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 00:00:25,695 --> 00:00:26,570 PROFESSOR: All right. 9 00:00:26,570 --> 00:00:27,995 Good afternoon, everyone, again. 10 00:00:30,870 --> 00:00:34,690 Today, our session is going to be on motion perception 11 00:00:34,690 --> 00:00:37,420 and pursuit eye movements. 12 00:00:37,420 --> 00:00:42,930 Now, I want to remind you that to our next class next Monday 13 00:00:42,930 --> 00:00:46,580 is going to be a whirlwind review. 14 00:00:46,580 --> 00:00:49,560 And then the next Wednesday we are 15 00:00:49,560 --> 00:00:53,020 going to have the midterm exam on what 16 00:00:53,020 --> 00:00:55,270 we have covered so far in the course. 17 00:00:55,270 --> 00:00:59,320 And as I've mentioned already to you a couple of times, 18 00:00:59,320 --> 00:01:04,209 that is going to consist of a series of multiple choice 19 00:01:04,209 --> 00:01:07,130 questions in which you have to circle 20 00:01:07,130 --> 00:01:10,950 the appropriate choice for each question. 21 00:01:10,950 --> 00:01:17,400 Then the following week, which is going to be November 28, 22 00:01:17,400 --> 00:01:18,206 is when Chris-- 23 00:01:18,206 --> 00:01:18,956 AUDIENCE: October. 24 00:01:18,956 --> 00:01:19,412 PROFESSOR: Huh? 25 00:01:19,412 --> 00:01:19,870 AUDIENCE: October 28. 26 00:01:19,870 --> 00:01:21,190 PROFESSOR: Sorry, October what? 27 00:01:21,190 --> 00:01:22,180 AUDIENCE: 28th. 28 00:01:22,180 --> 00:01:23,310 PROFESSOR: October 28. 29 00:01:23,310 --> 00:01:28,060 Yeah, October 28-- that's when Chris Brown 30 00:01:28,060 --> 00:01:31,475 is going to start the second half of the course on audition. 31 00:01:32,700 --> 00:01:35,270 So that is the general plan at this point. 32 00:01:35,270 --> 00:01:39,980 And again, I'm going to send around an attendance sheet. 33 00:01:39,980 --> 00:01:43,080 So just put your name on it and send it around, please, for me. 34 00:01:43,080 --> 00:01:43,650 Thank you. 35 00:01:45,680 --> 00:01:48,400 All right, so I think you're going 36 00:01:48,400 --> 00:01:51,821 to find talking about motion perception a lot of fun. 37 00:01:51,821 --> 00:01:53,070 I have several demonstrations. 38 00:01:54,210 --> 00:01:58,450 And it is certainly have a very interesting topic. 39 00:01:58,450 --> 00:02:02,490 And I'm going to highlight that by first of all telling you 40 00:02:02,490 --> 00:02:04,960 what we are going to cover. 41 00:02:04,960 --> 00:02:08,860 And then we are going to talk about the basic mechanisms 42 00:02:08,860 --> 00:02:13,240 of direction selectivity in the visual cortex. 43 00:02:13,240 --> 00:02:15,370 So first of all, as I've said, we're 44 00:02:15,370 --> 00:02:20,795 going to talk about the neurons in the brain that code motion. 45 00:02:21,910 --> 00:02:25,710 I'm going to talk about it mostly for in the cortex. 46 00:02:25,710 --> 00:02:30,310 But I would also say a few words about direction selective cells 47 00:02:30,310 --> 00:02:34,320 that you find in the retina that we had mentioned before. 48 00:02:34,320 --> 00:02:38,920 All right, so we are going to then discuss briefly 49 00:02:38,920 --> 00:02:43,360 the possible mechanisms that make 50 00:02:43,360 --> 00:02:47,370 it feasible to process motion. 51 00:02:47,370 --> 00:02:49,570 And then we are going to look at what 52 00:02:49,570 --> 00:02:55,650 happens when you make various brain lesions in determining 53 00:02:55,650 --> 00:02:58,300 how well you can process depth. 54 00:02:59,910 --> 00:03:03,190 Then we're going to talk about a very interesting topic, which 55 00:03:03,190 --> 00:03:05,260 is structure for motion. 56 00:03:05,260 --> 00:03:08,740 And then further pursuing that, we're 57 00:03:08,740 --> 00:03:11,340 going to talk about apparent motion. 58 00:03:11,340 --> 00:03:15,630 And then we are going to talk about an interesting temporal 59 00:03:15,630 --> 00:03:17,130 effect which is called metacontrast. 60 00:03:18,700 --> 00:03:23,950 And then we are going to talk about optokynetic nystagmus, 61 00:03:23,950 --> 00:03:29,430 where we link the visual input to the generation of pursuit 62 00:03:29,430 --> 00:03:30,810 eye movements. 63 00:03:30,810 --> 00:03:34,360 Then that is accomplished by the so-called accessory optic 64 00:03:34,360 --> 00:03:36,240 system that I will describe to you. 65 00:03:36,240 --> 00:03:40,350 And that is also the topic that, if you remember, 66 00:03:40,350 --> 00:03:42,300 you will be writing a paper on. 67 00:03:42,300 --> 00:03:44,400 Lastly then, we are going to have a summary. 68 00:03:44,400 --> 00:03:49,620 OK, so let's look first at the neural responses 69 00:03:49,620 --> 00:03:50,955 to motion in the cortex. 70 00:03:52,270 --> 00:03:55,080 Now, one of the most remarkable features 71 00:03:55,080 --> 00:03:58,560 of the visual cortex and extra cortical areas, 72 00:03:58,560 --> 00:04:02,200 as I had already mentioned, is that the majority 73 00:04:02,200 --> 00:04:07,530 of cells in these areas are direction selective, 74 00:04:07,530 --> 00:04:11,830 meaning that they respond to one particular direction of motion, 75 00:04:11,830 --> 00:04:14,220 but not to the opposite direction. 76 00:04:14,220 --> 00:04:18,890 Now, this is extremely evident in the visual cortex, where 77 00:04:18,890 --> 00:04:22,470 the so-called simple cells, virtually all of them 78 00:04:22,470 --> 00:04:24,070 are direction selective. 79 00:04:24,070 --> 00:04:25,940 So that's one of the major transforms, 80 00:04:25,940 --> 00:04:29,850 if you remember, of the input that you 81 00:04:29,850 --> 00:04:32,780 have from the lateral geniculate nucleus that 82 00:04:32,780 --> 00:04:34,560 goes to the cortex. 83 00:04:34,560 --> 00:04:39,470 That input is just from single cells, on and off single cells, 84 00:04:39,470 --> 00:04:43,330 midget and parasol cells that have circular receptive fields 85 00:04:43,330 --> 00:04:45,410 that centers around antagonism. 86 00:04:45,410 --> 00:04:47,630 And when they get up to the cortex, 87 00:04:47,630 --> 00:04:50,110 a transformation takes place by virtue 88 00:04:50,110 --> 00:04:56,900 of intracortical circuits that turns these cells 89 00:04:56,900 --> 00:05:00,060 into direction specific ones and cells that 90 00:05:00,060 --> 00:05:02,230 are orientation specific. 91 00:05:02,230 --> 00:05:08,990 So let's look at this first by starting with a method 92 00:05:08,990 --> 00:05:12,160 so that you'll understand exactly how this is done. 93 00:05:12,160 --> 00:05:14,040 In this case, what we are going to do 94 00:05:14,040 --> 00:05:16,041 is we are going to have a photo cell. 95 00:05:16,041 --> 00:05:17,540 Let's just look at the top one here, 96 00:05:17,540 --> 00:05:20,900 A. You're going to have a photo cell that responds both 97 00:05:20,900 --> 00:05:23,900 to a light edge and a dark edge. 98 00:05:23,900 --> 00:05:28,070 And so when this bar moves across this photo cell, 99 00:05:28,070 --> 00:05:29,236 it will activate it. 100 00:05:29,236 --> 00:05:30,610 And what they're going to do-- we 101 00:05:30,610 --> 00:05:32,486 are going to move it across in one direction. 102 00:05:32,486 --> 00:05:34,318 And then we are going to move it back across 103 00:05:34,318 --> 00:05:35,520 in the other direction. 104 00:05:35,520 --> 00:05:36,700 And this is shown here. 105 00:05:36,700 --> 00:05:38,570 This is a time scale. 106 00:05:38,570 --> 00:05:41,530 And it shows that when this edge goes across, 107 00:05:41,530 --> 00:05:44,530 it activates the photo cell, gives it the beep. 108 00:05:44,530 --> 00:05:46,460 And then when the trailing edge goes across, 109 00:05:46,460 --> 00:05:48,360 it also does the same thing. 110 00:05:48,360 --> 00:05:50,120 So this is the light edge response. 111 00:05:50,120 --> 00:05:51,790 This is the dark edge response. 112 00:05:51,790 --> 00:05:54,030 Now, when you're going back the other way, 113 00:05:54,030 --> 00:05:56,320 you're going to have, again, the same arrangement. 114 00:05:56,320 --> 00:05:59,280 In the opposite direction, the light edge responds first. 115 00:05:59,280 --> 00:06:01,630 And the dark edge responds second. 116 00:06:01,630 --> 00:06:06,620 Now we are going to make a change so that it mimics 117 00:06:06,620 --> 00:06:09,885 the organization of simple cells in the visual cortex. 118 00:06:09,885 --> 00:06:13,600 We're going to put two photo cells, one 119 00:06:13,600 --> 00:06:17,770 which responds to the light edge and a dark one which 120 00:06:17,770 --> 00:06:19,700 responds to the dark edge. 121 00:06:19,700 --> 00:06:26,840 So now when this bar goes across and activates these two photo 122 00:06:26,840 --> 00:06:30,640 cells, what happens is, because they are not all centered 123 00:06:30,640 --> 00:06:34,460 exactly in the middle, it takes a longer time 124 00:06:34,460 --> 00:06:37,840 to activate the two successive [INAUDIBLE] when the bar goes 125 00:06:37,840 --> 00:06:40,560 across this way and a short time when 126 00:06:40,560 --> 00:06:42,810 it goes across the other way. 127 00:06:42,810 --> 00:06:46,180 Now, the opposite happens when you have a dark bar, 128 00:06:46,180 --> 00:06:49,500 because then the dark bar doesn't activate 129 00:06:49,500 --> 00:06:52,920 the light sensitive photo cell. 130 00:06:52,920 --> 00:06:54,400 It activities the dark one. 131 00:06:54,400 --> 00:06:57,300 So it takes a little longer to start that. 132 00:06:57,300 --> 00:07:01,570 So that then shows you the temporal arrangements 133 00:07:01,570 --> 00:07:07,320 of these responses if when the photo cells 134 00:07:07,320 --> 00:07:08,530 are arranged in this fashion. 135 00:07:08,530 --> 00:07:10,360 Everybody understand that? 136 00:07:10,360 --> 00:07:13,420 OK, so now we are going to do this for the real. 137 00:07:13,420 --> 00:07:17,100 We are going to look at a cortical cell. 138 00:07:17,100 --> 00:07:21,150 And there are several different subclasses of simple cells. 139 00:07:21,150 --> 00:07:23,470 And the first subclass we can refer 140 00:07:23,470 --> 00:07:26,670 to as an S1, S for simple, OK? 141 00:07:26,670 --> 00:07:29,490 So the simple cell there is an example, 142 00:07:29,490 --> 00:07:31,200 very similar arrangement. 143 00:07:31,200 --> 00:07:34,170 Here's the receptive field of the cell, undefined. 144 00:07:34,170 --> 00:07:37,510 And you either put a light edge across back and forth 145 00:07:37,510 --> 00:07:39,210 or a dark edge back and forth. 146 00:07:39,210 --> 00:07:40,590 This is the light edge response. 147 00:07:40,590 --> 00:07:42,550 This is the dark edge response. 148 00:07:42,550 --> 00:07:44,390 Now, you can readily see what happens 149 00:07:44,390 --> 00:07:46,470 here, which is quite remarkable. 150 00:07:46,470 --> 00:07:52,440 The cell responds only when the bar moves upward 151 00:07:52,440 --> 00:07:54,860 across the receptive field and does not 152 00:07:54,860 --> 00:07:56,810 respond when it moves downward. 153 00:07:56,810 --> 00:08:00,310 So this cell is over 100% direction selective. 154 00:08:00,310 --> 00:08:01,410 Got it? 155 00:08:01,410 --> 00:08:02,160 All right. 156 00:08:02,160 --> 00:08:05,050 Now, the same thing happens to the dark edge, all right? 157 00:08:06,510 --> 00:08:12,120 And so this cell is one that responds only to the dark edge, 158 00:08:12,120 --> 00:08:12,750 OK? 159 00:08:12,750 --> 00:08:15,420 In this case, it's the trailing edge. 160 00:08:15,420 --> 00:08:17,670 In this case, it's the leading edge. 161 00:08:17,670 --> 00:08:20,940 Now, at the bottom here, you draw out 162 00:08:20,940 --> 00:08:23,790 what the receptive field looks like. 163 00:08:23,790 --> 00:08:26,610 This is in 10th of degrees. 164 00:08:26,610 --> 00:08:30,750 And so this is pretty much just the spatial response 165 00:08:30,750 --> 00:08:31,700 of the cell. 166 00:08:31,700 --> 00:08:33,970 And it only responds in this direction. 167 00:08:33,970 --> 00:08:35,990 There's nothing drawn in the other way. 168 00:08:35,990 --> 00:08:38,970 Now I'm going to show you another cell, which 169 00:08:38,970 --> 00:08:42,315 is going to be a so-called S2 cell. 170 00:08:42,315 --> 00:08:45,335 Now, this cell is different in that it responds 171 00:08:45,335 --> 00:08:48,060 both to the light and the dark edges. 172 00:08:48,060 --> 00:08:51,190 And the fact that they're-- in this case, a light bar, 173 00:08:51,190 --> 00:08:54,860 in this case, a dark bar, the fact that they're-- displaced 174 00:08:54,860 --> 00:08:58,250 temporally mean that they're side by side for the light 175 00:08:58,250 --> 00:09:01,110 and dark responses rather than overlapping. 176 00:09:01,110 --> 00:09:04,620 And that, of course, is the definition of a simple cell. 177 00:09:04,620 --> 00:09:08,070 And so if you draw out the receptive field of this cell, 178 00:09:08,070 --> 00:09:13,400 once again, the cell is almost 100% direction selective. 179 00:09:13,400 --> 00:09:17,685 And this is what its response looks like in space. 180 00:09:18,710 --> 00:09:20,710 Now we are going to look at yet another one. 181 00:09:20,710 --> 00:09:22,800 And this is going to be a surprise. 182 00:09:22,800 --> 00:09:25,280 It was certainly a surprise when this was discovered, 183 00:09:25,280 --> 00:09:26,780 because it's still not understood 184 00:09:26,780 --> 00:09:28,910 why we have these kinds of cells. 185 00:09:28,910 --> 00:09:31,570 Here is one-- you can look at that-- 186 00:09:31,570 --> 00:09:36,140 that responds to a light edge in this direction and a dark edge 187 00:09:36,140 --> 00:09:38,260 in this direction, OK? 188 00:09:38,260 --> 00:09:41,990 So both subfields are direction selective. 189 00:09:41,990 --> 00:09:45,070 But they're direction selective in opposite directions, 190 00:09:45,070 --> 00:09:47,000 as depicted here. 191 00:09:47,000 --> 00:09:52,430 Now why on earth would you have cells in the visual cortex that 192 00:09:52,430 --> 00:09:55,220 respond selectively to light and dark edges 193 00:09:55,220 --> 00:09:56,924 in opposite directions? 194 00:09:56,924 --> 00:09:58,465 That's something you can contemplate. 195 00:10:00,230 --> 00:10:03,190 There is no definitive answer to that, 196 00:10:03,190 --> 00:10:05,720 and so I'm not going to belabor it any further? 197 00:10:05,720 --> 00:10:08,600 Now what I can do is I can provide you with a summary. 198 00:10:08,600 --> 00:10:10,790 I showed you this one, this one, and this one. 199 00:10:10,790 --> 00:10:12,650 Now, you don't have to worry about those. 200 00:10:12,650 --> 00:10:15,700 But I would like you to look at this last one, which 201 00:10:15,700 --> 00:10:17,650 is a complex cell. 202 00:10:17,650 --> 00:10:21,510 In a complex cell, the light and dark edge responses 203 00:10:21,510 --> 00:10:25,360 are in the same location, space-wise, as shown here. 204 00:10:25,360 --> 00:10:29,430 And this particular complex cell that's a real cell 205 00:10:29,430 --> 00:10:32,730 is 100% direction selective. 206 00:10:32,730 --> 00:10:36,310 Now, it's true not for all complex cells, 207 00:10:36,310 --> 00:10:39,130 but the majority of them that they too 208 00:10:39,130 --> 00:10:40,515 are direction selective. 209 00:10:41,690 --> 00:10:45,150 So I can safely tell you that one 210 00:10:45,150 --> 00:10:49,920 of the major transforms that happens in the brain in vision 211 00:10:49,920 --> 00:10:54,490 is to convert the signals form the retina into direction 212 00:10:54,490 --> 00:10:55,885 selective cells. 213 00:10:56,920 --> 00:10:59,530 And therefore, we can only surmise 214 00:10:59,530 --> 00:11:02,490 that to have direction selective cells is extremely 215 00:11:02,490 --> 00:11:06,020 important for us to be able to process 216 00:11:06,020 --> 00:11:08,820 motion information in the world. 217 00:11:08,820 --> 00:11:12,110 So now we are going to move on and look 218 00:11:12,110 --> 00:11:15,900 at extrastriate cortex. 219 00:11:15,900 --> 00:11:19,170 And we already talked briefly about MT and MST. 220 00:11:19,170 --> 00:11:24,260 In these two areas, we have even an stronger disposition 221 00:11:24,260 --> 00:11:26,560 for cells to be direction selective. 222 00:11:26,560 --> 00:11:29,500 And here's an example of a cell in MT-- 223 00:11:29,500 --> 00:11:32,380 in this direction of motion, a vigorous response. 224 00:11:32,380 --> 00:11:33,830 This is a cumulative response. 225 00:11:33,830 --> 00:11:36,400 In the opposite direction, it actually has some inhibition, 226 00:11:36,400 --> 00:11:37,940 as you can see here and here. 227 00:11:39,390 --> 00:11:43,290 Then if you recall from MST, which is an area yet further 228 00:11:43,290 --> 00:11:50,050 removed from area V1-- it's the middle superior temporal area-- 229 00:11:50,050 --> 00:11:53,140 what you find is, again, very strongly direction 230 00:11:53,140 --> 00:11:54,430 selective cells. 231 00:11:54,430 --> 00:11:56,970 Most of them-- maybe not all of them, 232 00:11:56,970 --> 00:12:00,650 but the overwhelming majority of them-- are direction selective. 233 00:12:00,650 --> 00:12:03,640 But they have much, much larger receptive fields 234 00:12:03,640 --> 00:12:05,370 than cells do in MT. 235 00:12:06,440 --> 00:12:09,590 And as I've mentioned to you before, one of the basic rules 236 00:12:09,590 --> 00:12:19,310 is that as you progress from the V1 237 00:12:19,310 --> 00:12:22,520 to higher cortical areas, the receptive 238 00:12:22,520 --> 00:12:25,380 fields in these progressively higher areas become 239 00:12:25,380 --> 00:12:30,030 bigger and bigger and cover more and more of the visual space. 240 00:12:30,030 --> 00:12:32,880 So here you can see this is shown 241 00:12:32,880 --> 00:12:35,250 in terms of the visual field here. 242 00:12:36,580 --> 00:12:37,650 This is way out. 243 00:12:37,650 --> 00:12:39,820 This is 20 degrees just from here to here. 244 00:12:41,190 --> 00:12:43,375 The receptive is gigantic, actually. 245 00:12:44,560 --> 00:12:48,260 And there's a beautiful, very lawful arrangement 246 00:12:48,260 --> 00:12:52,900 and even though many, many cells feed into this particular cell 247 00:12:52,900 --> 00:12:58,760 from earlier portions of the visual cortex, 248 00:12:58,760 --> 00:13:04,220 they maintain or get a specific input from all of the cells 249 00:13:04,220 --> 00:13:05,880 that have this direction selectivity. 250 00:13:07,286 --> 00:13:11,410 Now, whether this direction selectivity is a product 251 00:13:11,410 --> 00:13:13,860 of the input from the cells, which have the same kind 252 00:13:13,860 --> 00:13:16,480 of directions, like many, many of these cells-- 253 00:13:16,480 --> 00:13:20,880 maybe 20 or 30 or 40 or 50 of them, OK?-- 254 00:13:20,880 --> 00:13:23,520 which have the same orientation and direction, 255 00:13:23,520 --> 00:13:26,510 or whether orientation and direction selectivity are 256 00:13:26,510 --> 00:13:31,805 created anew in MST is something that can be debated. 257 00:13:33,150 --> 00:13:35,400 It is still not 100% certain. 258 00:13:35,400 --> 00:13:38,220 So the only thing we have to worry about this time is 259 00:13:38,220 --> 00:13:40,680 to realize-- maybe not even worry. 260 00:13:40,680 --> 00:13:43,590 That's too strong a term-- is to be aware 261 00:13:43,590 --> 00:13:48,280 that there is prevalence of direction 262 00:13:48,280 --> 00:13:51,750 selectivity in both MT and MST. 263 00:13:51,750 --> 00:13:54,540 It's very beautiful and very, very specific, 264 00:13:54,540 --> 00:13:56,530 as you can see it here, even when 265 00:13:56,530 --> 00:13:59,320 you have a huge receptive field. 266 00:13:59,320 --> 00:14:01,460 So that is the basic story then. 267 00:14:01,460 --> 00:14:04,720 And now we can ask the question, how on earth 268 00:14:04,720 --> 00:14:06,900 is direction selectivity created? 269 00:14:06,900 --> 00:14:09,880 Now, there are several models that have been proposed. 270 00:14:09,880 --> 00:14:16,290 And most of these models-- I'll show you a simple one just so 271 00:14:16,290 --> 00:14:20,260 that you have a sense for it-- most of these models assume, 272 00:14:20,260 --> 00:14:22,790 I believe correctly, that you need 273 00:14:22,790 --> 00:14:30,410 to have inhibitory circuits that are selective for direction 274 00:14:30,410 --> 00:14:36,390 to produce this specific response in these cells. 275 00:14:36,390 --> 00:14:38,170 Now, the amazing thing about all this 276 00:14:38,170 --> 00:14:40,750 is that already in the retina you 277 00:14:40,750 --> 00:14:44,360 have some cells that are direction selective. 278 00:14:44,360 --> 00:14:47,630 And these cells I'll talk about a bit more 279 00:14:47,630 --> 00:14:51,410 when we talk about the accessory optic system. 280 00:14:51,410 --> 00:14:55,410 These cells in the retina are not very numerous. 281 00:14:55,410 --> 00:14:59,690 In the rabbit that has approximately 350,000 cells 282 00:14:59,690 --> 00:15:02,360 in each retina-- can you imagine a little rabbit? 283 00:15:02,360 --> 00:15:04,850 And it has 350,000 cells. 284 00:15:04,850 --> 00:15:07,380 We have a little over a million, OK? 285 00:15:07,380 --> 00:15:14,620 And among those 350,000 cells, about 7,000 286 00:15:14,620 --> 00:15:16,550 are the so-called cells of Dogiel. 287 00:15:17,810 --> 00:15:19,640 And actually, I'll come back to that. 288 00:15:19,640 --> 00:15:22,430 And those are direction selective cells. 289 00:15:22,430 --> 00:15:29,360 So a very clever experimentalist examined these direction 290 00:15:29,360 --> 00:15:32,910 selective cells in the retina, a fellow called Nigel Daw. 291 00:15:32,910 --> 00:15:36,340 And he asked the question, what can I 292 00:15:36,340 --> 00:15:39,570 do to determine what the mechanism is 293 00:15:39,570 --> 00:15:41,420 of direction selectivity? 294 00:15:41,420 --> 00:15:44,690 So he did an experiment that sounds simple. 295 00:15:44,690 --> 00:15:48,040 But let me tell you, it's a very difficult experiment. 296 00:15:48,040 --> 00:15:50,020 When you put your electrode into the retina, 297 00:15:50,020 --> 00:15:52,630 you can't see direction selective cells. 298 00:15:52,630 --> 00:15:54,990 So you're on a hunting expedition. 299 00:15:54,990 --> 00:15:58,090 So you keep recording from cell after cell after cell. 300 00:15:58,090 --> 00:16:02,830 And if indeed only 7,000 among the 350,000 301 00:16:02,830 --> 00:16:05,200 are direction selective, it will take you quite a while 302 00:16:05,200 --> 00:16:08,770 before you manage to record from a direction selective cell. 303 00:16:08,770 --> 00:16:11,360 So it's a lot of very hard work. 304 00:16:11,360 --> 00:16:15,790 So anyway, if you do that when your hypothesis is 305 00:16:15,790 --> 00:16:22,260 that some sort of inhibitory mechanism is involved, then 306 00:16:22,260 --> 00:16:24,170 you can do an experiment and say, 307 00:16:24,170 --> 00:16:26,880 well if it's inhibitory mechanisms, probably 308 00:16:26,880 --> 00:16:29,250 some sort of GABAergic system, we 309 00:16:29,250 --> 00:16:39,240 need to put some agent into the retina that blocks inhibition. 310 00:16:39,240 --> 00:16:41,820 We talked about that before, that that can happen. 311 00:16:41,820 --> 00:16:44,540 And so in this case, what they did-- let 312 00:16:44,540 --> 00:16:46,470 me come to that first. 313 00:16:46,470 --> 00:16:47,500 Let me see. 314 00:16:47,500 --> 00:16:48,794 Here we are. 315 00:16:48,794 --> 00:16:50,710 I'll come back to the others in just a minute. 316 00:16:52,120 --> 00:16:54,780 What we have here is one of these cells. 317 00:16:54,780 --> 00:16:58,650 And we move in this direction here 318 00:16:58,650 --> 00:17:00,230 and in this direction for this. 319 00:17:00,230 --> 00:17:03,090 And you can see that this cell is very strongly direction 320 00:17:03,090 --> 00:17:07,099 selective, maybe 100%, but maybe 80% or 90%. 321 00:17:08,160 --> 00:17:13,369 Then he injected picrotoxin into the eye that blocks GABA. 322 00:17:13,369 --> 00:17:15,059 And lo and behold, direction selectivity 323 00:17:15,059 --> 00:17:16,059 is virtually eliminated. 324 00:17:17,079 --> 00:17:20,420 And then when picrotoxin got washed out, 325 00:17:20,420 --> 00:17:23,310 there was recovery, all right? 326 00:17:23,310 --> 00:17:27,069 So now, on the basis of this finding-- 327 00:17:27,069 --> 00:17:33,590 it's a beautiful finding-- it was established that indeed, 328 00:17:33,590 --> 00:17:37,510 inhibitory circuits play a central role in giving rise 329 00:17:37,510 --> 00:17:39,870 to direction selectivity. 330 00:17:39,870 --> 00:17:42,110 And you can imagine that in the course of evolution, 331 00:17:42,110 --> 00:17:45,290 the pressures had to be tremendous to create 332 00:17:45,290 --> 00:17:49,592 these direction selective cells, because the circuitry is 333 00:17:49,592 --> 00:17:50,300 very complicated. 334 00:17:50,300 --> 00:17:54,090 You have to create these inhibitory neurons which 335 00:17:54,090 --> 00:17:56,425 inhibit in a particular direction rather than just 336 00:17:56,425 --> 00:17:56,925 randomly. 337 00:17:57,990 --> 00:17:59,570 So that's what was done. 338 00:17:59,570 --> 00:18:04,660 And so I'll show you one model that describes this. 339 00:18:04,660 --> 00:18:06,680 You have a bunch of-- this is for the retina. 340 00:18:06,680 --> 00:18:08,820 You have a bunch of receptors here. 341 00:18:08,820 --> 00:18:12,580 And they activate a retinal ganglion cell here. 342 00:18:12,580 --> 00:18:16,750 But they also connect with an inhibitory interneuron 343 00:18:16,750 --> 00:18:19,050 in one direction, in that direction. 344 00:18:19,050 --> 00:18:23,790 So now if you move a bar in this direction, as shown here, 345 00:18:23,790 --> 00:18:25,160 you get no response. 346 00:18:25,160 --> 00:18:28,380 But if you move a bar across the other way, 347 00:18:28,380 --> 00:18:30,710 you get a vigorous response, as shown by the action 348 00:18:30,710 --> 00:18:32,330 potentials here. 349 00:18:32,330 --> 00:18:38,650 So this is a fairly simple model of creating direction 350 00:18:38,650 --> 00:18:43,216 selectivity by virtue of inhibitory interneurons. 351 00:18:44,710 --> 00:18:47,810 So that then was also examined in the cortex. 352 00:18:47,810 --> 00:18:51,760 it was found that there too inhibitory interneurons 353 00:18:51,760 --> 00:18:57,660 play a central role when direction specificity is 354 00:18:57,660 --> 00:19:01,750 created in a cortical cell. 355 00:19:01,750 --> 00:19:06,000 So that now makes me come back to a further analysis 356 00:19:06,000 --> 00:19:13,720 of the basic nature of motion. 357 00:19:13,720 --> 00:19:18,490 Can we logically divide motion into several types, or what? 358 00:19:18,490 --> 00:19:24,750 And so it was proposed by a wonderful scientist called 359 00:19:24,750 --> 00:19:31,050 Wurtz at the National Institutes of Health that he could-- 360 00:19:31,050 --> 00:19:32,920 he decided that you can distinguish 361 00:19:32,920 --> 00:19:36,790 between planar motion, circular motion, and radial motion. 362 00:19:36,790 --> 00:19:39,070 Planar motion is when something moves straight. 363 00:19:39,070 --> 00:19:41,849 That's what we've talked about so far only. 364 00:19:41,849 --> 00:19:43,140 And of course, you can do that. 365 00:19:43,140 --> 00:19:47,602 In this case, he shows four different directions. 366 00:19:47,602 --> 00:19:49,060 But of course, there'd be all kinds 367 00:19:49,060 --> 00:19:53,010 of directions that are not vertical or horizontal 368 00:19:53,010 --> 00:19:54,374 but are diagonal. 369 00:19:54,374 --> 00:19:56,040 And there are many, many cortical cells. 370 00:19:56,040 --> 00:20:00,220 Some of those actually I showed you had orientation cell 371 00:20:00,220 --> 00:20:02,690 activity that was at an angle. 372 00:20:02,690 --> 00:20:06,150 So you have a huge number of different angles. 373 00:20:06,150 --> 00:20:09,850 And these cells hypothetically then 374 00:20:09,850 --> 00:20:13,340 would respond to planar motion. 375 00:20:13,340 --> 00:20:15,569 Then you have another type of motion, 376 00:20:15,569 --> 00:20:16,610 which is called circular. 377 00:20:17,960 --> 00:20:22,980 And let me tell you, when you see circular motion, obviously, 378 00:20:22,980 --> 00:20:26,360 if you were to be spun around or something, you'd see it. 379 00:20:26,360 --> 00:20:30,620 But it's not uncommon to see something rotate. 380 00:20:30,620 --> 00:20:35,935 So if you have a spinning wheel or if you watch a tire rotating 381 00:20:35,935 --> 00:20:38,820 in motion, that's circular motion. 382 00:20:38,820 --> 00:20:42,120 And so the hypothesis is that there 383 00:20:42,120 --> 00:20:47,210 should be cells in the brain that selectively respond 384 00:20:47,210 --> 00:20:51,120 to circular motion, either clockwise or counterclockwise. 385 00:20:51,120 --> 00:20:55,770 And lastly, of course, the so-called radial motion-- 386 00:20:55,770 --> 00:20:58,140 and you encounter that just about every day. 387 00:20:58,140 --> 00:21:01,980 Whenever you drive, you see trees. 388 00:21:01,980 --> 00:21:05,330 Or if you're driving out in the country, 389 00:21:05,330 --> 00:21:09,020 you see trees going past you, so to speak, as you drive forward. 390 00:21:10,270 --> 00:21:14,485 And that would be motion which would be outward. 391 00:21:15,570 --> 00:21:18,040 Or if you were to back up like this, 392 00:21:18,040 --> 00:21:21,040 then you would get this kind of motion 393 00:21:21,040 --> 00:21:23,810 where things would come together, OK? 394 00:21:23,810 --> 00:21:27,510 So that kind of motion is certainly one 395 00:21:27,510 --> 00:21:29,600 that you need to somehow code. 396 00:21:29,600 --> 00:21:32,060 And so another question came up once he came up 397 00:21:32,060 --> 00:21:36,490 with this hypothesis that these different kinds of motion-- 398 00:21:36,490 --> 00:21:42,280 the question is, are there cells in various parts of the brain, 399 00:21:42,280 --> 00:21:46,490 especially in higher areas like MST, 400 00:21:46,490 --> 00:21:49,865 that are selective for these three basic types of motion? 401 00:21:51,350 --> 00:21:53,450 And so they began to do an experiment. 402 00:21:53,450 --> 00:21:56,340 And I'll show you some summary data here. 403 00:21:56,340 --> 00:22:01,200 Here is a cell when you study planar motion. 404 00:22:01,200 --> 00:22:03,580 And these are the eight directions of motion 405 00:22:03,580 --> 00:22:06,330 that were studied-- sorry, four directions 406 00:22:06,330 --> 00:22:09,490 of motion on the top-- then the circular one on the next two, 407 00:22:09,490 --> 00:22:12,690 and the in and out motion at the bottom. 408 00:22:12,690 --> 00:22:16,820 This particular cell seems to be specific for planar motion 409 00:22:16,820 --> 00:22:19,180 from right to left. 410 00:22:19,180 --> 00:22:23,290 Now here is a cell that is selective to circular motion, 411 00:22:23,290 --> 00:22:24,110 OK? 412 00:22:24,110 --> 00:22:26,170 And lastly, here is a cell that's 413 00:22:26,170 --> 00:22:27,940 elective to radial motion. 414 00:22:29,090 --> 00:22:32,780 So if you only had recorded in your life 415 00:22:32,780 --> 00:22:37,550 from three cells in MST, and it happened 416 00:22:37,550 --> 00:22:39,630 to be these three cells, you would say, 417 00:22:39,630 --> 00:22:42,540 ah, my hypothesis is right. 418 00:22:42,540 --> 00:22:45,020 There are three kinds of cells in MST 419 00:22:45,020 --> 00:22:49,190 that respond selectively to these three classes of motion. 420 00:22:49,190 --> 00:22:52,240 So everything is incredibly beautiful and logical. 421 00:22:52,240 --> 00:22:55,060 But of course, that's not how you do your experiment, is it? 422 00:22:55,060 --> 00:22:57,330 The way you do your experiment is 423 00:22:57,330 --> 00:22:59,630 you record from MST in this case, 424 00:22:59,630 --> 00:23:02,180 and you record from many, many, many cells. 425 00:23:02,180 --> 00:23:05,170 And you test them just like this, OK? 426 00:23:05,170 --> 00:23:08,810 And if you do that, you realize that this is just 427 00:23:08,810 --> 00:23:10,930 an example of three cells that were specific. 428 00:23:12,160 --> 00:23:14,790 It was found that 40% of the cells 429 00:23:14,790 --> 00:23:17,830 responded to all three kinds of motion. 430 00:23:17,830 --> 00:23:20,660 305 responded to two types of motion. 431 00:23:20,660 --> 00:23:23,880 And only 20% of the cells responded 432 00:23:23,880 --> 00:23:26,550 to one type of motion. 433 00:23:26,550 --> 00:23:30,710 So therefore, these cells, as is so often 434 00:23:30,710 --> 00:23:36,700 the case in the visual cortex, are not unique and specific 435 00:23:36,700 --> 00:23:39,260 to analysis of one thing. 436 00:23:39,260 --> 00:23:42,820 These cells can tell you about all three of those. 437 00:23:42,820 --> 00:23:47,350 And what happens in reality then is that when something moves, 438 00:23:47,350 --> 00:23:51,100 you activate thousands of cells in MT, MST, 439 00:23:51,100 --> 00:23:53,740 and visual cortex, maybe tens of thousands. 440 00:23:53,740 --> 00:23:59,640 And it's some sort of complex relative activity 441 00:23:59,640 --> 00:24:02,960 that then eventually gives rise to the fact that you say, 442 00:24:02,960 --> 00:24:04,370 oh, yeah I see planar motion. 443 00:24:04,370 --> 00:24:06,390 Oh, yeah, I see a circular motion, 444 00:24:06,390 --> 00:24:09,530 or I see in and outward motion. 445 00:24:09,530 --> 00:24:12,990 So it's not a specific single cell 446 00:24:12,990 --> 00:24:15,400 that tells you or gives you that impression, 447 00:24:15,400 --> 00:24:19,260 but it's the concerted activity of thousands 448 00:24:19,260 --> 00:24:22,990 of cells even for the simplest of cases. 449 00:24:22,990 --> 00:24:24,960 So that's the way it works. 450 00:24:26,380 --> 00:24:29,270 It's a complex business. 451 00:24:29,270 --> 00:24:32,190 And the idea that was popular at one time, 452 00:24:32,190 --> 00:24:37,670 namely that you have feature detectors in the brain, 453 00:24:37,670 --> 00:24:40,780 has become downgraded progressively 454 00:24:40,780 --> 00:24:44,040 with increasing research. 455 00:24:45,080 --> 00:24:46,860 As I mentioned to you before, when 456 00:24:46,860 --> 00:24:49,740 we talked about to extrastriate cortex, 457 00:24:49,740 --> 00:24:53,950 the initial idea was when the various extrastriate areas 458 00:24:53,950 --> 00:24:58,670 had been discovered, that each area specializes 459 00:24:58,670 --> 00:25:00,480 in one particular thing. 460 00:25:00,480 --> 00:25:04,880 And then it turned out that that is not the case. 461 00:25:04,880 --> 00:25:10,750 There may be shifts or trends, but in each area, 462 00:25:10,750 --> 00:25:12,810 many different things are analyzed 463 00:25:12,810 --> 00:25:15,260 and thousands and thousands of cells 464 00:25:15,260 --> 00:25:22,480 are active whose relative activity is somehow 465 00:25:22,480 --> 00:25:26,310 what gives rise to the percept that you have. 466 00:25:26,310 --> 00:25:29,310 All right, so that then is the essence 467 00:25:29,310 --> 00:25:33,760 of the layout of the responses and the models 468 00:25:33,760 --> 00:25:38,660 that this model that [INAUDIBLE] told you about, 469 00:25:38,660 --> 00:25:40,450 or something similar to this. 470 00:25:40,450 --> 00:25:42,430 It may be a bit more complicated. 471 00:25:42,430 --> 00:25:44,250 And some people have hypothesized 472 00:25:44,250 --> 00:25:47,900 that there are temporal delays necessary to produce 473 00:25:47,900 --> 00:25:48,400 this thing. 474 00:25:48,400 --> 00:25:49,130 Yes? 475 00:25:49,130 --> 00:25:49,796 AUDIENCE: Sorry. 476 00:25:50,980 --> 00:25:53,728 There was a 10% missing in the breakdown. 477 00:25:54,762 --> 00:25:55,470 PROFESSOR: Sorry. 478 00:25:57,710 --> 00:25:58,970 Are you talking about this? 479 00:25:58,970 --> 00:25:59,960 AUDIENCE: Yeah. 480 00:25:59,960 --> 00:26:04,540 PROFESSOR: Well, that's-- 10% of the cells were unclassified 481 00:26:04,540 --> 00:26:07,370 or we're not direction selective. 482 00:26:09,800 --> 00:26:14,330 These they could with certainty say what the situation was. 483 00:26:14,330 --> 00:26:15,970 And maybe some of those cells they 484 00:26:15,970 --> 00:26:18,110 lost while that we're recording from 485 00:26:18,110 --> 00:26:20,341 or they didn't have any particular direction 486 00:26:20,341 --> 00:26:20,840 selectivity. 487 00:26:22,360 --> 00:26:25,170 And because of that, they couldn't come up 488 00:26:25,170 --> 00:26:27,145 with 100%, of course, yeah? 489 00:26:27,145 --> 00:26:28,145 Yes, you had a question. 490 00:26:28,145 --> 00:26:29,770 AUDIENCE: It was the same question. 491 00:26:29,770 --> 00:26:30,140 PROFESSOR: What? 492 00:26:30,140 --> 00:26:30,934 AUDIENCE: It was the same question. 493 00:26:30,934 --> 00:26:32,142 PROFESSOR: Same question, OK. 494 00:26:32,142 --> 00:26:34,390 Very observant-- good for you. 495 00:26:34,390 --> 00:26:40,580 All right, so now we are going to move forward 496 00:26:40,580 --> 00:26:44,200 and we're going to look at the effects of lesions 497 00:26:44,200 --> 00:26:45,273 on motion perception. 498 00:26:47,450 --> 00:26:54,600 One of the ready ways you can assess to what extent 499 00:26:54,600 --> 00:26:57,900 various higher cortical areas play 500 00:26:57,900 --> 00:27:04,320 a role in various neural mechanisms 501 00:27:04,320 --> 00:27:07,100 is to either reversibly inactivate them 502 00:27:07,100 --> 00:27:15,810 or to remove them by making lesions and then determining 503 00:27:15,810 --> 00:27:16,540 what happens. 504 00:27:16,540 --> 00:27:19,110 And if you remember, I talked about this a bit. 505 00:27:20,520 --> 00:27:23,450 I told you, for example, that the hypothesis that 506 00:27:23,450 --> 00:27:28,780 was made initially about area V4 is that it's a color area. 507 00:27:28,780 --> 00:27:32,050 That was a single function hypothesis. 508 00:27:32,050 --> 00:27:38,440 And then it turned out that when area V4 was removed, 509 00:27:38,440 --> 00:27:43,330 it was found that there was only a moderate to mild deficit 510 00:27:43,330 --> 00:27:44,560 in color vision. 511 00:27:44,560 --> 00:27:46,590 But there were all kinds of other deficits 512 00:27:46,590 --> 00:27:50,780 that we had discussed in previous sessions. 513 00:27:50,780 --> 00:27:52,430 So now we come to the question about, 514 00:27:52,430 --> 00:27:55,530 what about motion perception? 515 00:27:55,530 --> 00:28:00,250 When we talk about emotion perception, we of course 516 00:28:00,250 --> 00:28:03,010 are dealing with temporal things, 517 00:28:03,010 --> 00:28:08,490 because motion involves successive activation of cells. 518 00:28:08,490 --> 00:28:11,470 And because of that, it's also very important 519 00:28:11,470 --> 00:28:19,240 to study how well we can process rapid successive presentations 520 00:28:19,240 --> 00:28:21,010 as, for example, in flicker. 521 00:28:22,590 --> 00:28:25,160 And that brings me, of course, to the story-- 522 00:28:25,160 --> 00:28:27,240 and we will talk about that in a minute-- 523 00:28:27,240 --> 00:28:37,150 that in the olden days, when movies were first invented, 524 00:28:37,150 --> 00:28:40,670 the rate at which successive frames were presented 525 00:28:40,670 --> 00:28:48,860 was 16 Hertz, which was fairly close to having 526 00:28:48,860 --> 00:28:51,790 a sensational of flicker, all right? 527 00:28:51,790 --> 00:28:54,070 And because of that, what happened 528 00:28:54,070 --> 00:28:57,265 when a guy asked a girl to go the movies? 529 00:28:57,265 --> 00:28:59,390 How did the guy ask the person to go to the movies? 530 00:28:59,390 --> 00:29:00,660 Can anybody remember that? 531 00:29:00,660 --> 00:29:01,840 That's an olden thing. 532 00:29:04,240 --> 00:29:06,870 The guy would say, hey, how would you 533 00:29:06,870 --> 00:29:09,385 like to go to the flicks tonight? 534 00:29:09,385 --> 00:29:11,510 OK, you've heard that expression, go to the flicks, 535 00:29:11,510 --> 00:29:12,009 right? 536 00:29:13,320 --> 00:29:18,080 Well, nowadays we are not aware of flicks anymore. 537 00:29:18,080 --> 00:29:19,540 And there are two reasons for that. 538 00:29:19,540 --> 00:29:21,081 Anybody know what the two reasons are 539 00:29:21,081 --> 00:29:23,630 why no one can see clicks in the movies 540 00:29:23,630 --> 00:29:25,621 and you no longer go to the flicks anymore 541 00:29:25,621 --> 00:29:26,995 but you go to the movies instead? 542 00:29:28,680 --> 00:29:29,269 Yeah? 543 00:29:29,269 --> 00:29:30,310 What are the two reasons? 544 00:29:30,310 --> 00:29:31,940 Anybody know? 545 00:29:31,940 --> 00:29:33,350 OK, well, let me tell you. 546 00:29:33,350 --> 00:29:40,600 The first is that the rate of the number of frames per second 547 00:29:40,600 --> 00:29:45,310 had been increased to about 24, OK? 548 00:29:45,310 --> 00:29:50,551 But that's still slightly below the effusion. 549 00:29:50,551 --> 00:29:51,050 OK? 550 00:29:51,050 --> 00:29:52,800 You can still see some flicker. 551 00:29:52,800 --> 00:29:54,516 So how do you get rid of it? 552 00:29:54,516 --> 00:29:56,990 Well, let me tell you, some incredibly clever guy-- 553 00:29:56,990 --> 00:29:59,510 some things are really, really clever. 554 00:29:59,510 --> 00:30:01,930 I'll tell you what this person came up with. 555 00:30:01,930 --> 00:30:03,690 He said, all right, what we'll do 556 00:30:03,690 --> 00:30:08,390 is we're going to create a shutter which is round. 557 00:30:08,390 --> 00:30:10,420 And it's going to have one opening here. 558 00:30:11,630 --> 00:30:13,947 And then we're going to have a second opening here. 559 00:30:13,947 --> 00:30:16,030 And then we're going to have a third opening here. 560 00:30:16,030 --> 00:30:17,640 So this goes around. 561 00:30:17,640 --> 00:30:23,210 And each frame you're going to show three times. 562 00:30:23,210 --> 00:30:28,110 And then as this goes around, you go to the next frame, 563 00:30:28,110 --> 00:30:29,950 and you show it three times. 564 00:30:29,950 --> 00:30:35,590 So actually, if seeing frames 24 hertz, you see it 3 times 24, 565 00:30:35,590 --> 00:30:36,170 OK? 566 00:30:36,170 --> 00:30:38,800 But each frame is shown three times. 567 00:30:38,800 --> 00:30:42,230 Now, to be even more clever, the way they did this, 568 00:30:42,230 --> 00:30:46,940 they took two of these circular shutters. 569 00:30:46,940 --> 00:30:49,370 And one of them moved in this direction and the other 570 00:30:49,370 --> 00:30:50,720 in that direction. 571 00:30:50,720 --> 00:30:54,150 They sort of equalized from the center 572 00:30:54,150 --> 00:31:04,390 out as the light was expose from the projector. 573 00:31:04,390 --> 00:31:07,870 So that then reduced the rate of flicker. 574 00:31:07,870 --> 00:31:14,170 So therefore, we can next ask the question, 575 00:31:14,170 --> 00:31:19,240 what is the flicker rate at which you no longer can 576 00:31:19,240 --> 00:31:19,885 see anything? 577 00:31:21,480 --> 00:31:22,950 I mean, do you want to see flicker. 578 00:31:22,950 --> 00:31:25,820 Or do you not want to see flicker, and so on? 579 00:31:25,820 --> 00:31:28,620 And so experiments were done, hundreds 580 00:31:28,620 --> 00:31:33,640 of experiments in which a psychologist studied 581 00:31:33,640 --> 00:31:37,640 flicker rate, varying contrast, varying color, doing 582 00:31:37,640 --> 00:31:42,000 this, doing that, and then varying the spatial frequency 583 00:31:42,000 --> 00:31:45,590 and generating curves to see what they look like. 584 00:31:45,590 --> 00:31:47,160 So that's the process. 585 00:31:47,160 --> 00:31:52,450 And so what we can do next is to examine this. 586 00:31:52,450 --> 00:31:55,020 And one way to do this, for example, 587 00:31:55,020 --> 00:31:56,660 just to give you a sense, you can 588 00:31:56,660 --> 00:31:58,850 have, for example, a bunch of random dots, 589 00:31:58,850 --> 00:32:02,120 just like we had for the [INAUDIBLE] stereogram. 590 00:32:04,110 --> 00:32:06,790 And this would be done with a monkey, for example. 591 00:32:06,790 --> 00:32:08,210 The monkey fixates. 592 00:32:08,210 --> 00:32:12,470 And then you start something in motion-- ready?-- like that. 593 00:32:12,470 --> 00:32:16,240 And the monkey sees this and has to make a saccade to it. 594 00:32:16,240 --> 00:32:19,800 And when he does so, he gets a drop of apple juice for reward. 595 00:32:19,800 --> 00:32:21,910 And then what you can very are two things. 596 00:32:21,910 --> 00:32:23,330 You can vary the frequency. 597 00:32:23,330 --> 00:32:25,335 I'm showing another one, much higher frequency. 598 00:32:26,632 --> 00:32:28,590 Bang-- the monkey goes there and gets rewarded. 599 00:32:29,620 --> 00:32:33,590 Or you can-- and/or, actually-- you 600 00:32:33,590 --> 00:32:37,790 can vary the contrast of the display, OK? 601 00:32:37,790 --> 00:32:43,000 So once you have a sense of when you get fairly close to sort 602 00:32:43,000 --> 00:32:46,340 of the edge of things, then you can also vary the contrast. 603 00:32:46,340 --> 00:32:48,840 So now let's ask the question, what 604 00:32:48,840 --> 00:32:57,830 happens when you remove area MT and when you remove V4, OK? 605 00:32:57,830 --> 00:32:59,960 So think about that for a second. 606 00:32:59,960 --> 00:33:04,200 And then I will show you the actual data, all right? 607 00:33:07,770 --> 00:33:09,780 OK, here we go. 608 00:33:09,780 --> 00:33:10,880 Here is an example. 609 00:33:10,880 --> 00:33:12,550 This is motion detection. 610 00:33:12,550 --> 00:33:15,870 And here we vary percent luminance contrast 611 00:33:15,870 --> 00:33:17,650 at a fixed frequency. 612 00:33:17,650 --> 00:33:19,384 You get the same effect the other way. 613 00:33:19,384 --> 00:33:20,550 I'll just show you this one. 614 00:33:20,550 --> 00:33:23,920 And you can see very readily that the V4 lesion caused 615 00:33:23,920 --> 00:33:28,310 no deficit at all, but the MT lesion produced a huge deficit, 616 00:33:28,310 --> 00:33:29,690 OK? 617 00:33:29,690 --> 00:33:32,940 So clearly this establishes the fact 618 00:33:32,940 --> 00:33:37,060 that area MT plays a central role 619 00:33:37,060 --> 00:33:41,310 in the processing of flicker information 620 00:33:41,310 --> 00:33:45,590 in being able to analyze rapid temporal events. 621 00:33:45,590 --> 00:33:47,895 And now we can ask the next question. 622 00:33:49,710 --> 00:33:53,220 What happens when you actually do just stationary flicker, 623 00:33:53,220 --> 00:33:54,010 all right? 624 00:33:54,010 --> 00:33:56,150 So when you do stationary flicker, 625 00:33:56,150 --> 00:33:58,150 you can do that in two ways. 626 00:33:58,150 --> 00:34:01,700 You can do an on off flicker, or you 627 00:34:01,700 --> 00:34:07,650 can do a flicker between two colors, for example, all right? 628 00:34:07,650 --> 00:34:09,960 And you can see why that would be interesting. 629 00:34:09,960 --> 00:34:12,460 Because if you do flicker between two colors, 630 00:34:12,460 --> 00:34:17,610 like red and green, if area V4 were 631 00:34:17,610 --> 00:34:21,600 important for color processing, then a lesion in that area 632 00:34:21,600 --> 00:34:23,690 should mess up that kind of flicker. 633 00:34:23,690 --> 00:34:25,130 So that's the way to do it. 634 00:34:25,130 --> 00:34:28,790 So let me show you first of all how you do this experiment. 635 00:34:28,790 --> 00:34:29,860 Here we have it. 636 00:34:29,860 --> 00:34:31,679 You have a bunch of LED's. 637 00:34:31,679 --> 00:34:33,880 Why does one have LED's instead of just doing 638 00:34:33,880 --> 00:34:34,760 this on a monitor? 639 00:34:36,234 --> 00:34:36,775 Any thoughts? 640 00:34:39,440 --> 00:34:39,989 Yes. 641 00:34:39,989 --> 00:34:42,120 AUDIENCE: The frequency at which those [INAUDIBLE]. 642 00:34:42,120 --> 00:34:42,995 PROFESSOR: Very good. 643 00:34:42,995 --> 00:34:49,100 The reason you do it is because your typical frequency rate 644 00:34:49,100 --> 00:34:54,120 is the alternating current rate, meaning 60 Hertz. 645 00:34:54,120 --> 00:34:58,180 Now, nowadays you can of course buy 646 00:34:58,180 --> 00:35:02,580 devices that can double it or even quadruple it. 647 00:35:02,580 --> 00:35:07,070 But back then it was easier to do it this way. 648 00:35:07,070 --> 00:35:10,270 And it still is, because you can vary these flicker 649 00:35:10,270 --> 00:35:11,870 rates in small amounts. 650 00:35:11,870 --> 00:35:16,690 Now, each of these LED's, actually, 651 00:35:16,690 --> 00:35:18,910 for this particular set of experiments, 652 00:35:18,910 --> 00:35:25,170 is one that has both red and green in it, OK? 653 00:35:25,170 --> 00:35:29,510 So you can flicker between on and off, let's say for green. 654 00:35:29,510 --> 00:35:33,410 Or you can flicker between red and green, all right? 655 00:35:33,410 --> 00:35:35,890 So that's what we are going to find out. 656 00:35:35,890 --> 00:35:39,010 What happens when you do these kinds of flickers, red green 657 00:35:39,010 --> 00:35:42,600 as well as on and off? 658 00:35:42,600 --> 00:35:44,570 And ask the question, what happens 659 00:35:44,570 --> 00:35:49,360 when you make, again a V4 lesion or an MT lesion, OK? 660 00:35:50,550 --> 00:35:53,110 Now think about it for a minute, as 661 00:35:53,110 --> 00:35:55,320 to what you would expect in terms 662 00:35:55,320 --> 00:35:59,470 of the differences between on and off and red green flicker 663 00:35:59,470 --> 00:36:02,890 and possible differences in the lesions 664 00:36:02,890 --> 00:36:06,900 that cause a deficit in this, OK? 665 00:36:09,210 --> 00:36:10,570 All right, ready? 666 00:36:10,570 --> 00:36:11,286 Here we go. 667 00:36:11,286 --> 00:36:12,160 There is the flicker. 668 00:36:12,160 --> 00:36:13,010 You can see that. 669 00:36:13,010 --> 00:36:15,950 Now, the way that really works is that the overall flicker-- 670 00:36:15,950 --> 00:36:20,570 I'll show it to you again-- is done in such a fashion that 671 00:36:20,570 --> 00:36:26,920 the mean illumination level is the same so that you cannot 672 00:36:26,920 --> 00:36:32,830 tell the location of the flickering LED on the basis 673 00:36:32,830 --> 00:36:33,775 of a lumen exchange. 674 00:36:33,775 --> 00:36:37,480 You can only tell on the basis of the fact that it flickers 675 00:36:37,480 --> 00:36:39,910 or it doesn't flicker, OK? 676 00:36:39,910 --> 00:36:42,600 All right, so we are now ready for the experiment. 677 00:36:42,600 --> 00:36:43,960 And here it is. 678 00:36:43,960 --> 00:36:47,880 On the right, I'm showing you a green on and off flicker. 679 00:36:47,880 --> 00:36:51,050 And you can see that there's no deficit at the V4 lesion, 680 00:36:51,050 --> 00:36:57,710 but a gigantic, absolutely gigantic deficit at the MT 681 00:36:57,710 --> 00:36:59,450 lesion site. 682 00:36:59,450 --> 00:37:04,290 And then very surprisingly-- to some people at least who were 683 00:37:04,290 --> 00:37:07,430 convinced that V4 is a color area-- 684 00:37:07,430 --> 00:37:12,150 when you do that for a red green flicker, first of all, 685 00:37:12,150 --> 00:37:16,730 notice that the flicker rate that you can detect is much 686 00:37:16,730 --> 00:37:22,520 lower-- OK, this is the ratio of the flicker, much lower, OK?-- 687 00:37:22,520 --> 00:37:26,660 than it is for the on off flicker. 688 00:37:26,660 --> 00:37:30,790 And secondly, again, there's no significant deficit 689 00:37:30,790 --> 00:37:32,830 with the V4 lesion. 690 00:37:32,830 --> 00:37:35,325 But there's again a major deficit with MT lesion. 691 00:37:36,510 --> 00:37:41,610 So area MT is essential for being 692 00:37:41,610 --> 00:37:47,000 able to process both red and green flicker as well as on 693 00:37:47,000 --> 00:37:48,710 and off flicker, all right? 694 00:37:51,780 --> 00:37:55,330 So bear that in mind as we are going to now progress 695 00:37:55,330 --> 00:38:01,400 to the next stage of the experiment-- set of experiments 696 00:38:01,400 --> 00:38:02,300 I should say. 697 00:38:02,300 --> 00:38:04,380 And the next step I'm going to talk about 698 00:38:04,380 --> 00:38:05,630 is structure from motion. 699 00:38:06,830 --> 00:38:13,680 And that one is a very important attribute that we have. 700 00:38:13,680 --> 00:38:19,580 We can reconstruct structure by see nothing but motion. 701 00:38:19,580 --> 00:38:21,930 And I'm going to show you an example of that. 702 00:38:21,930 --> 00:38:26,770 OK, let me tell you, this here actually 703 00:38:26,770 --> 00:38:33,510 is a scene, a made-up scene, that's outside. 704 00:38:33,510 --> 00:38:38,615 And somewhere in here we have ladybug hiding. 705 00:38:39,810 --> 00:38:41,770 But the ladybug is not in motion. 706 00:38:41,770 --> 00:38:47,210 And you know, ladybugs have these spots on their back. 707 00:38:48,290 --> 00:38:49,910 And there's a ladybug here. 708 00:38:49,910 --> 00:38:52,780 And I'm going to tell you that there's a monkey watching this, 709 00:38:52,780 --> 00:38:56,510 hoping that he could catch the ladybug to have it, 710 00:38:56,510 --> 00:38:57,900 because he loves to eat ladybugs. 711 00:38:59,456 --> 00:39:00,580 Are you ready to catch her? 712 00:39:01,770 --> 00:39:02,700 OK, here we go. 713 00:39:02,700 --> 00:39:03,600 Ready? 714 00:39:03,600 --> 00:39:07,310 The ladybug after a while got tired of just sitting there 715 00:39:07,310 --> 00:39:09,510 and began to move. 716 00:39:09,510 --> 00:39:12,040 And that's a dead giveaway, because you 717 00:39:12,040 --> 00:39:14,100 have all these cells in your brain 718 00:39:14,100 --> 00:39:17,450 that selectivity see motion and flicker. 719 00:39:17,450 --> 00:39:18,590 Are you ready? 720 00:39:18,590 --> 00:39:19,200 Here we go. 721 00:39:21,140 --> 00:39:22,040 There's the ladybug. 722 00:39:22,040 --> 00:39:22,540 See her? 723 00:39:27,040 --> 00:39:31,330 OK, so the monkey is happily eating away the ladybug, 724 00:39:31,330 --> 00:39:32,360 all right? 725 00:39:32,360 --> 00:39:39,730 So here is a good example of how incredibly useful 726 00:39:39,730 --> 00:39:42,800 motion information is that you have 727 00:39:42,800 --> 00:39:44,330 in these visual cortical areas. 728 00:39:45,760 --> 00:39:52,209 Now, I would bet that if this monkey had had his MT removed, 729 00:39:52,209 --> 00:39:54,250 he wouldn't have been able to catch that ladybug. 730 00:39:55,310 --> 00:39:57,670 Now, another set of experiments carried out, 731 00:39:57,670 --> 00:40:00,450 beautiful experiments, were carried out 732 00:40:00,450 --> 00:40:04,300 by a fellow called Johanson-- I think 733 00:40:04,300 --> 00:40:09,660 that was in Sweden-- who did a very clever set of experiments. 734 00:40:09,660 --> 00:40:14,970 He wanted to analyze how well we can process motion information 735 00:40:14,970 --> 00:40:17,060 and derive structure from it. 736 00:40:17,060 --> 00:40:21,510 And what he did was that he put little teeny little light 737 00:40:21,510 --> 00:40:31,986 bulbs on about 13 points in the human's body. 738 00:40:31,986 --> 00:40:35,020 He put some here on the elbows, put some at the hand, 739 00:40:35,020 --> 00:40:39,090 put some at the head, and so on down, 13 of them. 740 00:40:39,090 --> 00:40:41,100 And then he turned off all the lights, 741 00:40:41,100 --> 00:40:43,920 so the only light was that little spot of light that 742 00:40:43,920 --> 00:40:48,770 was on these 13 spots of light that were on the person. 743 00:40:48,770 --> 00:40:53,300 And then he took a movie to see what 744 00:40:53,300 --> 00:40:56,260 happened when this person started to walk. 745 00:40:56,260 --> 00:40:58,370 And he discovered something amazing, 746 00:40:58,370 --> 00:41:03,690 that we are incredibly capable of reconstructing shape 747 00:41:03,690 --> 00:41:07,180 from minimal motion information. 748 00:41:07,180 --> 00:41:08,040 So are you ready? 749 00:41:09,050 --> 00:41:10,020 OK, here we go. 750 00:41:11,060 --> 00:41:11,970 So there it is. 751 00:41:11,970 --> 00:41:17,210 There are your 13 little spots of light. 752 00:41:17,210 --> 00:41:19,030 And now I'm going to set it in motion. 753 00:41:19,030 --> 00:41:19,530 Ready? 754 00:41:24,430 --> 00:41:25,550 So what do you see? 755 00:41:25,550 --> 00:41:30,060 You see a person walking on this very minimal information. 756 00:41:30,060 --> 00:41:34,140 And you can even tell-- who would say it's a male 757 00:41:34,140 --> 00:41:35,730 and who would say it's a female? 758 00:41:35,730 --> 00:41:37,890 Those of you who say it's a male, raise your hands. 759 00:41:39,810 --> 00:41:41,510 OK, it's virtually everybody. 760 00:41:41,510 --> 00:41:43,950 Right, very good. 761 00:41:43,950 --> 00:41:47,320 Now, so what he did which I don't have a movie of, 762 00:41:47,320 --> 00:41:50,220 which is quite remarkable, is that he 763 00:41:50,220 --> 00:41:53,140 took a couple, a male and a female 764 00:41:53,140 --> 00:41:56,580 and had them arranged in exactly the same way-- each 765 00:41:56,580 --> 00:42:02,685 had 13 of those bulbs-- and had them dance. 766 00:42:04,040 --> 00:42:06,990 And when you looked at that, you had no trouble 767 00:42:06,990 --> 00:42:09,770 telling which was the male and which was the female. 768 00:42:09,770 --> 00:42:10,730 Amazing. 769 00:42:10,730 --> 00:42:16,840 So we are able to reconstruct a tremendous amount of form 770 00:42:16,840 --> 00:42:20,520 detail on the basis of extremely limited information 771 00:42:20,520 --> 00:42:21,250 about motion. 772 00:42:23,660 --> 00:42:28,620 All right, so that then ends that session. 773 00:42:28,620 --> 00:42:32,690 And we're going to go next to what is called apparent motion, 774 00:42:32,690 --> 00:42:34,060 all right? 775 00:42:34,060 --> 00:42:36,934 So you're going to have various forms of apparent motion 776 00:42:36,934 --> 00:42:38,100 I'm going to tell you about. 777 00:42:38,100 --> 00:42:42,970 And the first one is that if you show these two dots, if I flick 778 00:42:42,970 --> 00:42:45,740 them back and forth, you have a very strong sense 779 00:42:45,740 --> 00:42:47,880 that the thing is going back and forth, right? 780 00:42:49,570 --> 00:42:51,990 You don't see one, two, going on and one going off, 781 00:42:51,990 --> 00:42:55,990 because of the right temporal succession. 782 00:42:55,990 --> 00:42:59,590 But you see them as just jumping back and forth. 783 00:42:59,590 --> 00:43:02,340 So now what we can do-- we can enlarge on that 784 00:43:02,340 --> 00:43:06,220 and carry out a series of experiments on the bases 785 00:43:06,220 --> 00:43:14,015 which can make some inferences as to whether areas V4 786 00:43:14,015 --> 00:43:16,750 or MT are contributing to this. 787 00:43:16,750 --> 00:43:18,590 So this is going be fun for you. 788 00:43:18,590 --> 00:43:20,180 And I want you to pay close attention. 789 00:43:21,230 --> 00:43:23,820 Now, in this kind of experiment when it was done, 790 00:43:23,820 --> 00:43:26,350 first, this stayed on all the time. 791 00:43:26,350 --> 00:43:28,480 And these two dots came on first. 792 00:43:28,480 --> 00:43:30,750 And then these two dots came on second. 793 00:43:30,750 --> 00:43:37,060 Now, if you do that, you can see either of two types of motion. 794 00:43:37,060 --> 00:43:40,940 You can either see what we call zigzag or seesaw. 795 00:43:40,940 --> 00:43:43,540 Seesaw is going to look like that, obviously, 796 00:43:43,540 --> 00:43:45,830 and zigzag like this, OK? 797 00:43:45,830 --> 00:43:47,840 So that's what I'm going to show you. 798 00:43:47,840 --> 00:43:52,320 And once I set this in motion, you 799 00:43:52,320 --> 00:43:54,630 will see it one way or the other. 800 00:43:54,630 --> 00:43:57,930 Then if you see it's, say, seesawing, 801 00:43:57,930 --> 00:44:00,480 you can take your finger and move it horizontally 802 00:44:00,480 --> 00:44:02,859 back and forth on top, say. 803 00:44:02,859 --> 00:44:05,150 And then in short order you're going to be seeing this. 804 00:44:05,150 --> 00:44:08,280 You can reverse the sense of motion, all right? 805 00:44:08,280 --> 00:44:09,700 So are you ready for that? 806 00:44:11,600 --> 00:44:13,250 OK, so here we go. 807 00:44:18,250 --> 00:44:20,270 If you see it one way or the other, 808 00:44:20,270 --> 00:44:23,830 try to move your finger back and forth to change it to see it. 809 00:44:23,830 --> 00:44:27,520 If you see it horizontally moving, move it vertically. 810 00:44:27,520 --> 00:44:29,710 If you see it vertically, move it horizontally. 811 00:44:29,710 --> 00:44:31,400 And then you can see it shift. 812 00:44:34,500 --> 00:44:36,190 Everybody able to reverse it? 813 00:44:37,546 --> 00:44:38,990 Well, keep doing it. 814 00:44:38,990 --> 00:44:41,466 And you should have no trouble reversing it 815 00:44:41,466 --> 00:44:43,090 if you move your finger back and forth. 816 00:44:47,930 --> 00:44:50,310 This is a bistable situation. 817 00:44:50,310 --> 00:44:53,320 And people can readily see the switch. 818 00:44:56,750 --> 00:44:58,325 OK, most of you seeing it switching? 819 00:44:59,750 --> 00:45:00,350 All right. 820 00:45:00,350 --> 00:45:02,795 So now we can ask a whole bunch of questions. 821 00:45:04,610 --> 00:45:07,200 And the way to put this question in general 822 00:45:07,200 --> 00:45:13,140 is we can ask, is first of all, is this something that 823 00:45:13,140 --> 00:45:16,390 happens locally, or does it happen everywhere 824 00:45:16,390 --> 00:45:17,630 when you're looking? 825 00:45:17,630 --> 00:45:19,210 So how can you do that? 826 00:45:19,210 --> 00:45:20,310 So let me show you. 827 00:45:20,310 --> 00:45:24,010 What they we do here instead is we can put up a bunch of them, 828 00:45:24,010 --> 00:45:24,510 OK? 829 00:45:24,510 --> 00:45:26,662 And they're quite far apart from each other. 830 00:45:26,662 --> 00:45:28,120 And depending on how close you are, 831 00:45:28,120 --> 00:45:30,110 they may be as far as a centimeter 832 00:45:30,110 --> 00:45:33,840 apart in the visual cortex for the activation. 833 00:45:33,840 --> 00:45:35,760 So now what we're going to do is we're 834 00:45:35,760 --> 00:45:39,119 going to move them back and forth just like I did before. 835 00:45:39,119 --> 00:45:40,660 And the question that you want to ask 836 00:45:40,660 --> 00:45:43,150 yourself-- am I going to see these all 837 00:45:43,150 --> 00:45:46,260 move the same way vertically, say, or see all of them 838 00:45:46,260 --> 00:45:47,950 move horizontally, OK? 839 00:45:49,060 --> 00:45:49,957 So here we go. 840 00:45:49,957 --> 00:45:50,540 Are you ready? 841 00:45:54,630 --> 00:45:59,170 Basically, the rule is that they all move the same way. 842 00:45:59,170 --> 00:46:00,360 And then they shift. 843 00:46:00,360 --> 00:46:02,710 Again, you can use your thumb or your forefinger. 844 00:46:03,740 --> 00:46:05,980 And if you manage to make it switch, 845 00:46:05,980 --> 00:46:07,710 they all will switch at the same time. 846 00:46:11,990 --> 00:46:13,450 OK, everybody manage that? 847 00:46:21,260 --> 00:46:27,200 All right, so now the rule is that even though the activation 848 00:46:27,200 --> 00:46:32,150 is very far apart in V1 from here to here, 849 00:46:32,150 --> 00:46:35,020 there's some organizing principle that 850 00:46:35,020 --> 00:46:39,150 makes you see everything a particular given way. 851 00:46:39,150 --> 00:46:40,220 And they shift. 852 00:46:40,220 --> 00:46:41,870 They all shift. 853 00:46:41,870 --> 00:46:45,590 Now, that's a very interesting arrangement. 854 00:46:45,590 --> 00:46:48,690 And you can wonder, what is there 855 00:46:48,690 --> 00:46:52,760 that can tell the visual cortex how to see it and insist 856 00:46:52,760 --> 00:46:54,260 that you got to see it the same way? 857 00:46:55,300 --> 00:46:57,120 Now we can move on. 858 00:46:57,120 --> 00:46:59,230 And you can ask the question-- so far, 859 00:46:59,230 --> 00:47:04,830 what I've shown you is they're all the same-- sorry. 860 00:47:04,830 --> 00:47:05,700 That's my fault. 861 00:47:07,340 --> 00:47:14,960 They are all the same size. 862 00:47:14,960 --> 00:47:19,130 And they are geometrically arranged, OK? 863 00:47:19,130 --> 00:47:23,310 So now what happens if instead, we do it differently 864 00:47:23,310 --> 00:47:27,300 and we make them different sizes and random locations? 865 00:47:27,300 --> 00:47:29,380 What happens under these conditions? 866 00:47:29,380 --> 00:47:30,769 Now think about it for a minute. 867 00:47:30,769 --> 00:47:32,310 And then I'm going to show it to you. 868 00:47:34,800 --> 00:47:36,970 And once again, even though they're different sizes, 869 00:47:36,970 --> 00:47:39,880 different locations, they all go the same way. 870 00:47:39,880 --> 00:47:43,380 And when they shift-- again, use your forefinger 871 00:47:43,380 --> 00:47:45,010 to try to make it shift. 872 00:47:45,010 --> 00:47:49,245 And when they shift, they all shift at the same time, OK? 873 00:47:53,190 --> 00:47:57,120 All right, so now we can move on and ask 874 00:47:57,120 --> 00:47:59,830 some interesting additional questions. 875 00:47:59,830 --> 00:48:02,370 We can ask the question, how important 876 00:48:02,370 --> 00:48:07,450 is color information for you to see it go one way or the other? 877 00:48:07,450 --> 00:48:08,160 All right? 878 00:48:08,160 --> 00:48:13,350 OK, so to do that, here we have the top two are colored red 879 00:48:13,350 --> 00:48:15,430 and the bottom two are colored green. 880 00:48:15,430 --> 00:48:17,460 Now we're going to move these back and forth. 881 00:48:18,530 --> 00:48:22,940 And the question is, will you see them move horizontally 882 00:48:22,940 --> 00:48:25,600 because of the color information or what? 883 00:48:25,600 --> 00:48:28,370 So the question is, again, are you 884 00:48:28,370 --> 00:48:31,200 going to see zigzag or seesaw? 885 00:48:31,200 --> 00:48:32,050 OK, are you ready? 886 00:48:35,310 --> 00:48:38,920 And you can readily see if, again, you can readily shift it 887 00:48:38,920 --> 00:48:44,210 again with your forefinger, that color doesn't seem 888 00:48:44,210 --> 00:48:50,740 to be an important determinant of the direction in which 889 00:48:50,740 --> 00:48:54,330 the apparent motion will take place. 890 00:48:54,330 --> 00:48:56,140 Everybody follow that? 891 00:48:56,140 --> 00:48:58,310 Well, now think about that. 892 00:48:58,310 --> 00:49:02,120 How can we do this experiment to be 893 00:49:02,120 --> 00:49:04,870 more convinced that color for some reason 894 00:49:04,870 --> 00:49:08,410 doesn't seem to be important in how we see apparent motion? 895 00:49:08,410 --> 00:49:10,280 Well, let me tell you how you do it. 896 00:49:10,280 --> 00:49:11,090 Here is an example. 897 00:49:12,300 --> 00:49:18,940 Here we have again four rows and four columns, all right? 898 00:49:18,940 --> 00:49:23,020 And the first, one A and C, the colors are horizontal. 899 00:49:23,020 --> 00:49:26,160 And in the second one, B and D, they are vertical. 900 00:49:27,370 --> 00:49:29,330 So if color plays a role, we should 901 00:49:29,330 --> 00:49:32,200 be able to see these, the first and the third rows 902 00:49:32,200 --> 00:49:34,180 go in one direction and the second 903 00:49:34,180 --> 00:49:36,055 and the fourth row go in the other direction. 904 00:49:37,240 --> 00:49:38,990 So let's examine that. 905 00:49:40,210 --> 00:49:41,540 Ready? 906 00:49:41,540 --> 00:49:44,940 And you can see that they're all going the same direction 907 00:49:44,940 --> 00:49:45,440 anyway. 908 00:49:45,440 --> 00:49:48,410 When they switch, they switch in the same direction, 909 00:49:48,410 --> 00:49:52,460 meaning that indeed you pay no attention to the color itself, 910 00:49:52,460 --> 00:49:53,130 OK? 911 00:49:53,130 --> 00:49:55,580 Some mechanism is involved in which 912 00:49:55,580 --> 00:49:58,670 the apparent motion you see does not 913 00:49:58,670 --> 00:50:00,880 play a significant role in color. 914 00:50:00,880 --> 00:50:05,000 Now we can do another experiment and see 915 00:50:05,000 --> 00:50:07,710 what happens when instead of colors, we 916 00:50:07,710 --> 00:50:11,050 use small shapes, X's and O's which 917 00:50:11,050 --> 00:50:13,270 are horizontal in the first and the third row 918 00:50:13,270 --> 00:50:16,940 and vertical in the second and the fourth row. 919 00:50:16,940 --> 00:50:23,460 Will this succeed in breaking up the unity of apparent motion? 920 00:50:23,460 --> 00:50:24,180 So here we go. 921 00:50:24,180 --> 00:50:25,500 Ready? 922 00:50:25,500 --> 00:50:28,680 And you still see them all the same way, 923 00:50:28,680 --> 00:50:30,180 meaning that you don't pay attention 924 00:50:30,180 --> 00:50:32,295 to the small differences in shape. 925 00:50:33,780 --> 00:50:38,510 OK, so now what I have shown you so far, 926 00:50:38,510 --> 00:50:43,220 remember, has something to do with the midget and the parasol 927 00:50:43,220 --> 00:50:45,130 systems for the test. 928 00:50:45,130 --> 00:50:48,140 The midget system can process fine detail. 929 00:50:48,140 --> 00:50:50,720 And the midget system can process color. 930 00:50:50,720 --> 00:50:54,000 The parasol system does not process fine detail 931 00:50:54,000 --> 00:50:56,540 and does not separately process color. 932 00:50:56,540 --> 00:50:59,590 So we'll come to that in a minute. 933 00:50:59,590 --> 00:51:01,030 Now let's ask the next question. 934 00:51:01,030 --> 00:51:03,750 What happens if you vary the size? 935 00:51:03,750 --> 00:51:06,880 OK, again, the first and third row, the size difference 936 00:51:06,880 --> 00:51:08,010 is horizontal. 937 00:51:08,010 --> 00:51:10,590 And the second and fourth row, it's vertical. 938 00:51:10,590 --> 00:51:12,540 OK, so let's see what happens. 939 00:51:12,540 --> 00:51:15,280 Here it's a bit unclear, OK? 940 00:51:15,280 --> 00:51:17,344 So as you see, some of them go one way. 941 00:51:17,344 --> 00:51:18,260 Some go the other way. 942 00:51:18,260 --> 00:51:20,220 And some of you may see it all go the same way, 943 00:51:20,220 --> 00:51:22,090 depending on how far you are. 944 00:51:22,090 --> 00:51:25,340 So let's change the ratio so it's much, much bigger. 945 00:51:25,340 --> 00:51:28,600 And if you do it now, what you see is dramatic. 946 00:51:28,600 --> 00:51:31,460 You see horizontal motion in the first and third rows 947 00:51:31,460 --> 00:51:34,950 and vertical motion in the second and fourth rows. 948 00:51:34,950 --> 00:51:35,830 Everybody see that? 949 00:51:36,850 --> 00:51:40,380 OK, so we can enhance on this maybe a little 950 00:51:40,380 --> 00:51:46,520 bit by adding these together, adding size and color 951 00:51:46,520 --> 00:51:52,505 and shape, quite large differences in size and shape 952 00:51:52,505 --> 00:51:53,430 and color. 953 00:51:53,430 --> 00:51:55,760 Then if you do this, you will indeed 954 00:51:55,760 --> 00:52:01,020 have fully succeeded in breaking up the unity of motion, 955 00:52:01,020 --> 00:52:03,411 with the first and third rows going horizontally 956 00:52:03,411 --> 00:52:05,410 and the second and fourth rows going vertically. 957 00:52:07,000 --> 00:52:09,680 Now then there's one other important factor 958 00:52:09,680 --> 00:52:11,740 that we need to consider about this. 959 00:52:11,740 --> 00:52:13,436 And that has to do with proximity. 960 00:52:14,830 --> 00:52:16,900 So in this case, what we're doing now-- 961 00:52:16,900 --> 00:52:18,930 in the first and the third rows, we 962 00:52:18,930 --> 00:52:22,540 have the horizontal dots closer to each other. 963 00:52:22,540 --> 00:52:26,200 he second and fourth rows, we have the vertical dots closer 964 00:52:26,200 --> 00:52:27,550 to each other. 965 00:52:27,550 --> 00:52:30,050 So if you do that, you're going to ask the question, 966 00:52:30,050 --> 00:52:31,780 how important is proximity? 967 00:52:31,780 --> 00:52:35,490 And if you do that, you don't see much so far. 968 00:52:35,490 --> 00:52:37,660 But now what I'm going to do is I'm 969 00:52:37,660 --> 00:52:40,340 going to make the difference even greater, OK? 970 00:52:40,340 --> 00:52:43,890 The first and third are very close horizontally. 971 00:52:43,890 --> 00:52:47,190 And the second and fourth are very close vertically. 972 00:52:47,190 --> 00:52:49,740 If you do that, you readily can see 973 00:52:49,740 --> 00:52:53,840 that indeed the first and third rows move horizontally 974 00:52:53,840 --> 00:52:57,870 and the second and fourth rows move vertically. 975 00:52:57,870 --> 00:53:00,860 So proximity is a very important factor. 976 00:53:00,860 --> 00:53:06,190 And so that then brings me to an interesting question 977 00:53:06,190 --> 00:53:09,610 that probably none of you have given much thought. 978 00:53:10,746 --> 00:53:13,610 But I will make you more aware of it. 979 00:53:15,090 --> 00:53:19,435 The fact is that when you see movies, 980 00:53:19,435 --> 00:53:28,267 you very often see a car or a chariot or something move 981 00:53:28,267 --> 00:53:30,100 forward, but the wheels are going backwards. 982 00:53:32,290 --> 00:53:34,640 And you certainly do that when you 983 00:53:34,640 --> 00:53:38,970 watch those endless ads on television about cars. 984 00:53:38,970 --> 00:53:41,770 And you see the wheels going backwards most of the time. 985 00:53:41,770 --> 00:53:43,490 And the car is going forward. 986 00:53:43,490 --> 00:53:45,670 And some of you say, well, why would I 987 00:53:45,670 --> 00:53:47,570 buy a car whose wheels are going backwards? 988 00:53:49,070 --> 00:53:53,200 So at any rate, let's examine that question. 989 00:53:53,200 --> 00:53:54,590 And what I'm going to do first is 990 00:53:54,590 --> 00:53:55,890 I'm going to show you a movie. 991 00:53:55,890 --> 00:53:56,700 Ready? 992 00:53:56,700 --> 00:53:58,290 This a famous movie, Ben-Hur. 993 00:53:58,290 --> 00:53:59,690 You've probably seen it. 994 00:53:59,690 --> 00:54:01,270 Look at the wheels going backwards. 995 00:54:01,270 --> 00:54:02,133 Can you see that? 996 00:54:05,590 --> 00:54:08,210 Now, of course since I have given this a lot of thought, 997 00:54:08,210 --> 00:54:11,600 it drives me nuts when I see wheels going backwards when 998 00:54:11,600 --> 00:54:13,780 the chariot or the car goes forward. 999 00:54:13,780 --> 00:54:15,610 All right, so now the question is, 1000 00:54:15,610 --> 00:54:19,760 why do we see wheels going backwards in the movies 1001 00:54:19,760 --> 00:54:26,620 or on the ads on television, whereas when you see it 1002 00:54:26,620 --> 00:54:29,240 in the real, you never see them going backwards? 1003 00:54:32,640 --> 00:54:42,330 Well, remember that when you look at a movie or television, 1004 00:54:42,330 --> 00:54:44,640 you always see apparent motion. 1005 00:54:44,640 --> 00:54:47,730 You never see real motion on television, 1006 00:54:47,730 --> 00:54:50,320 never see real motion in the movies. 1007 00:54:50,320 --> 00:54:52,352 You're not aware of it, but you don't. 1008 00:54:55,300 --> 00:54:57,490 You only see apparent motion because you 1009 00:54:57,490 --> 00:54:59,940 have successive frames in each of which there 1010 00:54:59,940 --> 00:55:03,890 is a specific, say, if it's wheels, rotation, OK? 1011 00:55:05,060 --> 00:55:07,000 So let's look at that analyze this. 1012 00:55:10,410 --> 00:55:14,040 So I'm going to show you a wheel that is rotating slowly. 1013 00:55:14,040 --> 00:55:15,000 Are you ready? 1014 00:55:15,000 --> 00:55:15,950 it's a very brief. 1015 00:55:18,170 --> 00:55:21,260 OK, you can see that wheel is going forward, right? 1016 00:55:21,260 --> 00:55:24,340 Now I'm going to show you the same and arrangement. 1017 00:55:24,340 --> 00:55:27,480 But I'm going to have it move fast. 1018 00:55:27,480 --> 00:55:27,980 Ready? 1019 00:55:30,230 --> 00:55:31,030 Ready, go. 1020 00:55:32,740 --> 00:55:34,520 That time you saw it go backwards, right? 1021 00:55:35,610 --> 00:55:37,670 OK, so let's analyze this. 1022 00:55:37,670 --> 00:55:39,800 What happens when you have slow motion? 1023 00:55:39,800 --> 00:55:43,760 Let's imagine that the red line is 1024 00:55:43,760 --> 00:55:47,075 designating one of the spokes. 1025 00:55:48,120 --> 00:55:51,270 Then if it moves slowly, the next frame 1026 00:55:51,270 --> 00:55:55,380 is going to show that displaced only a little bit from the red, 1027 00:55:55,380 --> 00:55:56,290 and so on. 1028 00:55:56,290 --> 00:55:59,550 And because of that, you're going to see forward motion. 1029 00:55:59,550 --> 00:56:03,390 But now if what you do instead is moving fast, what happens 1030 00:56:03,390 --> 00:56:08,190 is that this spoke here is going to be closer 1031 00:56:08,190 --> 00:56:09,970 to this one in the next frame. 1032 00:56:09,970 --> 00:56:11,595 And because of that-- of course, that's 1033 00:56:11,595 --> 00:56:13,100 all of them all the way around-- you 1034 00:56:13,100 --> 00:56:16,350 will see them go backwards, because of the principle 1035 00:56:16,350 --> 00:56:21,210 of proximity that I showed you with the apparent motion 1036 00:56:21,210 --> 00:56:25,840 of the four moving dots. 1037 00:56:27,020 --> 00:56:35,390 OK, so that then explains to you why you see wheels rotating 1038 00:56:35,390 --> 00:56:38,015 backwards in the movies and on television. 1039 00:56:39,210 --> 00:56:43,540 And it's due to a fact of the proximity 1040 00:56:43,540 --> 00:56:45,930 of successive elements, in this case, 1041 00:56:45,930 --> 00:56:53,710 spokes that are rotating as the picture is taken with a camera. 1042 00:56:55,080 --> 00:56:55,970 All right. 1043 00:56:55,970 --> 00:56:59,830 So now I'm going to come up and tell you briefly 1044 00:56:59,830 --> 00:57:02,870 about another very interesting phenomenon that 1045 00:57:02,870 --> 00:57:06,570 is closely related to motion. 1046 00:57:06,570 --> 00:57:13,990 When two successive stimuli are presented, 1047 00:57:13,990 --> 00:57:17,390 there can be an interaction between them, all right? 1048 00:57:17,390 --> 00:57:24,040 And in a famous series of experiments back in the 1930s, 1049 00:57:24,040 --> 00:57:26,555 they did what is called a metacontrast experiment. 1050 00:57:27,570 --> 00:57:31,030 First, a disc was presented like this. 1051 00:57:31,030 --> 00:57:32,950 And then it was followed by a ring. 1052 00:57:32,950 --> 00:57:36,180 And they shared the same contour, OK? 1053 00:57:36,180 --> 00:57:38,050 This is it, OK? 1054 00:57:38,050 --> 00:57:39,550 So now what you could do-- you could 1055 00:57:39,550 --> 00:57:42,850 vary the temporal interval between them. 1056 00:57:42,850 --> 00:57:44,756 So if you present these two simultaneously, 1057 00:57:44,756 --> 00:57:47,620 since they have overlapping contours, 1058 00:57:47,620 --> 00:57:49,450 that's all you would see. 1059 00:57:49,450 --> 00:57:52,720 But then if you present them sequentially-- ready? 1060 00:57:52,720 --> 00:57:55,340 Like that, then the question is, to what degree 1061 00:57:55,340 --> 00:57:57,780 did you see that first one? 1062 00:57:57,780 --> 00:57:59,160 So let's look at that. 1063 00:57:59,160 --> 00:58:01,970 And I want you to pay close attention 1064 00:58:01,970 --> 00:58:04,685 to when I present this to you, OK? 1065 00:58:04,685 --> 00:58:05,814 Are you ready? 1066 00:58:05,814 --> 00:58:07,730 You're going to do it now so that they're only 1067 00:58:07,730 --> 00:58:10,650 60 milliseconds apart, OK? 1068 00:58:12,450 --> 00:58:13,010 Here we go. 1069 00:58:15,600 --> 00:58:18,010 You barely see the first one, right? 1070 00:58:18,010 --> 00:58:19,540 You could just barely make it out. 1071 00:58:20,570 --> 00:58:22,090 But it's not gone. 1072 00:58:22,090 --> 00:58:25,090 Some people thought it would cause it to be gone. 1073 00:58:25,090 --> 00:58:26,830 But that's not the case. 1074 00:58:26,830 --> 00:58:29,060 So what you do then is the following. 1075 00:58:30,740 --> 00:58:33,050 If you fixate on the red cross, I'm 1076 00:58:33,050 --> 00:58:35,230 going to present the stimuli. 1077 00:58:35,230 --> 00:58:39,200 And I want you to tell me which of the two stimuli 1078 00:58:39,200 --> 00:58:41,920 I present is preceded by the disc. 1079 00:58:41,920 --> 00:58:42,420 Ready? 1080 00:58:43,620 --> 00:58:44,310 Could you tell? 1081 00:58:45,700 --> 00:58:47,120 Which one, left or right? 1082 00:58:48,380 --> 00:58:49,540 Left, good. 1083 00:58:49,540 --> 00:58:51,204 Again, repeat. 1084 00:58:51,204 --> 00:58:52,626 AUDIENCE: Right. 1085 00:58:52,626 --> 00:58:54,050 PROFESSOR: Oh. 1086 00:58:54,050 --> 00:58:55,330 Ready? 1087 00:58:55,330 --> 00:58:55,954 AUDIENCE: Left. 1088 00:58:58,762 --> 00:59:00,170 Left. 1089 00:59:00,170 --> 00:59:01,580 PROFESSOR: OK, very good. 1090 00:59:01,580 --> 00:59:05,840 So the fact is that when you present two side by side where 1091 00:59:05,840 --> 00:59:07,840 one of them is preceded by the disk, 1092 00:59:07,840 --> 00:59:09,740 you can tell which one came first. 1093 00:59:09,740 --> 00:59:12,680 So somehow this didn't obliterate your ability 1094 00:59:12,680 --> 00:59:13,880 to see it. 1095 00:59:13,880 --> 00:59:15,850 Now, the other interesting fact is 1096 00:59:15,850 --> 00:59:20,260 that if now you flicker back and forth between them, 1097 00:59:20,260 --> 00:59:23,630 at equal rates, you see both equally. 1098 00:59:24,950 --> 00:59:31,570 So therefore, what happened is you did not truly 1099 00:59:31,570 --> 00:59:32,910 obliterate the percept. 1100 00:59:32,910 --> 00:59:35,280 You just somehow weren't too much aware of it 1101 00:59:35,280 --> 00:59:39,270 because of the temporal offset between 1102 00:59:39,270 --> 00:59:40,970 the successive presentations. 1103 00:59:40,970 --> 00:59:43,360 Now, you can do this slightly differently. 1104 00:59:43,360 --> 00:59:49,960 Instead of having two stimuli that do not overlap but have 1105 00:59:49,960 --> 00:59:52,600 adjacent contours, we can overlap them. 1106 00:59:52,600 --> 00:59:54,790 So here I'm going to show you something and ask you 1107 00:59:54,790 --> 00:59:55,610 what you see. 1108 00:59:55,610 --> 00:59:57,010 Ready? 1109 00:59:57,010 --> 00:59:58,390 What do you see? 1110 00:59:58,390 --> 00:59:59,371 A big disk? 1111 00:59:59,371 --> 00:59:59,870 OK? 1112 01:00:01,510 --> 01:00:02,140 There it is. 1113 01:00:03,230 --> 01:00:06,750 Now I show it to you. 1114 01:00:06,750 --> 01:00:09,220 As to what I really showed you, was first 1115 01:00:09,220 --> 01:00:14,660 I presented a dim small disc followed by a brighter higher 1116 01:00:14,660 --> 01:00:17,820 contrast bigger disc. 1117 01:00:17,820 --> 01:00:21,970 And that obliterated the perception of the first one. 1118 01:00:21,970 --> 01:00:24,570 And that is truly obliterated. 1119 01:00:24,570 --> 01:00:27,830 And the interesting experiment you can do in this regard-- 1120 01:00:27,830 --> 01:00:29,130 you can do this interocularly. 1121 01:00:30,920 --> 01:00:35,080 In other words, you can present one of the stimuli 1122 01:00:35,080 --> 01:00:37,740 to the left eye and the other to the right eye. 1123 01:00:37,740 --> 01:00:41,040 If you do that, you don't get a masking effect. 1124 01:00:41,040 --> 01:00:45,130 But if you go back and do the metacontrast, 1125 01:00:45,130 --> 01:00:49,030 you still get the same pseudo blockage 1126 01:00:49,030 --> 01:00:51,130 without actually eliminating it, but sort 1127 01:00:51,130 --> 01:00:54,570 of making you less aware of the first stimulus. 1128 01:00:54,570 --> 01:00:58,360 OK, so now if one examines this, the way 1129 01:00:58,360 --> 01:01:00,620 to do the experiment right is you 1130 01:01:00,620 --> 01:01:05,300 vary the temporal offset between the two stimuli. 1131 01:01:05,300 --> 01:01:07,840 And when this was done systematically 1132 01:01:07,840 --> 01:01:11,580 with the brightness masking that I showed you later, 1133 01:01:11,580 --> 01:01:15,790 secondly, what happens is that the effect rapidly declines. 1134 01:01:15,790 --> 01:01:17,920 It's the biggest when the two are just 1135 01:01:17,920 --> 01:01:19,380 right next to each other. 1136 01:01:19,380 --> 01:01:22,450 By contrast, when you do the metacontrast, 1137 01:01:22,450 --> 01:01:27,030 the biggest effect is when they're about 62 to 100 1138 01:01:27,030 --> 01:01:29,790 milliseconds separated, because that's 1139 01:01:29,790 --> 01:01:33,350 when apparent motion is most effective. 1140 01:01:33,350 --> 01:01:39,430 And then if you do an experiment to see where this happens 1141 01:01:39,430 --> 01:01:43,280 for the brightness masking, here's 1142 01:01:43,280 --> 01:01:46,920 what happens in the retina recording from a cell. 1143 01:01:46,920 --> 01:01:49,800 This is a dim stimulus, a smaller response, 1144 01:01:49,800 --> 01:01:51,520 and with a longer latency. 1145 01:01:51,520 --> 01:01:54,180 And the mask alone will get a much shorter response 1146 01:01:54,180 --> 01:01:55,780 and a much more vigorous one. 1147 01:01:55,780 --> 01:02:02,130 And then when you systematically vary the temporal interval 1148 01:02:02,130 --> 01:02:03,940 between the two of them, you can see 1149 01:02:03,940 --> 01:02:05,380 when the interval becomes short. 1150 01:02:06,460 --> 01:02:08,890 The second stimulus obliterates the first one 1151 01:02:08,890 --> 01:02:11,960 because it's faster and it has a bigger response. 1152 01:02:11,960 --> 01:02:14,370 So this explains what happens, because it happens 1153 01:02:14,370 --> 01:02:17,220 in the retina by virtue of transmission. 1154 01:02:17,220 --> 01:02:20,150 And the rule of the fact that the rate at which information 1155 01:02:20,150 --> 01:02:24,390 is transmitted from the four receptors 1156 01:02:24,390 --> 01:02:28,635 to the retinal ganglion cells is contrast dependent. 1157 01:02:28,635 --> 01:02:32,140 The higher the contrast, the faster the response. 1158 01:02:32,140 --> 01:02:35,780 All right, so now does anybody have 1159 01:02:35,780 --> 01:02:39,720 any questions about the masking or the metacontrast 1160 01:02:39,720 --> 01:02:41,130 that we talked about so far? 1161 01:02:42,850 --> 01:02:44,980 All right, in that case, we are now 1162 01:02:44,980 --> 01:02:51,960 going to move on and talk about another important motion 1163 01:02:51,960 --> 01:02:56,720 perception effect which links one to eye movement generation. 1164 01:02:56,720 --> 01:02:58,460 And that falls into the category of 1165 01:02:58,460 --> 01:02:59,880 so-called optokinetic nystagmus. 1166 01:03:01,040 --> 01:03:04,050 So how do you study optokinetic nystagmus 1167 01:03:04,050 --> 01:03:05,420 at the very basic level? 1168 01:03:05,420 --> 01:03:10,350 Well, the way you study it is, for example, in the olden days, 1169 01:03:10,350 --> 01:03:14,590 you could take a large lamp shade like that 1170 01:03:14,590 --> 01:03:16,930 and put vertical bars on int on the inside, 1171 01:03:16,930 --> 01:03:19,040 have a person have his head inside. 1172 01:03:19,040 --> 01:03:23,050 And then you rotate the lamp shape 1173 01:03:23,050 --> 01:03:25,450 so that the bars go across like that. 1174 01:03:25,450 --> 01:03:27,250 Or you can do that more simply nowadays. 1175 01:03:27,250 --> 01:03:32,200 You just go put a bunch of vertical bars on a monitor 1176 01:03:32,200 --> 01:03:33,250 and have them drift. 1177 01:03:33,250 --> 01:03:35,750 And the person is asked to look at it. 1178 01:03:35,750 --> 01:03:38,710 So when the person does that, what you get 1179 01:03:38,710 --> 01:03:41,480 is the so-called optokinetic nystagmus. 1180 01:03:41,480 --> 01:03:43,000 It has two phases. 1181 01:03:43,000 --> 01:03:45,220 It has a fast and a slow phase. 1182 01:03:45,220 --> 01:03:49,400 The fast phase is a saccade, a resetting saccade, if you will. 1183 01:03:49,400 --> 01:03:50,710 So you watch an edge. 1184 01:03:50,710 --> 01:03:52,370 You follow it. 1185 01:03:52,370 --> 01:03:54,360 And then once it gets to the periphery, 1186 01:03:54,360 --> 01:03:57,100 you make a saccade back and you watch the next one. 1187 01:03:57,100 --> 01:03:58,580 And you keep doing it. 1188 01:03:58,580 --> 01:04:00,600 And this shows you the three different rates. 1189 01:04:01,670 --> 01:04:03,590 Now, this is quite an interesting mechanism. 1190 01:04:04,750 --> 01:04:11,150 And it was determined that an important contribution 1191 01:04:11,150 --> 01:04:15,040 to executing these kinds of eye movements, which 1192 01:04:15,040 --> 01:04:17,660 is very important, because when you track anything, if you 1193 01:04:17,660 --> 01:04:21,210 track a bird of something, you keep your eye on it. 1194 01:04:21,210 --> 01:04:24,230 And then once it gets past you, make this a saccade back 1195 01:04:24,230 --> 01:04:26,670 and take the next bird and track it. 1196 01:04:26,670 --> 01:04:28,300 So that's the mechanism. 1197 01:04:28,300 --> 01:04:31,840 And it's been shown that several neural mechanisms play 1198 01:04:31,840 --> 01:04:32,700 a role in it. 1199 01:04:32,700 --> 01:04:41,520 And one of those belongs to the so-called accessory optic 1200 01:04:41,520 --> 01:04:46,180 system, which has those cells of Dogiel 1201 01:04:46,180 --> 01:04:48,869 in the retina that connect to it. 1202 01:04:48,869 --> 01:04:50,160 So we're going to look at that. 1203 01:04:50,160 --> 01:04:53,000 Some beautiful work has been done about that. 1204 01:04:53,000 --> 01:04:55,790 But before I do that, let me show you 1205 01:04:55,790 --> 01:04:56,880 a very interesting factor. 1206 01:04:58,260 --> 01:05:03,440 Here is a typical optokinetic nystagmus 1207 01:05:03,440 --> 01:05:07,850 induced by a drifting vertical set of bars at various rates, 1208 01:05:07,850 --> 01:05:09,370 as you can see here. 1209 01:05:09,370 --> 01:05:11,190 And this is a rabbit. 1210 01:05:11,190 --> 01:05:15,860 And in the rabbit, what you have is a left eye and a right eye. 1211 01:05:15,860 --> 01:05:20,690 And you can activate them separately, as shown here. 1212 01:05:20,690 --> 01:05:25,570 Now, a remarkable discovery that was made 1213 01:05:25,570 --> 01:05:29,680 is that the accessory optic system-- that there's 1214 01:05:29,680 --> 01:05:34,460 a primary role in generating this optokinetic 1215 01:05:34,460 --> 01:05:39,360 nystagmus that is found initially 1216 01:05:39,360 --> 01:05:43,450 in among the cells of Dogiel, of which in the rabbit, as I've 1217 01:05:43,450 --> 01:05:48,760 said, there are about 7,000 out of 350,000 in the eye. 1218 01:05:48,760 --> 01:05:55,280 And it was found-- which was incredibly surprising-- 1219 01:05:55,280 --> 01:06:01,550 that this system originating with the cells of Dogiel, 1220 01:06:01,550 --> 01:06:08,470 the cells of Dogiel themselves get input only from ON bipolar 1221 01:06:08,470 --> 01:06:10,780 cells, OK? 1222 01:06:10,780 --> 01:06:17,410 So what can we do to prove that this system here indeed 1223 01:06:17,410 --> 01:06:21,620 is one that is activated by the cells of Dogiel, 1224 01:06:21,620 --> 01:06:24,730 since the cells of Dogiel in the rabbit 1225 01:06:24,730 --> 01:06:30,102 get input only from the ON bipolar cells? 1226 01:06:30,102 --> 01:06:31,560 Anybody can think of an experiment? 1227 01:06:33,760 --> 01:06:37,150 What did I tell you about the on and off systems 1228 01:06:37,150 --> 01:06:39,240 and how they can be manipulated? 1229 01:06:39,240 --> 01:06:39,965 Anybody remember? 1230 01:06:41,762 --> 01:06:43,345 Now, see you've got to remember facts. 1231 01:06:44,400 --> 01:06:48,090 And one of the facts was I told you about what happens when you 1232 01:06:48,090 --> 01:06:53,080 use 2-amino-4-phosphonobutyric acid, APB. 1233 01:06:53,080 --> 01:06:54,210 Anybody remember that? 1234 01:06:55,695 --> 01:06:58,143 If not, you'd better study it, because there 1235 01:06:58,143 --> 01:07:00,880 are going to be questions on that on the exam, of course. 1236 01:07:00,880 --> 01:07:03,300 OK, so if you put APB into the eye, 1237 01:07:03,300 --> 01:07:08,230 you block the ON bipolar cells. 1238 01:07:08,230 --> 01:07:11,890 And therefore, if only ON bipolar cells 1239 01:07:11,890 --> 01:07:14,480 feed into the cells of Dogiel, you 1240 01:07:14,480 --> 01:07:17,470 would expect that if you block those bipolar cells, 1241 01:07:17,470 --> 01:07:24,000 you should eliminate the optokinetic nystagmus 1242 01:07:24,000 --> 01:07:28,960 in the eye that had been injected, if it's indeed true 1243 01:07:28,960 --> 01:07:34,090 that the cells of Dogiel are involved, OK? 1244 01:07:34,090 --> 01:07:35,850 So that's a beautiful experiment, 1245 01:07:35,850 --> 01:07:39,320 because it's a definite one if indeed the data come out 1246 01:07:39,320 --> 01:07:40,740 the way one had hoped. 1247 01:07:40,740 --> 01:07:42,310 And guess what happens? 1248 01:07:42,310 --> 01:07:43,500 Here are the data. 1249 01:07:44,840 --> 01:07:49,080 The red ones are from the eye that 1250 01:07:49,080 --> 01:07:50,970 had been injected with APB. 1251 01:07:50,970 --> 01:07:54,130 You know, you blocked the ON cells. 1252 01:07:54,130 --> 01:07:57,080 And hence, you blocked the responses 1253 01:07:57,080 --> 01:08:02,330 of the cells of Dogiel, all right? 1254 01:08:08,150 --> 01:08:13,593 So this is a rather nice, luckily definitive, experiment. 1255 01:08:13,593 --> 01:08:17,370 Now, the curious thing is, if you did the same experiment 1256 01:08:17,370 --> 01:08:21,602 in a monkey, you would get nothing. 1257 01:08:21,602 --> 01:08:24,460 Here you do the same experiment in a monkey. 1258 01:08:24,460 --> 01:08:26,270 Here is your OKN. 1259 01:08:26,270 --> 01:08:28,359 And here is when you put in APB. 1260 01:08:28,359 --> 01:08:31,250 And still after putting in APB, the monkey 1261 01:08:31,250 --> 01:08:33,620 still does optokinetic nystagmus. 1262 01:08:35,100 --> 01:08:41,600 That's because the rabbit is a very unusual species. 1263 01:08:41,600 --> 01:08:43,870 And god only knows why it happened 1264 01:08:43,870 --> 01:08:47,939 that only On bipolar cells fed into the cells of Dogiel. 1265 01:08:47,939 --> 01:08:52,800 In the monkey, you get both ON and OFF cells 1266 01:08:52,800 --> 01:08:54,479 feeding into the cells of Dogiel. 1267 01:08:56,340 --> 01:09:00,625 All right, so now another interesting factor 1268 01:09:00,625 --> 01:09:03,069 that I want to mention just briefly 1269 01:09:03,069 --> 01:09:08,520 is that it's indeed is true that the feedback mechanism is 1270 01:09:08,520 --> 01:09:12,430 essential for you to do optokinetic nystagmus right. 1271 01:09:12,430 --> 01:09:16,020 Here is a monkey in which one eye has been immobilized 1272 01:09:16,020 --> 01:09:17,290 and the other eye is normal. 1273 01:09:17,290 --> 01:09:21,189 So if you look at the normal eye and you 1274 01:09:21,189 --> 01:09:25,680 move a set of vertical bars very slowly, 1275 01:09:25,680 --> 01:09:28,770 this is what your optokinetic nystagmus 1276 01:09:28,770 --> 01:09:32,370 looks like-- just a very slow tracking. 1277 01:09:33,390 --> 01:09:37,649 But then if you do the same thing in the mobilized eye, 1278 01:09:37,649 --> 01:09:42,010 and you, of course, record the eye movement of the other eye 1279 01:09:42,010 --> 01:09:45,439 which is blocked, what you get is a tremendously rapid 1280 01:09:45,439 --> 01:09:48,799 increase in the pursuit movement. 1281 01:09:50,670 --> 01:09:53,050 This can be called open loop. 1282 01:09:53,050 --> 01:09:55,640 And this can be called closed loop. 1283 01:09:55,640 --> 01:09:57,870 Now, the mechanism involved here is 1284 01:09:57,870 --> 01:10:00,880 that your intent is to track. 1285 01:10:00,880 --> 01:10:03,850 But when you realize-- or not you realize, 1286 01:10:03,850 --> 01:10:05,180 but the brain realizes. 1287 01:10:05,180 --> 01:10:08,820 You're not even aware of it-- that you are not 1288 01:10:08,820 --> 01:10:11,800 following fast enough to keep up with the motion 1289 01:10:11,800 --> 01:10:15,530 of the vertical bar, you order the eyes 1290 01:10:15,530 --> 01:10:17,580 to move faster, faster, and faster, 1291 01:10:17,580 --> 01:10:19,610 and faster and faster, and it practically 1292 01:10:19,610 --> 01:10:21,337 runs away like crazy. 1293 01:10:21,337 --> 01:10:23,170 And it still-- of course, because the eye is 1294 01:10:23,170 --> 01:10:28,470 immobilized-- hasn't been able to catch up to the moving bar. 1295 01:10:28,470 --> 01:10:33,340 So this further highlights the importance of this mechanism. 1296 01:10:33,340 --> 01:10:36,740 And it has to be a closed loop system, obviously, 1297 01:10:36,740 --> 01:10:38,130 for it to work properly. 1298 01:10:38,130 --> 01:10:40,090 So there's a feedback telling you 1299 01:10:40,090 --> 01:10:42,370 whether you had correctly tracked something. 1300 01:10:43,600 --> 01:10:48,880 OK, so now I come to the last point today, 1301 01:10:48,880 --> 01:10:54,450 which is now the mechanism that is involved in the motion 1302 01:10:54,450 --> 01:10:56,500 analysis of accessory optic system. 1303 01:10:56,500 --> 01:10:58,380 I'll talk about this only briefly, 1304 01:10:58,380 --> 01:11:02,950 because this is going to be part of your paper. 1305 01:11:02,950 --> 01:11:05,570 All right, now, they remarkable discovery 1306 01:11:05,570 --> 01:11:07,950 that was made-- amazing discovery-- 1307 01:11:07,950 --> 01:11:12,750 was that when people recorded from the cells of Dogiel, 1308 01:11:12,750 --> 01:11:21,270 they found that these cells come in three different types 1309 01:11:21,270 --> 01:11:25,550 with three different axes of orientation, these three axes. 1310 01:11:25,550 --> 01:11:27,900 And when that was first found, people 1311 01:11:27,900 --> 01:11:29,540 said, what on earth is going on? 1312 01:11:29,540 --> 01:11:32,100 Why would we have three like that? 1313 01:11:32,100 --> 01:11:34,040 I would have thought we'd have four, 1314 01:11:34,040 --> 01:11:36,470 vertical and horizontal at least, or something. 1315 01:11:36,470 --> 01:11:38,130 So this was really puzzling. 1316 01:11:38,130 --> 01:11:40,710 But then they began to do some more experiments. 1317 01:11:40,710 --> 01:11:43,080 And they made a truly remarkable discovery. 1318 01:11:43,080 --> 01:11:45,530 And their discovery is shown here 1319 01:11:45,530 --> 01:11:49,600 in a summary diagram of the accessory optic system. 1320 01:11:49,600 --> 01:11:53,010 OK, here are these cells of Dogiel 1321 01:11:53,010 --> 01:11:55,890 with the three axes of direction selectivity. 1322 01:11:55,890 --> 01:12:01,105 They respond to rather slow motion velocities. 1323 01:12:02,300 --> 01:12:04,012 So what happens here-- I'm not going 1324 01:12:04,012 --> 01:12:06,220 to go into detail, because you're going to, as I say, 1325 01:12:06,220 --> 01:12:07,390 report on this. 1326 01:12:07,390 --> 01:12:10,320 But let me tell you about the essence of it. 1327 01:12:10,320 --> 01:12:16,320 The cells of Dogiel project into the so-called terminal nuclei. 1328 01:12:16,320 --> 01:12:19,030 And then they project up to the cerebellum. 1329 01:12:19,030 --> 01:12:21,414 And from the cerebellum, there's a projection down 1330 01:12:21,414 --> 01:12:22,455 to the vestibular nuclei. 1331 01:12:23,900 --> 01:12:27,710 And then it was found, as it was known already, 1332 01:12:27,710 --> 01:12:30,690 that the semicircular canals come 1333 01:12:30,690 --> 01:12:34,610 in three orientations which are the same as those 1334 01:12:34,610 --> 01:12:39,630 of the orientations of the cells of Dogiel. 1335 01:12:39,630 --> 01:12:45,822 And so then they feed into the ocular motor system 1336 01:12:45,822 --> 01:12:53,850 to move the eyes so as to keep the eyes in accurate 1337 01:12:53,850 --> 01:12:58,530 contact with whatever you're intending to look at. 1338 01:12:58,530 --> 01:13:02,960 So that is the very essence of the accessory optic system. 1339 01:13:02,960 --> 01:13:04,850 And I just thought this will help you 1340 01:13:04,850 --> 01:13:08,090 in putting your paper together. 1341 01:13:08,090 --> 01:13:10,990 All right, so now I am ready to summarize for you. 1342 01:13:10,990 --> 01:13:14,760 First of all, motion has been classified 1343 01:13:14,760 --> 01:13:20,655 into several different types that include planar, circular, 1344 01:13:20,655 --> 01:13:21,155 and radial. 1345 01:13:23,000 --> 01:13:26,310 And one thing we talked about earlier that I mentioned here, 1346 01:13:26,310 --> 01:13:28,640 but I didn't mention in this presentation-- 1347 01:13:28,640 --> 01:13:30,470 when we talked about depth perception, 1348 01:13:30,470 --> 01:13:32,460 we talked about motion parallax. 1349 01:13:32,460 --> 01:13:35,005 That's a very important aspect of motion analysis. 1350 01:13:36,200 --> 01:13:39,160 Then secondly, the majority of cells in V1 1351 01:13:39,160 --> 01:13:43,650 and most cells in MT are directionally selective, OK? 1352 01:13:43,650 --> 01:13:46,440 And an addition which I mentioned earlier 1353 01:13:46,440 --> 01:13:49,410 is that they also show velocity selectivity. 1354 01:13:51,050 --> 01:13:54,180 Some V1 cells respond to different directions 1355 01:13:54,180 --> 01:13:57,020 of movement for light and dark edges, like the S2 type cell 1356 01:13:57,020 --> 01:13:57,870 I showed you. 1357 01:13:57,870 --> 01:14:01,270 And some cells are sensitive to differential velocities 1358 01:14:01,270 --> 01:14:03,570 of movement, which I didn't show. 1359 01:14:03,570 --> 01:14:06,190 Then the accessory optic system that 1360 01:14:06,190 --> 01:14:11,200 begins with the retinal ganglion cells of the cells of Dogiel 1361 01:14:11,200 --> 01:14:15,370 form three axes of direction selectivity 1362 01:14:15,370 --> 01:14:17,120 that correspond to the three axes 1363 01:14:17,120 --> 01:14:19,000 of the semicircular canals. 1364 01:14:19,000 --> 01:14:22,090 And they are involved in generating pursuit eye 1365 01:14:22,090 --> 01:14:24,180 movements for image stabilization. 1366 01:14:25,560 --> 01:14:28,770 Then one of the most important tasks of motion analysis 1367 01:14:28,770 --> 01:14:31,470 is motion parallax, as it provides 1368 01:14:31,470 --> 01:14:32,620 information about depth. 1369 01:14:33,630 --> 01:14:36,005 Motion cues can provide important information 1370 01:14:36,005 --> 01:14:38,820 for object recognition, often referred 1371 01:14:38,820 --> 01:14:41,470 to as structure for motion. 1372 01:14:41,470 --> 01:14:43,800 And stationary stimuli that flicker 1373 01:14:43,800 --> 01:14:47,440 with various temporary signals induce apparent motion, 1374 01:14:47,440 --> 01:14:49,580 of which I showed you a lot of examples. 1375 01:14:49,580 --> 01:14:52,750 And then we talked about metacontrast masking, 1376 01:14:52,750 --> 01:14:54,970 which is a phenomenon that occurs 1377 01:14:54,970 --> 01:14:59,280 when two stimuli share contours. 1378 01:14:59,280 --> 01:15:03,300 And when they're presented with an interval which corresponds 1379 01:15:03,300 --> 01:15:06,550 to apparent motion, you tend not to see the first one. 1380 01:15:07,640 --> 01:15:11,540 Lastly, brightness masking arises with overlapping stimuli 1381 01:15:11,540 --> 01:15:13,340 appearing in rapid succession. 1382 01:15:13,340 --> 01:15:17,020 And this does not occur with interocular presentation, 1383 01:15:17,020 --> 01:15:19,830 proving that it's a record of phenomenon. 1384 01:15:19,830 --> 01:15:24,840 And it has to do with the fact that the velocity with which 1385 01:15:24,840 --> 01:15:31,950 a stimulus activates the retinal ganglion cells and hence, 1386 01:15:31,950 --> 01:15:35,980 of course, the cortex, is much much slower with dim stimuli 1387 01:15:35,980 --> 01:15:37,160 then with bright ones. 1388 01:15:37,160 --> 01:15:39,610 And so bright stimulus, in a sense 1389 01:15:39,610 --> 01:15:44,270 overtakes the responses to a dim and weak stimulus 1390 01:15:44,270 --> 01:15:47,900 when the two successive presentations are very rapid. 1391 01:15:47,900 --> 01:15:50,600 OK, so that's the end of it here, OK? 1392 01:15:50,600 --> 01:15:53,009 And I'll show you one more motion effect. 1393 01:15:53,009 --> 01:15:53,508 Ready? 1394 01:15:56,690 --> 01:16:00,310 All right, so that is the story for today. 1395 01:16:00,310 --> 01:16:03,260 And next time, you're going to have, 1396 01:16:03,260 --> 01:16:07,040 as I said, a rapid summary of what 1397 01:16:07,040 --> 01:16:10,360 we have covered in the first half of this course. 1398 01:16:10,360 --> 01:16:14,590 And then the following Wednesday, we 1399 01:16:14,590 --> 01:16:18,495 are going to have the multiple choice exam for the midterms. 1400 01:16:20,240 --> 01:16:24,540 OK, thank you kindly.