39,270 quite well. 562 00:26:39,270 --> 00:26:40,980 And they found that there was a center 563 00:26:40,980 --> 00:26:45,970 region in this kind of simple cell that gave an on response, 564 00:26:45,970 --> 00:26:48,830 and a surround region that gave an off response. 565 00:26:50,060 --> 00:26:51,570 And then they found another class 566 00:26:51,570 --> 00:26:56,940 of simple cells, in which the left half in this case was on, 567 00:26:56,940 --> 00:26:58,980 and the right half was off. 568 00:26:58,980 --> 00:27:01,010 And there were variations of this, 569 00:27:01,010 --> 00:27:05,010 I'll mention those in a bit more detail later, 570 00:27:05,010 --> 00:27:09,990 indicating that the input from the on and off systems 571 00:27:09,990 --> 00:27:13,160 may be separate, or they [? sub ?] some interaction 572 00:27:13,160 --> 00:27:16,800 that makes for their specificity for orientation. 573 00:27:18,500 --> 00:27:22,500 Now then, they also discovered another class of cells called 574 00:27:22,500 --> 00:27:24,640 complex that on the whole tended to have 575 00:27:24,640 --> 00:27:27,500 a bit larger receptive fields. 576 00:27:27,500 --> 00:27:30,270 And in these cells, even when you use small spots, 577 00:27:30,270 --> 00:27:32,780 no matter where you stimulate it, 578 00:27:32,780 --> 00:27:35,430 you got both on and off responses intermingled. 579 00:27:36,540 --> 00:27:40,770 So you did not have a spatial separation 580 00:27:40,770 --> 00:27:42,615 of the on and the off responses. 581 00:27:43,720 --> 00:27:46,640 Now what exactly we mean by on and off responses 582 00:27:46,640 --> 00:27:50,690 we will discuss for the whole session next time. 583 00:27:50,690 --> 00:27:51,190 OK. 584 00:27:51,190 --> 00:27:55,190 So now, let me explain to you how you do these experiments. 585 00:27:55,190 --> 00:27:58,680 Once you get a little bit more sophisticated, and instead 586 00:27:58,680 --> 00:28:01,510 of just using a projector that you reflect the light 587 00:28:01,510 --> 00:28:04,590 with by hand, you can do this on a computer. 588 00:28:04,590 --> 00:28:05,230 OK? 589 00:28:05,230 --> 00:28:08,320 And that indeed was a big step forward. 590 00:28:08,320 --> 00:28:15,620 I would say in the early 1960s, maybe late '50s or early '60s, 591 00:28:15,620 --> 00:28:20,340 people became computerized as we had especially here at MIT. 592 00:28:20,340 --> 00:28:24,710 And the first set of computers that were used to enable us 593 00:28:24,710 --> 00:28:29,031 to quantitatively measure these attributes in the visual cortex 594 00:28:29,031 --> 00:28:30,280 were the so-called PDP-11/10s. 595 00:28:32,400 --> 00:28:33,600 Subsequently, PDP-11/20s. 596 00:28:34,760 --> 00:28:36,225 Subsequently then PDP-11/34s. 597 00:28:37,570 --> 00:28:42,650 Some of you may have heard about these ancient old computers 598 00:28:42,650 --> 00:28:44,820 that practically nobody uses anymore, 599 00:28:44,820 --> 00:28:46,800 because of all the advances that have 600 00:28:46,800 --> 00:28:49,130 been made in computer technology. 601 00:28:49,130 --> 00:28:52,020 So what you did then using a variety of means, 602 00:28:52,020 --> 00:28:54,330 I don't have to describe that in detail, is 603 00:28:54,330 --> 00:28:58,160 to move a bar of light across the receptive 604 00:28:58,160 --> 00:29:00,180 field in different orientations. 605 00:29:00,180 --> 00:29:01,740 So here's an example here. 606 00:29:01,740 --> 00:29:04,030 You first use very small spots, and you 607 00:29:04,030 --> 00:29:06,160 find the receptive field. 608 00:29:06,160 --> 00:29:08,030 And then after, because the cells 609 00:29:08,030 --> 00:29:10,120 do respond to very small spots of light. 610 00:29:10,120 --> 00:29:12,940 But they don't respond at all to larger spots. 611 00:29:12,940 --> 00:29:15,320 And once you've found the receptive field, what 612 00:29:15,320 --> 00:29:17,240 you can do is you can take a bar, 613 00:29:17,240 --> 00:29:19,940 and move it across it in different orientations, 614 00:29:19,940 --> 00:29:20,830 like this. 615 00:29:20,830 --> 00:29:22,565 So here's a direct example. 616 00:29:25,060 --> 00:29:25,690 Like that. 617 00:29:25,690 --> 00:29:26,300 OK? 618 00:29:26,300 --> 00:29:28,150 So as soon as it's been across the field, 619 00:29:28,150 --> 00:29:29,445 the cell responded vigorously. 620 00:29:30,860 --> 00:29:33,280 And so if we then did it systematically, 621 00:29:33,280 --> 00:29:39,150 as I've shown here, you can generate an orientation tuning 622 00:29:39,150 --> 00:29:40,960 function, it is called. 623 00:29:40,960 --> 00:29:43,470 So you can quantitatively establish just 624 00:29:43,470 --> 00:29:46,410 how sharply specific a cell like this 625 00:29:46,410 --> 00:29:48,830 is to different orientations. 626 00:29:48,830 --> 00:29:51,690 So to do that here is an example of that in a second. 627 00:29:52,970 --> 00:29:54,460 Here is an example. 628 00:29:54,460 --> 00:29:58,270 What you can see here is when you move the ball down, 629 00:29:58,270 --> 00:30:00,400 you get a huge response here. 630 00:30:00,400 --> 00:30:04,020 And then we get rapidly progressively less 631 00:30:04,020 --> 00:30:05,340 of a response. 632 00:30:05,340 --> 00:30:08,960 And then you can establish by taking the half height, 633 00:30:08,960 --> 00:30:12,815 or whatever point here, what the width is of this function. 634 00:30:12,815 --> 00:30:15,720 And then you can plot that for hundreds of cells 635 00:30:15,720 --> 00:30:19,390 to establish just how sharply they are tuned, 636 00:30:19,390 --> 00:30:23,030 or how different types of cells, say simple and complex, how 637 00:30:23,030 --> 00:30:24,790 sharp they are tuned. 638 00:30:24,790 --> 00:30:27,120 Now, that was only quite amazing. 639 00:30:27,120 --> 00:30:29,740 But then if you look at this figure, 640 00:30:29,740 --> 00:30:32,470 I want you guys to be my best detectives. 641 00:30:32,470 --> 00:30:35,080 I bet you all of you are best detectives, right? 642 00:30:35,080 --> 00:30:37,310 So if you are, you have to tell me 643 00:30:37,310 --> 00:30:40,270 what else is in this figure that tells you something 644 00:30:40,270 --> 00:30:42,890 more about this cell that is not just orientation. 645 00:30:50,050 --> 00:30:50,550 Yes. 646 00:30:50,550 --> 00:30:52,554 AUDIENCE: Also, the direction as well. 647 00:30:52,554 --> 00:30:53,220 PROFESSOR: A-ha. 648 00:30:53,220 --> 00:30:54,150 There we go. 649 00:30:54,150 --> 00:30:55,010 Best detective. 650 00:30:55,010 --> 00:30:55,510 OK. 651 00:30:55,510 --> 00:30:56,750 Is that your middle name? 652 00:30:56,750 --> 00:30:57,560 Best detective? 653 00:30:57,560 --> 00:30:57,960 AUDIENCE: Sure. 654 00:30:57,960 --> 00:30:58,330 PROFESSOR: OK. 655 00:30:58,330 --> 00:30:58,940 Very good. 656 00:30:58,940 --> 00:31:01,600 So what you see here, if this cell only 657 00:31:01,600 --> 00:31:05,687 were orientation selective, you should get another peak here, 658 00:31:05,687 --> 00:31:07,270 because you have the same orientation, 659 00:31:07,270 --> 00:31:09,330 but moving upward, like this. 660 00:31:09,330 --> 00:31:09,830 OK? 661 00:31:09,830 --> 00:31:11,660 But you don't have anything like that. 662 00:31:11,660 --> 00:31:12,990 So what does that mean? 663 00:31:12,990 --> 00:31:15,850 That means that this particular cell is not 664 00:31:15,850 --> 00:31:19,995 only orientation specific, but it is also direction specific. 665 00:31:21,400 --> 00:31:22,910 Now that's quite something. 666 00:31:22,910 --> 00:31:25,440 And the overwhelming majority of cells 667 00:31:25,440 --> 00:31:30,459 in the visual cortex in V1 are direction selective, as well, 668 00:31:30,459 --> 00:31:32,250 in addition to being orientation selective. 669 00:31:33,960 --> 00:31:35,640 And then when you go to other areas, 670 00:31:35,640 --> 00:31:37,480 as we'll see in just a minute, there 671 00:31:37,480 --> 00:31:40,270 are some areas in which virtually all the cells 672 00:31:40,270 --> 00:31:42,330 are direction selective. 673 00:31:42,330 --> 00:31:45,160 So direction cell activity becomes 674 00:31:45,160 --> 00:31:50,970 an inherent central attribute of visual processing, 675 00:31:50,970 --> 00:31:54,850 as revealed by the fact that all the cells, I shouldn't say all, 676 00:31:54,850 --> 00:31:57,890 but so many of the cells in the visual cortex 677 00:31:57,890 --> 00:31:59,581 have that attribute. 678 00:31:59,581 --> 00:32:00,080 OK. 679 00:32:00,080 --> 00:32:04,410 So now, what we can do is ask the next question. 680 00:32:04,410 --> 00:32:08,680 Now, are there other attributes in the cells 681 00:32:08,680 --> 00:32:13,500 of the visual cortex that are selective [INAUDIBLE] 682 00:32:13,500 --> 00:32:16,590 some other attribute of the visual scene? 683 00:32:16,590 --> 00:32:19,450 So far we have orientation, and we have direction. 684 00:32:19,450 --> 00:32:22,230 So another one that has been studied 685 00:32:22,230 --> 00:32:26,520 is spatial frequency selectivity. 686 00:32:26,520 --> 00:32:28,170 Now, how do you do that? 687 00:32:28,170 --> 00:32:31,220 The most common way this had been done 688 00:32:31,220 --> 00:32:33,000 is to use sinusoidal gratings. 689 00:32:33,000 --> 00:32:35,067 An example of that is shown here. 690 00:32:35,067 --> 00:32:37,150 But you could use also [? square-wave ?] gradings, 691 00:32:37,150 --> 00:32:40,610 or sinusoidal gratings work extremely well for a variety 692 00:32:40,610 --> 00:32:41,220 of reasons. 693 00:32:41,220 --> 00:32:42,960 We will talk about that later. 694 00:32:42,960 --> 00:32:45,630 And what you do then on each trial, 695 00:32:45,630 --> 00:32:48,130 as you move this back and forth across the receptive 696 00:32:48,130 --> 00:32:50,060 field in the optimal orientation, 697 00:32:50,060 --> 00:32:52,930 you can vary the spatial frequency. 698 00:32:52,930 --> 00:32:55,090 You can make it extremely high, or extremely low. 699 00:32:55,090 --> 00:32:56,730 This is obviously very low. 700 00:32:56,730 --> 00:32:57,560 OK? 701 00:32:57,560 --> 00:33:00,510 And if you do that systematically once again, 702 00:33:00,510 --> 00:33:04,080 using a computer system that can establish to what degree 703 00:33:04,080 --> 00:33:07,480 are cells specific to different spatial frequencies. 704 00:33:07,480 --> 00:33:09,000 Everybody understand that? 705 00:33:09,000 --> 00:33:09,570 OK. 706 00:33:09,570 --> 00:33:11,540 So let me show you what happens. 707 00:33:11,540 --> 00:33:13,480 Here is an example of a simple cell, 708 00:33:13,480 --> 00:33:15,660 and here's an example of a complex cell. 709 00:33:15,660 --> 00:33:19,690 Here we have different spatial frequencies on each line, 710 00:33:19,690 --> 00:33:21,540 as specified in here. 711 00:33:21,540 --> 00:33:24,740 And it shows that both of these types of cells 712 00:33:24,740 --> 00:33:28,350 respond to some in-between level of spatial frequency 713 00:33:28,350 --> 00:33:31,980 selectivity, and don't respond at the extremes. 714 00:33:31,980 --> 00:33:35,630 So that means that these cortical cells, in addition 715 00:33:35,630 --> 00:33:37,930 to being orientation and direction selective, 716 00:33:37,930 --> 00:33:40,300 are also spatial frequency selective. 717 00:33:41,480 --> 00:33:44,840 And that had lead to many interesting ideas 718 00:33:44,840 --> 00:33:49,680 about what is the coding operation in the visual cortex 719 00:33:49,680 --> 00:33:51,200 that enables you to see. 720 00:33:51,200 --> 00:33:53,525 And that we will discuss some more later. 721 00:33:54,680 --> 00:33:55,240 OK. 722 00:33:55,240 --> 00:33:58,320 The other important feature here that is worth mentioning 723 00:33:58,320 --> 00:34:05,070 is that simple cells respond selectively to each phase, 724 00:34:05,070 --> 00:34:08,827 whereas complex cells respond in a more general fashion. 725 00:34:08,827 --> 00:34:10,250 OK? 726 00:34:10,250 --> 00:34:10,750 All right. 727 00:34:10,750 --> 00:34:14,823 So now, we can summarize, and I'm 728 00:34:14,823 --> 00:34:16,489 going to add a couple more things to it, 729 00:34:16,489 --> 00:34:19,964 of what the major so-called transforms are 730 00:34:19,964 --> 00:34:22,040 in the visual cortex. 731 00:34:22,040 --> 00:34:24,330 And this is remarkable as you go from the geniculate 732 00:34:24,330 --> 00:34:27,189 to the visual cortex, and suddenly there 733 00:34:27,189 --> 00:34:30,860 are all these transforms that somehow 734 00:34:30,860 --> 00:34:33,400 will enable you to see things better. 735 00:34:33,400 --> 00:34:35,710 So the first one of course, we talked about a lot, 736 00:34:35,710 --> 00:34:36,940 is orientation. 737 00:34:36,940 --> 00:34:38,370 Next one is direction. 738 00:34:38,370 --> 00:34:39,739 Then spatial frequency. 739 00:34:39,739 --> 00:34:41,670 Those are the three I gave you examples of. 740 00:34:41,670 --> 00:34:44,690 But now, one thing I haven't mentioned so far 741 00:34:44,690 --> 00:34:47,570 is that you have also a transform 742 00:34:47,570 --> 00:34:49,400 in terms of binocularity. 743 00:34:49,400 --> 00:34:52,790 As I mentioned to you that already, in the geniculate 744 00:34:52,790 --> 00:34:56,080 you have specificity by layers of input 745 00:34:56,080 --> 00:34:57,860 from the left and right eyes, which 746 00:34:57,860 --> 00:35:01,160 means that those cells in each of those layers 747 00:35:01,160 --> 00:35:02,701 are driven only by one eye. 748 00:35:02,701 --> 00:35:03,200 OK? 749 00:35:03,200 --> 00:35:04,170 They have a monocular. 750 00:35:05,380 --> 00:35:07,940 But now when you come up to the cortex, 751 00:35:07,940 --> 00:35:12,590 the numerous cells once you get above the major input 752 00:35:12,590 --> 00:35:15,910 layers, 4CL from there or below, most of the cells 753 00:35:15,910 --> 00:35:18,750 there can be driven binocularly, meaning 754 00:35:18,750 --> 00:35:22,585 they get an input from both eyes. 755 00:35:23,940 --> 00:35:27,020 And we shall examine what that is. 756 00:35:27,020 --> 00:35:30,010 And another thing that we have that you already 757 00:35:30,010 --> 00:35:32,400 must have inferred on the basis looking at the receptive 758 00:35:32,400 --> 00:35:36,850 fields, that there is a dramatic ON/OFF convergence. 759 00:35:36,850 --> 00:35:38,434 In complex terms, it's so complete 760 00:35:38,434 --> 00:35:39,725 that it's totally intermingled. 761 00:35:40,790 --> 00:35:43,860 So the cell responds to on and off by the way you present it. 762 00:35:43,860 --> 00:35:46,010 The simple cells also respond to on and off, 763 00:35:46,010 --> 00:35:48,550 but there's a small spatial separation 764 00:35:48,550 --> 00:35:51,680 between the on and off sub region. 765 00:35:51,680 --> 00:35:58,054 Now in addition, that we will encounter also later, 766 00:35:58,054 --> 00:35:59,720 some of the cells, not all of the cells, 767 00:35:59,720 --> 00:36:02,420 but some of the cells in the visual cortex 768 00:36:02,420 --> 00:36:06,280 get a convergent in both the midget and parasol cells, 769 00:36:06,280 --> 00:36:09,900 while others get an input only from one or the other. 770 00:36:09,900 --> 00:36:16,190 So that further highlights the extensive increase 771 00:36:16,190 --> 00:36:18,470 of complexity, and analysis that is 772 00:36:18,470 --> 00:36:21,610 being performed in the visual cortex. 773 00:36:21,610 --> 00:36:22,230 OK. 774 00:36:22,230 --> 00:36:26,970 So now, to sort of schematize about this, what we can say 775 00:36:26,970 --> 00:36:28,010 is the following. 776 00:36:28,010 --> 00:36:31,710 We can say here is that we have a cell 777 00:36:31,710 --> 00:36:35,150 in primary visual cortex, in V1. 778 00:36:35,150 --> 00:36:35,950 OK? 779 00:36:35,950 --> 00:36:37,730 And many of these cells, most of them, 780 00:36:37,730 --> 00:36:41,530 get an input from both the left and the right eyes, 781 00:36:41,530 --> 00:36:43,400 as shown here. 782 00:36:43,400 --> 00:36:46,540 They also get an input, many of them, 783 00:36:46,540 --> 00:36:49,790 that are convergent from the parasol and midget system. 784 00:36:51,680 --> 00:36:54,360 And then the output of these cells, 785 00:36:54,360 --> 00:36:56,800 in some fashion or other, should be 786 00:36:56,800 --> 00:36:58,890 able to tell us, either from the same cell 787 00:36:58,890 --> 00:37:01,250 or from separate cells that we've examined, 788 00:37:01,250 --> 00:37:05,400 tell us about luminance, color, orientation, spatial frequency, 789 00:37:05,400 --> 00:37:07,360 depth and motion. 790 00:37:07,360 --> 00:37:12,790 So those are some of the major tasks a cortical cell performs, 791 00:37:12,790 --> 00:37:14,540 so that you can see. 792 00:37:15,700 --> 00:37:18,820 So that then is a very general scheme 793 00:37:18,820 --> 00:37:20,760 that we are then going to examine 794 00:37:20,760 --> 00:37:22,055 in considerably more detail. 795 00:37:25,090 --> 00:37:25,590 All right. 796 00:37:25,590 --> 00:37:28,350 So now, we take another step forward. 797 00:37:30,020 --> 00:37:31,700 The question came up. 798 00:37:33,200 --> 00:37:39,190 How are these things in general, these attributes and whatnot, 799 00:37:39,190 --> 00:37:41,700 arranged in the visual cortex? 800 00:37:41,700 --> 00:37:43,710 Is everything just helter skelter, 801 00:37:43,710 --> 00:37:47,110 or is there spatial separation among various attributes? 802 00:37:47,110 --> 00:37:50,800 And this is yet another major line of research 803 00:37:50,800 --> 00:37:54,289 that Hubel and Wiesel had done, for which they got the Nobel 804 00:37:54,289 --> 00:37:56,205 Prize, as well as for their other discoveries. 805 00:37:58,850 --> 00:38:05,331 I think they got the Nobel Prize in 1983, I believe it was. 806 00:38:05,331 --> 00:38:05,830 OK. 807 00:38:05,830 --> 00:38:07,957 So this is called cytoarchitecture. 808 00:38:10,010 --> 00:38:10,510 OK? 809 00:38:10,510 --> 00:38:17,960 Your ability to understand what the layout is of the functional 810 00:38:17,960 --> 00:38:22,211 attributes of the cells in the visual cortex. 811 00:38:22,211 --> 00:38:22,710 All right. 812 00:38:22,710 --> 00:38:24,870 So let me, first of all, tell you 813 00:38:24,870 --> 00:38:27,700 the initial experiment Hubel and Wiesel had done. 814 00:38:27,700 --> 00:38:33,100 What they did is they would inject a label into one eye. 815 00:38:33,100 --> 00:38:36,780 And then if they waited a week or so, 816 00:38:36,780 --> 00:38:40,070 you would get transneuronal transport, 817 00:38:40,070 --> 00:38:43,080 meaning that the label would go to the geniculate, 818 00:38:43,080 --> 00:38:47,250 and that label would transport to the cells in the geniculate, 819 00:38:47,250 --> 00:38:50,440 and then would go up to the visual cortex. 820 00:38:50,440 --> 00:38:53,930 And so when they looked at this, put the input 821 00:38:53,930 --> 00:38:57,290 into the left eye, what they found 822 00:38:57,290 --> 00:39:02,220 was in the cortex of the cat in this case, 823 00:39:02,220 --> 00:39:04,250 you see these alternations, you can 824 00:39:04,250 --> 00:39:06,500 label them unlabeled regions. 825 00:39:06,500 --> 00:39:08,185 And the labels in unlabeled regions 826 00:39:08,185 --> 00:39:09,685 are pretty much equally distributed. 827 00:39:10,870 --> 00:39:15,300 The assumption therefore is, and that was proven of course, 828 00:39:15,300 --> 00:39:17,510 that if you were to label both sides, 829 00:39:17,510 --> 00:39:20,390 then you would get a continuous label. 830 00:39:21,490 --> 00:39:25,135 These include areas, or if you labeled the other eye, 831 00:39:25,135 --> 00:39:28,710 you would get the dark areas labeled. 832 00:39:28,710 --> 00:39:33,730 So this established then that we have an orderly arrangement 833 00:39:33,730 --> 00:39:39,679 of, if you will, ocular dominance columns. 834 00:39:39,679 --> 00:39:40,720 That's what it is called. 835 00:39:40,720 --> 00:39:43,160 Ocular dominance columns. 836 00:39:43,160 --> 00:39:45,250 Now, this is a cross section. 837 00:39:45,250 --> 00:39:48,640 So now, what you can do is do a similar experiment. 838 00:39:48,640 --> 00:39:50,850 Unless you're doing a cross section, 839 00:39:50,850 --> 00:39:56,150 you can take the brain, and make horizontal cuts across it. 840 00:39:56,150 --> 00:39:58,190 Now, the brain is slightly curved, of course, 841 00:39:58,190 --> 00:40:00,040 depending on the species how much. 842 00:40:00,040 --> 00:40:04,160 And so, when you make these very thin cross sections, 843 00:40:04,160 --> 00:40:07,750 they will only tell you about that particular cross section. 844 00:40:07,750 --> 00:40:09,470 But then you can put them together. 845 00:40:09,470 --> 00:40:12,190 You can take a montage. 846 00:40:12,190 --> 00:40:13,710 So let me explain that to you. 847 00:40:13,710 --> 00:40:18,530 Here is a single section of the left eye 848 00:40:18,530 --> 00:40:21,520 columns that are labeled in the monkey. 849 00:40:21,520 --> 00:40:24,830 So in other words to repeat, you inject a label 850 00:40:24,830 --> 00:40:26,970 into the left eye. 851 00:40:26,970 --> 00:40:29,620 Then that is transported transneuronally 852 00:40:29,620 --> 00:40:32,480 to the visual cortex, and this is how it lights up. 853 00:40:34,094 --> 00:40:35,510 But now what you can do is you can 854 00:40:35,510 --> 00:40:37,500 take each of these sections. 855 00:40:37,500 --> 00:40:38,680 There are many sections. 856 00:40:38,680 --> 00:40:42,060 I told you the cortex is about 2 millimeters in thickness. 857 00:40:42,060 --> 00:40:42,780 OK? 858 00:40:42,780 --> 00:40:45,220 So we can take a huge number of sections 859 00:40:45,220 --> 00:40:49,090 across, depending on how thick or thin you want to cut it. 860 00:40:49,090 --> 00:40:52,660 And then you can label each of these, 861 00:40:52,660 --> 00:40:54,680 and then you can superimpose them. 862 00:40:54,680 --> 00:40:57,020 And this is what we call a montage. 863 00:40:57,020 --> 00:41:03,560 In this case we have five layers that have been superimposed, 864 00:41:03,560 --> 00:41:05,250 so you can see it. 865 00:41:05,250 --> 00:41:08,950 Now, these columns here, so-called, it's 866 00:41:08,950 --> 00:41:12,590 a column because it goes through the thickness of the cortex, 867 00:41:12,590 --> 00:41:18,050 are then ocular dominance columns. 868 00:41:18,050 --> 00:41:18,550 OK? 869 00:41:18,550 --> 00:41:20,200 Ocular dominance columns. 870 00:41:21,890 --> 00:41:24,850 We will next examine whether they're also 871 00:41:24,850 --> 00:41:27,080 columns for orientation. 872 00:41:27,080 --> 00:41:28,861 That's why I said orientation. 873 00:41:28,861 --> 00:41:29,360 OK. 874 00:41:29,360 --> 00:41:31,650 Anyway, so we have this arrangement. 875 00:41:31,650 --> 00:41:36,600 And these often have been called zebra stripes, 876 00:41:36,600 --> 00:41:38,550 because it looks like a zebra. 877 00:41:38,550 --> 00:41:42,440 And that helps as a mnemonic device to remember this. 878 00:41:42,440 --> 00:41:45,659 Now, what you can do is to examine 879 00:41:45,659 --> 00:41:46,950 this a bit more systematically. 880 00:41:48,000 --> 00:41:51,730 And if you do that, we can draw together 881 00:41:51,730 --> 00:41:54,300 a huge section of the visual cortex. 882 00:41:54,300 --> 00:41:56,880 This is the fovea again, in this case. 883 00:41:56,880 --> 00:41:59,220 This is about 80 degrees out. 884 00:41:59,220 --> 00:42:03,310 And so this is a section that was put together in the Hubel 885 00:42:03,310 --> 00:42:06,480 Wiesel Laboratory at Harvard, showing you 886 00:42:06,480 --> 00:42:09,195 that the thickness of them is constant throughout. 887 00:42:10,390 --> 00:42:13,360 And then to understand what they're talking about in terms 888 00:42:13,360 --> 00:42:17,470 of size, here we have David Hubel's thumbprint. 889 00:42:17,470 --> 00:42:18,070 OK? 890 00:42:18,070 --> 00:42:19,670 So if you look at your thumbprint, 891 00:42:19,670 --> 00:42:22,900 the spatial frequency here is about twice roughly, 892 00:42:22,900 --> 00:42:25,330 twice that you have here. 893 00:42:25,330 --> 00:42:28,050 So again, if you look at your thumb like this, 894 00:42:28,050 --> 00:42:32,050 it can give you a sense of how unbelievably fine 895 00:42:32,050 --> 00:42:34,350 these columns are in the visual cortex. 896 00:42:34,350 --> 00:42:34,850 OK? 897 00:42:35,940 --> 00:42:37,840 So that's basically the layout. 898 00:42:37,840 --> 00:42:42,380 So now, the next big question came up in the work 899 00:42:42,380 --> 00:42:45,790 that Hubel and Wiesel originally did, 900 00:42:45,790 --> 00:42:56,096 was whether you also had columns for orientation. 901 00:42:58,390 --> 00:42:59,370 So how do you do that? 902 00:43:00,700 --> 00:43:03,840 Well, another technique has evolved, 903 00:43:03,840 --> 00:43:06,990 and everything that you discover in this business 904 00:43:06,990 --> 00:43:11,010 heavily depends on you being able to come up 905 00:43:11,010 --> 00:43:12,680 with a new technique, or somebody else, 906 00:43:12,680 --> 00:43:14,790 and use a technique they invented. 907 00:43:14,790 --> 00:43:17,350 And when you do that, almost inevitably you 908 00:43:17,350 --> 00:43:19,410 can make a major discovery. 909 00:43:19,410 --> 00:43:23,380 But if you're not sensitive to new technologies, 910 00:43:23,380 --> 00:43:24,880 then it's unlikely that you're going 911 00:43:24,880 --> 00:43:26,130 to make a major new discovery. 912 00:43:27,260 --> 00:43:29,970 So you've got to keep your nose to the grindstone, 913 00:43:29,970 --> 00:43:31,370 and constantly look, and say what 914 00:43:31,370 --> 00:43:33,270 is the latest discovery in many techniques. 915 00:43:34,401 --> 00:43:34,900 All right. 916 00:43:34,900 --> 00:43:37,980 So if you do that, one of the remarkable techniques that 917 00:43:37,980 --> 00:43:41,110 had been developed is to use a substance 918 00:43:41,110 --> 00:43:45,314 called 2-deoxyglucose that I'm not 919 00:43:45,314 --> 00:43:46,730 going to go into details about it. 920 00:43:46,730 --> 00:43:50,950 It's radioactively labeled as well, 921 00:43:50,950 --> 00:43:53,540 and what you inject in the bloodstream. 922 00:43:53,540 --> 00:43:58,370 And then what happens is that those cells in the brain that 923 00:43:58,370 --> 00:44:03,010 are highly active, absorb more of this 2-deoxyglucose 924 00:44:03,010 --> 00:44:05,110 than those cells that are inactive. 925 00:44:05,110 --> 00:44:08,260 And so if you do that, the experiment 926 00:44:08,260 --> 00:44:12,710 you would do to look at the selectivity for orientation, 927 00:44:12,710 --> 00:44:15,380 is you take an animal, unless it is paralyzed, 928 00:44:15,380 --> 00:44:22,420 and present for an hour vertically oriented set 929 00:44:22,420 --> 00:44:25,580 of bars, or signs of gratings that keep moving, 930 00:44:25,580 --> 00:44:27,040 and keep moving, and keep moving. 931 00:44:28,780 --> 00:44:35,190 And then those cells that are, in this case going this way, 932 00:44:35,190 --> 00:44:37,400 have selected the horizontal orientation. 933 00:44:37,400 --> 00:44:40,190 Those will take up a lot more of this 2-deoxyglucose, 934 00:44:40,190 --> 00:44:42,730 and therefore that would be heavily labeled. 935 00:44:42,730 --> 00:44:45,880 So then the question is, what does that look like? 936 00:44:45,880 --> 00:44:47,730 And so if you do that kind of experiment, 937 00:44:47,730 --> 00:44:51,150 let me skip that, here's an experiment like that, 938 00:44:51,150 --> 00:44:53,130 and this happens to be the tree shrew. 939 00:44:53,130 --> 00:44:55,530 It shows that in this case, when you 940 00:44:55,530 --> 00:44:58,520 use one particular orientation that you 941 00:44:58,520 --> 00:45:02,940 get these stripes, which reflect those cells 942 00:45:02,940 --> 00:45:06,720 in the visual cortex in the column arrangement 943 00:45:06,720 --> 00:45:09,420 that respond best to that orientation. 944 00:45:09,420 --> 00:45:12,140 So therefore, if this is horizontal, 945 00:45:12,140 --> 00:45:15,200 the ones in between with these cells are vertical. 946 00:45:15,200 --> 00:45:16,040 All right? 947 00:45:16,040 --> 00:45:19,990 So if you then do this systematically in the monkey, 948 00:45:19,990 --> 00:45:23,060 you can see exactly what that looks like. 949 00:45:23,060 --> 00:45:26,250 And then you can ask the big question that was posed. 950 00:45:26,250 --> 00:45:31,510 What is the relationship between orientation selectivity columns 951 00:45:31,510 --> 00:45:34,350 and ocular dominance columns? 952 00:45:34,350 --> 00:45:38,280 So that was then done by a set of experiments, 953 00:45:38,280 --> 00:45:40,340 also in the Hubel-Wiesel Laboratory. 954 00:45:40,340 --> 00:45:43,980 And I'm proud to say that the prime person who 955 00:45:43,980 --> 00:45:47,460 did that experiment was one of my former students, who 956 00:45:47,460 --> 00:45:50,200 got his PhD in my laboratory, Michael Stryker. 957 00:45:50,200 --> 00:45:52,040 And so he did these experiments. 958 00:45:53,210 --> 00:45:58,190 And so in a way, the experiment was 959 00:45:58,190 --> 00:46:02,130 done to test a hypothesis that was 960 00:46:02,130 --> 00:46:04,510 proposed by Huber and Wiesel. 961 00:46:04,510 --> 00:46:07,174 And let me interject at this point, that one 962 00:46:07,174 --> 00:46:08,590 of the important things to realize 963 00:46:08,590 --> 00:46:12,470 is that you do not want to fall in love with your hypotheses, 964 00:46:12,470 --> 00:46:17,560 because in most cases the hypothesis that you dream up, 965 00:46:17,560 --> 00:46:20,520 so to speak, out of thin air, will end up 966 00:46:20,520 --> 00:46:23,050 being wrong when it comes to the brain. 967 00:46:23,050 --> 00:46:25,350 So in this case, this also happened. 968 00:46:25,350 --> 00:46:27,330 This is the model they came up with. 969 00:46:27,330 --> 00:46:31,650 They proposed that you had what we call an ice cube 970 00:46:31,650 --> 00:46:37,260 model, which in one direction specifies orientation columns, 971 00:46:37,260 --> 00:46:39,905 and in the other direction, ocular dominance columns. 972 00:46:41,060 --> 00:46:42,620 So this is the ice cube model. 973 00:46:42,620 --> 00:46:45,480 And now that you had this model, this hypothesis, 974 00:46:45,480 --> 00:46:46,710 you could test it. 975 00:46:46,710 --> 00:46:52,110 And the way this was tested was to use the same procedures 976 00:46:52,110 --> 00:46:55,560 I described to you, but do it in the same brain. 977 00:46:55,560 --> 00:46:56,190 All right? 978 00:46:56,190 --> 00:47:02,490 And so here's an example of here we have a part of the brain, 979 00:47:02,490 --> 00:47:06,140 again looking at it from the top down, which 980 00:47:06,140 --> 00:47:08,820 had been labeled for both orientation and ocular 981 00:47:08,820 --> 00:47:10,120 dominance columns. 982 00:47:10,120 --> 00:47:12,250 And then you take that same brain, 983 00:47:12,250 --> 00:47:15,335 and you label it for orientation columns. 984 00:47:17,540 --> 00:47:23,170 So now, if your hypothesis is correct, let me clarify this, 985 00:47:23,170 --> 00:47:25,850 if you say that they are right angles to each other, 986 00:47:25,850 --> 00:47:29,980 then you would expect that if you have a bunch of orientation 987 00:47:29,980 --> 00:47:34,870 columns, the ocular dominance columns 988 00:47:34,870 --> 00:47:36,290 should be at right angles to it. 989 00:47:36,290 --> 00:47:37,190 Right? 990 00:47:37,190 --> 00:47:39,910 And so if you label this, and then can draw it out, 991 00:47:39,910 --> 00:47:43,320 you can see whether or not that hypothesis is correct. 992 00:47:43,320 --> 00:47:48,050 And when you did that, here is a real example of what they did. 993 00:47:48,050 --> 00:47:51,470 One of these is orientation, and the thick ones 994 00:47:51,470 --> 00:47:54,730 and the thin ones is ocular dominant, maybe the other way 995 00:47:54,730 --> 00:47:55,440 around. 996 00:47:55,440 --> 00:47:56,210 It doesn't matter. 997 00:47:56,210 --> 00:47:58,800 But at any rate, if the hypothesis is correct, 998 00:47:58,800 --> 00:48:00,780 you'll expect something like what you see here. 999 00:48:00,780 --> 00:48:02,210 They're at right angles to each other. 1000 00:48:02,210 --> 00:48:02,790 OK? 1001 00:48:02,790 --> 00:48:05,010 But if you look at it carefully, you 1002 00:48:05,010 --> 00:48:07,230 can see the endless locations where they're not 1003 00:48:07,230 --> 00:48:09,050 at right angles to each other. 1004 00:48:09,050 --> 00:48:09,550 All right? 1005 00:48:10,660 --> 00:48:13,800 And because of that, even though that's a very attractive 1006 00:48:13,800 --> 00:48:18,170 hypothesis, people have come up with other hypotheses saying 1007 00:48:18,170 --> 00:48:20,880 that this hypothesis is really questionable. 1008 00:48:20,880 --> 00:48:24,650 And an alternative hypothesis that had been proposed 1009 00:48:24,650 --> 00:48:29,650 arose in part by yet another new discovery that had been made. 1010 00:48:29,650 --> 00:48:32,125 And this discovery was made by a woman 1011 00:48:32,125 --> 00:48:33,250 called Margaret Wong-Riley. 1012 00:48:34,590 --> 00:48:40,520 But this particular picture has been done by processing, 1013 00:48:40,520 --> 00:48:44,170 again this is a horizontal view looking down at the brain, 1014 00:48:44,170 --> 00:48:46,970 by Marge Livingstone, who is at Harvard, 1015 00:48:46,970 --> 00:48:50,330 and had done all the collaborative work with Hubel. 1016 00:48:50,330 --> 00:48:55,376 So what you see here, are the so-called cytochrome oxidase 1017 00:48:55,376 --> 00:48:55,875 patches. 1018 00:48:56,900 --> 00:48:58,150 OK? 1019 00:48:58,150 --> 00:49:00,970 Now, when this was first discovered, 1020 00:49:00,970 --> 00:49:02,414 people said, my god. 1021 00:49:02,414 --> 00:49:04,330 They had never seen anything like this before. 1022 00:49:04,330 --> 00:49:09,180 This is the only kind of stain that showed this up. 1023 00:49:09,180 --> 00:49:10,710 Let me explain to you. 1024 00:49:10,710 --> 00:49:13,060 This is also an activity label. 1025 00:49:13,060 --> 00:49:14,420 The cytochrome oxidase. 1026 00:49:14,420 --> 00:49:18,469 And we selectively label those cells that are most active, 1027 00:49:18,469 --> 00:49:20,260 rather than those cells which are inactive. 1028 00:49:22,720 --> 00:49:26,180 And so this is what these patches look like. 1029 00:49:26,180 --> 00:49:28,770 And so people kept saying, why do we have this patches? 1030 00:49:28,770 --> 00:49:31,570 We've never seen them before in the visual cortex. 1031 00:49:31,570 --> 00:49:32,750 What on earth are they for? 1032 00:49:32,750 --> 00:49:34,180 What do they tell us about? 1033 00:49:34,180 --> 00:49:36,590 So all kinds of hypotheses evolved. 1034 00:49:36,590 --> 00:49:40,100 And people asked the question, since we have this thing, what 1035 00:49:40,100 --> 00:49:42,860 if we recorded in one of these patches, 1036 00:49:42,860 --> 00:49:45,140 as opposed to outside of them. 1037 00:49:45,140 --> 00:49:48,890 And so when this was done by several people-- 1038 00:49:48,890 --> 00:49:52,780 let me just show you if you have a much larger section. 1039 00:49:52,780 --> 00:49:57,450 Here again is a fovea that's about 80 degrees out. 1040 00:49:57,450 --> 00:49:59,810 This is what these patches look like when 1041 00:49:59,810 --> 00:50:02,170 you do it on the high contrast. 1042 00:50:02,170 --> 00:50:02,710 OK? 1043 00:50:02,710 --> 00:50:04,460 So indeed, they're extremely frequent. 1044 00:50:04,460 --> 00:50:05,335 They're very orderly. 1045 00:50:06,250 --> 00:50:08,670 And so the question, of course, is whether these batches 1046 00:50:08,670 --> 00:50:12,370 are relative to the columns that we talked about. 1047 00:50:12,370 --> 00:50:14,290 And what was discovered, it if you 1048 00:50:14,290 --> 00:50:18,050 look at the ocular dominance columns, 1049 00:50:18,050 --> 00:50:20,780 these patches are always in the middle of a column. 1050 00:50:20,780 --> 00:50:25,560 So in other words, to make that clear, 1051 00:50:25,560 --> 00:50:32,280 if you have here a column like that, 1052 00:50:32,280 --> 00:50:34,991 the patches would be like this. 1053 00:50:34,991 --> 00:50:35,490 OK? 1054 00:50:36,540 --> 00:50:40,463 And the next one over, again the patches would be like that. 1055 00:50:40,463 --> 00:50:40,962 OK? 1056 00:50:42,570 --> 00:50:44,380 They would not be between patches. 1057 00:50:44,380 --> 00:50:47,820 So there's a direct relationship between the ocular dominance 1058 00:50:47,820 --> 00:50:52,110 columns, and the layout of these patches. 1059 00:50:52,110 --> 00:50:52,610 OK? 1060 00:50:53,830 --> 00:50:54,330 All right. 1061 00:50:54,330 --> 00:50:57,559 So now, as a result of this work, 1062 00:50:57,559 --> 00:50:59,100 people have come up with yet, and let 1063 00:50:59,100 --> 00:51:04,700 me finish, another hypothesis, which 1064 00:51:04,700 --> 00:51:06,750 is called the radio model. 1065 00:51:06,750 --> 00:51:09,230 They argued that the cells that are 1066 00:51:09,230 --> 00:51:14,610 in the center of these patches are largely unoriented, 1067 00:51:14,610 --> 00:51:18,800 and that the orientation selectivity goes around 1068 00:51:18,800 --> 00:51:26,000 in a radio fashion, as indicated here in the visual cortex 1069 00:51:26,000 --> 00:51:29,580 within each of the ocular dominance columns. 1070 00:51:29,580 --> 00:51:30,080 OK? 1071 00:51:32,004 --> 00:51:32,970 AUDIENCE: I'm good now. 1072 00:51:32,970 --> 00:51:33,690 PROFESSOR: Are you good now? 1073 00:51:33,690 --> 00:51:34,000 AUDIENCE: Thank you. 1074 00:51:34,000 --> 00:51:34,875 PROFESSOR: All right. 1075 00:51:34,875 --> 00:51:36,855 So that's the second model that has emerged. 1076 00:51:38,180 --> 00:51:41,350 And this model also was highly questioned, 1077 00:51:41,350 --> 00:51:43,530 because the evidence was not that you 1078 00:51:43,530 --> 00:51:50,490 had these nice clear radio orientations for orientation 1079 00:51:50,490 --> 00:51:50,990 selectivity. 1080 00:51:52,150 --> 00:51:55,470 So then finally another technique 1081 00:51:55,470 --> 00:51:58,260 had emerged, which was carried out again 1082 00:51:58,260 --> 00:52:02,420 at Harvard, at Harvard Medical School by a fellow called 1083 00:52:02,420 --> 00:52:02,920 Blaisdell. 1084 00:52:03,950 --> 00:52:07,330 He developed a technique that actually other people 1085 00:52:07,330 --> 00:52:08,740 have used, but then he applied it 1086 00:52:08,740 --> 00:52:13,380 to the visual cortex, that is called optical recording. 1087 00:52:14,460 --> 00:52:20,310 So you could record optically from the visual cortex 1088 00:52:20,310 --> 00:52:25,860 straight down, and then it could vary orientation, 1089 00:52:25,860 --> 00:52:27,170 and ocular dominance. 1090 00:52:27,170 --> 00:52:29,290 And when you did that, this is what 1091 00:52:29,290 --> 00:52:31,350 a typical example looks like. 1092 00:52:31,350 --> 00:52:32,080 OK? 1093 00:52:32,080 --> 00:52:37,030 Those red lines show what the layout is of the orientation 1094 00:52:37,030 --> 00:52:41,560 selectivity of the cells, and the patches here 1095 00:52:41,560 --> 00:52:43,150 are those regions where you don't 1096 00:52:43,150 --> 00:52:45,200 have orientation selectivity. 1097 00:52:45,200 --> 00:52:46,310 OK? 1098 00:52:46,310 --> 00:52:52,230 So this then indicated that you do not 1099 00:52:52,230 --> 00:52:55,760 have a computer like layout in the visual cortex going 1100 00:52:55,760 --> 00:52:58,200 one way, and the other way for orientation direction. 1101 00:52:59,360 --> 00:53:02,650 But you have it a bit messier. 1102 00:53:02,650 --> 00:53:06,680 And so this particular arrangement then was called, 1103 00:53:06,680 --> 00:53:09,250 or at least I called it, the swirl model. 1104 00:53:09,250 --> 00:53:09,750 OK? 1105 00:53:09,750 --> 00:53:13,130 But actually, to call it a model is incorrect. 1106 00:53:13,130 --> 00:53:14,270 This is a fact. 1107 00:53:14,270 --> 00:53:15,730 The others are models. 1108 00:53:15,730 --> 00:53:16,230 OK? 1109 00:53:16,230 --> 00:53:19,090 So now to put it all together, here is the Hubel and Wiesel 1110 00:53:19,090 --> 00:53:23,710 ice-cube model, where the orientation selectivity 1111 00:53:23,710 --> 00:53:26,050 [INAUDIBLE] right angles to each other. 1112 00:53:26,050 --> 00:53:27,710 Then we have the radio model. 1113 00:53:27,710 --> 00:53:29,710 And finally we have the swirl model, 1114 00:53:29,710 --> 00:53:32,340 which is the way it really is. 1115 00:53:32,340 --> 00:53:33,340 All right. 1116 00:53:33,340 --> 00:53:38,000 So that then in essence tells you 1117 00:53:38,000 --> 00:53:43,110 about the cytoarchitecture of the visual cortex. 1118 00:53:43,110 --> 00:53:45,470 Now we are going to take another big step forward. 1119 00:53:45,470 --> 00:53:48,850 We are going to look at extrastriate cortex. 1120 00:53:48,850 --> 00:53:51,860 As I've mentioned to you before, when 1121 00:53:51,860 --> 00:53:55,270 you look at the visual cortex, it 1122 00:53:55,270 --> 00:53:57,830 comes actually in many subdivisions, and maybe 1123 00:53:57,830 --> 00:54:02,832 as many as 30 visual cortical areas, 1124 00:54:02,832 --> 00:54:04,790 of which we will only talk about a few of them, 1125 00:54:04,790 --> 00:54:06,206 because it's totally overwhelming. 1126 00:54:08,030 --> 00:54:16,780 And these areas were presumed to involve, maybe correctly so, 1127 00:54:16,780 --> 00:54:21,680 initially at least, increasingly complex visual analyses. 1128 00:54:21,680 --> 00:54:22,180 All right? 1129 00:54:24,180 --> 00:54:28,230 Now, one of the prime ideas that, 1130 00:54:28,230 --> 00:54:31,390 or hypotheses again if you will, to my mind 1131 00:54:31,390 --> 00:54:36,200 turned out to be incorrect, is that each of these areas 1132 00:54:36,200 --> 00:54:40,620 in the brain specializes in analyzing 1133 00:54:40,620 --> 00:54:42,305 a certain aspect of vision. 1134 00:54:43,370 --> 00:54:45,065 So what are certain aspects of vision? 1135 00:54:46,440 --> 00:54:49,090 Just a few examples are analyzing color, 1136 00:54:49,090 --> 00:54:54,460 analyzing shape, analyzing motion, and so on. 1137 00:54:54,460 --> 00:54:57,400 And so they argue that all these visual areas 1138 00:54:57,400 --> 00:55:00,010 specialize in one of those. 1139 00:55:00,010 --> 00:55:03,690 So that was an interesting idea, and so people 1140 00:55:03,690 --> 00:55:07,010 began to study extrastriate cortex in more detail. 1141 00:55:07,010 --> 00:55:08,810 And I'm going to say few words about it, 1142 00:55:08,810 --> 00:55:11,790 but let me go back even further in history. 1143 00:55:11,790 --> 00:55:19,520 Let's go back to a time in the early 1800s. 1144 00:55:19,520 --> 00:55:22,250 There were two very famous people there, 1145 00:55:22,250 --> 00:55:24,640 which were called Gall and Spurzheim. 1146 00:55:25,720 --> 00:55:29,180 And they came up with the idea which 1147 00:55:29,180 --> 00:55:31,350 eventually was called phrenology. 1148 00:55:31,350 --> 00:55:32,790 How many of know phrenology? 1149 00:55:33,840 --> 00:55:34,470 Oh, well. 1150 00:55:34,470 --> 00:55:35,240 My goodness. 1151 00:55:35,240 --> 00:55:36,210 All of you do. 1152 00:55:36,210 --> 00:55:36,960 OK. 1153 00:55:36,960 --> 00:55:42,920 So phrenology, in essence, claimed 1154 00:55:42,920 --> 00:55:48,100 that there are specific areas in the brain, as shown here, 1155 00:55:48,100 --> 00:55:51,825 that specialize in certain aspects of the information 1156 00:55:51,825 --> 00:55:52,325 processing. 1157 00:55:53,750 --> 00:55:55,760 And how did they come up with this? 1158 00:55:55,760 --> 00:56:01,380 The way they came up with it was to palpate the skull. 1159 00:56:01,380 --> 00:56:03,560 And wherever there were big bulges, 1160 00:56:03,560 --> 00:56:05,270 they felt that there was a lot of that 1161 00:56:05,270 --> 00:56:07,980 attribute that a person had. 1162 00:56:07,980 --> 00:56:09,740 And if it was small, then the person 1163 00:56:09,740 --> 00:56:11,260 didn't have too much of it. 1164 00:56:11,260 --> 00:56:13,800 And so they played with it, and played with it, 1165 00:56:13,800 --> 00:56:18,130 and they came up in an 1812 publication 1166 00:56:18,130 --> 00:56:25,800 with essentially 35 basic visual attributes of processing 1167 00:56:25,800 --> 00:56:32,490 in the brain, for the cortex of the brain in humans. 1168 00:56:32,490 --> 00:56:34,080 And the interesting thing about this 1169 00:56:34,080 --> 00:56:35,639 is it gives you a sense of history 1170 00:56:35,639 --> 00:56:39,100 and how much we have changed in our lives, 1171 00:56:39,100 --> 00:56:45,220 as to what these specialized areas were conceived to be. 1172 00:56:45,220 --> 00:56:48,490 Now, to make that a little bit clearer for you, 1173 00:56:48,490 --> 00:56:49,880 let me stop here for a second. 1174 00:56:51,950 --> 00:56:55,000 And what I want to do here is I want to enlarge this, 1175 00:56:55,000 --> 00:56:56,270 so you can see it better. 1176 00:57:00,380 --> 00:57:00,880 OK. 1177 00:57:00,880 --> 00:57:02,750 So here are some of these areas. 1178 00:57:02,750 --> 00:57:05,440 And if you look at these, you will be almost taken aback 1179 00:57:05,440 --> 00:57:08,270 as to what the hell the names are here. 1180 00:57:08,270 --> 00:57:12,670 I mean, some of the common names here that you can see would be, 1181 00:57:12,670 --> 00:57:17,660 for example, let's see if I can put them in the right order 1182 00:57:17,660 --> 00:57:18,160 here. 1183 00:57:19,710 --> 00:57:23,430 Actually, the thing was that they called these not only 1184 00:57:23,430 --> 00:57:25,910 propensities, but they called them sentiments. 1185 00:57:27,050 --> 00:57:29,080 So in those days sentiments were very important. 1186 00:57:30,240 --> 00:57:33,200 There's not much brain allocated to sentiments anymore. 1187 00:57:33,200 --> 00:57:35,690 But at any rate, such things that 1188 00:57:35,690 --> 00:57:41,360 existed as amativeness, cautiousness, benevolence, 1189 00:57:41,360 --> 00:57:43,070 veneration, wonder and ideality. 1190 00:57:44,900 --> 00:57:47,250 Those are just a few examples here that you can spot. 1191 00:57:47,250 --> 00:57:49,430 Amativeness right there. 1192 00:57:49,430 --> 00:57:51,200 And indeed you can ask the question, 1193 00:57:51,200 --> 00:57:55,380 really would that much brain be devoted to amativeness? 1194 00:57:55,380 --> 00:57:57,570 And certainly that's not the case anymore. 1195 00:57:57,570 --> 00:58:00,480 So maybe our brains are totally different from the way 1196 00:58:00,480 --> 00:58:03,660 they were back in 1812. 1197 00:58:03,660 --> 00:58:04,670 I doubt it. 1198 00:58:04,670 --> 00:58:09,630 And so therefore, I think these things are rather fanciful, 1199 00:58:09,630 --> 00:58:11,705 and very far-fetched hypotheses. 1200 00:58:12,830 --> 00:58:16,530 And as I mentioned before, that's par for the course. 1201 00:58:16,530 --> 00:58:19,280 So many hypotheses end up being wrong, 1202 00:58:19,280 --> 00:58:23,010 and the best way to overcome that is 1203 00:58:23,010 --> 00:58:28,077 to do solid experiments to find out what is really going on. 1204 00:58:49,830 --> 00:58:50,330 OK. 1205 00:58:50,330 --> 00:58:52,450 So now, we move on. 1206 00:58:52,450 --> 00:59:00,960 And we need to understand what the more modern techniques are 1207 00:59:00,960 --> 00:59:05,160 that enable us to specify things about these higher 1208 00:59:05,160 --> 00:59:06,850 cortical areas. 1209 00:59:06,850 --> 00:59:09,600 The first one of these is architectonics. 1210 00:59:09,600 --> 00:59:10,900 That's obvious what it is. 1211 00:59:10,900 --> 00:59:13,390 That's simply to look at the brain, 1212 00:59:13,390 --> 00:59:19,020 and identify the various areas in a systematic fashion. 1213 00:59:19,020 --> 00:59:21,100 The second one is to look at connections. 1214 00:59:21,100 --> 00:59:23,580 You do anatomical studies to determine 1215 00:59:23,580 --> 00:59:27,240 which areas to connect to what areas in what fashion. 1216 00:59:27,240 --> 00:59:29,530 Another one is topographic mapping. 1217 00:59:29,530 --> 00:59:31,450 We talked about that already. 1218 00:59:31,450 --> 00:59:34,260 One of those was to actually do single cell recording 1219 00:59:34,260 --> 00:59:36,760 systematically moving the electrode across, 1220 00:59:36,760 --> 00:59:39,490 or to do the two digit type studies. 1221 00:59:39,490 --> 00:59:41,900 Another one we can do is what we call 1222 00:59:41,900 --> 00:59:43,940 physiological characterization. 1223 00:59:43,940 --> 00:59:45,920 That's also known to you. 1224 00:59:45,920 --> 00:59:49,564 When you looked at the cells in V1, 1225 00:59:49,564 --> 00:59:51,230 you established that they're orientation 1226 00:59:51,230 --> 00:59:52,940 direction selective. 1227 00:59:52,940 --> 00:59:55,170 That's your physiological characterization 1228 00:59:55,170 --> 00:59:56,534 of the cells in V1. 1229 00:59:56,534 --> 00:59:58,200 And then you can ask the question, well, 1230 00:59:58,200 --> 01:00:00,950 what about V2, V4, and so on, these other areas. 1231 01:00:02,300 --> 01:00:03,550 What are the cells like there? 1232 01:00:03,550 --> 01:00:05,290 What do they respond to well? 1233 01:00:05,290 --> 01:00:06,680 And so on. 1234 01:00:06,680 --> 01:00:09,040 And then, another very important technique 1235 01:00:09,040 --> 01:00:14,510 is to say, OK, what if we removed a particular area that 1236 01:00:14,510 --> 01:00:16,100 had been identified. 1237 01:00:16,100 --> 01:00:19,820 What kind of loss do you have in vision, 1238 01:00:19,820 --> 01:00:21,640 or in general in the brain? 1239 01:00:21,640 --> 01:00:25,430 If you move a particular area, what kind of deficit arises? 1240 01:00:25,430 --> 01:00:27,480 And if that's a specific deficit, 1241 01:00:27,480 --> 01:00:30,090 you can infer that that particular area plays 1242 01:00:30,090 --> 01:00:32,400 an important role in the analysis 1243 01:00:32,400 --> 01:00:33,530 that you can no longer do. 1244 01:00:34,780 --> 01:00:40,900 That can also be done instead of making specific lesions. 1245 01:00:40,900 --> 01:00:43,040 It's less accurate, but you can do it 1246 01:00:43,040 --> 01:00:46,510 by studying humans who have had various kinds 1247 01:00:46,510 --> 01:00:48,150 of febrile accidents. 1248 01:00:48,150 --> 01:00:51,990 And that's one of the things that our former chairman, 1249 01:00:51,990 --> 01:00:54,570 Hans-Lukas Teuber, has done extensively. 1250 01:00:54,570 --> 01:00:56,860 Studying people after the Second World 1251 01:00:56,860 --> 01:01:00,090 War who have sustained specific brain injuries to see 1252 01:01:00,090 --> 01:01:03,640 what kinds of deficits they had suffered on the basis of that 1253 01:01:03,640 --> 01:01:07,250 you could infer what various brain areas do. 1254 01:01:07,250 --> 01:01:09,350 And the last one here that I want to mention 1255 01:01:09,350 --> 01:01:10,430 is imaging, of course. 1256 01:01:10,430 --> 01:01:11,810 We've talked about that. 1257 01:01:11,810 --> 01:01:17,410 You can present certain specific stimuli to activate the brain. 1258 01:01:17,410 --> 01:01:21,010 And then once you process that using magnetic resonance 1259 01:01:21,010 --> 01:01:24,770 imaging, for example, functional magnetic resonance imaging can 1260 01:01:24,770 --> 01:01:28,400 tell you how important, for example, 1261 01:01:28,400 --> 01:01:31,510 areas are in recognising faces. 1262 01:01:31,510 --> 01:01:35,790 And some of that work is being done, actually here, 1263 01:01:35,790 --> 01:01:38,820 by Nancy Kanwisher, for example, in our department. 1264 01:01:38,820 --> 01:01:39,650 OK. 1265 01:01:39,650 --> 01:01:42,540 So to do this then systematically, 1266 01:01:42,540 --> 01:01:48,560 you have to have a sort of idea of what kinds of functions 1267 01:01:48,560 --> 01:01:50,090 do we want to study. 1268 01:01:50,090 --> 01:01:53,460 And so you want to break it down visual functions. 1269 01:01:53,460 --> 01:01:55,640 And one way to break them down, I'm not 1270 01:01:55,640 --> 01:02:00,390 saying this is overall satisfactory to everyone, 1271 01:02:00,390 --> 01:02:03,700 they can talk about so-called basic visual capacities, 1272 01:02:03,700 --> 01:02:05,180 and more higher level ones. 1273 01:02:05,180 --> 01:02:07,900 When you talk about basic visual capacities, 1274 01:02:07,900 --> 01:02:09,660 they say well how well can you see color. 1275 01:02:10,720 --> 01:02:13,780 How well can you distinguish differences in brightness? 1276 01:02:13,780 --> 01:02:17,610 How good are you at seeing basic patterns? 1277 01:02:17,610 --> 01:02:18,610 Textures? 1278 01:02:18,610 --> 01:02:19,630 Motion? 1279 01:02:19,630 --> 01:02:20,560 Depth? 1280 01:02:20,560 --> 01:02:21,170 OK? 1281 01:02:21,170 --> 01:02:23,760 So those would be your basic visual functions. 1282 01:02:23,760 --> 01:02:26,570 And then when you come to intermediate visual capacities, 1283 01:02:26,570 --> 01:02:29,210 things become much more complex, of course. 1284 01:02:29,210 --> 01:02:31,250 You come up with constancy. 1285 01:02:31,250 --> 01:02:33,790 How come that when I look at something that's nearby, 1286 01:02:33,790 --> 01:02:35,330 and something that's further way, 1287 01:02:35,330 --> 01:02:37,570 I can recognize it's the same thing? 1288 01:02:37,570 --> 01:02:38,200 OK? 1289 01:02:38,200 --> 01:02:43,220 Or how can we select things in the visual scene? 1290 01:02:43,220 --> 01:02:46,430 How can we recognize things, just like recognizing faces? 1291 01:02:48,542 --> 01:02:49,833 How can we make transpositions? 1292 01:02:51,330 --> 01:02:53,046 How can we make comparisons? 1293 01:02:54,160 --> 01:02:57,480 And how can we locate things in space? 1294 01:02:57,480 --> 01:03:00,170 So those would be some of those so-called intermediate visual 1295 01:03:00,170 --> 01:03:01,420 capacities. 1296 01:03:01,420 --> 01:03:03,930 And I'm not even going to mention high level 1297 01:03:03,930 --> 01:03:07,240 visual capacities, because they are even more complicated, 1298 01:03:07,240 --> 01:03:09,380 and we know even less about them. 1299 01:03:09,380 --> 01:03:09,880 OK. 1300 01:03:09,880 --> 01:03:14,980 So now, let's go and lay out the visual areas. 1301 01:03:14,980 --> 01:03:16,830 Just the very basics of it, because it's 1302 01:03:16,830 --> 01:03:18,740 unbelievably complicated. 1303 01:03:18,740 --> 01:03:20,480 Here we have a human brain. 1304 01:03:20,480 --> 01:03:23,270 And back when this was put together, 1305 01:03:23,270 --> 01:03:26,960 this posterior area here, which is the primary visual cortex, 1306 01:03:26,960 --> 01:03:30,980 called area 17 then, we now call that V1. 1307 01:03:30,980 --> 01:03:33,140 And 18 would be V2. 1308 01:03:33,140 --> 01:03:35,461 And 19, V3, and so on. 1309 01:03:35,461 --> 01:03:35,960 OK. 1310 01:03:35,960 --> 01:03:39,390 So now again, we go back to the monkey brain. 1311 01:03:39,390 --> 01:03:43,620 And when we look at the monkey brain, what you see here 1312 01:03:43,620 --> 01:03:45,056 is again the central sulcus. 1313 01:03:45,056 --> 01:03:46,930 You're going to see this over and over again, 1314 01:03:46,930 --> 01:03:49,420 and it's going to stick into your head. 1315 01:03:49,420 --> 01:03:51,430 Here's the lunate sulcus. 1316 01:03:51,430 --> 01:03:54,030 And here is area V1. 1317 01:03:54,030 --> 01:03:58,570 This is where we had examined the properties of these cells 1318 01:03:58,570 --> 01:03:59,990 that we had just talked about. 1319 01:03:59,990 --> 01:04:04,000 Then if you move on, right at the edge here of the lunate, 1320 01:04:04,000 --> 01:04:08,650 V2 starts, and then goes from under the gyrus. 1321 01:04:09,970 --> 01:04:14,290 And then in this region here, we have what is called area V4. 1322 01:04:15,570 --> 01:04:17,339 So those are some of the basic areas. 1323 01:04:17,339 --> 01:04:19,380 And you are going to see a few things about them. 1324 01:04:19,380 --> 01:04:21,830 Now, people have studied this extensively, 1325 01:04:21,830 --> 01:04:24,720 and they came up with frightening diagrams of this. 1326 01:04:24,720 --> 01:04:27,330 This is a flattened monkey brain. 1327 01:04:27,330 --> 01:04:28,540 This is V1. 1328 01:04:28,540 --> 01:04:31,180 And then you have a whole bunch of areas 1329 01:04:31,180 --> 01:04:33,820 that are following succession here. 1330 01:04:33,820 --> 01:04:36,040 I'm not going to label many of these. 1331 01:04:36,040 --> 01:04:38,135 And we talked about V2, V4. 1332 01:04:39,560 --> 01:04:41,690 But there are many, many more. 1333 01:04:41,690 --> 01:04:43,650 I'll bring up a few more in a minute. 1334 01:04:43,650 --> 01:04:46,280 And then, if you look at the interconnections that 1335 01:04:46,280 --> 01:04:49,740 had been made, it's totally frightening. 1336 01:04:49,740 --> 01:04:51,990 There's so many hundreds, and hundreds, and hundreds 1337 01:04:51,990 --> 01:04:55,370 of connections going every which way. 1338 01:04:55,370 --> 01:05:00,527 So it is very difficult to say that particular area receives 1339 01:05:00,527 --> 01:05:02,860 inputs only from one other area, or something like that. 1340 01:05:02,860 --> 01:05:06,940 There's just a tremendous amount of interconnections, 1341 01:05:06,940 --> 01:05:10,820 indicating that any analysis is likely to take place involving 1342 01:05:10,820 --> 01:05:14,270 thousands, and tens of thousands of thousands of neurons being 1343 01:05:14,270 --> 01:05:17,080 active in many, many different brain areas. 1344 01:05:17,080 --> 01:05:17,640 OK. 1345 01:05:17,640 --> 01:05:21,850 Now the major cortical visual areas that we shall consider 1346 01:05:21,850 --> 01:05:27,230 are V1, that we already did, V2, V3, V4 and MT. 1347 01:05:27,230 --> 01:05:30,440 Then in the temporal region of the brain, 1348 01:05:30,440 --> 01:05:32,540 we have inferotemporal cortex. 1349 01:05:32,540 --> 01:05:34,380 And then in the private region, we 1350 01:05:34,380 --> 01:05:39,780 have the lateral intraparietal area, the ventral area, 1351 01:05:39,780 --> 01:05:44,240 and MST, which is called the Medial Superior Temporal area. 1352 01:05:44,240 --> 01:05:46,350 And then lastly, in the frontal lobe, 1353 01:05:46,350 --> 01:05:47,925 we have the frontal eye fields. 1354 01:05:47,925 --> 01:05:49,590 Now, we will talk about each of these 1355 01:05:49,590 --> 01:05:52,210 at various levels in the course. 1356 01:05:52,210 --> 01:05:55,465 Today, we will just briefly talk about V2, V4 and MT. 1357 01:05:56,620 --> 01:05:57,120 OK. 1358 01:05:57,120 --> 01:06:02,370 So now, a couple of general principles 1359 01:06:02,370 --> 01:06:05,050 that emerge from this kind of work. 1360 01:06:05,050 --> 01:06:09,150 First of all is that the size of the receptive 1361 01:06:09,150 --> 01:06:13,770 fields in these different areas changes dramatically. 1362 01:06:13,770 --> 01:06:16,580 In the visual cortex there's aptitudes that are very small. 1363 01:06:17,654 --> 01:06:19,070 I mean, bigger than in the retina, 1364 01:06:19,070 --> 01:06:21,890 or in the geniculate, but still very, very small. 1365 01:06:21,890 --> 01:06:24,190 But then when you get to V2, they're 1366 01:06:24,190 --> 01:06:27,550 about three times bigger on the average in diameter. 1367 01:06:27,550 --> 01:06:29,730 And when you come to V4, they're huge. 1368 01:06:29,730 --> 01:06:30,230 OK? 1369 01:06:30,230 --> 01:06:34,540 And that's true throughout, that the specificity 1370 01:06:34,540 --> 01:06:40,620 of the location of the receptive fields 1371 01:06:40,620 --> 01:06:46,770 decreases as you progress up into higher cortical areas. 1372 01:06:46,770 --> 01:06:48,650 Now, there's another very interesting fact 1373 01:06:48,650 --> 01:06:52,956 that had been discovered, which again nobody hypothesized. 1374 01:06:54,080 --> 01:06:59,280 Namely, that the way these areas are laid out next to each other 1375 01:06:59,280 --> 01:07:03,060 is not what you would have thought. 1376 01:07:03,060 --> 01:07:04,570 So let me give you an example. 1377 01:07:04,570 --> 01:07:09,350 Suppose what we do is we take an electrode, and record in V1, 1378 01:07:09,350 --> 01:07:14,040 and then we go in progressive steps across V1, 1379 01:07:14,040 --> 01:07:19,060 where the receptive fields move out from the fovea 1380 01:07:19,060 --> 01:07:20,175 into the periphery. 1381 01:07:21,420 --> 01:07:23,450 And if you do that, here's a receptive field 1382 01:07:23,450 --> 01:07:25,470 of that cell. [INAUDIBLE], right there. 1383 01:07:25,470 --> 01:07:29,190 The next one is a little bit further removed. 1384 01:07:29,190 --> 01:07:31,510 The next one is a little bit further removed. 1385 01:07:31,510 --> 01:07:35,570 And here's the next one, and that one is there. 1386 01:07:35,570 --> 01:07:36,270 OK? 1387 01:07:36,270 --> 01:07:38,370 So what you have is a progression 1388 01:07:38,370 --> 01:07:39,770 of receptive fields. 1389 01:07:39,770 --> 01:07:43,330 They get bigger, and they move from the center out. 1390 01:07:43,330 --> 01:07:47,000 So now the big question is what happens 1391 01:07:47,000 --> 01:07:48,845 when you now come in to V2. 1392 01:07:51,540 --> 01:07:52,890 And think about it for a minute. 1393 01:07:52,890 --> 01:07:53,727 What do you think? 1394 01:07:53,727 --> 01:07:55,310 Where do you think the receptive field 1395 01:07:55,310 --> 01:07:59,509 will be when you just ride across from V1 to V2 in the map 1396 01:07:59,509 --> 01:08:00,675 that I had shown you before? 1397 01:08:02,110 --> 01:08:04,540 Well, you will be surprised. 1398 01:08:04,540 --> 01:08:06,170 I think you will be surprised. 1399 01:08:06,170 --> 01:08:09,690 You will find that the next receptive field, close as this, 1400 01:08:09,690 --> 01:08:12,965 is at the same location, almost, very close the same location. 1401 01:08:15,010 --> 01:08:16,750 And of course, the receptive field 1402 01:08:16,750 --> 01:08:18,680 is about three times bigger. 1403 01:08:18,680 --> 01:08:21,790 Now, what happens when you now move one over? 1404 01:08:21,790 --> 01:08:24,819 Well, the receptive field comes back towards the center. 1405 01:08:24,819 --> 01:08:28,140 And again, and again like that. 1406 01:08:28,140 --> 01:08:35,100 So these areas connect to each other in reverse, so to speak. 1407 01:08:35,100 --> 01:08:38,200 In other words, if they didn't one would connect to here. 1408 01:08:38,200 --> 01:08:40,510 Two would connect to there, and so on. 1409 01:08:40,510 --> 01:08:43,729 But instead, four and five could have short connections. 1410 01:08:43,729 --> 01:08:46,529 Three and six, longer, and so on. 1411 01:08:46,529 --> 01:08:49,330 So we have this very curious arrangement. 1412 01:08:49,330 --> 01:08:51,340 And to this day I don't have a clue 1413 01:08:51,340 --> 01:08:54,600 as to why in the course of evolution this happened. 1414 01:08:54,600 --> 01:08:56,930 I mean, there all kinds of hypotheses. 1415 01:08:56,930 --> 01:09:00,000 This may take less elaborate wiring. 1416 01:09:01,260 --> 01:09:05,825 Not as long wiring overall, or something, 1417 01:09:05,825 --> 01:09:07,200 because if it were the other way, 1418 01:09:07,200 --> 01:09:09,540 then all these wires would be long. 1419 01:09:09,540 --> 01:09:12,970 So that could be a reason, but one is not sure. 1420 01:09:12,970 --> 01:09:15,260 And to my knowledge, no experiment 1421 01:09:15,260 --> 01:09:19,819 has yet been done to truly explain 1422 01:09:19,819 --> 01:09:22,319 why we have this curious arrangement. 1423 01:09:22,319 --> 01:09:25,060 Now, when you look at area V2, we 1424 01:09:25,060 --> 01:09:27,279 have yet another very interesting factor. 1425 01:09:27,279 --> 01:09:31,700 When you come back to area V1, this would use V1. 1426 01:09:31,700 --> 01:09:34,209 We have those famous cytochrome oxidase patches. 1427 01:09:35,359 --> 01:09:39,640 But then when you get to area V2, instead of what you have, 1428 01:09:39,640 --> 01:09:44,840 you have these elongated bars. 1429 01:09:44,840 --> 01:09:45,470 OK? 1430 01:09:45,470 --> 01:09:47,000 And if you look closely, you can see 1431 01:09:47,000 --> 01:09:49,060 that there's a tendency for the bars 1432 01:09:49,060 --> 01:09:54,320 to go from thick to thin with an inter-bar area. 1433 01:09:54,320 --> 01:09:55,980 And so when this was discovered, people 1434 01:09:55,980 --> 01:09:59,820 began to record to say why is this different from here. 1435 01:09:59,820 --> 01:10:01,390 What does it signify? 1436 01:10:01,390 --> 01:10:03,300 And so when they did that systematically, 1437 01:10:03,300 --> 01:10:06,840 they came up with a model, which may not be entirely correct. 1438 01:10:06,840 --> 01:10:10,550 But the claim was that the thin stripes in V2 1439 01:10:10,550 --> 01:10:13,800 get inputs predominantly from there 1440 01:10:13,800 --> 01:10:16,330 patches that you have in V1. 1441 01:10:16,330 --> 01:10:24,130 And the thick stripes get input from the so-called parasol 1442 01:10:24,130 --> 01:10:28,640 system, meaning that they would be heavily involved 1443 01:10:28,640 --> 01:10:31,180 in things like what the parasol cells do, 1444 01:10:31,180 --> 01:10:34,490 namely play an important role in emotion perception. 1445 01:10:34,490 --> 01:10:36,225 And the interstripes here called, 1446 01:10:36,225 --> 01:10:40,360 or the pale stripes sometimes called, they get input 1447 01:10:40,360 --> 01:10:44,770 in this particular model from the orientation specific cells. 1448 01:10:44,770 --> 01:10:45,450 OK? 1449 01:10:45,450 --> 01:10:49,430 So now, once this was done and proposed, 1450 01:10:49,430 --> 01:10:51,780 people begin to do experiments to record 1451 01:10:51,780 --> 01:10:54,610 in the thick and thin stripes of V2 1452 01:10:54,610 --> 01:10:57,150 to see what the properties are of the cells 1453 01:10:57,150 --> 01:11:02,010 there, recognizing of course that the size of the receptive 1454 01:11:02,010 --> 01:11:06,190 fields is uniformly about three times bigger than in V1. 1455 01:11:06,190 --> 01:11:09,210 So when they did that, they looked at orientation. 1456 01:11:09,210 --> 01:11:13,860 And they found that in V2, most of the cells 1457 01:11:13,860 --> 01:11:17,280 in the thick stripes are orientation specific, but also 1458 01:11:17,280 --> 01:11:20,790 many in the thin stripes, and many also in the pale stripes. 1459 01:11:20,790 --> 01:11:24,130 So there didn't seem to be a huge distinction. 1460 01:11:24,130 --> 01:11:25,699 Then when it came to end stopping, 1461 01:11:25,699 --> 01:11:26,740 I didn't talk about that. 1462 01:11:26,740 --> 01:11:28,500 Let me mention what end stopping means. 1463 01:11:29,510 --> 01:11:32,190 If you take a bar of light, and you move it 1464 01:11:32,190 --> 01:11:34,650 across a receptive field, you get a vigorous response 1465 01:11:34,650 --> 01:11:36,500 when it's a fairly short bar. 1466 01:11:36,500 --> 01:11:38,760 But then when you make the bar a lot longer, 1467 01:11:38,760 --> 01:11:40,670 you get less of a response, because 1468 01:11:40,670 --> 01:11:42,420 of some surrounding [INAUDIBLE]. 1469 01:11:42,420 --> 01:11:45,640 So that's what's called, by Hubel and Wiesel, end stopping. 1470 01:11:45,640 --> 01:11:49,390 And so that attribute is one that 1471 01:11:49,390 --> 01:11:52,580 is again fairly similar in the three areas. 1472 01:11:52,580 --> 01:11:54,840 When you come to color, in the thin stripes 1473 01:11:54,840 --> 01:11:56,640 there is much more color sensitivity 1474 01:11:56,640 --> 01:11:57,970 than the fixed stripes. 1475 01:11:57,970 --> 01:12:00,740 When it comes to direction, again not that much difference. 1476 01:12:00,740 --> 01:12:02,660 Maybe more in the thick stripes. 1477 01:12:02,660 --> 01:12:05,380 Disparity, more in thickness. 1478 01:12:05,380 --> 01:12:06,970 Disparity refers to depth perception, 1479 01:12:06,970 --> 01:12:08,200 as we have talked about. 1480 01:12:08,200 --> 01:12:14,070 So the prime message here is that you do not 1481 01:12:14,070 --> 01:12:19,950 have a complete clear separation of function in those three 1482 01:12:19,950 --> 01:12:26,450 stripes that had been identified in V2, 1483 01:12:26,450 --> 01:12:27,610 virtually inputs from V1. 1484 01:12:29,250 --> 01:12:34,240 So now, we are going to move on to V4. 1485 01:12:34,240 --> 01:12:36,070 And when you talk about area of V4, 1486 01:12:36,070 --> 01:12:40,520 we'll talk about that again in more detail later on. 1487 01:12:40,520 --> 01:12:43,440 We have a huge increase of complexity 1488 01:12:43,440 --> 01:12:46,530 of the response properties of cells. 1489 01:12:46,530 --> 01:12:49,960 And here is area V4, as you can see. 1490 01:12:49,960 --> 01:12:52,070 A lot of recording has been done in this area. 1491 01:12:52,070 --> 01:12:54,190 Hundreds of papers have been published. 1492 01:12:54,190 --> 01:12:56,920 And I can conclude on the basis of that. 1493 01:12:56,920 --> 01:12:58,810 First of all, here the receptive fields 1494 01:12:58,810 --> 01:13:02,410 are even a lot bigger than in V2. 1495 01:13:02,410 --> 01:13:04,270 And the receptive field properties 1496 01:13:04,270 --> 01:13:07,665 are far more complex than in either V1 or V2. 1497 01:13:09,020 --> 01:13:12,900 And the response properties are dynamic. 1498 01:13:12,900 --> 01:13:15,060 If the monkey is paying attention, 1499 01:13:15,060 --> 01:13:17,580 or is looking for something, the cells 1500 01:13:17,580 --> 01:13:21,160 will find a lot more than when he's not looking for something. 1501 01:13:22,610 --> 01:13:25,490 And yet another discovery was made 1502 01:13:25,490 --> 01:13:27,910 that the amount of activity is also 1503 01:13:27,910 --> 01:13:30,870 modulated by how you move your eyes as you're looking around 1504 01:13:30,870 --> 01:13:31,750 in the visual scene. 1505 01:13:33,010 --> 01:13:35,000 And then, an important conclusion from that 1506 01:13:35,000 --> 01:13:37,780 is that this is not just a color area. 1507 01:13:37,780 --> 01:13:41,270 And the reason I mention that is because a good many years 1508 01:13:41,270 --> 01:13:45,420 ago an experiment was done in England. 1509 01:13:45,420 --> 01:13:48,520 I shan't name the person who was doing it, 1510 01:13:48,520 --> 01:13:52,310 who claimed that this is a color area. 1511 01:13:52,310 --> 01:13:55,280 Now, one of the reasons he made that claim 1512 01:13:55,280 --> 01:13:59,870 was that he removed area V4, or studied 1513 01:13:59,870 --> 01:14:02,770 humans who lacked area V4. 1514 01:14:02,770 --> 01:14:06,155 And he found that they had difficulties in telling colors. 1515 01:14:08,290 --> 01:14:09,660 Well, that was nice. 1516 01:14:09,660 --> 01:14:13,010 But the problem was, and this is a good lesson for all of us, 1517 01:14:13,010 --> 01:14:16,050 that the only thing he tested was 1518 01:14:16,050 --> 01:14:18,675 for color, because that was his hypothesis. 1519 01:14:19,770 --> 01:14:23,580 Had he tested for a multitude of other functions, 1520 01:14:23,580 --> 01:14:26,920 he would have found that area V4 has 1521 01:14:26,920 --> 01:14:28,830 much more significant deficits when 1522 01:14:28,830 --> 01:14:33,240 it's removed for many other visual functions 1523 01:14:33,240 --> 01:14:34,900 than for color. 1524 01:14:37,020 --> 01:14:41,320 So again, we'll talk about that input with more detail 1525 01:14:41,320 --> 01:14:44,320 later on when we talk about very specific processing 1526 01:14:44,320 --> 01:14:45,430 of digital attributes. 1527 01:14:46,600 --> 01:14:48,900 Now then the next areas I want to deal with, 1528 01:14:48,900 --> 01:14:51,220 the last areas we are going to discuss, 1529 01:14:51,220 --> 01:14:53,580 are going to be areas MT and MST. 1530 01:14:53,580 --> 01:14:57,040 MT stands for Middle Temporal area. 1531 01:14:57,040 --> 01:15:01,860 And MST stands for Middle Superior Temporal area. 1532 01:15:01,860 --> 01:15:04,980 And again, to look at the brain, what you can see here 1533 01:15:04,980 --> 01:15:08,040 is a superior temporal sulcus. 1534 01:15:08,040 --> 01:15:13,080 And if you went into it, it's about 13 millimeters deep. 1535 01:15:13,080 --> 01:15:16,090 On the posterior side, we have area MT. 1536 01:15:16,090 --> 01:15:18,770 And on the interior part of it we have area MST. 1537 01:15:19,870 --> 01:15:21,490 Now, this is a remarkable area. 1538 01:15:21,490 --> 01:15:24,010 An incredible amount of work had been done on it. 1539 01:15:24,010 --> 01:15:26,600 And the major discovery that had been made 1540 01:15:26,600 --> 01:15:32,200 is that the cells in MT and MST respond predominantly 1541 01:15:32,200 --> 01:15:33,355 to direction. 1542 01:15:34,360 --> 01:15:36,500 Virtually, all the cells are direction specific. 1543 01:15:37,600 --> 01:15:42,240 And also, this is in part believed to be due to the fact 1544 01:15:42,240 --> 01:15:47,510 that this area gets half their input from the parasol system. 1545 01:15:47,510 --> 01:15:49,660 So to look at this in more detail, 1546 01:15:49,660 --> 01:15:52,990 here is a receptive field, quite large and empty. 1547 01:15:52,990 --> 01:15:55,490 And if you move a bar of light across in this direction, 1548 01:15:55,490 --> 01:15:57,100 you get a huge response. 1549 01:15:57,100 --> 01:15:59,140 This is the cumulative response histogram. 1550 01:15:59,140 --> 01:16:01,020 But if you move it in the opposite direction, 1551 01:16:01,020 --> 01:16:03,490 you won't actually get an inhibition. 1552 01:16:03,490 --> 01:16:05,710 Tremendous direction specificity. 1553 01:16:05,710 --> 01:16:08,660 And you get the same in MST, but there the receptive fields 1554 01:16:08,660 --> 01:16:09,606 are just gigantic. 1555 01:16:10,630 --> 01:16:14,760 But again, direction specificity is just as specific 1556 01:16:14,760 --> 01:16:16,690 as it is in MT. 1557 01:16:16,690 --> 01:16:20,350 So these two areas play a central role 1558 01:16:20,350 --> 01:16:28,010 in motion analysis, as we shall see in more detail later on. 1559 01:16:28,010 --> 01:16:28,640 All right. 1560 01:16:28,640 --> 01:16:32,130 So now, if you look at the spatial layout of this, 1561 01:16:32,130 --> 01:16:37,860 what you do is you can move an electrode into the area. 1562 01:16:37,860 --> 01:16:43,720 And so you move it across the represented area 1563 01:16:43,720 --> 01:16:46,390 by going into the sulcus. 1564 01:16:46,390 --> 01:16:50,440 And what you can see here is it's a systematic progression 1565 01:16:50,440 --> 01:16:53,520 of directions as you move the electrodes. 1566 01:16:53,520 --> 01:16:58,650 Here the distance is expressed in micrometers. 1567 01:16:58,650 --> 01:16:59,150 OK? 1568 01:17:01,290 --> 01:17:10,210 Now, if you map that out you can create a layout of the area, 1569 01:17:10,210 --> 01:17:11,160 area MT. 1570 01:17:11,160 --> 01:17:14,110 And what you can see is there's a systematic kilometer 1571 01:17:14,110 --> 01:17:18,030 arrangement, or different direction specificities. 1572 01:17:19,350 --> 01:17:21,990 So this is a general principle, by the way, 1573 01:17:21,990 --> 01:17:24,310 of the way the cortex is organized. 1574 01:17:24,310 --> 01:17:27,920 It first actually was discovered by Mountcastle 1575 01:17:27,920 --> 01:17:30,030 in the somatosensory system. 1576 01:17:30,030 --> 01:17:30,630 OK? 1577 01:17:30,630 --> 01:17:33,230 Showing that there's a kilometer arrangement there, 1578 01:17:33,230 --> 01:17:35,750 and then subsequently it was shown that this is also 1579 01:17:35,750 --> 01:17:40,761 true for many other areas, including vision and audition. 1580 01:17:40,761 --> 01:17:41,260 All right. 1581 01:17:41,260 --> 01:17:44,370 So now, the last thing I want to mention very briefly, 1582 01:17:44,370 --> 01:17:47,060 we'll get back to it, is so-called inferotemporal 1583 01:17:47,060 --> 01:17:48,140 cortex. 1584 01:17:48,140 --> 01:17:51,950 And here once again, we have a map of the monkey cortex. 1585 01:17:51,950 --> 01:17:55,850 And this down here is a temporal area here. 1586 01:17:55,850 --> 01:17:58,780 And this is where the inferotemporal cortex resides. 1587 01:17:58,780 --> 01:18:01,480 And this has been studied extensively 1588 01:18:01,480 --> 01:18:06,340 by people like Charlie Gross, and more recently 1589 01:18:06,340 --> 01:18:10,260 with imaging techniques by people in the department here. 1590 01:18:10,260 --> 01:18:12,440 It was shown that this area has a lot 1591 01:18:12,440 --> 01:18:15,760 to do with object recognition, face recognition. 1592 01:18:15,760 --> 01:18:18,840 A very complex high level area. 1593 01:18:18,840 --> 01:18:20,100 OK. 1594 01:18:20,100 --> 01:18:24,610 So now, I want to summarize what I have covered today. 1595 01:18:24,610 --> 01:18:27,990 First of all, the contralateral visual hemifield 1596 01:18:27,990 --> 01:18:31,385 is laid out topographically in V1 in each hemisphere. 1597 01:18:32,470 --> 01:18:34,340 You know that by now, cold. 1598 01:18:34,340 --> 01:18:37,140 Secondly, the major transforms in V1 1599 01:18:37,140 --> 01:18:40,969 are orientation, direction, spatial frequency, selectivity, 1600 01:18:40,969 --> 01:18:41,510 binocularity. 1601 01:18:42,830 --> 01:18:45,390 You have an ON/OFF convergence, and you 1602 01:18:45,390 --> 01:18:47,210 have also in many cells a convergence 1603 01:18:47,210 --> 01:18:50,140 of the midget and parasol systems. 1604 01:18:50,140 --> 01:18:55,136 Then V1 is organized in a modular fashion. 1605 01:18:55,136 --> 01:18:56,510 Another way to put it is that you 1606 01:18:56,510 --> 01:19:00,750 have a kilometer organization, and that we 1607 01:19:00,750 --> 01:19:05,839 talked about three models that had been proposed. 1608 01:19:05,839 --> 01:19:07,922 One of them, the original one by Hubel and Wiesel, 1609 01:19:07,922 --> 01:19:09,400 is the ice cube model. 1610 01:19:09,400 --> 01:19:11,940 Then we had the radio model, and the swirl model. 1611 01:19:11,940 --> 01:19:14,450 And I pointed out to you that the swirl model is not really 1612 01:19:14,450 --> 01:19:15,140 a model. 1613 01:19:15,140 --> 01:19:16,380 It's a fact. 1614 01:19:16,380 --> 01:19:18,260 That's how it's laid out. 1615 01:19:18,260 --> 01:19:19,050 OK. 1616 01:19:19,050 --> 01:19:22,970 And then I said also that there are more than 30 visual areas 1617 01:19:22,970 --> 01:19:24,860 that make more than 300 interconnections. 1618 01:19:27,060 --> 01:19:29,100 Extrastriate areas do not specialize 1619 01:19:29,100 --> 01:19:31,820 in any one single function, contrary to what 1620 01:19:31,820 --> 01:19:36,532 had been very, very popular, maybe even 10 or 15 years ago. 1621 01:19:36,532 --> 01:19:38,670 The receptive field's size in neurons 1622 01:19:38,670 --> 01:19:41,100 increases greatly in progressively higher 1623 01:19:41,100 --> 01:19:42,180 visual areas. 1624 01:19:42,180 --> 01:19:46,680 That is a very highly solid fact. 1625 01:19:46,680 --> 01:19:50,570 Then, area MT is involved in the analysis of motion. 1626 01:19:50,570 --> 01:19:53,370 And as we shall see later on, it also 1627 01:19:53,370 --> 01:19:55,320 contributes to perception of depth, 1628 01:19:55,320 --> 01:19:57,385 and to flickering stimuli. 1629 01:19:58,580 --> 01:20:02,580 Area V4 engages in many aspects of visual analysis, 1630 01:20:02,580 --> 01:20:04,740 and neurons have dynamic properties. 1631 01:20:04,740 --> 01:20:09,750 Attention and eye movements modulate 1632 01:20:09,750 --> 01:20:11,680 the way those cells respond. 1633 01:20:11,680 --> 01:20:14,500 And then in inferotemporal cortex, high level 1634 01:20:14,500 --> 01:20:18,750 visual analysis takes place that includes object recognition, 1635 01:20:18,750 --> 01:20:21,960 and therefore also the recognition of faces. 1636 01:20:21,960 --> 01:20:24,490 So that then summarizes what I had to say. 1637 01:20:25,590 --> 01:20:28,530 And to conclude in general, we can 1638 01:20:28,530 --> 01:20:35,870 say that the cells in the cortex have 1639 01:20:35,870 --> 01:20:41,310 one very important increase in their mode of functioning 1640 01:20:41,310 --> 01:20:44,520 compared to lower areas, in that the cells are multifunctional. 1641 01:20:45,820 --> 01:20:49,030 So any one cell can tell you information 1642 01:20:49,030 --> 01:20:54,510 already in V1 about direction, as well as 1643 01:20:54,510 --> 01:20:58,370 orientation and spatial frequency. 1644 01:20:58,370 --> 01:21:01,750 So if one cell can do many, many different things then 1645 01:21:01,750 --> 01:21:04,860 that's a good thing, because if each cell in the brain 1646 01:21:04,860 --> 01:21:07,160 specialized in one thing, it used 1647 01:21:07,160 --> 01:21:11,410 to be the hypothesis that was called your famous grandmother 1648 01:21:11,410 --> 01:21:12,330 cell. 1649 01:21:12,330 --> 01:21:14,160 That there was a cell in the brain 1650 01:21:14,160 --> 01:21:16,580 that represented your grandmother. 1651 01:21:16,580 --> 01:21:21,090 Now, if that were the case, that individual cells were 1652 01:21:21,090 --> 01:21:23,880 future selective, you would need a brain 1653 01:21:23,880 --> 01:21:28,390 as big as this room to accommodate the ability to what 1654 01:21:28,390 --> 01:21:31,110 you can do that you can do anyway. 1655 01:21:31,110 --> 01:21:33,970 So cells in the brain are multifunctional. 1656 01:21:33,970 --> 01:21:36,490 They carry out many, many different analyses, 1657 01:21:36,490 --> 01:21:41,610 just like when computers do complex mathematical analyses. 1658 01:21:41,610 --> 01:21:43,650 And because of that it's very difficult, 1659 01:21:43,650 --> 01:21:46,940 of course, especially when you study with high visual areas, 1660 01:21:46,940 --> 01:21:51,530 to learn just how these cells function, 1661 01:21:51,530 --> 01:21:54,020 because to understand that you would need to record 1662 01:21:54,020 --> 01:21:58,310 at the same time from virtually all the cells in an area 1663 01:21:58,310 --> 01:22:01,910 to see what each of them does, from which you can derive what 1664 01:22:01,910 --> 01:22:03,900 the actual analysis is. 1665 01:22:03,900 --> 01:22:07,380 So indeed, we have a very complex task ahead of us 1666 01:22:07,380 --> 01:22:11,130 in trying to understand how this very complex interaction 1667 01:22:11,130 --> 01:22:15,440 among neurons eventually results in your ability 1668 01:22:15,440 --> 01:22:18,670 to analyze various aspects of the visual scene. 1669 01:22:18,670 --> 01:22:20,170 And that's true not only for vision. 1670 01:22:20,170 --> 01:22:22,030 It's true for many other areas. 1671 01:22:22,030 --> 01:22:25,450 In vision it's very, very complex, 1672 01:22:25,450 --> 01:22:29,760 because you have all these incredible number of cells. 1673 01:22:29,760 --> 01:22:35,550 As I mentioned before, you have more than a million 1674 01:22:35,550 --> 01:22:38,510 retinal ganglion cells that project 1675 01:22:38,510 --> 01:22:42,100 from the one eye, from each eye, into the brain. 1676 01:22:42,100 --> 01:22:48,520 And then this multiplies in these higher visual areas. 1677 01:22:48,520 --> 01:22:52,540 Then you have millions and millions, billions of cells 1678 01:22:52,540 --> 01:22:55,640 to perform all of these analyses. 1679 01:22:55,640 --> 01:22:59,500 Now by contrast, when you look at the auditory system, 1680 01:22:59,500 --> 01:23:01,260 and I'm not saying this to belittle it, 1681 01:23:01,260 --> 01:23:04,560 I'm just saying it's probably easier to understand it, 1682 01:23:04,560 --> 01:23:08,530 because in the auditory system the fibers that 1683 01:23:08,530 --> 01:23:16,760 project from the cochlear nucleus amount to about 30,000 1684 01:23:16,760 --> 01:23:21,680 in each cochlear to the central nervous system. 1685 01:23:21,680 --> 01:23:28,240 So we're talking about 20, 30-fold higher level 1686 01:23:28,240 --> 01:23:31,680 of number of neurons involved in the visual system. 1687 01:23:31,680 --> 01:23:35,920 And that's of course because vision for us 1688 01:23:35,920 --> 01:23:41,350 is just a very, very important attribute, 1689 01:23:41,350 --> 01:23:43,890 perhaps more so than many other attributes 1690 01:23:43,890 --> 01:23:45,550 that we have in the nervous system. 1691 01:23:45,550 --> 01:23:46,050 OK. 1692 01:23:46,050 --> 01:23:48,460 So that's what I have to cover today. 1693 01:23:48,460 --> 01:23:51,559 And then next time, you're going to sort of move back. 1694 01:23:51,559 --> 01:23:53,600 We're going to talk about the on and off channels 1695 01:23:53,600 --> 01:23:56,580 to try to understand why on earth did they evolve. 1696 01:23:56,580 --> 01:23:58,130 OK?