1 00:00:00,060 --> 00:00:02,400 The following content is provided under a Creative 2 00:00:02,400 --> 00:00:03,820 Commons license. 3 00:00:03,820 --> 00:00:06,030 Your support will help MIT OpenCourseWare 4 00:00:06,030 --> 00:00:10,120 continue to offer high quality educational resources for free. 5 00:00:10,120 --> 00:00:12,660 To make a donation or to view additional materials 6 00:00:12,660 --> 00:00:16,620 from hundreds of MIT courses, visit MIT OpenCourseWare 7 00:00:16,620 --> 00:00:17,926 at ocw.mit.edu. 8 00:00:20,700 --> 00:00:22,170 ABBY NOYCE: So yesterday, we talked 9 00:00:22,170 --> 00:00:27,280 about theories and models of attention. 10 00:00:27,280 --> 00:00:30,090 Today, we're going to get down into the nitty gritty 11 00:00:30,090 --> 00:00:33,090 a little bit more and actually talk about some 12 00:00:33,090 --> 00:00:36,900 of the neuroscience that underlies it. 13 00:00:36,900 --> 00:00:39,900 So what were some of the attention models 14 00:00:39,900 --> 00:00:41,419 we talked about yesterday? 15 00:00:44,413 --> 00:00:45,411 AUDIENCE: Spotlight. 16 00:00:45,411 --> 00:00:48,310 ABBY NOYCE: The spotlight one is one of the older models, 17 00:00:48,310 --> 00:00:51,612 that attention focuses on a particular location. 18 00:00:51,612 --> 00:00:54,070 We talked about some of the evidence showing that attention 19 00:00:54,070 --> 00:00:57,400 seems to be more object-focused than space-focused, 20 00:00:57,400 --> 00:00:59,390 which is one of the big flaws in that theory. 21 00:00:59,390 --> 00:01:01,292 What's another one? 22 00:01:01,292 --> 00:01:02,822 AUDIENCE: The feature integration? 23 00:01:02,822 --> 00:01:03,530 ABBY NOYCE: Yeah. 24 00:01:03,530 --> 00:01:06,516 What's the feature integration one? 25 00:01:06,516 --> 00:01:08,679 AUDIENCE: Wasn't it like each feature in the brain 26 00:01:08,679 --> 00:01:09,484 has a certain map? 27 00:01:09,484 --> 00:01:11,275 And when you're searching for that feature, 28 00:01:11,275 --> 00:01:13,817 that map is pulled up when you're looking for it. 29 00:01:13,817 --> 00:01:15,441 And then when there's two new features, 30 00:01:15,441 --> 00:01:18,340 you line the maps next to each other? 31 00:01:18,340 --> 00:01:19,286 ABBY NOYCE: Yeah. 32 00:01:19,286 --> 00:01:20,596 Good. 33 00:01:20,596 --> 00:01:21,970 The feature integration theory is 34 00:01:21,970 --> 00:01:24,386 trying to explain the difference between those disjunctive 35 00:01:24,386 --> 00:01:26,470 single-feature different searches, where you're 36 00:01:26,470 --> 00:01:28,690 looking for the black circle among white circles 37 00:01:28,690 --> 00:01:31,240 and the conjunctive, multiple-feature black circle 38 00:01:31,240 --> 00:01:33,670 among white circles and black squares 39 00:01:33,670 --> 00:01:36,320 and the difference between those tasks. 40 00:01:36,320 --> 00:01:38,712 What's another one? 41 00:01:38,712 --> 00:01:40,630 AUDIENCE: Did someone mention spotlight? 42 00:01:40,630 --> 00:01:41,500 ABBY NOYCE: Spotlight we got. 43 00:01:41,500 --> 00:01:42,392 AUDIENCE: Thank you. 44 00:01:42,392 --> 00:01:43,730 Early selection. 45 00:01:43,730 --> 00:01:46,156 ABBY NOYCE: So early selection, which is kind of from way 46 00:01:46,156 --> 00:01:47,530 back in the day when people first 47 00:01:47,530 --> 00:01:48,620 started thinking about this. 48 00:01:48,620 --> 00:01:51,161 It says that attention happens really early in the processing 49 00:01:51,161 --> 00:01:52,480 stream. 50 00:01:52,480 --> 00:01:55,720 And the counterexample to that is the cocktail party effect, 51 00:01:55,720 --> 00:01:58,080 where someone says your name from across the room 52 00:01:58,080 --> 00:01:58,340 and you go, "Huh? 53 00:01:58,340 --> 00:01:58,840 What? 54 00:01:58,840 --> 00:02:00,464 They're talking about me," even if you 55 00:02:00,464 --> 00:02:03,130 hadn't been paying any attention to the conversation previously. 56 00:02:03,130 --> 00:02:04,546 So you're switching your attention 57 00:02:04,546 --> 00:02:06,340 based on the semantic content of that, 58 00:02:06,340 --> 00:02:08,580 not just from a low level feature. 59 00:02:12,280 --> 00:02:12,780 Good. 60 00:02:12,780 --> 00:02:13,680 OK. 61 00:02:13,680 --> 00:02:17,910 So today, we're going to start with looking 62 00:02:17,910 --> 00:02:20,197 at what do we know. 63 00:02:20,197 --> 00:02:21,780 Well, we know that attention has to do 64 00:02:21,780 --> 00:02:24,240 with an awful lot of the brain. 65 00:02:24,240 --> 00:02:26,370 Somewhere or other, when you are paying attention 66 00:02:26,370 --> 00:02:30,270 to a task or stimulus, lots of different regions get involved. 67 00:02:30,270 --> 00:02:32,760 There's two big cortical areas. 68 00:02:32,760 --> 00:02:35,610 There's this posterior attention network, 69 00:02:35,610 --> 00:02:37,860 which is kind of like the dorsal parietal lobe here. 70 00:02:37,860 --> 00:02:40,370 Remember dorsal is towards the top of the head? 71 00:02:40,370 --> 00:02:42,600 And there's the anterior attention network, 72 00:02:42,600 --> 00:02:45,910 which is way down here in the prefrontal area. 73 00:02:55,580 --> 00:02:58,040 Prefrontal is, of course, kind of a silly name-- in front 74 00:02:58,040 --> 00:03:01,040 of the frontal-- but it means the very forward-most part 75 00:03:01,040 --> 00:03:03,980 of the frontal lobe, where it kind of ducks underneath there. 76 00:03:13,650 --> 00:03:15,780 So these are areas that have been 77 00:03:15,780 --> 00:03:20,370 shown to be involved in both PET and fMRI imaging, which we know 78 00:03:20,370 --> 00:03:24,930 are imaging techniques that are good for finding areas that 79 00:03:24,930 --> 00:03:25,860 are involved in stuff. 80 00:03:25,860 --> 00:03:27,318 They're not so good at figuring out 81 00:03:27,318 --> 00:03:29,115 exactly what those areas are doing. 82 00:03:29,115 --> 00:03:30,990 We'll talk today about evidence from a couple 83 00:03:30,990 --> 00:03:32,412 of different methodologies. 84 00:03:32,412 --> 00:03:34,620 We talked yesterday also about evidence from patients 85 00:03:34,620 --> 00:03:36,510 with lesions-- patients who have holes 86 00:03:36,510 --> 00:03:40,060 in their brain in certain areas. 87 00:03:40,060 --> 00:03:41,510 So these two. 88 00:03:41,510 --> 00:03:43,200 The posterior attention network. 89 00:03:43,200 --> 00:03:45,330 Remember this is the dorsal parietal area. 90 00:03:45,330 --> 00:03:48,120 It seems to be really involved in spatial attention. 91 00:03:48,120 --> 00:03:51,390 When you shift your attention to a particular region, 92 00:03:51,390 --> 00:03:54,670 like on a display, the posterior attention network 93 00:03:54,670 --> 00:03:57,670 seems to be the part that's involved in this. 94 00:03:57,670 --> 00:04:01,200 This is involved in like visual search tasks. 95 00:04:01,200 --> 00:04:03,750 Again, think of looking for that black circle 96 00:04:03,750 --> 00:04:08,460 among the black boxes and the white circles. 97 00:04:08,460 --> 00:04:10,745 Then your posterior attention network 98 00:04:10,745 --> 00:04:12,120 is the part of your cortex that's 99 00:04:12,120 --> 00:04:16,360 directing your attention to each of the shapes 100 00:04:16,360 --> 00:04:19,200 there one at a time, checking to see if they're 101 00:04:19,200 --> 00:04:22,230 going to be the correct target and shifting it 102 00:04:22,230 --> 00:04:24,570 back and forth between these. 103 00:04:24,570 --> 00:04:27,580 And this is our hemispatial neglect drawing from yesterday. 104 00:04:27,580 --> 00:04:30,840 So remember that patients with hemispatial neglect 105 00:04:30,840 --> 00:04:35,400 can't or are biased against, at the very least, 106 00:04:35,400 --> 00:04:38,684 directing their attention to usually the left-hand side 107 00:04:38,684 --> 00:04:39,600 of their visual field. 108 00:04:39,600 --> 00:04:44,080 Usually you see this with damage to the right parietal lobe. 109 00:04:44,080 --> 00:04:47,970 And so this posterior attention network 110 00:04:47,970 --> 00:04:51,570 seems to be what's damaged in these patients who 111 00:04:51,570 --> 00:04:53,880 are biased against shifting their attention 112 00:04:53,880 --> 00:04:54,750 to certain regions. 113 00:04:58,030 --> 00:04:59,960 This posterior attention network-- 114 00:04:59,960 --> 00:05:02,210 which I wrote "parietal," but I'm wrong-- 115 00:05:02,210 --> 00:05:08,260 also seems to be linked to the specific processing 116 00:05:08,260 --> 00:05:14,732 modules of color and feature and shape in the visual system. 117 00:05:14,732 --> 00:05:16,690 We talked last week about how the visual system 118 00:05:16,690 --> 00:05:19,870 has areas that work on processing color, areas that 119 00:05:19,870 --> 00:05:22,180 work on processing shape, areas that work on processing 120 00:05:22,180 --> 00:05:23,656 direction of movement. 121 00:05:23,656 --> 00:05:25,030 When you're doing a visual search 122 00:05:25,030 --> 00:05:26,696 task that looks for one of those things, 123 00:05:26,696 --> 00:05:28,750 the posterior attention network seems 124 00:05:28,750 --> 00:05:32,020 to be influencing the responsiveness of neurons 125 00:05:32,020 --> 00:05:33,630 in those individual modules. 126 00:05:33,630 --> 00:05:35,880 AUDIENCE: Is that supposed to be "posterior" and not-- 127 00:05:35,880 --> 00:05:36,588 ABBY NOYCE: Yeah. 128 00:05:36,588 --> 00:05:38,372 AUDIENCE: OK. 129 00:05:38,372 --> 00:05:39,580 ABBY NOYCE: Too many P words. 130 00:05:44,180 --> 00:05:46,180 And so what it will do is, if you're 131 00:05:46,180 --> 00:05:48,130 looking for a target that's a black circle, 132 00:05:48,130 --> 00:05:50,920 it'll mean that the neurons that respond 133 00:05:50,920 --> 00:05:54,150 to things that are black are going to be more responsive. 134 00:05:54,150 --> 00:05:57,640 The posterior attention network connects to those neurons 135 00:05:57,640 --> 00:06:01,640 and makes them a little bit more excitable 136 00:06:01,640 --> 00:06:04,720 than neurons that are tuned to less relevant features, 137 00:06:04,720 --> 00:06:07,540 like blueness or redness or whiteness. 138 00:06:12,409 --> 00:06:14,200 So attention is one of these functions that 139 00:06:14,200 --> 00:06:16,750 seems to really take advantage of the interconnectedness 140 00:06:16,750 --> 00:06:19,680 of all our different cortical abilities. 141 00:06:25,290 --> 00:06:26,950 We're going to do this, but it's not 142 00:06:26,950 --> 00:06:30,790 going to link for me because it's not a pretty link. 143 00:06:30,790 --> 00:06:32,566 Copy. 144 00:06:32,566 --> 00:06:33,915 Firefox. 145 00:06:33,915 --> 00:06:34,414 Nope. 146 00:06:37,190 --> 00:06:39,550 This is a kind of classic attentional task, 147 00:06:39,550 --> 00:06:41,380 the Stroop test. 148 00:06:41,380 --> 00:06:43,430 Just be quiet. 149 00:06:43,430 --> 00:06:44,505 Go on. 150 00:06:44,505 --> 00:06:46,760 This is not my Firefox. 151 00:06:46,760 --> 00:06:48,762 Are you going to be trouble? 152 00:06:48,762 --> 00:06:50,190 AUDIENCE: Safari runs faster. 153 00:06:50,190 --> 00:06:51,065 ABBY NOYCE: Yeah, OK. 154 00:06:51,065 --> 00:06:51,620 It works. 155 00:06:51,620 --> 00:06:53,910 Safari does not run faster. 156 00:06:53,910 --> 00:06:57,260 So this is a task where it's going to show you 157 00:06:57,260 --> 00:07:01,970 a set of words printed in different colors 158 00:07:01,970 --> 00:07:05,810 and the goal is to say the color that the word is printed in, 159 00:07:05,810 --> 00:07:07,709 not the word itself. 160 00:07:07,709 --> 00:07:09,500 Some of you may have seen this demo before. 161 00:07:09,500 --> 00:07:10,565 It's kind of a classic. 162 00:07:13,714 --> 00:07:15,130 So what we're going to do is we're 163 00:07:15,130 --> 00:07:18,972 going to have somebody read the list aloud as fast as they can. 164 00:07:18,972 --> 00:07:20,930 We're going to do two rounds of it, a congruent 165 00:07:20,930 --> 00:07:23,500 and an incongruent round. 166 00:07:23,500 --> 00:07:25,115 Anyone want to volunteer to go first? 167 00:07:28,260 --> 00:07:30,320 All right. 168 00:07:30,320 --> 00:07:32,760 Jen, ready? 169 00:07:35,390 --> 00:07:37,525 Read the color the word is printed in. 170 00:07:37,525 --> 00:07:38,066 AUDIENCE: OK. 171 00:07:38,066 --> 00:07:38,912 [INAUDIBLE] Up? 172 00:07:38,912 --> 00:07:39,760 Down? 173 00:07:39,760 --> 00:07:40,551 ABBY NOYCE: Across. 174 00:07:40,551 --> 00:07:41,868 AUDIENCE: OK. 175 00:07:41,868 --> 00:07:42,368 Red. 176 00:07:42,368 --> 00:07:42,846 Green. 177 00:07:42,846 --> 00:07:43,324 Blue. 178 00:07:43,324 --> 00:07:43,824 Yellow. 179 00:07:43,824 --> 00:07:45,236 Pink. 180 00:07:45,236 --> 00:07:46,648 Salmon? 181 00:07:46,648 --> 00:07:47,148 Blue. 182 00:07:47,148 --> 00:07:47,648 Green. 183 00:07:47,648 --> 00:07:48,582 Blue. 184 00:07:48,582 --> 00:07:49,495 Is it orange? 185 00:07:49,495 --> 00:07:50,870 ABBY NOYCE: It's probably orange. 186 00:07:50,870 --> 00:07:51,355 AUDIENCE: Orange. 187 00:07:51,355 --> 00:07:51,840 Blue. 188 00:07:51,840 --> 00:07:52,325 Green. 189 00:07:52,325 --> 00:07:52,810 Blue. 190 00:07:52,810 --> 00:07:53,295 White. 191 00:07:53,295 --> 00:07:53,780 Green. 192 00:07:53,780 --> 00:07:54,265 Yellow. 193 00:07:54,265 --> 00:07:54,750 Orange. 194 00:07:54,750 --> 00:07:55,250 Blue. 195 00:07:55,250 --> 00:07:56,190 White. 196 00:07:56,190 --> 00:07:56,690 Brown. 197 00:07:56,690 --> 00:07:57,175 Red. 198 00:07:57,175 --> 00:07:57,417 Blue. 199 00:07:57,417 --> 00:07:57,660 Yellow. 200 00:07:57,660 --> 00:07:58,145 Green. 201 00:07:58,145 --> 00:07:58,387 Pink. 202 00:07:58,387 --> 00:07:58,630 Yellow. 203 00:07:58,630 --> 00:07:58,880 Green. 204 00:07:58,880 --> 00:07:59,380 Blue. 205 00:07:59,380 --> 00:08:00,172 Red. 206 00:08:00,172 --> 00:08:01,280 ABBY NOYCE: All right. 207 00:08:01,280 --> 00:08:03,051 23.241 seconds, but there was a lot 208 00:08:03,051 --> 00:08:05,300 of hemming and hawing about what we're doing in there. 209 00:08:05,300 --> 00:08:07,548 Yeah, I think that color is meant to be orange. 210 00:08:07,548 --> 00:08:08,629 AUDIENCE: Or salmon? 211 00:08:08,629 --> 00:08:10,420 AUDIENCE: You talking about the orange one? 212 00:08:10,420 --> 00:08:11,740 ABBY NOYCE: Yeah. 213 00:08:11,740 --> 00:08:12,280 All right. 214 00:08:15,114 --> 00:08:17,030 Anyone want to volunteer to do the next round? 215 00:08:17,030 --> 00:08:18,230 AUDIENCE: All of those ones are the same. 216 00:08:18,230 --> 00:08:18,938 ABBY NOYCE: Yeah. 217 00:08:18,938 --> 00:08:20,330 So this is the congruent case. 218 00:08:20,330 --> 00:08:22,760 Anyone want to volunteer to read for the next round? 219 00:08:22,760 --> 00:08:23,260 Sarah? 220 00:08:26,760 --> 00:08:29,206 Ready? 221 00:08:29,206 --> 00:08:30,867 Go. 222 00:08:30,867 --> 00:08:31,533 AUDIENCE: Green. 223 00:08:31,533 --> 00:08:32,512 Yellow. 224 00:08:32,512 --> 00:08:33,012 White. 225 00:08:33,012 --> 00:08:33,998 Pink. 226 00:08:33,998 --> 00:08:35,470 Orange. 227 00:08:35,470 --> 00:08:35,970 Blue. 228 00:08:35,970 --> 00:08:36,470 Red. 229 00:08:36,470 --> 00:08:37,442 Yellow. 230 00:08:37,442 --> 00:08:37,942 Green. 231 00:08:37,942 --> 00:08:38,442 Blue. 232 00:08:38,442 --> 00:08:39,414 Green. 233 00:08:39,414 --> 00:08:39,914 Green. 234 00:08:39,914 --> 00:08:40,893 Red. 235 00:08:40,893 --> 00:08:41,393 Blue. 236 00:08:41,393 --> 00:08:42,372 Green. 237 00:08:42,372 --> 00:08:42,872 Blue. 238 00:08:42,872 --> 00:08:43,372 Pink. 239 00:08:43,372 --> 00:08:44,844 ABBY NOYCE: That's not blue. 240 00:08:44,844 --> 00:08:45,830 AUDIENCE: Ah! 241 00:08:45,830 --> 00:08:47,309 Red. 242 00:08:47,309 --> 00:08:48,788 Pink. 243 00:08:48,788 --> 00:08:50,760 Ah, I keep reading. 244 00:08:50,760 --> 00:08:51,746 Blue. 245 00:08:51,746 --> 00:08:52,732 Yellow. 246 00:08:52,732 --> 00:08:53,718 Pink. 247 00:08:53,718 --> 00:08:54,704 Green. 248 00:08:54,704 --> 00:08:55,690 Orange. 249 00:08:55,690 --> 00:08:57,170 AUDIENCE: That wasn't-- 250 00:08:57,170 --> 00:08:59,646 ABBY NOYCE: That was totally orange. 251 00:08:59,646 --> 00:09:00,300 She's right. 252 00:09:00,300 --> 00:09:00,776 AUDIENCE: Is that orange? 253 00:09:00,776 --> 00:09:01,621 ABBY NOYCE: Yeah. 254 00:09:01,621 --> 00:09:02,204 AUDIENCE: Red. 255 00:09:02,204 --> 00:09:03,608 Blue. 256 00:09:03,608 --> 00:09:04,108 Pink. 257 00:09:04,108 --> 00:09:06,488 No, that's orange. 258 00:09:06,488 --> 00:09:07,916 Red. 259 00:09:07,916 --> 00:09:09,360 Green. 260 00:09:09,360 --> 00:09:10,920 ABBY NOYCE: Very nice. 261 00:09:10,920 --> 00:09:11,947 Is there a difference? 262 00:09:11,947 --> 00:09:12,572 AUDIENCE: Yeah. 263 00:09:12,572 --> 00:09:13,885 It made quite a difference. 264 00:09:13,885 --> 00:09:15,176 ABBY NOYCE: Which one's harder? 265 00:09:15,176 --> 00:09:16,762 AUDIENCE: The discongruent one. 266 00:09:16,762 --> 00:09:17,470 ABBY NOYCE: Yeah. 267 00:09:17,470 --> 00:09:19,790 The incongruent task is a lot harder. 268 00:09:19,790 --> 00:09:22,470 So what you're doing here is your brain-- 269 00:09:22,470 --> 00:09:26,830 reading for most literate adults is a really automatic task. 270 00:09:26,830 --> 00:09:27,460 You see text. 271 00:09:27,460 --> 00:09:28,752 You have to read it. 272 00:09:28,752 --> 00:09:30,460 And what you're trying to do in this task 273 00:09:30,460 --> 00:09:33,880 is inhibit that reading response and do this other, less 274 00:09:33,880 --> 00:09:34,600 familiar thing. 275 00:09:34,600 --> 00:09:36,130 We don't spend a lot of time trying 276 00:09:36,130 --> 00:09:40,006 to read off colors of ink. 277 00:09:40,006 --> 00:09:42,300 AUDIENCE: If you squint your eyes 278 00:09:42,300 --> 00:09:44,259 enough so that you can't read the text anymore, 279 00:09:44,259 --> 00:09:45,966 you can actually do it with [INAUDIBLE].. 280 00:09:45,966 --> 00:09:47,090 ABBY NOYCE: It's easier. 281 00:09:47,090 --> 00:09:48,820 You get better at it with practice, too. 282 00:09:48,820 --> 00:09:49,690 This is definitely something in which 283 00:09:49,690 --> 00:09:50,807 there's a practice effect. 284 00:09:50,807 --> 00:09:52,390 It's also something on which women are 285 00:09:52,390 --> 00:09:55,096 dramatically better than men. 286 00:09:55,096 --> 00:09:57,220 And this goes all the way back to when Stroop first 287 00:09:57,220 --> 00:09:59,620 demonstrated this in the '30s. 288 00:09:59,620 --> 00:10:01,720 And his theory was that, well, because women 289 00:10:01,720 --> 00:10:03,760 are the domestic members of the household, 290 00:10:03,760 --> 00:10:06,650 they spend time thinking about colors of fabric and furniture 291 00:10:06,650 --> 00:10:08,530 and clothing and has so much more 292 00:10:08,530 --> 00:10:11,920 practice thinking about colors and naming colors than men do. 293 00:10:11,920 --> 00:10:13,210 Who knows? 294 00:10:13,210 --> 00:10:16,390 I think that's a bit of a stretch. 295 00:10:16,390 --> 00:10:18,110 AUDIENCE: It has some merit. 296 00:10:18,110 --> 00:10:19,150 AUDIENCE: Ha, ha. 297 00:10:19,150 --> 00:10:20,710 ABBY NOYCE: But there's definitely 298 00:10:20,710 --> 00:10:22,150 a significant difference. 299 00:10:22,150 --> 00:10:25,360 Anyway, this is a task that requires a lot of attention, 300 00:10:25,360 --> 00:10:27,696 but it's not attention that's in the sense of focusing 301 00:10:27,696 --> 00:10:30,070 on a specific spatial location in the way a visual search 302 00:10:30,070 --> 00:10:30,906 task is. 303 00:10:33,880 --> 00:10:41,120 This is a task that requires the other cortical attention 304 00:10:41,120 --> 00:10:41,620 network. 305 00:10:41,620 --> 00:10:43,870 This is the anterior attention network 306 00:10:43,870 --> 00:10:46,200 that isn't so spatially-oriented, 307 00:10:46,200 --> 00:10:50,380 but on tasks like this, where you're really 308 00:10:50,380 --> 00:10:52,420 working on a response and inhibiting something 309 00:10:52,420 --> 00:10:54,910 and really having to focus on what you're doing, 310 00:10:54,910 --> 00:10:56,470 the anterior attention network, which 311 00:10:56,470 --> 00:10:59,050 is right up there in that prefrontal area, 312 00:10:59,050 --> 00:10:59,890 is really involved. 313 00:11:07,122 --> 00:11:08,330 Anyone want to read this one? 314 00:11:12,551 --> 00:11:13,990 Who hasn't done one yet? 315 00:11:13,990 --> 00:11:15,055 Naman, go for it. 316 00:11:15,055 --> 00:11:19,560 AUDIENCE: Red, bue, yellow, purple, yellow, blue, green, 317 00:11:19,560 --> 00:11:26,976 red, green, red, purple, blue, blue, red, green, yellow, red, 318 00:11:26,976 --> 00:11:34,371 green, yellow, purple, yellow, blue, white, purple. 319 00:11:34,371 --> 00:11:36,836 green. purple. 320 00:11:36,836 --> 00:11:38,315 [INAUDIBLE] 321 00:11:38,315 --> 00:11:45,579 So green, purple, green, yellow, red, purple, red, blue, green. 322 00:11:45,579 --> 00:11:47,370 ABBY NOYCE: This is also one of those tasks 323 00:11:47,370 --> 00:11:49,290 where you can start out just fine and get about halfway 324 00:11:49,290 --> 00:11:51,780 through it and then you stumble, and it throws you all off. 325 00:11:51,780 --> 00:11:53,151 I don't know. 326 00:11:53,151 --> 00:11:55,650 But it's definitely easier to do for a shorter list of words 327 00:11:55,650 --> 00:11:56,566 than for a longer one. 328 00:11:56,566 --> 00:12:00,450 It gets harder the more you continue doing it, 329 00:12:00,450 --> 00:12:03,246 again, fitting into this idea that attention is limited 330 00:12:03,246 --> 00:12:05,370 and you're kind of using it up on a task like this. 331 00:12:10,110 --> 00:12:13,800 So how do we measure what the brain is doing if we 332 00:12:13,800 --> 00:12:16,110 don't want to use PET or fMRI? 333 00:12:16,110 --> 00:12:20,280 Because PET and fMRI have lousy temporal resolution. 334 00:12:20,280 --> 00:12:22,680 They can tell you what portions of the brain are active, 335 00:12:22,680 --> 00:12:24,540 but it takes them a few seconds to get 336 00:12:24,540 --> 00:12:27,739 that response and neurons work a whole lot faster 337 00:12:27,739 --> 00:12:28,530 than a few seconds. 338 00:12:28,530 --> 00:12:30,270 You'll see neural responses on the order 339 00:12:30,270 --> 00:12:32,790 of tens of milliseconds. 340 00:12:32,790 --> 00:12:34,610 So this is a picture of the lab I'm 341 00:12:34,610 --> 00:12:36,360 going to be working in when I start school 342 00:12:36,360 --> 00:12:41,730 in the fall of an EEG, electroencephalogram, setup. 343 00:12:41,730 --> 00:12:45,735 So this is a net of electrodes that covers the scalp. 344 00:12:48,720 --> 00:12:53,670 And an EEG can measure the electrical activity 345 00:12:53,670 --> 00:12:56,220 of the brain. 346 00:12:56,220 --> 00:12:59,520 All the electrical stuff that your neurons are doing actually 347 00:12:59,520 --> 00:13:02,820 emits a very subtle electrical field around your head, 348 00:13:02,820 --> 00:13:08,100 and the electrodes in this can pick it up. 349 00:13:08,100 --> 00:13:10,350 And usually, EEGs are used in particular 350 00:13:10,350 --> 00:13:13,110 for what's called an ERP study, where they're measuring 351 00:13:13,110 --> 00:13:18,840 the event-related potential, so changes in the electrical field 352 00:13:18,840 --> 00:13:20,574 that the brain generates in response 353 00:13:20,574 --> 00:13:21,615 to a particular stimulus. 354 00:13:24,180 --> 00:13:28,740 EEGs are not good for high spatial detail the way fMRI is, 355 00:13:28,740 --> 00:13:30,970 but they're really good for time detail. 356 00:13:30,970 --> 00:13:32,640 You can present a stimulus and see 357 00:13:32,640 --> 00:13:38,605 these very quick changes in how the brain responds to them. 358 00:13:38,605 --> 00:13:40,230 And you get to look like a space alien. 359 00:13:45,609 --> 00:13:48,089 AUDIENCE: Can you get money for participating? 360 00:13:48,089 --> 00:13:49,630 ABBY NOYCE: Usually you can get money 361 00:13:49,630 --> 00:13:51,255 for participating in studies like this, 362 00:13:51,255 --> 00:13:53,560 yeah, if only because they take a couple of hours 363 00:13:53,560 --> 00:13:55,840 because getting one of these things set up 364 00:13:55,840 --> 00:13:56,846 is annoying and fiddly. 365 00:13:56,846 --> 00:13:58,720 You've got to check each individual electrode 366 00:13:58,720 --> 00:14:00,160 and make sure it's getting adequate signal, 367 00:14:00,160 --> 00:14:02,110 and there's a little gel you have to squirt into it 368 00:14:02,110 --> 00:14:04,510 to make sure it's getting good contact with the scalp. 369 00:14:04,510 --> 00:14:06,850 And then you've got gel in your hair. 370 00:14:06,850 --> 00:14:09,910 It's not harmful in any way or risky in any way 371 00:14:09,910 --> 00:14:13,360 to the participants, but it's kind of a pain. 372 00:14:13,360 --> 00:14:15,660 AUDIENCE: What was it-- oh, never mind. 373 00:14:15,660 --> 00:14:17,880 ABBY NOYCE: Electroencephalograph 374 00:14:17,880 --> 00:14:21,870 or electroencephalogram, depending on who you ask. 375 00:14:21,870 --> 00:14:25,300 It's like an electrocardiogram is an EKG. 376 00:14:25,300 --> 00:14:27,000 It measures how your heart is doing. 377 00:14:27,000 --> 00:14:30,156 This measures what your brain is doing. 378 00:14:30,156 --> 00:14:34,677 AUDIENCE: [INAUDIBLE] 379 00:14:34,677 --> 00:14:36,260 ABBY NOYCE: So the way they're working 380 00:14:36,260 --> 00:14:38,300 is because they can only pick up the electrical field 381 00:14:38,300 --> 00:14:39,470 on the surface of the scalp. 382 00:14:39,470 --> 00:14:43,520 It's really hard to tell from an EEG where exactly in the brain 383 00:14:43,520 --> 00:14:44,834 that signal is coming from. 384 00:14:44,834 --> 00:14:46,250 You can certainly do broad signal. 385 00:14:46,250 --> 00:14:48,375 So the signal is strongest in the back of the head, 386 00:14:48,375 --> 00:14:50,960 and it's weaker in the anterior areas. 387 00:14:50,960 --> 00:14:55,950 Some people will tell you that with lots of electrodes 388 00:14:55,950 --> 00:14:56,700 capped like that-- 389 00:14:56,700 --> 00:15:01,240 I think that's the 256-electrode cap they have-- 390 00:15:01,240 --> 00:15:01,820 that's a lot. 391 00:15:01,820 --> 00:15:04,640 And you can get better spatial information 392 00:15:04,640 --> 00:15:08,510 with more electrodes, but it's still not as accurate as fMRI 393 00:15:08,510 --> 00:15:11,570 because you're not getting it in those three-dimensional slices 394 00:15:11,570 --> 00:15:12,430 that fMRI gets. 395 00:15:17,800 --> 00:15:20,897 AUDIENCE: [INAUDIBLE] the chart will sort of even out when 396 00:15:20,897 --> 00:15:21,730 it gets [INAUDIBLE]? 397 00:15:21,730 --> 00:15:22,600 ABBY NOYCE: Right. 398 00:15:22,600 --> 00:15:23,530 Well, kind of. 399 00:15:23,530 --> 00:15:25,060 That's why you want to make sure your electrodes have 400 00:15:25,060 --> 00:15:26,920 good contact with the skull and stuff. 401 00:15:26,920 --> 00:15:29,524 Mostly it's this kind of triangulation problem. 402 00:15:29,524 --> 00:15:31,690 If that signal's coming from deep inside your brain, 403 00:15:31,690 --> 00:15:35,140 then it's going to activate a wide range of electrodes 404 00:15:35,140 --> 00:15:36,547 on the surface. 405 00:15:36,547 --> 00:15:39,130 So this is the kind of data you get out of these things, where 406 00:15:39,130 --> 00:15:41,620 you've got a track for each electrode showing 407 00:15:41,620 --> 00:15:45,541 how the charge in its spot is going up and down. 408 00:15:45,541 --> 00:15:47,290 And then people analyze these and come out 409 00:15:47,290 --> 00:15:48,951 with meaningful changes. 410 00:15:48,951 --> 00:15:50,450 What they're usually looking at is-- 411 00:15:53,050 --> 00:15:55,244 this red vertical line is a point in time. 412 00:15:55,244 --> 00:15:57,410 Time's going along this axis, and then each of these 413 00:15:57,410 --> 00:15:59,470 is an individual electrode. 414 00:15:59,470 --> 00:16:02,620 And they're looking at some of these characteristic changes 415 00:16:02,620 --> 00:16:08,930 in the spiking that happened after a stimulus is presented. 416 00:16:08,930 --> 00:16:10,597 And negative is up, and positive is down 417 00:16:10,597 --> 00:16:12,596 is the other thing you need to know about these. 418 00:16:12,596 --> 00:16:14,350 So there's this positive spike, which 419 00:16:14,350 --> 00:16:17,080 is P1, a big negative spike, which 420 00:16:17,080 --> 00:16:22,260 you can see is labeled here N1, P2, N2, and so on. 421 00:16:22,260 --> 00:16:25,629 P3, this big positive drop, you can see there. 422 00:16:25,629 --> 00:16:27,670 And you can see that across different electrodes, 423 00:16:27,670 --> 00:16:28,878 they're pretty much the same. 424 00:16:28,878 --> 00:16:30,402 You can pull out these big patterns. 425 00:16:30,402 --> 00:16:32,110 And so what people are doing when they're 426 00:16:32,110 --> 00:16:34,630 doing ERP measurements is they're looking mostly 427 00:16:34,630 --> 00:16:40,060 for changes in the sizes of these particular potential 428 00:16:40,060 --> 00:16:43,510 changes you see after the stimulus happens. 429 00:16:43,510 --> 00:16:45,230 And I don't know a whole lot about ERP. 430 00:16:45,230 --> 00:16:46,600 I just know the basics. 431 00:16:46,600 --> 00:16:48,480 It's one of the things I want to learn to do. 432 00:16:51,610 --> 00:16:53,960 So ERP and attention. 433 00:16:53,960 --> 00:16:57,220 Here's a couple of classic ERP tasks. 434 00:16:57,220 --> 00:16:59,470 This is an old study, a 1985 study. 435 00:16:59,470 --> 00:17:03,190 So when ERP had really terrible spatial detail because you 436 00:17:03,190 --> 00:17:06,480 were using like 60 electrodes nets or something. 437 00:17:06,480 --> 00:17:11,349 Sams et al had their subjects listen to a set of headphones. 438 00:17:11,349 --> 00:17:15,609 And the headphones would go (LOW-PITCHED) beep beep, beep, 439 00:17:15,609 --> 00:17:17,589 and occasionally they'd go (HIGH-PITCHED) beep, 440 00:17:17,589 --> 00:17:18,940 (LOW-PITCHED) beep. 441 00:17:18,940 --> 00:17:21,970 And they'd have them listen for that higher pitch tone 442 00:17:21,970 --> 00:17:22,967 and press a button. 443 00:17:22,967 --> 00:17:24,550 And then the control case was subjects 444 00:17:24,550 --> 00:17:26,515 who had the exact same set of cones, 445 00:17:26,515 --> 00:17:28,960 but they were just told to ignore the high-pitched tone 446 00:17:28,960 --> 00:17:29,876 and just listen to it. 447 00:17:33,040 --> 00:17:36,410 So this is a task of central control of attention. 448 00:17:36,410 --> 00:17:41,140 You're deciding whether or not your goals in the task 449 00:17:41,140 --> 00:17:44,530 are controlling whether or not the higher-pitched tone is 450 00:17:44,530 --> 00:17:46,900 salient, whether it's important. 451 00:17:46,900 --> 00:17:51,130 And they're measuring what happens what those ERPs-- 452 00:17:51,130 --> 00:17:54,010 Event-Related Potentials-- look like when subjects 453 00:17:54,010 --> 00:17:55,930 heard that higher tone. 454 00:17:55,930 --> 00:17:58,420 And what they found is that, when subjects were attending 455 00:17:58,420 --> 00:18:01,420 to that particular tone, the stimulus-- 456 00:18:01,420 --> 00:18:03,040 the high-pitched tone-- 457 00:18:03,040 --> 00:18:06,520 caused a bigger change in the electrical fields 458 00:18:06,520 --> 00:18:11,180 in their brains than when they were instructed to ignore it. 459 00:18:11,180 --> 00:18:14,412 And that's a whole brain kind of measure. 460 00:18:14,412 --> 00:18:17,328 AUDIENCE: Were those significantly higher? 461 00:18:17,328 --> 00:18:20,250 Like, you'd be able to tell [INAUDIBLE]?? 462 00:18:20,250 --> 00:18:22,400 ABBY NOYCE: Yeah. 463 00:18:22,400 --> 00:18:23,599 AUDIENCE: [INAUDIBLE] 464 00:18:23,599 --> 00:18:24,640 ABBY NOYCE: I don't know. 465 00:18:29,835 --> 00:18:31,610 It was different enough that-- 466 00:18:31,610 --> 00:18:34,130 no subjects said they had any difficulty distinguishing 467 00:18:34,130 --> 00:18:35,060 the tones. 468 00:18:35,060 --> 00:18:36,500 This is one of the things you check for in a debrief. 469 00:18:36,500 --> 00:18:38,660 If you've got a subject who's so tone-deaf they can't hear 470 00:18:38,660 --> 00:18:41,150 the difference in the two tones you're having them listen to, 471 00:18:41,150 --> 00:18:43,441 you want to catch that data and get it out of the pile. 472 00:18:43,441 --> 00:18:44,160 It's not useful. 473 00:18:51,490 --> 00:18:53,530 There's lots of studies like this showing this. 474 00:18:53,530 --> 00:18:56,590 Here's another one that was using exogenously cued 475 00:18:56,590 --> 00:18:59,080 attention-- attention that is being driven 476 00:18:59,080 --> 00:19:00,160 by something [INAUDIBLE]. 477 00:19:00,160 --> 00:19:04,060 They'd have subjects look at a display, fixate on the X 478 00:19:04,060 --> 00:19:06,400 at the bottom, and they would briefly 479 00:19:06,400 --> 00:19:09,070 flash a Q, which in this case is just going 480 00:19:09,070 --> 00:19:11,590 to be the little X in one box. 481 00:19:11,590 --> 00:19:14,830 You're going to keep your eyes fixated on the plots, 482 00:19:14,830 --> 00:19:17,080 but you're going to direct your attention to whichever 483 00:19:17,080 --> 00:19:19,300 box is cued, and then they'd flash 484 00:19:19,300 --> 00:19:23,740 you a stimulus in the same place usually, 485 00:19:23,740 --> 00:19:26,950 and ask you to determine is this a green star or a red star? 486 00:19:31,171 --> 00:19:35,260 And this is kind of my mocked-up model of the type of tasks 487 00:19:35,260 --> 00:19:37,335 that they were doing. 488 00:19:37,335 --> 00:19:38,710 And sometimes people will do this 489 00:19:38,710 --> 00:19:40,334 and they'll cue you to one box and then 490 00:19:40,334 --> 00:19:42,940 they'll flash the stimulus in the other box. 491 00:19:42,940 --> 00:19:44,140 Sneaky. 492 00:19:44,140 --> 00:19:47,635 And what they're comparing is these correctly cued responses 493 00:19:47,635 --> 00:19:49,510 to the stimulus when your attention's already 494 00:19:49,510 --> 00:19:53,230 there versus the opposite. 495 00:19:53,230 --> 00:19:59,990 And again, just like the endogenously cued tone task, 496 00:19:59,990 --> 00:20:03,250 you get ERPs that are bigger for the cued 497 00:20:03,250 --> 00:20:04,930 than the un-cued location. 498 00:20:04,930 --> 00:20:06,882 But it's a short effect. 499 00:20:06,882 --> 00:20:08,840 You've got to have the cue and then the target. 500 00:20:08,840 --> 00:20:11,470 They've got to be 300 milliseconds or less apart. 501 00:20:11,470 --> 00:20:14,050 If it's longer than that, then the effect dies off. 502 00:20:14,050 --> 00:20:18,777 So exogenous cuing of location is a short-term effect. 503 00:20:18,777 --> 00:20:20,110 It doesn't last for a long time. 504 00:20:20,110 --> 00:20:23,930 People's attention leaves that in shifts again. 505 00:20:23,930 --> 00:20:26,410 And somebody ran this more recently 506 00:20:26,410 --> 00:20:29,110 in a better EEG setup, a better one that 507 00:20:29,110 --> 00:20:31,470 gives you some spatial detail and shows 508 00:20:31,470 --> 00:20:36,340 that where you're actually seeing the increased ERP 509 00:20:36,340 --> 00:20:38,170 is in the early processing areas. 510 00:20:38,170 --> 00:20:42,790 So the one kind of primary visual processing jumps 511 00:20:42,790 --> 00:20:48,840 is more responsive to stimuli that are in the attended space 512 00:20:48,840 --> 00:20:53,040 than if it was in the unattended space. 513 00:20:53,040 --> 00:20:56,430 So this is one of the pieces of information showing 514 00:20:56,430 --> 00:20:59,105 that those two attention networks, the posterior 515 00:20:59,105 --> 00:21:01,590 and the anterior attention networks, 516 00:21:01,590 --> 00:21:04,740 actually have effects on a lot of other parts of your brain. 517 00:21:09,210 --> 00:21:10,220 Brains. 518 00:21:10,220 --> 00:21:12,890 OK. 519 00:21:12,890 --> 00:21:19,390 As for some of the fMRI data, again, 520 00:21:19,390 --> 00:21:24,110 fMRI, good spatial resolution, lousy temporal resolution. 521 00:21:24,110 --> 00:21:28,100 And they found that these endogenous and exogenous 522 00:21:28,100 --> 00:21:29,600 attentional shifts again are coming 523 00:21:29,600 --> 00:21:33,100 from very different areas. 524 00:21:33,100 --> 00:21:38,060 Endogenous are coming from way down here in the frontal lobe 525 00:21:38,060 --> 00:21:41,210 and in the dorsal parietal lobe, so towards the top 526 00:21:41,210 --> 00:21:42,860 of the parietal lobe. 527 00:21:42,860 --> 00:21:44,590 And endogenous attentional shifts 528 00:21:44,590 --> 00:21:47,780 are happening down here in the temporal parietal junction 529 00:21:47,780 --> 00:21:50,190 right down into the ventral part of the frontal lobe. 530 00:21:50,190 --> 00:21:53,990 So there's these two distinct attention networks, 531 00:21:53,990 --> 00:21:56,880 and they seem to be independent of one another, 532 00:21:56,880 --> 00:21:59,960 but you can activate one more strongly 533 00:21:59,960 --> 00:22:02,150 than the other or vise versa. 534 00:22:02,150 --> 00:22:05,210 Activating one doesn't necessarily activate the other, 535 00:22:05,210 --> 00:22:07,220 but they interact. 536 00:22:07,220 --> 00:22:08,885 They share a lot of connections. 537 00:22:42,260 --> 00:22:46,010 One more new methodology thing we haven't talked about, which 538 00:22:46,010 --> 00:22:47,420 is this stuff called TMS-- 539 00:22:47,420 --> 00:22:50,302 Transcranial Magnetic Stimulation. 540 00:22:50,302 --> 00:22:52,760 This is a methodology that's really freaky and really cool. 541 00:22:52,760 --> 00:22:54,874 What they do is they've got this little wand. 542 00:22:54,874 --> 00:22:56,540 It looks like it's got a figure 8 on it. 543 00:22:56,540 --> 00:23:00,100 It's got an electromagnetic coil inside there, 544 00:23:00,100 --> 00:23:02,660 so it can create this localized, pretty strong-- 545 00:23:02,660 --> 00:23:04,700 it doesn't have to be super strong-- 546 00:23:04,700 --> 00:23:08,760 magnetic field, a fairly small one, about a cubic centimeter. 547 00:23:08,760 --> 00:23:11,957 And you hold this thing up against somebody's head, 548 00:23:11,957 --> 00:23:13,290 and you send a pulse through it. 549 00:23:13,290 --> 00:23:15,440 And the magnetic field it generates 550 00:23:15,440 --> 00:23:18,590 stops the neurons in that particular region of cortex 551 00:23:18,590 --> 00:23:19,880 from firing. 552 00:23:19,880 --> 00:23:21,650 They just get all thrown out of whack 553 00:23:21,650 --> 00:23:25,520 by the magnetic field messing up their electrical stuff, 554 00:23:25,520 --> 00:23:28,481 and they just don't go. 555 00:23:28,481 --> 00:23:30,230 People have done a lot of really cool work 556 00:23:30,230 --> 00:23:36,470 with this because it lets you take healthy patients, patients 557 00:23:36,470 --> 00:23:39,590 who don't have real lesions or brain injuries 558 00:23:39,590 --> 00:23:42,500 or any kind of neurological complications, 559 00:23:42,500 --> 00:23:45,260 and you can disrupt very particular portions 560 00:23:45,260 --> 00:23:48,380 of their brain, very particular cognitive abilities, 561 00:23:48,380 --> 00:23:50,630 and you can do it quickly or you can 562 00:23:50,630 --> 00:23:52,520 give kind of a sustained series of pulses 563 00:23:52,520 --> 00:23:54,645 and knock the neurons in that particular region out 564 00:23:54,645 --> 00:23:56,090 for a longer time. 565 00:23:56,090 --> 00:23:58,220 And as far as anyone can tell, it 566 00:23:58,220 --> 00:24:00,335 doesn't have any lasting side effects. 567 00:24:00,335 --> 00:24:02,210 You can knock your neurons out for 15 minutes 568 00:24:02,210 --> 00:24:04,126 and then they come back and they're just fine. 569 00:24:06,124 --> 00:24:06,748 AUDIENCE: Wait. 570 00:24:06,748 --> 00:24:08,610 Does it hurt? 571 00:24:08,610 --> 00:24:11,890 ABBY NOYCE: It doesn't hurt. 572 00:24:11,890 --> 00:24:13,580 I haven't gotten to play with this. 573 00:24:13,580 --> 00:24:15,440 I'm a little bit freaked out by it. 574 00:24:15,440 --> 00:24:17,690 But it's very cool. 575 00:24:17,690 --> 00:24:19,880 I've heard people say that if you are doing it 576 00:24:19,880 --> 00:24:21,546 kind of along the sides of the forehead, 577 00:24:21,546 --> 00:24:24,650 it can make the muscles there twitch, which is uncomfortable. 578 00:24:24,650 --> 00:24:27,565 And again, it's the magnetic pulses 579 00:24:27,565 --> 00:24:29,690 interfering with the electrical parts of your body. 580 00:24:34,868 --> 00:24:37,031 AUDIENCE: Can it damage things? 581 00:24:37,031 --> 00:24:38,280 ABBY NOYCE: Can damage things? 582 00:24:41,490 --> 00:24:44,850 If it's used incorrectly, like very incorrectly, 583 00:24:44,850 --> 00:24:47,010 it can cause seizures, and seizures can 584 00:24:47,010 --> 00:24:49,350 cause you to damage yourself. 585 00:24:49,350 --> 00:24:51,660 But if it's done right, it doesn't 586 00:24:51,660 --> 00:24:54,169 seem to have any kind of long-term side effects. 587 00:24:54,169 --> 00:24:55,622 AUDIENCE: Is it like an X-ray? 588 00:24:55,622 --> 00:24:57,163 Like, you can't have too many X-rays. 589 00:24:57,163 --> 00:24:58,787 Can you not have too many [INAUDIBLE]?? 590 00:24:58,787 --> 00:25:02,250 ABBY NOYCE: Well, it's not giving you radiation. 591 00:25:02,250 --> 00:25:04,870 X-rays are exposed to radiation, right? 592 00:25:04,870 --> 00:25:05,840 It's harmful. 593 00:25:05,840 --> 00:25:08,340 AUDIENCE: [INAUDIBLE] too many of them? 594 00:25:08,340 --> 00:25:11,490 ABBY NOYCE: It's been around for only a very little while. 595 00:25:11,490 --> 00:25:15,690 Like five to 10 years this technology has been around, 596 00:25:15,690 --> 00:25:18,774 so nobody's managed to find any long-term side effects. 597 00:25:18,774 --> 00:25:20,940 It may turn out that if you do this to yourself once 598 00:25:20,940 --> 00:25:23,310 a week for 10 years, bad stuff happens, 599 00:25:23,310 --> 00:25:25,909 but nobody's been exposed to that level yet. 600 00:25:25,909 --> 00:25:26,700 It's kind of scary. 601 00:25:26,700 --> 00:25:28,283 There was a point where people thought 602 00:25:28,283 --> 00:25:30,330 that X-rays were harmless, too. 603 00:25:30,330 --> 00:25:31,830 Definitely something to think about. 604 00:25:31,830 --> 00:25:32,450 AUDIENCE: [INAUDIBLE]? 605 00:25:32,450 --> 00:25:33,260 ABBY NOYCE: Yeah. 606 00:25:33,260 --> 00:25:35,670 In like the '40s, you could go into the shoe store 607 00:25:35,670 --> 00:25:37,620 and try on your shoes and they'd put your feet 608 00:25:37,620 --> 00:25:39,744 into the little X-ray and they checked to make sure 609 00:25:39,744 --> 00:25:42,060 your shoes were fitting with an X-ray. 610 00:25:42,060 --> 00:25:43,964 Nobody knew that it was bad for you. 611 00:25:43,964 --> 00:25:44,839 AUDIENCE: [INAUDIBLE] 612 00:25:44,839 --> 00:25:46,110 AUDIENCE: [INAUDIBLE] 613 00:25:46,110 --> 00:25:48,609 ABBY NOYCE: Like an upscale shoe store, a classy shoe store. 614 00:25:48,609 --> 00:25:49,620 AUDIENCE: [INAUDIBLE] 615 00:25:49,620 --> 00:25:50,310 ABBY NOYCE: Now it would be. 616 00:25:50,310 --> 00:25:51,840 And it would be dumb because we know 617 00:25:51,840 --> 00:25:56,550 that cumulative exposure to X-rays is bad. 618 00:25:56,550 --> 00:25:57,890 But nobody knew it then. 619 00:25:57,890 --> 00:25:58,980 Try them on. 620 00:25:58,980 --> 00:26:01,878 You could see if your toes were scooched up together badly. 621 00:26:01,878 --> 00:26:03,954 AUDIENCE: Can't you just [INAUDIBLE],, though? 622 00:26:03,954 --> 00:26:05,370 ABBY NOYCE: But it was scientific. 623 00:26:05,370 --> 00:26:05,911 AUDIENCE: Oh. 624 00:26:05,911 --> 00:26:06,828 AUDIENCE: A-ha. 625 00:26:06,828 --> 00:26:08,924 AUDIENCE: Technological, it's more attractive. 626 00:26:08,924 --> 00:26:11,340 ABBY NOYCE: There's a lot of good data showing that people 627 00:26:11,340 --> 00:26:12,960 are more responsive to information 628 00:26:12,960 --> 00:26:15,090 that seems to be presented in a scientific manner. 629 00:26:15,090 --> 00:26:18,420 Somebody did a study about like, science 630 00:26:18,420 --> 00:26:20,280 has shown that something or other, 631 00:26:20,280 --> 00:26:23,340 and there was some vaguely neurosciency babble that 632 00:26:23,340 --> 00:26:25,320 didn't really mean anything in with it, 633 00:26:25,320 --> 00:26:27,120 and had people read this little paragraph 634 00:26:27,120 --> 00:26:31,201 with or without a picture of your classic fMRI brain 635 00:26:31,201 --> 00:26:33,450 with the pretty colors lighting up in a certain area-- 636 00:26:37,050 --> 00:26:39,690 that had nothing to do with the study-- 637 00:26:39,690 --> 00:26:42,930 information that didn't even really make sense, 638 00:26:42,930 --> 00:26:47,040 then had them rate how believable and convincing 639 00:26:47,040 --> 00:26:48,480 this article was. 640 00:26:48,480 --> 00:26:51,480 Adding the picture of the brain brought it up significantly. 641 00:26:51,480 --> 00:26:53,942 AUDIENCE: Is it like them feeling like [INAUDIBLE] 642 00:26:53,942 --> 00:26:55,650 ABBY NOYCE: Or it just seems like there's 643 00:26:55,650 --> 00:26:58,290 "more data supporting it" even if you're not 644 00:26:58,290 --> 00:27:03,660 qualified to assess the data. 645 00:27:03,660 --> 00:27:04,530 It's interesting. 646 00:27:04,530 --> 00:27:06,630 It's definitely a matter of if you frame this data like it 647 00:27:06,630 --> 00:27:08,850 is being presented by somebody who knows what they're 648 00:27:08,850 --> 00:27:13,410 talking about, then people are more likely to buy into it. 649 00:27:13,410 --> 00:27:18,312 Things to think about when you write your research articles. 650 00:27:18,312 --> 00:27:19,770 So transcranial magnetic simulation 651 00:27:19,770 --> 00:27:23,496 lets you knock out neurons in a particular region of cortex. 652 00:27:23,496 --> 00:27:25,620 It only works for stuff that's right on the surface 653 00:27:25,620 --> 00:27:26,220 of the brain. 654 00:27:26,220 --> 00:27:28,178 It won't work for any sort of subcortical areas 655 00:27:28,178 --> 00:27:32,610 or cortex that's down in the gap between the hemispheres. 656 00:27:32,610 --> 00:27:37,570 But people have used this to look at, OK, we 657 00:27:37,570 --> 00:27:39,790 know that patients with parietal lesions 658 00:27:39,790 --> 00:27:44,750 have difficulty in especially visual attention tasks. 659 00:27:44,750 --> 00:27:49,090 What happens if we do the same thing with normal patients 660 00:27:49,090 --> 00:27:53,030 and try and knock out that portion of the parietal lobe? 661 00:27:53,030 --> 00:27:57,430 And you get very similar to patients with parietal lesions 662 00:27:57,430 --> 00:28:03,340 so that they take more time on conjunctive tasks, where 663 00:28:03,340 --> 00:28:06,820 they've got to match more than one feature, which, remember, 664 00:28:06,820 --> 00:28:11,650 takes visual attention to check each individual member 665 00:28:11,650 --> 00:28:14,920 of the set. 666 00:28:14,920 --> 00:28:18,640 But on disjunctive tasks, which are pre-attentive-- 667 00:28:18,640 --> 00:28:20,740 they don't require visual attention 668 00:28:20,740 --> 00:28:22,000 to jump between things. 669 00:28:22,000 --> 00:28:25,390 You can just see which one is different. 670 00:28:25,390 --> 00:28:28,670 So TMS meshes up the attention requiring task 671 00:28:28,670 --> 00:28:31,300 but not the pre-attentive task when 672 00:28:31,300 --> 00:28:33,942 it's applied to the parietal lobe, 673 00:28:33,942 --> 00:28:35,650 particularly to the dorsal parietal lobe, 674 00:28:35,650 --> 00:28:37,660 that parietal attention network region. 675 00:28:42,760 --> 00:28:45,285 We've been talking about the brain, the whole brain, bits 676 00:28:45,285 --> 00:28:47,080 of the brain. 677 00:28:47,080 --> 00:28:49,650 Let's talk for a minute about what 678 00:28:49,650 --> 00:28:52,290 happens at the cellular level. 679 00:28:52,290 --> 00:28:55,470 We've been talking about neurons. 680 00:28:55,470 --> 00:28:58,560 So one of the things that's really true 681 00:28:58,560 --> 00:29:01,950 is that cells early in the visual processing stream-- 682 00:29:01,950 --> 00:29:04,020 ganglion cells, cells in primary visual cortex-- 683 00:29:04,020 --> 00:29:06,210 have these very small receptive fields. 684 00:29:06,210 --> 00:29:11,980 They're focusing on a particular region of your visual input. 685 00:29:11,980 --> 00:29:14,910 But as you move up into higher processing areas, 686 00:29:14,910 --> 00:29:16,680 each of those higher cells is taking input 687 00:29:16,680 --> 00:29:20,910 from several or even many lower-level cells, 688 00:29:20,910 --> 00:29:24,600 and their receptive fields get bigger as you go up. 689 00:29:24,600 --> 00:29:27,100 So by the time you're looking at cells in, say, the fusiform 690 00:29:27,100 --> 00:29:31,560 face area or in the parietal place, whatever 691 00:29:31,560 --> 00:29:34,620 it's called, the place-monitoring cells-- 692 00:29:34,620 --> 00:29:37,335 cells that respond to pictures of houses and landscapes, 693 00:29:37,335 --> 00:29:39,210 these cells have very large receptive fields. 694 00:29:39,210 --> 00:29:41,820 They'll respond to their preferred stimulus 695 00:29:41,820 --> 00:29:44,010 over a wide range of locations. 696 00:29:46,860 --> 00:29:50,580 And somebody said, well, what happens if we take that range 697 00:29:50,580 --> 00:29:55,080 of locations that the cell will respond to-- maybe the cell 698 00:29:55,080 --> 00:29:58,290 responds to the entire left-hand side of your visual field-- 699 00:29:58,290 --> 00:30:01,770 and we have you attend to a particular, smaller 700 00:30:01,770 --> 00:30:05,580 location within that big receptive field? 701 00:30:05,580 --> 00:30:07,770 Does the cell notice? 702 00:30:07,770 --> 00:30:11,130 Do you see the effects of attention on the cellular level 703 00:30:11,130 --> 00:30:12,430 here? 704 00:30:12,430 --> 00:30:13,860 And the answer is yes. 705 00:30:13,860 --> 00:30:15,150 Yes, you do. 706 00:30:15,150 --> 00:30:17,050 So people doing single-cell recordings-- 707 00:30:17,050 --> 00:30:19,710 so that's with an electrode dropped straight 708 00:30:19,710 --> 00:30:20,835 into the neuron. 709 00:30:20,835 --> 00:30:22,030 You can't do this in people. 710 00:30:22,030 --> 00:30:23,640 You do it in monkeys-- 711 00:30:23,640 --> 00:30:30,150 shows that the cells are way more responsive to stimuli 712 00:30:30,150 --> 00:30:31,770 in the attended area. 713 00:30:31,770 --> 00:30:34,710 They respond more strongly to stimuli 714 00:30:34,710 --> 00:30:39,980 in the attended location than they do normally. 715 00:30:39,980 --> 00:30:42,480 And if you're attending to one location within the receptive 716 00:30:42,480 --> 00:30:47,370 fields, they are less responsive than the baseline 717 00:30:47,370 --> 00:30:49,550 to stimuli elsewhere in the receptive field. 718 00:30:49,550 --> 00:30:53,030 The cell kind of actively throws that information away 719 00:30:53,030 --> 00:30:55,530 and doesn't respond to it. 720 00:30:55,530 --> 00:30:59,010 So attention is modulating how responsive 721 00:30:59,010 --> 00:31:01,786 these cells are to stimuli in different locations 722 00:31:01,786 --> 00:31:02,910 in their respective fields. 723 00:31:22,254 --> 00:31:25,226 [CELL PHONE RINGING] 724 00:31:25,726 --> 00:31:29,198 Tsk, tsk, tsk. 725 00:31:29,198 --> 00:31:30,190 [INAUDIBLE] 726 00:31:45,570 --> 00:31:47,570 So remember yesterday, I said we'd come back 727 00:31:47,570 --> 00:31:50,850 to talking about this competition model of attention, 728 00:31:50,850 --> 00:31:55,220 which is what especially most people with a neuro perspective 729 00:31:55,220 --> 00:31:58,470 on attention think is really what's going on here. 730 00:31:58,470 --> 00:32:02,090 So attention is basically a competition 731 00:32:02,090 --> 00:32:04,610 between all of the different inputs coming into your brain 732 00:32:04,610 --> 00:32:06,350 at any given time. 733 00:32:06,350 --> 00:32:10,417 You can get low-level competition very early 734 00:32:10,417 --> 00:32:11,000 in processing. 735 00:32:11,000 --> 00:32:15,050 You can get high-level competition. 736 00:32:15,050 --> 00:32:17,180 At any of these points in times, one input 737 00:32:17,180 --> 00:32:19,550 can kind of elbow another out of the way 738 00:32:19,550 --> 00:32:22,340 and say, hey, no, I'm getting major processing 739 00:32:22,340 --> 00:32:23,210 resources right now. 740 00:32:26,850 --> 00:32:31,010 So these guys say that inputs that are the most salient-- 741 00:32:31,010 --> 00:32:32,160 and saliency-- 742 00:32:32,160 --> 00:32:32,960 AUDIENCE: Can you go back to [INAUDIBLE]?? 743 00:32:32,960 --> 00:32:34,239 ABBY NOYCE: I can go back one. 744 00:32:38,505 --> 00:32:40,880 So we were talking yesterday about like bottleneck models 745 00:32:40,880 --> 00:32:43,010 that think of attention as being limited, 746 00:32:43,010 --> 00:32:46,280 and different tasks call on different amounts of attention, 747 00:32:46,280 --> 00:32:48,860 and they're using up your attentional resources. 748 00:32:48,860 --> 00:32:50,870 Competition model people say, it's not so much 749 00:32:50,870 --> 00:32:53,000 that you have attentional resources. 750 00:32:53,000 --> 00:32:55,880 You've got honest to goodness processing resources, 751 00:32:55,880 --> 00:32:59,806 and attention is a means of directing processing resources 752 00:32:59,806 --> 00:33:01,430 to different tasks or different inputs. 753 00:33:11,410 --> 00:33:15,150 So inputs with the most saliency can claim the most resources. 754 00:33:15,150 --> 00:33:17,340 They can push other inputs out of the way. 755 00:33:17,340 --> 00:33:19,860 If you know you're looking for something red, 756 00:33:19,860 --> 00:33:23,730 then informational about red things that's 757 00:33:23,730 --> 00:33:26,490 coming in kind of trumps information about other things, 758 00:33:26,490 --> 00:33:30,600 and you assign more resources to things that are red. 759 00:33:30,600 --> 00:33:33,780 And so the red ones get processed the most deeply. 760 00:33:33,780 --> 00:33:37,770 Think about the basketball passing video 761 00:33:37,770 --> 00:33:39,540 that we looked at yesterday. 762 00:33:39,540 --> 00:33:42,030 You're busy watching the white team pass the ball, 763 00:33:42,030 --> 00:33:45,810 so all of the information about the people wearing white 764 00:33:45,810 --> 00:33:50,220 got assigned more saliency endogenously 765 00:33:50,220 --> 00:33:52,380 by your central control system. 766 00:33:52,380 --> 00:33:54,630 And so they're shoving aside the information 767 00:33:54,630 --> 00:33:58,590 about the people wearing black so completely that you don't 768 00:33:58,590 --> 00:34:00,480 even notice this guy in a gorilla suit 769 00:34:00,480 --> 00:34:04,260 walking into the middle of the screen and being like, rahh. 770 00:34:04,260 --> 00:34:06,630 You watch it again when you're not focusing so 771 00:34:06,630 --> 00:34:09,210 explicitly on the white ones, then 772 00:34:09,210 --> 00:34:11,489 the saliency of, hey, wait. 773 00:34:11,489 --> 00:34:15,179 Something weird just walked into the frame. 774 00:34:15,179 --> 00:34:20,130 Now stimulus that's interesting in some way 775 00:34:20,130 --> 00:34:21,996 can grab your attention and direct it to it. 776 00:34:26,560 --> 00:34:29,620 So in early stages of processing, 777 00:34:29,620 --> 00:34:32,650 saliency depends on things like color. 778 00:34:32,650 --> 00:34:35,590 In a scene of mostly dark things with one bright thing, 779 00:34:35,590 --> 00:34:38,110 or vise versa, your attention is going 780 00:34:38,110 --> 00:34:41,632 to be drawn to the one that's different in some way. 781 00:34:41,632 --> 00:34:43,840 We've said the brain likes things that are different. 782 00:34:43,840 --> 00:34:45,876 It likes changes. 783 00:34:45,876 --> 00:34:47,500 Your attention is drawn to these places 784 00:34:47,500 --> 00:34:51,760 where things are different, especially 785 00:34:51,760 --> 00:34:53,300 in low-level processing. 786 00:34:53,300 --> 00:34:56,080 So areas that have edges get processed more heavily 787 00:34:56,080 --> 00:34:57,070 in early vision. 788 00:34:57,070 --> 00:34:59,200 We know that. 789 00:34:59,200 --> 00:35:00,820 In later processing, this competition 790 00:35:00,820 --> 00:35:04,930 is affected by things like what do I want to do? 791 00:35:04,930 --> 00:35:05,950 What's around it? 792 00:35:05,950 --> 00:35:09,580 It's affected by all of this more semantic, 793 00:35:09,580 --> 00:35:13,570 higher-level input that's coming in from you, from what 794 00:35:13,570 --> 00:35:16,270 you want to get out of the information 795 00:35:16,270 --> 00:35:18,860 that you're looking at. 796 00:35:18,860 --> 00:35:21,250 And so the thing is that all through the stages 797 00:35:21,250 --> 00:35:23,080 of processing, the different things 798 00:35:23,080 --> 00:35:27,670 that are being thought about are competing. 799 00:35:27,670 --> 00:35:33,460 So if you're looking at, say, the red chair, 800 00:35:33,460 --> 00:35:36,290 people who are really into competition models will say, 801 00:35:36,290 --> 00:35:39,730 OK, we're already thinking about your perception of color 802 00:35:39,730 --> 00:35:42,970 as depending on how activated different color 803 00:35:42,970 --> 00:35:45,110 representations are. 804 00:35:45,110 --> 00:35:48,010 You can say that even just seeing the color red 805 00:35:48,010 --> 00:35:50,500 is a result of having the redness model be 806 00:35:50,500 --> 00:35:53,830 more strongly activated than the greenness or the blueness 807 00:35:53,830 --> 00:35:56,320 or the brownness representation. 808 00:35:56,320 --> 00:35:59,200 And this is a similar kind of processing 809 00:35:59,200 --> 00:36:00,290 as it goes up the chain. 810 00:36:00,290 --> 00:36:02,260 And because there's so much connection 811 00:36:02,260 --> 00:36:04,150 between different regions of the brain, 812 00:36:04,150 --> 00:36:08,680 connections up and down between different levels of processing, 813 00:36:08,680 --> 00:36:12,370 then each kind of local competition for attention 814 00:36:12,370 --> 00:36:16,480 for processing manages to influence-- 815 00:36:16,480 --> 00:36:20,800 when one piece of information wins a local competition, 816 00:36:20,800 --> 00:36:23,260 then it can influence other competitions. 817 00:36:23,260 --> 00:36:26,500 So if you're trying to split your attention 818 00:36:26,500 --> 00:36:28,624 between the water bottle and the red chair, 819 00:36:28,624 --> 00:36:30,790 if you're paying more attention to the water bottle, 820 00:36:30,790 --> 00:36:32,748 that's going to actually go back down the chain 821 00:36:32,748 --> 00:36:34,840 and affect how your color processing is working. 822 00:36:34,840 --> 00:36:37,550 At a lower level, you're going to put more effort 823 00:36:37,550 --> 00:36:41,620 into the blue processing than the red processing. 824 00:36:41,620 --> 00:36:45,130 But you'll also see, for example, 825 00:36:45,130 --> 00:36:47,270 there's connections between sensory modes. 826 00:36:47,270 --> 00:36:49,029 So if you hear a loud noise behind you, 827 00:36:49,029 --> 00:36:50,820 you're going to turn around and look at it. 828 00:36:50,820 --> 00:36:52,450 That's, again, dependent on connections 829 00:36:52,450 --> 00:36:56,380 between the parts of your brain that analyze 830 00:36:56,380 --> 00:36:58,427 auditory information and try and figure out 831 00:36:58,427 --> 00:37:00,010 where it is in the parts of your brain 832 00:37:00,010 --> 00:37:03,340 that direct your visual gaze. 833 00:37:03,340 --> 00:37:06,040 So the brain is very interconnected. 834 00:37:06,040 --> 00:37:09,160 The competition model here really takes advantage of that. 835 00:37:14,925 --> 00:37:16,800 Another good piece of evidence for this model 836 00:37:16,800 --> 00:37:20,070 is what's called competition effects. 837 00:37:20,070 --> 00:37:23,188 This is an fMRI study. 838 00:37:23,188 --> 00:37:27,090 It had subjects look at individual, moderately complex 839 00:37:27,090 --> 00:37:28,410 shapes on a screen. 840 00:37:28,410 --> 00:37:32,700 It measured how strongly their primary visual cortex 841 00:37:32,700 --> 00:37:35,940 responded to these shapes and found 842 00:37:35,940 --> 00:37:40,310 that, compared to how much it responded for shapes one 843 00:37:40,310 --> 00:37:43,080 at a time, if they showed four shapes at once, 844 00:37:43,080 --> 00:37:46,440 response actually went down, which at first seems 845 00:37:46,440 --> 00:37:47,490 like an odd effect. 846 00:37:47,490 --> 00:37:49,320 You're looking at more things. 847 00:37:49,320 --> 00:37:52,380 You'd think you'd be getting more processing going on. 848 00:37:52,380 --> 00:37:56,700 And the theory is that when you've 849 00:37:56,700 --> 00:37:58,680 got multiple things competing for attention 850 00:37:58,680 --> 00:38:00,840 and competing for resources, each of them 851 00:38:00,840 --> 00:38:03,990 kind of inhibits the processing of the other things that 852 00:38:03,990 --> 00:38:06,040 are in the competition. 853 00:38:06,040 --> 00:38:09,690 So when there are four objects and not one of them 854 00:38:09,690 --> 00:38:13,830 has any kind of edge up in this competition, 855 00:38:13,830 --> 00:38:16,170 then they all inhibit each other and none of them 856 00:38:16,170 --> 00:38:21,270 gets processed as fully as one object at a time would. 857 00:38:21,270 --> 00:38:22,910 And if the experimenter says, OK, I 858 00:38:22,910 --> 00:38:26,490 want you to attend to the object in the upper left-hand corner, 859 00:38:26,490 --> 00:38:28,860 all of the sudden you see the V1 activity go right back 860 00:38:28,860 --> 00:38:32,430 up to what it was when you were just looking at one object. 861 00:38:32,430 --> 00:38:37,200 The suppression object wins the competition between the four 862 00:38:37,200 --> 00:38:42,540 objects and gets to be fully processed. 863 00:38:45,660 --> 00:38:49,030 So your option of attending to a particular location-- 864 00:38:49,030 --> 00:38:50,505 the upper left-hand corner-- 865 00:38:50,505 --> 00:38:52,630 then makes the items in that location more salient, 866 00:38:52,630 --> 00:38:56,660 and they get more processing. 867 00:38:56,660 --> 00:38:58,637 So what's our general consensus? 868 00:38:58,637 --> 00:39:00,970 Does the competition model explain decreased performance 869 00:39:00,970 --> 00:39:02,590 on divided attention tasks? 870 00:39:02,590 --> 00:39:03,940 AUDIENCE: Yes. 871 00:39:03,940 --> 00:39:05,950 ABBY NOYCE: Why? 872 00:39:05,950 --> 00:39:07,450 In terms of the competition model. 873 00:39:07,450 --> 00:39:08,020 Sorry, Sarah. 874 00:39:08,020 --> 00:39:09,814 I missed you entirely. 875 00:39:09,814 --> 00:39:11,230 In terms of the competition model, 876 00:39:11,230 --> 00:39:12,688 what's going on when you are trying 877 00:39:12,688 --> 00:39:15,875 to do two things at once? 878 00:39:15,875 --> 00:39:17,458 AUDIENCE: The object is like competing 879 00:39:17,458 --> 00:39:20,086 for your attention [INAUDIBLE]. 880 00:39:20,086 --> 00:39:23,402 And therefore, if there's two things that-- 881 00:39:23,402 --> 00:39:26,986 equal things are going to have 50% of your attention, I guess? 882 00:39:26,986 --> 00:39:30,906 And they won't be done so efficiently 883 00:39:30,906 --> 00:39:32,252 as if it was just one thing. 884 00:39:32,252 --> 00:39:33,710 ABBY NOYCE: Yeah, although remember 885 00:39:33,710 --> 00:39:35,750 that this model talks about attention kind of 886 00:39:35,750 --> 00:39:38,960 is that competition, and what they're competing for 887 00:39:38,960 --> 00:39:42,320 is kind of processing power. 888 00:39:42,320 --> 00:39:42,890 But yeah. 889 00:39:42,890 --> 00:39:44,210 So you've got two things going on, 890 00:39:44,210 --> 00:39:46,543 and they're both kind of trying to get processing power. 891 00:39:46,543 --> 00:39:49,280 So does either one of them get all of it? 892 00:39:49,280 --> 00:39:50,808 Probably not. 893 00:39:50,808 --> 00:39:54,084 AUDIENCE: I think the other thing's that [INAUDIBLE] 894 00:39:54,084 --> 00:39:58,766 differentiate between stimuli versus actually a [INAUDIBLE] 895 00:39:58,766 --> 00:40:02,330 perception of or processing of [INAUDIBLE]---- 896 00:40:02,330 --> 00:40:04,957 say driving and talking on the phone. 897 00:40:04,957 --> 00:40:07,285 There's stimuli by light and color 898 00:40:07,285 --> 00:40:10,945 of what you see on the road, because a bit 899 00:40:10,945 --> 00:40:12,941 from your ears [INAUDIBLE]. 900 00:40:16,933 --> 00:40:19,927 You also have stimuli in what you're hearing, 901 00:40:19,927 --> 00:40:24,020 and you have to process what the person's saying on the phone. 902 00:40:24,020 --> 00:40:26,396 And then you also have to form words in your mind 903 00:40:26,396 --> 00:40:29,992 and, I guess, make sure your grammar's correct or something, 904 00:40:29,992 --> 00:40:31,590 pronunciation. 905 00:40:31,590 --> 00:40:33,338 And that's not even two stimuli. 906 00:40:33,338 --> 00:40:35,290 That's a whole bunch. 907 00:40:35,290 --> 00:40:36,320 ABBY NOYCE: Yeah. 908 00:40:36,320 --> 00:40:39,890 These two tasks are both ones that need a lot of resources. 909 00:40:39,890 --> 00:40:41,877 You can't half-ass either one of them. 910 00:40:41,877 --> 00:40:44,210 You've got to put a lot of processing into both of them. 911 00:40:44,210 --> 00:40:46,460 So it's probably a more demanding divided attention 912 00:40:46,460 --> 00:40:47,922 task than some others. 913 00:40:47,922 --> 00:40:48,630 And this is true. 914 00:40:48,630 --> 00:40:50,420 I actually have a friend who works on attention. 915 00:40:50,420 --> 00:40:52,878 And one of the things she does is, when she designs a task, 916 00:40:52,878 --> 00:40:54,712 she's got to run them through a pilot group, 917 00:40:54,712 --> 00:40:57,002 like two or three people, just to make sure they're not 918 00:40:57,002 --> 00:40:59,390 too easy because they've got to be a challenging task 919 00:40:59,390 --> 00:41:01,010 or you're not going to get any interesting results 920 00:41:01,010 --> 00:41:02,551 across different experimental groups. 921 00:41:02,551 --> 00:41:05,355 Everyone's just going to get 100%. 922 00:41:05,355 --> 00:41:06,300 And they can't be-- 923 00:41:06,300 --> 00:41:07,550 well, they can be pretty hard. 924 00:41:07,550 --> 00:41:09,290 I've been a subject for her from time to time, 925 00:41:09,290 --> 00:41:10,850 and it's usually really frustrating 926 00:41:10,850 --> 00:41:14,620 because you're like, this is an easy task and I am failing! 927 00:41:14,620 --> 00:41:16,040 OK. 928 00:41:16,040 --> 00:41:17,250 Thank you very much. 929 00:41:17,250 --> 00:41:19,666 Moving along, the other thing I wanted to talk about today 930 00:41:19,666 --> 00:41:25,280 was ADHD really briefly, in part because it's 931 00:41:25,280 --> 00:41:31,460 a really common diagnosis, especially among kids. 932 00:41:31,460 --> 00:41:32,810 So what is ADHD? 933 00:41:32,810 --> 00:41:39,380 This is the DSM IV text revision criteria for diagnosis of ADHD. 934 00:41:39,380 --> 00:41:43,880 The DSM is the Diagnostic and Statistic Manual, 935 00:41:43,880 --> 00:41:46,430 and it is the book that psychologists 936 00:41:46,430 --> 00:41:48,240 use to make diagnoses. 937 00:41:48,240 --> 00:41:53,180 So for everything that is an "official" disorder of one 938 00:41:53,180 --> 00:41:56,090 sort or another, there's a page of information roughly 939 00:41:56,090 --> 00:41:59,240 like this that says, in order to be diagnosed with this thing, 940 00:41:59,240 --> 00:42:02,580 you've got to have these criteria. 941 00:42:02,580 --> 00:42:06,530 So for ADHD, to be diagnosed with it, 942 00:42:06,530 --> 00:42:09,530 you've got to have six or more symptoms of inattention. 943 00:42:09,530 --> 00:42:12,070 So this would include things like spacing out in school. 944 00:42:12,070 --> 00:42:13,486 It would include things like being 945 00:42:13,486 --> 00:42:16,260 needed to be given directions more than once, 946 00:42:16,260 --> 00:42:19,910 difficulty staying on task, distractibility. 947 00:42:19,910 --> 00:42:21,770 Six or more symptoms of hyperactivity. 948 00:42:25,096 --> 00:42:27,470 Again, most of these are over an extended period of time. 949 00:42:27,470 --> 00:42:30,860 This can't be just like you have one bad day. 950 00:42:30,860 --> 00:42:34,880 Inattention, hyperactivity, and impulsivity are maladaptive. 951 00:42:34,880 --> 00:42:37,070 They're causing you problems. 952 00:42:37,070 --> 00:42:39,710 And they are inconsistent with developmental level. 953 00:42:39,710 --> 00:42:41,420 Clearly, your attention span expectations 954 00:42:41,420 --> 00:42:42,919 for a three-year-old are going to be 955 00:42:42,919 --> 00:42:45,530 different than your attention span expectations 956 00:42:45,530 --> 00:42:47,970 for a seven-year-old. 957 00:42:47,970 --> 00:42:50,540 Saying a three-year-old is being inattentive because they 958 00:42:50,540 --> 00:42:53,300 can't sit still as long as a seven-year-old can is not 959 00:42:53,300 --> 00:42:54,420 productive. 960 00:42:54,420 --> 00:42:57,650 So it's got to be based around how old the kid is. 961 00:42:57,650 --> 00:43:00,830 Some symptoms present pretty young. 962 00:43:00,830 --> 00:43:03,710 Impairment from symptoms present in at least two settings, 963 00:43:03,710 --> 00:43:06,080 so at school or at home. 964 00:43:06,080 --> 00:43:09,660 And to be diagnosed, it's got to be impairing functioning. 965 00:43:09,660 --> 00:43:12,980 And this is pretty typical for DSM diagnoses 966 00:43:12,980 --> 00:43:15,200 is that, if you meet all the other symptoms 967 00:43:15,200 --> 00:43:17,390 but you are functioning just fine in your life 968 00:43:17,390 --> 00:43:20,060 and it's not causing you any problems, 969 00:43:20,060 --> 00:43:21,529 then you do not have a disorder. 970 00:43:21,529 --> 00:43:23,570 It's got to be causing you trouble for you to get 971 00:43:23,570 --> 00:43:26,480 a diagnosis of a disorder. 972 00:43:26,480 --> 00:43:30,680 ADHD is pretty common. 973 00:43:30,680 --> 00:43:32,750 It's estimated to occur in about one 974 00:43:32,750 --> 00:43:36,740 in every 16 or 17 grade-school-age kids. 975 00:43:36,740 --> 00:43:43,022 And we don't know a whole lot about what causes it. 976 00:43:43,022 --> 00:43:44,480 There's been a little bit of stuff. 977 00:43:44,480 --> 00:43:46,100 There's some subtle differences. 978 00:43:46,100 --> 00:43:50,150 Kids with ADHD tend to have slightly smaller overall brains 979 00:43:50,150 --> 00:43:52,670 than kids without. 980 00:43:52,670 --> 00:43:54,890 Particularly frontal cortex and basal ganglia 981 00:43:54,890 --> 00:43:57,980 are areas that tend to be smaller. 982 00:43:57,980 --> 00:43:59,480 Mothers who smoke during pregnancy 983 00:43:59,480 --> 00:44:03,020 are three times less likely to have a kid with ADHD. 984 00:44:03,020 --> 00:44:05,390 But it hasn't been shown that that's a causal link, 985 00:44:05,390 --> 00:44:07,700 that smoking causes ADHD. 986 00:44:07,700 --> 00:44:09,440 There could be other factors involved 987 00:44:09,440 --> 00:44:11,210 in both of those things. 988 00:44:11,210 --> 00:44:12,710 Remember that one of the things that 989 00:44:12,710 --> 00:44:15,140 happens in people with ADHD is they're more impulsive, 990 00:44:15,140 --> 00:44:18,500 and smoking links pretty well to impulsivity. 991 00:44:18,500 --> 00:44:21,410 Doing things that you know are bad for you often 992 00:44:21,410 --> 00:44:23,815 link to impulsivity. 993 00:44:26,316 --> 00:44:27,190 What else can we say? 994 00:44:33,320 --> 00:44:35,660 Kids in a family where there is one child with ADHD 995 00:44:35,660 --> 00:44:39,050 are more likely to have other kids with ADHD. 996 00:44:39,050 --> 00:44:42,350 Kids in families with more than one kid with ADHD 997 00:44:42,350 --> 00:44:45,590 also tend to be, on the whole, again, 998 00:44:45,590 --> 00:44:47,960 less functional than families with kids without. 999 00:44:47,960 --> 00:44:49,370 It's a complicated disorder. 1000 00:44:49,370 --> 00:44:52,735 Nobody really knows what causes it for sure. 1001 00:44:52,735 --> 00:44:54,110 But one of the things we've found 1002 00:44:54,110 --> 00:44:57,350 is there's this set of drugs that treat it really, really-- 1003 00:44:57,350 --> 00:45:00,180 well, moderately, anyway-- well. 1004 00:45:00,180 --> 00:45:05,000 And these drugs are a group of drugs called psychostimulants. 1005 00:45:05,000 --> 00:45:06,920 They're closely related to cocaine. 1006 00:45:06,920 --> 00:45:09,080 They're closely related to amphetamines. 1007 00:45:14,984 --> 00:45:19,810 And the way they work is, so if you're at your dopamine synapse 1008 00:45:19,810 --> 00:45:21,810 here. 1009 00:45:21,810 --> 00:45:23,070 There's our synapse, right? 1010 00:45:23,070 --> 00:45:25,380 Cell 1, cell 2. 1011 00:45:28,230 --> 00:45:34,380 So at the dopamine synapse, the presynaptic cell, remember, 1012 00:45:34,380 --> 00:45:37,920 has these transporter channels that 1013 00:45:37,920 --> 00:45:41,670 are going to vacuum up extra dopamine from the synapse. 1014 00:45:41,670 --> 00:45:45,600 And these psychostimulants get picked up 1015 00:45:45,600 --> 00:45:49,050 by the channel, brought into this cell, 1016 00:45:49,050 --> 00:45:52,710 and they bind with the vesicles that contain 1017 00:45:52,710 --> 00:45:54,930 the dopamine neurotransmitter. 1018 00:45:54,930 --> 00:45:56,870 And these drugs actually cause those 1019 00:45:56,870 --> 00:46:00,580 vesicles to go ahead and release dopamine into the synapse. 1020 00:46:00,580 --> 00:46:04,140 So the effect of the drug is that more dopamine 1021 00:46:04,140 --> 00:46:08,460 gets released into the synapse at a dopaminergic cell. 1022 00:46:08,460 --> 00:46:11,820 Now, we'll remember that most of the dopamine in your brain 1023 00:46:11,820 --> 00:46:15,950 is coming from right here in the midbrain, 1024 00:46:15,950 --> 00:46:18,780 in the very upper portion of the brain stem. 1025 00:46:18,780 --> 00:46:20,040 And there's two main pathways. 1026 00:46:20,040 --> 00:46:22,956 There's this one that goes to the nucleus accumbens 1027 00:46:22,956 --> 00:46:25,080 and up into cortex, and there's this other one that 1028 00:46:25,080 --> 00:46:26,655 goes to like motor control. 1029 00:46:30,880 --> 00:46:33,720 So dopamine has effects on lots of portions of your brain. 1030 00:46:33,720 --> 00:46:36,060 Changing how dopamine is behaving at the synapse 1031 00:46:36,060 --> 00:46:38,100 has effects on large portions of the brain. 1032 00:46:41,100 --> 00:46:44,059 There's some evidence that ADHD patients start out 1033 00:46:44,059 --> 00:46:46,350 with a different form of the dopamine transmitter, kind 1034 00:46:46,350 --> 00:46:50,640 of a more efficient transporter, a more efficient form of this, 1035 00:46:50,640 --> 00:46:52,800 so that it kind of vacuums dopamine out 1036 00:46:52,800 --> 00:46:55,680 of the synapse more enthusiastically 1037 00:46:55,680 --> 00:46:58,320 than the normal form so that you start out 1038 00:46:58,320 --> 00:47:00,720 with a shortage of dopamine in the synapse. 1039 00:47:00,720 --> 00:47:03,540 And one of the things the psychostimulants are doing 1040 00:47:03,540 --> 00:47:07,800 is causing there to be enough stimulants. 1041 00:47:07,800 --> 00:47:11,969 Stimulants usually cause increased motor activity. 1042 00:47:11,969 --> 00:47:13,260 It's one of the things they do. 1043 00:47:13,260 --> 00:47:15,450 They stimulate all this motor cortex. 1044 00:47:15,450 --> 00:47:16,580 They get you up and going. 1045 00:47:21,270 --> 00:47:25,470 So having Ritalin or Adderall or one of these drugs actually 1046 00:47:25,470 --> 00:47:29,280 work to calm patients with ADHD was originally 1047 00:47:29,280 --> 00:47:32,100 called a "paradoxical" response, and people 1048 00:47:32,100 --> 00:47:36,120 thought it was something different about how ADHD brains 1049 00:47:36,120 --> 00:47:37,510 work. 1050 00:47:37,510 --> 00:47:39,010 More recently, people have tried it, 1051 00:47:39,010 --> 00:47:40,426 because one of things that happens 1052 00:47:40,426 --> 00:47:42,600 is the dose of an amphetamine you take if you're 1053 00:47:42,600 --> 00:47:45,720 trying to get high off of it is a lot higher than the dose that 1054 00:47:45,720 --> 00:47:48,450 is prescribed for a kid with ADHD. 1055 00:47:48,450 --> 00:47:51,960 And it's actually turned out that both in animal models 1056 00:47:51,960 --> 00:47:54,510 and in humans, if you give one of these stimulants 1057 00:47:54,510 --> 00:47:56,460 in that really low dose, it actually 1058 00:47:56,460 --> 00:48:00,270 does decrease locomotor activity in the animal models. 1059 00:48:00,270 --> 00:48:03,690 It's calming on both kids and adults, 1060 00:48:03,690 --> 00:48:05,550 again, in the low doses-- 1061 00:48:05,550 --> 00:48:09,780 not just ADHD kids and adults, but even 1062 00:48:09,780 --> 00:48:13,021 people who are not diagnosed with the disorder. 1063 00:48:13,021 --> 00:48:15,270 One of the big problems with most of these stimulants, 1064 00:48:15,270 --> 00:48:18,240 on the other hand, is that things like methylphenidate, 1065 00:48:18,240 --> 00:48:20,610 they're upping the amount of dopamine in your brain. 1066 00:48:20,610 --> 00:48:24,540 Dopamine is the reward pathway. 1067 00:48:24,540 --> 00:48:27,660 Drugs that up the amount of dopamine in your brain 1068 00:48:27,660 --> 00:48:29,205 are prone to addiction and abuse. 1069 00:48:32,430 --> 00:48:33,750 Methylphenidate is Ritalin. 1070 00:48:33,750 --> 00:48:35,670 It's the actual drug name for it. 1071 00:48:35,670 --> 00:48:41,142 Rates of abuse of Ritalin and other drugs are on the rise. 1072 00:48:41,142 --> 00:48:42,600 They've been going kind of steadily 1073 00:48:42,600 --> 00:48:46,100 up since this drug hit the market, jeez, like 15 years 1074 00:48:46,100 --> 00:48:46,600 ago now. 1075 00:48:51,717 --> 00:48:54,300 You'll see like college students use them as like study drugs. 1076 00:48:54,300 --> 00:48:56,340 There's a black market for these. 1077 00:48:56,340 --> 00:48:58,380 But don't do that. 1078 00:48:58,380 --> 00:49:02,632 Taking things that are not prescribed to you is dumb. 1079 00:49:02,632 --> 00:49:06,083 AUDIENCE: [INAUDIBLE] the entire dorm at school 1080 00:49:06,083 --> 00:49:08,060 was like kicked out because of this one kid. 1081 00:49:08,060 --> 00:49:11,368 He had ADHD and gave the drugs to his friend, who sold it 1082 00:49:11,368 --> 00:49:13,200 to like the rest of their dorm. 1083 00:49:13,200 --> 00:49:16,298 And it's like a high-pressure environment [INAUDIBLE].. 1084 00:49:16,298 --> 00:49:18,530 They all got kicked out. 1085 00:49:18,530 --> 00:49:21,320 ABBY NOYCE: Yeah. 1086 00:49:21,320 --> 00:49:25,872 Yeah, it's prone to abuse, partly because it's so widely 1087 00:49:25,872 --> 00:49:28,080 prescribed now that there's a lot of kids who can get 1088 00:49:28,080 --> 00:49:36,360 their hands on it, and kids are lax about prescription drugs 1089 00:49:36,360 --> 00:49:39,435 in a way that adults or people-- and I think this may be 1090 00:49:39,435 --> 00:49:41,310 a generational thing more than an age thing-- 1091 00:49:41,310 --> 00:49:43,620 aren't and are much more likely to be like, "Oh yeah, 1092 00:49:43,620 --> 00:49:44,550 a doctor prescribed it to me. 1093 00:49:44,550 --> 00:49:45,360 It must be safe. 1094 00:49:45,360 --> 00:49:48,090 Here, have some," and have that kind of train of thought 1095 00:49:48,090 --> 00:49:51,000 that says that if it came from a doctor, it must be safe. 1096 00:49:51,000 --> 00:49:55,270 Street drugs are bad, but prescription drugs are OK. 1097 00:49:55,270 --> 00:49:55,854 This is dumb. 1098 00:49:55,854 --> 00:49:57,270 The reason they're prescription is 1099 00:49:57,270 --> 00:50:00,820 because they can mess you up. 1100 00:50:00,820 --> 00:50:03,390 If they were harmless, you could get them over the counter. 1101 00:50:03,390 --> 00:50:04,359 More harmless. 1102 00:50:04,359 --> 00:50:06,150 Anyway, because these are psychostimulants, 1103 00:50:06,150 --> 00:50:08,850 because they trigger that reward pathway, 1104 00:50:08,850 --> 00:50:13,590 that dopamine pathway that goes up into the nucleus accumbens-- 1105 00:50:13,590 --> 00:50:17,460 that's this pathway that goes right in here-- 1106 00:50:17,460 --> 00:50:18,770 they are prone to abuse. 1107 00:50:18,770 --> 00:50:22,470 And there is a second, newish ADHD drug 1108 00:50:22,470 --> 00:50:25,570 called Strattera is the market name for it, 1109 00:50:25,570 --> 00:50:27,330 and it actually doesn't work on dopamine. 1110 00:50:27,330 --> 00:50:29,550 It works on norepinephrine. 1111 00:50:29,550 --> 00:50:32,460 So dopamine goes up into the nucleus accumbens here. 1112 00:50:32,460 --> 00:50:35,160 Cells in the nucleus accumbens, then, are mostly 1113 00:50:35,160 --> 00:50:36,220 norepinephrine cells. 1114 00:50:36,220 --> 00:50:40,930 And they also go up and have wide links to cortical areas. 1115 00:50:40,930 --> 00:50:43,614 And so often, you see a chain where 1116 00:50:43,614 --> 00:50:45,530 you've got groups of one neurotransmitter that 1117 00:50:45,530 --> 00:50:48,090 affect then a group of cells of another. 1118 00:50:48,090 --> 00:50:50,760 And you can affect the broad outcome 1119 00:50:50,760 --> 00:50:53,440 by intervening at different places along the chain. 1120 00:50:53,440 --> 00:50:55,110 So there's some people looking to see 1121 00:50:55,110 --> 00:51:01,440 if the seemingly dopaminergic portions of ADHD 1122 00:51:01,440 --> 00:51:04,950 are actually because of how dopamine then adequately 1123 00:51:04,950 --> 00:51:08,400 or inadequately activates the norepinephrine 1124 00:51:08,400 --> 00:51:10,980 at the next stage.