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 --> 00:00:16,790 from hundreds of MIT courses, visit MIT OpenCourseWare 7 00:00:16,790 --> 00:00:19,800 at ocw.mit.edu. 8 00:00:19,800 --> 00:00:24,360 ABBY NOYCE: Today, we are talking about attention. 9 00:00:24,360 --> 00:00:27,090 [HUMMING] And the the goal for this week, 10 00:00:27,090 --> 00:00:30,930 we've spent a week looking at how information 11 00:00:30,930 --> 00:00:37,082 gets into your brain through the visual modality in some detail. 12 00:00:37,082 --> 00:00:39,540 So this week, we're going to talk about what the brain does 13 00:00:39,540 --> 00:00:41,380 with all of this information. 14 00:00:41,380 --> 00:00:43,680 So we're going to talk today and tomorrow 15 00:00:43,680 --> 00:00:48,570 about attention today in an abstract from the cognitive 16 00:00:48,570 --> 00:00:50,070 perspective sense tomorrow, actually 17 00:00:50,070 --> 00:00:52,080 start looking at what goes on in your brain when 18 00:00:52,080 --> 00:00:53,990 you attend to information. 19 00:00:53,990 --> 00:00:56,490 I will probably talk a little bit about ADT and ADHD, 20 00:00:56,490 --> 00:00:58,830 and some of the medications that people use for those, 21 00:00:58,830 --> 00:01:01,180 and what we know about how they work, 22 00:01:01,180 --> 00:01:05,200 which is not nearly as much as one might like. 23 00:01:05,200 --> 00:01:07,500 And then on Wednesday and Thursday 24 00:01:07,500 --> 00:01:09,570 we'll talk about working memory, which 25 00:01:09,570 --> 00:01:12,390 is the holding information, what you're doing 26 00:01:12,390 --> 00:01:14,190 with it portion of memory. 27 00:01:14,190 --> 00:01:17,765 And then next week, we'll segue into longer term storage 28 00:01:17,765 --> 00:01:19,140 of knowledge and working with it. 29 00:01:19,140 --> 00:01:22,110 So attention. 30 00:01:22,110 --> 00:01:25,770 William James has this lovely description of attention. 31 00:01:25,770 --> 00:01:28,410 William James, remember, is the philosopher and psychologist 32 00:01:28,410 --> 00:01:30,140 from the late 1800s. 33 00:01:30,140 --> 00:01:33,300 He says attention is the taking possession by the mind 34 00:01:33,300 --> 00:01:36,210 in clear and vivid form of one out of what 35 00:01:36,210 --> 00:01:39,210 seems several simultaneously possible objects or trains 36 00:01:39,210 --> 00:01:40,230 of thought. 37 00:01:40,230 --> 00:01:42,120 It implies withdrawal from some things 38 00:01:42,120 --> 00:01:45,700 in order to deal effectively with others. 39 00:01:45,700 --> 00:01:48,660 And this is a good description of what it subjectively 40 00:01:48,660 --> 00:01:52,020 feels like to attend to a stimulus of some kind. 41 00:01:52,020 --> 00:01:53,924 We all know what paying attention feels like, 42 00:01:53,924 --> 00:01:56,340 or what it feels like when you don't have enough attention 43 00:01:56,340 --> 00:02:01,020 to go around all the different tasks you have to do. 44 00:02:01,020 --> 00:02:04,130 But James is very good at taking those subjective experiences we 45 00:02:04,130 --> 00:02:08,699 all have and putting them into words. 46 00:02:08,699 --> 00:02:11,090 A more modern definition of attention 47 00:02:11,090 --> 00:02:14,100 is basically just that you are working 48 00:02:14,100 --> 00:02:17,490 with one source of information that's coming in. 49 00:02:17,490 --> 00:02:19,290 So we talk about focused attention 50 00:02:19,290 --> 00:02:22,220 when you're really only dealing with one source of information. 51 00:02:22,220 --> 00:02:24,930 We'll talk about undivided attention, 52 00:02:24,930 --> 00:02:27,840 when you're trying to attend to multiple things at once. 53 00:02:27,840 --> 00:02:31,350 Usually when you try and do that, your performance on both 54 00:02:31,350 --> 00:02:32,070 falls off. 55 00:02:34,680 --> 00:02:37,630 You can't fully attend to two things at the same time. 56 00:02:37,630 --> 00:02:39,960 So you've got a limited amount of attention 57 00:02:39,960 --> 00:02:42,670 that you can give to the world around you. 58 00:02:42,670 --> 00:02:46,865 And when it runs out, you've got to drop something. 59 00:02:46,865 --> 00:02:49,360 AUDIENCE: It's like multi-threading. 60 00:02:49,360 --> 00:02:53,080 ABBY NOYCE: Yeah, except it's limited differently. 61 00:02:56,230 --> 00:02:58,931 But it's a reasonable analogy. 62 00:02:58,931 --> 00:02:59,930 Just be careful with it. 63 00:03:02,880 --> 00:03:04,540 What's multi-threading, [? Naman? ?] 64 00:03:04,540 --> 00:03:07,686 Would you like to explain to the rest of the class? 65 00:03:07,686 --> 00:03:10,460 AUDIENCE: It's when a crossover can only 66 00:03:10,460 --> 00:03:12,690 do one function at a time. 67 00:03:12,690 --> 00:03:14,640 So what the operating system does 68 00:03:14,640 --> 00:03:18,470 is [INAUDIBLE] programs at the same time, 69 00:03:18,470 --> 00:03:20,750 it switches between the programs really, really, fast. 70 00:03:20,750 --> 00:03:21,458 ABBY NOYCE: Yeah. 71 00:03:23,347 --> 00:03:25,430 So you guys probably know your computer slows down 72 00:03:25,430 --> 00:03:27,639 if you try to do a bajillion things at the same time. 73 00:03:27,639 --> 00:03:29,055 And that's one of the reasons why. 74 00:03:29,055 --> 00:03:31,510 It's working on task A, and then task B, and then task C, 75 00:03:31,510 --> 00:03:34,070 and cycling back to the beginning, which isn't quite 76 00:03:34,070 --> 00:03:38,540 how divided attention tends to work in humans. 77 00:03:38,540 --> 00:03:40,040 But it's the same problem, that when 78 00:03:40,040 --> 00:03:42,668 you're trying to do a lot of things at once, it falls apart. 79 00:03:42,668 --> 00:03:45,660 AUDIENCE: Is there a gender difference in that? 80 00:03:45,660 --> 00:03:47,870 ABBY NOYCE: On divided attention? 81 00:03:47,870 --> 00:03:48,495 AUDIENCE: Yeah. 82 00:03:48,495 --> 00:03:50,328 ABBY NOYCE: I don't know about that, per se. 83 00:03:50,328 --> 00:03:52,380 But there's definitely-- there are really 84 00:03:52,380 --> 00:03:54,180 well-documented gender differences 85 00:03:54,180 --> 00:03:57,450 in various kinds of spatial visual attention tasks, 86 00:03:57,450 --> 00:04:01,630 looking for a stimulus and a field of other things. 87 00:04:01,630 --> 00:04:03,540 And there's some really interesting research 88 00:04:03,540 --> 00:04:06,870 on this that says it correlates really heavily 89 00:04:06,870 --> 00:04:11,310 with the fact that way more guys than girls play video games. 90 00:04:11,310 --> 00:04:14,280 So if you take a male and a female subject pool, 91 00:04:14,280 --> 00:04:16,959 like standard college freshmen, you'll 92 00:04:16,959 --> 00:04:18,750 see the guys do a lot better than the girls 93 00:04:18,750 --> 00:04:21,000 on these kinds of tasks. 94 00:04:21,000 --> 00:04:22,950 If you then have them play-- 95 00:04:22,950 --> 00:04:25,650 I think honestly, they used a first-person shooter game, 96 00:04:25,650 --> 00:04:26,760 like "Quake" or something, where you're 97 00:04:26,760 --> 00:04:28,230 going around looking for enemies, 98 00:04:28,230 --> 00:04:31,240 jumping out from behind other targets, and things like that. 99 00:04:31,240 --> 00:04:34,350 You've got to really learn to be quick to respond 100 00:04:34,350 --> 00:04:35,820 to visual stimuli. 101 00:04:35,820 --> 00:04:38,490 And testing them again, the differences in performance 102 00:04:38,490 --> 00:04:39,760 were a lot smaller. 103 00:04:39,760 --> 00:04:42,000 So spatial attention is definitely 104 00:04:42,000 --> 00:04:47,024 something that can be learned and can be improved on. 105 00:04:47,024 --> 00:04:49,440 So we're going to start out talking about some cases where 106 00:04:49,440 --> 00:04:53,269 people's attention manages to drop an important piece 107 00:04:53,269 --> 00:04:54,310 of information coming in. 108 00:04:54,310 --> 00:04:57,480 This is a classic study from 1998. 109 00:04:57,480 --> 00:05:00,840 The experimenters, one of them went up 110 00:05:00,840 --> 00:05:03,360 to a pedestrian on campus and said, 111 00:05:03,360 --> 00:05:07,050 hey, can you tell me how to get to the student center, 112 00:05:07,050 --> 00:05:10,680 or I don't remember what the exact location 113 00:05:10,680 --> 00:05:11,790 they were looking for was. 114 00:05:11,790 --> 00:05:14,070 They asked for directions, and the passer by 115 00:05:14,070 --> 00:05:15,420 started giving these directions. 116 00:05:15,420 --> 00:05:19,710 And mid-conversation, two people carrying a door 117 00:05:19,710 --> 00:05:21,330 cut right in between the conversations 118 00:05:21,330 --> 00:05:23,160 so that they were cut off from each other 119 00:05:23,160 --> 00:05:25,620 for a couple of seconds. 120 00:05:25,620 --> 00:05:27,120 And the two people carrying the door 121 00:05:27,120 --> 00:05:28,710 were, in fact, part of the experiment. 122 00:05:28,710 --> 00:05:30,960 And the person who'd originally stopped the pedestrian 123 00:05:30,960 --> 00:05:33,030 and asked the question swapped places 124 00:05:33,030 --> 00:05:34,920 with one of the people who was with the door 125 00:05:34,920 --> 00:05:38,700 so that when the door had left and the random passerby who 126 00:05:38,700 --> 00:05:43,140 was being asked for directions could continue his explanation. 127 00:05:43,140 --> 00:05:44,610 Then the person he was talking to 128 00:05:44,610 --> 00:05:46,746 was a completely different person. 129 00:05:46,746 --> 00:05:48,570 AUDIENCE: In the same clothes? 130 00:05:48,570 --> 00:05:51,161 ABBY NOYCE: Roughly, but not identical. 131 00:05:51,161 --> 00:05:53,160 You can see the photo on the bottom right there. 132 00:05:53,160 --> 00:05:54,077 So there are two guys. 133 00:05:54,077 --> 00:05:56,034 They're both dressed in like a dark outer layer 134 00:05:56,034 --> 00:05:56,980 and a white shirt on. 135 00:05:56,980 --> 00:05:58,730 But they really don't look anything alike. 136 00:05:58,730 --> 00:05:59,882 AUDIENCE: Yeah, they do. 137 00:05:59,882 --> 00:06:01,590 ABBY NOYCE: They look a little bit alike. 138 00:06:01,590 --> 00:06:03,672 They don't look a lot alike. 139 00:06:03,672 --> 00:06:06,172 AUDIENCE: You'd be able to tell the difference [INAUDIBLE].. 140 00:06:06,172 --> 00:06:08,422 ABBY NOYCE: I think you'd be able to tell a difference 141 00:06:08,422 --> 00:06:09,410 in a color photo too. 142 00:06:09,410 --> 00:06:11,326 One of them's wearing a jacket and one of them 143 00:06:11,326 --> 00:06:12,380 is wearing a sweater. 144 00:06:12,380 --> 00:06:13,671 AUDIENCE: One of them's balder. 145 00:06:13,671 --> 00:06:15,730 ABBY NOYCE: One of them is way balder. 146 00:06:15,730 --> 00:06:17,238 So anyway, so about-- 147 00:06:17,238 --> 00:06:18,494 [INTERPOSING VOICES] 148 00:06:18,494 --> 00:06:20,910 ABBY NOYCE: When they finished you know getting directions 149 00:06:20,910 --> 00:06:23,439 from the pedestrian, they said, OK, so 150 00:06:23,439 --> 00:06:25,980 we're required to tell you we're doing a psychological study, 151 00:06:25,980 --> 00:06:27,090 blah, blah, blah. 152 00:06:27,090 --> 00:06:29,800 If you don't want us to use your data, we won't use your data. 153 00:06:29,800 --> 00:06:30,670 And with some questions. 154 00:06:30,670 --> 00:06:31,950 One of the questions asked was, did 155 00:06:31,950 --> 00:06:34,680 you notice that I'm not the same person who originally stopped 156 00:06:34,680 --> 00:06:36,430 you and ask for directions. 157 00:06:36,430 --> 00:06:40,000 And about 50% of pedestrians said, yeah, I noticed. 158 00:06:40,000 --> 00:06:43,290 And the other half said, nope. 159 00:06:43,290 --> 00:06:48,660 So this is a classic example of your attentional system 160 00:06:48,660 --> 00:06:51,314 disregarding a big chunk of information, 161 00:06:51,314 --> 00:06:53,730 the appearance of the person, the appearance of the person 162 00:06:53,730 --> 00:06:57,000 you're talking to in order because it's 163 00:06:57,000 --> 00:06:59,490 irrelevant in some ways. 164 00:06:59,490 --> 00:07:02,725 Anyone here work retail behind a register? 165 00:07:02,725 --> 00:07:04,350 I've done that, and I know that there's 166 00:07:04,350 --> 00:07:06,120 been days when it's busy, and you're just pushing people 167 00:07:06,120 --> 00:07:08,560 through as fast as you can, and you ring somebody up, 168 00:07:08,560 --> 00:07:10,270 and you hand them your change, and you say, have a nice day. 169 00:07:10,270 --> 00:07:12,400 And right about then when I make eye contact, 170 00:07:12,400 --> 00:07:14,631 I notice that this is my mom's best friend. 171 00:07:14,631 --> 00:07:16,380 Just didn't see it before because I was so 172 00:07:16,380 --> 00:07:18,120 focused on other stuff. 173 00:07:18,120 --> 00:07:20,460 So one of the big things that attention does 174 00:07:20,460 --> 00:07:23,312 is it disregards some of the information that's coming in 175 00:07:23,312 --> 00:07:25,770 because otherwise, there's just way too much for your brain 176 00:07:25,770 --> 00:07:28,860 to process at any point in time. 177 00:07:28,860 --> 00:07:31,110 All right, so this is a phenomenon 178 00:07:31,110 --> 00:07:33,300 called change blindness, where something changes 179 00:07:33,300 --> 00:07:35,988 and you just don't see it happen. 180 00:07:35,988 --> 00:07:38,954 AUDIENCE: Is there any way to build your attention span or-- 181 00:07:38,954 --> 00:07:40,620 ABBY NOYCE: What sort of attention span? 182 00:07:43,428 --> 00:07:47,020 AUDIENCE: Is there a way to be able to multitask accurately, 183 00:07:47,020 --> 00:07:48,450 or-- 184 00:07:48,450 --> 00:07:50,400 ABBY NOYCE: Better. 185 00:07:50,400 --> 00:07:52,340 So when you're doing divided attention tasks-- 186 00:07:52,340 --> 00:07:54,340 and we'll come around to this again in a minute. 187 00:07:54,340 --> 00:07:57,510 But one of the things that makes a lot of difference 188 00:07:57,510 --> 00:08:01,860 is how well you know the tasks we're doing. 189 00:08:01,860 --> 00:08:05,974 So, for example, novice versus experienced drivers, 190 00:08:05,974 --> 00:08:07,890 novice drivers who are asked to change a radio 191 00:08:07,890 --> 00:08:10,348 station or something will look away from the road for a lot 192 00:08:10,348 --> 00:08:12,600 longer than experienced drivers will. 193 00:08:12,600 --> 00:08:16,290 And that's partly because just one more familiarity with one 194 00:08:16,290 --> 00:08:19,590 or better with both tasks means that it requires less attention 195 00:08:19,590 --> 00:08:20,640 to do them. 196 00:08:20,640 --> 00:08:22,260 And it also gives you better judgment 197 00:08:22,260 --> 00:08:24,821 about how to switch your attention back and forth 198 00:08:24,821 --> 00:08:25,320 optimally. 199 00:08:25,320 --> 00:08:28,710 So experienced drivers have a good sense of how long 200 00:08:28,710 --> 00:08:31,623 can I look away from the road before bad stuff happens more 201 00:08:31,623 --> 00:08:33,789 so than someone who's only been driving for a couple 202 00:08:33,789 --> 00:08:36,389 months or a year or whatever. 203 00:08:36,389 --> 00:08:39,559 AUDIENCE: [INAUDIBLE]. 204 00:08:39,559 --> 00:08:42,059 ABBY NOYCE: Anyone here know anyone who texts while driving? 205 00:08:42,059 --> 00:08:43,010 My sister does that. 206 00:08:43,010 --> 00:08:43,687 And I will not-- 207 00:08:43,687 --> 00:08:45,020 she does it when I'm in the car. 208 00:08:45,020 --> 00:08:46,772 I take the phone away from her. 209 00:08:46,772 --> 00:08:48,740 AUDIENCE: It's scary. 210 00:08:48,740 --> 00:08:50,180 ABBY NOYCE: I don't get it. 211 00:08:50,180 --> 00:08:52,610 OK, so this is one kind of change blindness. 212 00:08:52,610 --> 00:08:56,180 Another kind of change blindness, information 213 00:08:56,180 --> 00:08:57,730 is relevant and ignores other input. 214 00:08:57,730 --> 00:08:59,540 This is what we just said. 215 00:08:59,540 --> 00:09:01,880 So that's attention that's centrally controlled, where 216 00:09:01,880 --> 00:09:03,547 you're deciding what sort of stimulus 217 00:09:03,547 --> 00:09:05,630 you're looking for in some way, either consciously 218 00:09:05,630 --> 00:09:07,100 or unconsciously. 219 00:09:07,100 --> 00:09:09,009 If you're at a crowded party, and you're 220 00:09:09,009 --> 00:09:10,550 looking for your best friend, and you 221 00:09:10,550 --> 00:09:12,074 know that she has a red shirt on, 222 00:09:12,074 --> 00:09:14,240 then you're going to be looking, scanning the crowd, 223 00:09:14,240 --> 00:09:15,960 looking for something red. 224 00:09:15,960 --> 00:09:19,574 And if she has unbeknownst to you changed into a black shirt, 225 00:09:19,574 --> 00:09:21,740 you're going to have a much harder time finding her. 226 00:09:21,740 --> 00:09:24,281 You're not going to be attending to the people wearing black. 227 00:09:24,281 --> 00:09:25,980 You're just going to disregard them. 228 00:09:29,326 --> 00:09:35,320 So top down in this case means centrally controlled, right? 229 00:09:35,320 --> 00:09:39,430 It's not being-- as opposed to a stimulus driven attention. 230 00:09:39,430 --> 00:09:41,620 So alternatively, if you guys were all sitting here 231 00:09:41,620 --> 00:09:44,530 focusing on me, and the tripod in the back of the classroom 232 00:09:44,530 --> 00:09:46,840 fell over and went bang, you'd probably all turn around 233 00:09:46,840 --> 00:09:47,830 and look, right? 234 00:09:47,830 --> 00:09:50,812 So that's a stimulus, grabbing onto your attention, 235 00:09:50,812 --> 00:09:52,270 and pulling it away from whatever's 236 00:09:52,270 --> 00:09:54,311 going on that you were trying to focus on before. 237 00:09:57,190 --> 00:10:00,640 All right, here's a task for you guys. 238 00:10:00,640 --> 00:10:04,450 We're going to see a white team and a black team 239 00:10:04,450 --> 00:10:07,090 dribbling some basketballs. 240 00:10:07,090 --> 00:10:10,120 And I want you guys to watch the video 241 00:10:10,120 --> 00:10:14,287 and count how many times the kids wearing white 242 00:10:14,287 --> 00:10:15,120 pass the basketball. 243 00:10:15,120 --> 00:10:15,760 Are you ready? 244 00:10:18,732 --> 00:10:19,232 Go. 245 00:10:44,550 --> 00:10:45,637 How many times? 246 00:10:45,637 --> 00:10:46,551 AUDIENCE: 14. 247 00:10:46,551 --> 00:10:48,330 ABBY NOYCE: 14 [INAUDIBLE]? 248 00:10:48,330 --> 00:10:51,030 Who's seen this demo before? 249 00:10:51,030 --> 00:10:52,770 OK, anyone notice anything weird? 250 00:10:55,670 --> 00:10:56,610 Watch it again. 251 00:10:56,610 --> 00:10:58,110 Don't try and count the basketballs. 252 00:11:01,860 --> 00:11:04,300 Just watch, see if you notice anything odd going on. 253 00:11:12,940 --> 00:11:14,860 [LAUGHTER] 254 00:11:14,860 --> 00:11:17,710 Ready? 255 00:11:17,710 --> 00:11:19,174 [LAUGHTER] 256 00:11:19,174 --> 00:11:21,620 AUDIENCE: Wow. 257 00:11:21,620 --> 00:11:24,376 ABBY NOYCE: Your attentional processes 258 00:11:24,376 --> 00:11:26,750 decide that they want to watch this piece of information, 259 00:11:26,750 --> 00:11:29,420 and they just throw out everything else. 260 00:11:29,420 --> 00:11:31,100 AUDIENCE: That was there the whole time? 261 00:11:31,100 --> 00:11:32,850 ABBY NOYCE: That was the exact same video. 262 00:11:32,850 --> 00:11:34,750 AUDIENCE: Since it was black, and all the other [INAUDIBLE] 263 00:11:34,750 --> 00:11:37,333 you got confused because you're not paying attention to those. 264 00:11:37,333 --> 00:11:39,350 ABBY NOYCE: Yeah, you disregarded all 265 00:11:39,350 --> 00:11:40,980 of the people wearing black. 266 00:11:40,980 --> 00:11:42,650 And that's, I think, a big part of what happens there. 267 00:11:42,650 --> 00:11:43,316 I think it was-- 268 00:11:43,316 --> 00:11:46,577 I don't know-- a white gorilla, or a snow leopard costume, 269 00:11:46,577 --> 00:11:48,910 or something, then you'd be a lot more likely to see it. 270 00:11:48,910 --> 00:11:50,868 But because it fits in with this other category 271 00:11:50,868 --> 00:11:52,730 that you've decided to disregard, 272 00:11:52,730 --> 00:11:56,030 then your system just throws it out. 273 00:11:56,030 --> 00:11:58,220 So, again, disregarding a good chunk 274 00:11:58,220 --> 00:12:01,532 of the input that comes in. 275 00:12:01,532 --> 00:12:05,600 [INAUDIBLE] grabs onto to information 276 00:12:05,600 --> 00:12:10,030 that's relevant in some way and ignores other stuff. 277 00:12:10,030 --> 00:12:12,920 It breaks when you've got lots of different stimuli trying 278 00:12:12,920 --> 00:12:14,930 to get your attention at the same time. 279 00:12:14,930 --> 00:12:17,150 If you're trying to do your homework, 280 00:12:17,150 --> 00:12:18,650 and holler at your little brother, 281 00:12:18,650 --> 00:12:22,490 and watch a TV show all at once, probably one of these things 282 00:12:22,490 --> 00:12:25,730 is not going to get done very well. 283 00:12:25,730 --> 00:12:26,501 Everybody here? 284 00:12:26,501 --> 00:12:27,000 Good. 285 00:12:27,000 --> 00:12:27,530 Thank you. 286 00:12:31,240 --> 00:12:33,260 The other interesting question for attention 287 00:12:33,260 --> 00:12:38,750 is what constitutes information that's relevant? 288 00:12:38,750 --> 00:12:43,610 So one kind of immediate [? instances, ?] OK, probably 289 00:12:43,610 --> 00:12:46,070 in an evolutionary sense, information that in some way 290 00:12:46,070 --> 00:12:48,480 affects your survival, right? 291 00:12:48,480 --> 00:12:51,109 We know that our attention is grabbed by unexpected stimuli 292 00:12:51,109 --> 00:12:51,650 of some sort. 293 00:12:51,650 --> 00:12:53,441 If there's a loud noise, or something moves 294 00:12:53,441 --> 00:12:55,160 in your peripheral vision, all of these 295 00:12:55,160 --> 00:12:58,121 are things that will cause you to direct your attention. 296 00:12:58,121 --> 00:12:59,870 On the other hand, there's a lot of things 297 00:12:59,870 --> 00:13:01,370 that are really relevant to your survival 298 00:13:01,370 --> 00:13:02,828 that we don't pay any attention to, 299 00:13:02,828 --> 00:13:04,430 your breathing, your heart rate. 300 00:13:04,430 --> 00:13:06,638 All of these are things that are, in fact, controlled 301 00:13:06,638 --> 00:13:07,490 by your brain. 302 00:13:07,490 --> 00:13:10,430 But we don't put any of our conscious attention 303 00:13:10,430 --> 00:13:13,532 into making sure that these things keep happening. 304 00:13:13,532 --> 00:13:16,780 AUDIENCE: Would it help us if we had to. 305 00:13:16,780 --> 00:13:17,760 ABBY NOYCE: Yeah, see? 306 00:13:17,760 --> 00:13:22,520 You wouldn't have any intention left over to do anything else. 307 00:13:22,520 --> 00:13:26,970 So to some extent, weird stimuli that are relevant to survival, 308 00:13:26,970 --> 00:13:31,180 but because we are creatures of these great big 309 00:13:31,180 --> 00:13:32,830 powerful cognitive capabilities, we 310 00:13:32,830 --> 00:13:35,740 can direct our attention to whatever we want, 311 00:13:35,740 --> 00:13:38,830 which is not directly-- 312 00:13:38,830 --> 00:13:40,690 this attention got grabbed-- 313 00:13:40,690 --> 00:13:42,985 is not directly relevant to-- 314 00:13:42,985 --> 00:13:44,580 AUDIENCE: Was that [INAUDIBLE]? 315 00:13:44,580 --> 00:13:45,205 ABBY NOYCE: No. 316 00:13:49,219 --> 00:13:50,280 AUDIENCE: That's strange. 317 00:13:50,280 --> 00:13:51,363 ABBY NOYCE: It was clever. 318 00:13:53,478 --> 00:13:56,180 And my train of thought is broken. 319 00:13:56,180 --> 00:13:58,930 So attention that's relevant to survival, 320 00:13:58,930 --> 00:14:01,180 but we also have these big, strong powerful brains. 321 00:14:01,180 --> 00:14:05,416 And having all of these abilities is adaptive, 322 00:14:05,416 --> 00:14:06,790 but we can do things to them that 323 00:14:06,790 --> 00:14:09,041 are not necessarily adaptive. 324 00:14:09,041 --> 00:14:10,790 Yes, that's where we were going with this. 325 00:14:10,790 --> 00:14:13,330 OK. 326 00:14:13,330 --> 00:14:16,170 AUDIENCE: Like how you have a flashing light 327 00:14:16,170 --> 00:14:18,954 to grab someone's attention [INAUDIBLE].. 328 00:14:18,954 --> 00:14:21,460 If you flash it too rapidly, then it's 329 00:14:21,460 --> 00:14:23,800 almost as if it was just shining constantly, 330 00:14:23,800 --> 00:14:26,002 and then attention fades. 331 00:14:26,002 --> 00:14:27,960 ABBY NOYCE: Well, is that an intentional thing, 332 00:14:27,960 --> 00:14:30,024 or is that a perceptual limit? 333 00:14:30,024 --> 00:14:31,440 Because that's going to have to do 334 00:14:31,440 --> 00:14:34,770 in part with how fast a stimulus your rods and cones can 335 00:14:34,770 --> 00:14:35,270 perceive. 336 00:14:35,270 --> 00:14:37,290 And there is certainly something to that too. 337 00:14:37,290 --> 00:14:39,150 A computer monitor, where if you look at it sideways, 338 00:14:39,150 --> 00:14:40,233 you can see it flickering. 339 00:14:40,233 --> 00:14:42,800 But if you look straight at it, it looks OK. 340 00:14:42,800 --> 00:14:46,300 Less of a problem with flat screens, but with old monitors. 341 00:14:46,300 --> 00:14:50,400 AUDIENCE: [INAUDIBLE] you can see change-- 342 00:14:50,400 --> 00:14:54,750 ABBY NOYCE: Yeah, because your rods reset. 343 00:14:54,750 --> 00:14:56,250 They can respond to a new stimulus 344 00:14:56,250 --> 00:14:59,310 faster than the cones can. 345 00:14:59,310 --> 00:15:03,630 So and we know our peripheral vision is 346 00:15:03,630 --> 00:15:05,800 more sensitive to motion in a lot of ways. 347 00:15:05,800 --> 00:15:08,310 So this is one that goes with that. 348 00:15:08,310 --> 00:15:11,274 AUDIENCE: So if someone's in front of you, 349 00:15:11,274 --> 00:15:13,818 or if someone is attacking in front of you, 350 00:15:13,818 --> 00:15:16,248 would you respond as fast if someone tries 351 00:15:16,248 --> 00:15:18,680 to attack you from the side? 352 00:15:18,680 --> 00:15:20,780 ABBY NOYCE: Probably not on an order that 353 00:15:20,780 --> 00:15:27,139 makes any kind of difference in that real life circumstance, 354 00:15:27,139 --> 00:15:29,430 in part, because a person who's already in front of you 355 00:15:29,430 --> 00:15:31,960 probably has more of your attention in the first place 356 00:15:31,960 --> 00:15:33,735 than somebody on your side. 357 00:15:33,735 --> 00:15:36,360 But we're talking differences on the orders of a couple hundred 358 00:15:36,360 --> 00:15:39,380 milliseconds. 359 00:15:39,380 --> 00:15:40,460 Less than that for this. 360 00:15:40,460 --> 00:15:43,913 We're talking 70 hertz monitor or something 361 00:15:43,913 --> 00:15:45,954 that looks good when you look at it and flickers. 362 00:15:45,954 --> 00:15:47,890 AUDIENCE: If you look at that with peripheral vision, 363 00:15:47,890 --> 00:15:48,389 would you-- 364 00:15:48,389 --> 00:15:51,310 ABBY NOYCE: 70, for me. 365 00:15:51,310 --> 00:15:51,940 Yeah. 366 00:15:51,940 --> 00:15:55,730 70 for me is about where I stop having flicker issues and CRTs. 367 00:15:55,730 --> 00:15:58,700 Old school 60 hertz ones give me headaches. 368 00:15:58,700 --> 00:16:03,470 Anyway, so come in. 369 00:16:03,470 --> 00:16:05,090 There we go. 370 00:16:05,090 --> 00:16:08,060 So that's a demo of what's called attentional blink. 371 00:16:08,060 --> 00:16:12,470 So it is a short period of time where incoming information just 372 00:16:12,470 --> 00:16:14,360 doesn't get registered. 373 00:16:14,360 --> 00:16:16,520 So in this case, it's after there's one. 374 00:16:16,520 --> 00:16:18,200 After you see one target, there's 375 00:16:18,200 --> 00:16:20,120 a period at which that information 376 00:16:20,120 --> 00:16:24,530 about the next target that's coming in does not get seen. 377 00:16:24,530 --> 00:16:27,290 Another good example of this is that as we look at things, 378 00:16:27,290 --> 00:16:29,510 we move our eyes around, right? 379 00:16:29,510 --> 00:16:31,010 We have what are called saccades, 380 00:16:31,010 --> 00:16:34,760 where you jump from one focal point to another focal point. 381 00:16:34,760 --> 00:16:36,884 Your eyes, you don't register any 382 00:16:36,884 --> 00:16:39,050 of the information that comes in while your eyes are 383 00:16:39,050 --> 00:16:41,330 moving like that. 384 00:16:41,330 --> 00:16:44,535 And a good demo of this is if you 385 00:16:44,535 --> 00:16:46,160 look at the person sitting next to you, 386 00:16:46,160 --> 00:16:48,531 and have them look at one of your eyes, and the other 387 00:16:48,531 --> 00:16:50,030 of your eyes, and go back and forth, 388 00:16:50,030 --> 00:16:52,350 and you'll see their eyes moving just fine. 389 00:16:52,350 --> 00:16:54,050 But if you try and do this with yourself 390 00:16:54,050 --> 00:16:55,014 in the mirror at home-- 391 00:16:55,014 --> 00:16:56,930 look at your left eye, look at your right eye, 392 00:16:56,930 --> 00:16:59,427 look at your left eye in the mirror in your reflection, 393 00:16:59,427 --> 00:17:01,010 you won't actually see your eyes move. 394 00:17:01,010 --> 00:17:03,590 You'll just see yourself looking at each of those. 395 00:17:03,590 --> 00:17:06,200 Your brain just throws out all of the visual information 396 00:17:06,200 --> 00:17:08,339 that's coming in during that jump 397 00:17:08,339 --> 00:17:10,730 while the world is sliding across your retina at really 398 00:17:10,730 --> 00:17:12,200 high speed. 399 00:17:12,200 --> 00:17:14,191 It doesn't use it. 400 00:17:14,191 --> 00:17:14,690 Try it. 401 00:17:14,690 --> 00:17:16,565 Turn around and look at somebody next to you, 402 00:17:16,565 --> 00:17:19,349 and have them move their eyes back and forth. 403 00:17:19,349 --> 00:17:22,669 Like here, Helen and Natasha, Natasha, 404 00:17:22,669 --> 00:17:24,710 look at Helen's left eye, and then her right eye, 405 00:17:24,710 --> 00:17:27,126 and then her left eye, and go back and forth between them. 406 00:17:27,126 --> 00:17:28,980 Helen, can you see her eyes moving? 407 00:17:28,980 --> 00:17:29,940 AUDIENCE: Yes. 408 00:17:29,940 --> 00:17:31,070 ABBY NOYCE: Try it in front of the mirror at home. 409 00:17:31,070 --> 00:17:32,320 You won't see it. 410 00:17:32,320 --> 00:17:35,260 It's cool. 411 00:17:35,260 --> 00:17:40,940 So this is a case of attentional blink where your brain just-- 412 00:17:40,940 --> 00:17:44,270 not because there's a target that is eating your attention, 413 00:17:44,270 --> 00:17:46,970 but because your visual system, for whatever reason, 414 00:17:46,970 --> 00:17:50,660 always just throws out whatever is coming in 415 00:17:50,660 --> 00:17:54,420 while you're saccading from one point of focus to another. 416 00:17:58,354 --> 00:18:00,270 Why do you think your visual system does that? 417 00:18:09,713 --> 00:18:12,149 AUDIENCE: Wait, what was the question? 418 00:18:12,149 --> 00:18:13,690 AUDIENCE: For computational purposes? 419 00:18:13,690 --> 00:18:15,648 ABBY NOYCE: Maybe, so that it has some downtime 420 00:18:15,648 --> 00:18:17,854 to think about things, you mean? 421 00:18:17,854 --> 00:18:20,800 AUDIENCE: Well, even if it kept information, 422 00:18:20,800 --> 00:18:23,600 wouldn't you just see a really big-- a blur or something 423 00:18:23,600 --> 00:18:28,112 because your eyes would be [INAUDIBLE] so fast, 424 00:18:28,112 --> 00:18:30,500 [INAUDIBLE] keep track of it. 425 00:18:30,500 --> 00:18:31,310 ABBY NOYCE: Yeah. 426 00:18:31,310 --> 00:18:33,740 So the usual conventional wisdom about this 427 00:18:33,740 --> 00:18:37,340 is that tracking an object as it moves 428 00:18:37,340 --> 00:18:39,620 across your field of vision is-- it takes some work. 429 00:18:39,620 --> 00:18:41,660 It takes some matching up, and that it just 430 00:18:41,660 --> 00:18:44,840 isn't worth the processing power that it would take to track 431 00:18:44,840 --> 00:18:47,120 everything as it moves around. 432 00:18:47,120 --> 00:18:49,010 But you do it OK. 433 00:18:49,010 --> 00:18:50,546 You just do a piecewise. 434 00:18:50,546 --> 00:18:51,920 You match up an object over here, 435 00:18:51,920 --> 00:18:53,461 and then if I move my eyes over here, 436 00:18:53,461 --> 00:18:55,770 I can still match the same objects up. 437 00:18:55,770 --> 00:18:59,405 So we don't really know for sure why the visual system does it. 438 00:18:59,405 --> 00:19:01,280 But it's pretty well documented that it does. 439 00:19:01,280 --> 00:19:03,560 If you have people who are reading something, 440 00:19:03,560 --> 00:19:06,644 when you read, your eyes jump from word to word. 441 00:19:06,644 --> 00:19:08,810 And if you have them fitted with an eye tracker that 442 00:19:08,810 --> 00:19:12,870 can see them do this that's fast responding enough to project 443 00:19:12,870 --> 00:19:15,800 something like mid-saccade, then you can have 444 00:19:15,800 --> 00:19:17,900 their story that they're reading be interrupted 445 00:19:17,900 --> 00:19:21,390 by giant presentation of an apple or something, 446 00:19:21,390 --> 00:19:24,110 and they just won't see it, as long 447 00:19:24,110 --> 00:19:26,669 as it's gone back to the original story that 448 00:19:26,669 --> 00:19:28,960 was on the screen by the time that saccade is finished. 449 00:19:28,960 --> 00:19:33,930 So we're talking single-frame presentation. 450 00:19:33,930 --> 00:19:35,870 OK, so we have attentional blank, 451 00:19:35,870 --> 00:19:39,432 which is this temporal breaking of attention. 452 00:19:39,432 --> 00:19:41,792 AUDIENCE: Is that how subliminal messages are sometimes 453 00:19:41,792 --> 00:19:43,564 put in television? 454 00:19:43,564 --> 00:19:44,980 ABBY NOYCE: So subliminal messages 455 00:19:44,980 --> 00:19:46,438 aren't necessarily taking advantage 456 00:19:46,438 --> 00:19:48,240 of that particular saccading. 457 00:19:48,240 --> 00:19:53,520 They're just too quick for your brain to see and process. 458 00:19:53,520 --> 00:19:56,460 So getting stuff, a stimulus into your system 459 00:19:56,460 --> 00:19:58,380 takes a certain amount of time. 460 00:19:58,380 --> 00:20:00,600 And people who believe that subliminal, 461 00:20:00,600 --> 00:20:04,830 that the single frame subliminal messaging works, 462 00:20:04,830 --> 00:20:08,310 believe that it gets input and processed at some level, 463 00:20:08,310 --> 00:20:10,672 but doesn't make it into conscious perception. 464 00:20:10,672 --> 00:20:11,910 AUDIENCE: Wait, what is subliminal messages? 465 00:20:11,910 --> 00:20:13,951 AUDIENCE: So there's kind of the classic version. 466 00:20:13,951 --> 00:20:17,580 The classic urban legend story of this is the movie theater 467 00:20:17,580 --> 00:20:22,400 that during the previews enters [INAUDIBLE] single frames that 468 00:20:22,400 --> 00:20:25,950 say popcorn is delicious, and you're thirsty, drink Coke, 469 00:20:25,950 --> 00:20:28,890 and whatever else, and sees their concession sales-- 470 00:20:28,890 --> 00:20:31,170 concession stand sales go way up, 471 00:20:31,170 --> 00:20:34,056 which I'm pretty sure is just an urban legend. 472 00:20:34,056 --> 00:20:37,542 AUDIENCE: [INAUDIBLE]. 473 00:20:37,542 --> 00:20:40,640 Oh, there was a [? stat ?] problem 474 00:20:40,640 --> 00:20:42,682 they had back during the school year. 475 00:20:42,682 --> 00:20:46,424 And it mentioned that. 476 00:20:46,424 --> 00:20:48,340 How would you design an experiment with that-- 477 00:20:48,340 --> 00:20:49,270 ABBY NOYCE: To see if this works? 478 00:20:49,270 --> 00:20:49,895 AUDIENCE: Yeah. 479 00:20:49,895 --> 00:20:51,936 ABBY NOYCE: Yeah, and it's a worthwhile question. 480 00:20:51,936 --> 00:20:53,740 But there isn't any really good solid data 481 00:20:53,740 --> 00:20:55,700 showing that it does work. 482 00:20:55,700 --> 00:20:56,890 It's like one of those-- 483 00:20:56,890 --> 00:20:57,980 oh, that would be cool. 484 00:20:57,980 --> 00:20:59,229 That would seem to make sense. 485 00:20:59,229 --> 00:21:02,040 But it doesn't seem to actually happen cases. 486 00:21:02,040 --> 00:21:03,540 AUDIENCE: Well, that's good to know. 487 00:21:03,540 --> 00:21:04,956 ABBY NOYCE: Isn't it good to know? 488 00:21:06,996 --> 00:21:08,370 Subtler kind of subliminal stuff, 489 00:21:08,370 --> 00:21:11,670 like product placement in movies, does seem to work. 490 00:21:11,670 --> 00:21:16,140 So if you have a video, and the main character in it 491 00:21:16,140 --> 00:21:18,762 is drinking a soda, and it happens to be a Coke, 492 00:21:18,762 --> 00:21:21,220 and I give you a choice between Coke and Sprite afterwards, 493 00:21:21,220 --> 00:21:23,010 you're more likely to pick the Coke. 494 00:21:23,010 --> 00:21:24,944 But that's a very different kind of-- 495 00:21:24,944 --> 00:21:25,860 AUDIENCE: [INAUDIBLE]. 496 00:21:25,860 --> 00:21:26,550 ABBY NOYCE: Yeah. 497 00:21:26,550 --> 00:21:28,008 So that seems to actually influence 498 00:21:28,008 --> 00:21:30,060 the choices people make, if it's a character 499 00:21:30,060 --> 00:21:31,744 that they like and identify with. 500 00:21:31,744 --> 00:21:33,660 I don't know if anyone's done any studies were 501 00:21:33,660 --> 00:21:37,980 a villain drinks Coke and does it make their sales go down. 502 00:21:37,980 --> 00:21:39,930 But probably Coke would not pay to get 503 00:21:39,930 --> 00:21:43,340 their products placed in that kind of context, so we'll see. 504 00:21:43,340 --> 00:21:46,470 Anyway, so attention breaks-- 505 00:21:46,470 --> 00:21:47,586 yes? 506 00:21:47,586 --> 00:21:52,260 AUDIENCE: I have a question about subliminal messaging 507 00:21:52,260 --> 00:21:53,346 with audio. 508 00:21:53,346 --> 00:21:54,205 [INAUDIBLE]? 509 00:21:54,205 --> 00:21:55,830 ABBY NOYCE: If you play this backwards, 510 00:21:55,830 --> 00:21:56,520 you can hear something? 511 00:21:56,520 --> 00:21:57,155 AUDIENCE: Yeah. 512 00:21:57,155 --> 00:21:57,655 Yeah. 513 00:21:57,655 --> 00:22:00,926 So does that actually work, or does that-- 514 00:22:00,926 --> 00:22:01,800 AUDIENCE: Wait, what? 515 00:22:01,800 --> 00:22:03,780 ABBY NOYCE: The majority of the, if you play this backwards 516 00:22:03,780 --> 00:22:05,310 you can hear something things, I think 517 00:22:05,310 --> 00:22:07,518 can get chalked up to the fact that humans are really 518 00:22:07,518 --> 00:22:09,420 good pattern recognizers. 519 00:22:09,420 --> 00:22:11,220 And so we tend to find patterns even 520 00:22:11,220 --> 00:22:13,380 when there isn't one there. 521 00:22:13,380 --> 00:22:13,962 AUDIENCE: OK. 522 00:22:13,962 --> 00:22:15,420 ABBY NOYCE: I don't think it works, 523 00:22:15,420 --> 00:22:17,086 because anything played backwards sounds 524 00:22:17,086 --> 00:22:18,480 like gibberish, right? 525 00:22:18,480 --> 00:22:20,350 You can't make sense out of-- 526 00:22:20,350 --> 00:22:24,970 AUDIENCE: I can speak backwards if I really try to. 527 00:22:24,970 --> 00:22:26,060 ABBY NOYCE: Yeah? 528 00:22:26,060 --> 00:22:28,244 I think this might require demonstration. 529 00:22:28,244 --> 00:22:30,970 AUDIENCE: Do it. 530 00:22:30,970 --> 00:22:34,692 AUDIENCE: OK, the backwards is composed of two phonemes. 531 00:22:34,692 --> 00:22:36,478 So you just say phoneme backwards. 532 00:22:36,478 --> 00:22:37,644 ABBY NOYCE: So it's just of? 533 00:22:37,644 --> 00:22:38,700 AUDIENCE: Of. 534 00:22:38,700 --> 00:22:40,200 ABBY NOYCE: Is that what it actually 535 00:22:40,200 --> 00:22:42,110 sounds like if you run it backwards on an audio tape? 536 00:22:42,110 --> 00:22:42,693 AUDIENCE: Yes. 537 00:22:42,693 --> 00:22:43,871 ABBY NOYCE: OK. 538 00:22:43,871 --> 00:22:46,437 AUDIENCE: You have to say the phoneme backwards, basically. 539 00:22:46,437 --> 00:22:47,770 AUDIENCE: [INAUDIBLE] backwards? 540 00:22:47,770 --> 00:22:49,600 ABBY NOYCE: Phonemes are the units of sound that make up-- 541 00:22:49,600 --> 00:22:51,390 AUDIENCE: Split it up into each sound. 542 00:22:51,390 --> 00:22:57,044 So the is made up of two sounds, the th sound and the uh sound. 543 00:22:57,044 --> 00:22:58,470 So yous say it backwards. 544 00:22:58,470 --> 00:23:02,162 So [INAUDIBLE] would be the backwards. 545 00:23:02,162 --> 00:23:03,120 ABBY NOYCE: How about-- 546 00:23:03,120 --> 00:23:05,271 AUDIENCE: Wait, isn't it [INAUDIBLE]?? 547 00:23:05,271 --> 00:23:07,020 ABBY NOYCE: That's the spelling backwards. 548 00:23:07,020 --> 00:23:09,930 But even in languages that are not English-- 549 00:23:09,930 --> 00:23:12,240 English is all messed up spelling of its own-- 550 00:23:12,240 --> 00:23:14,296 the relationship between a sound and its spelling 551 00:23:14,296 --> 00:23:16,170 are loose, especially where you've got things 552 00:23:16,170 --> 00:23:19,080 like the th where the two letters are actually making 553 00:23:19,080 --> 00:23:21,030 a completely separate sound. 554 00:23:21,030 --> 00:23:25,050 Anyway, yes, cool. 555 00:23:25,050 --> 00:23:32,220 Moving along, so what happens when your attention 556 00:23:32,220 --> 00:23:34,330 fails at some of these tasks? 557 00:23:34,330 --> 00:23:38,640 So there's some possible modes of failure, possibilities, 558 00:23:38,640 --> 00:23:40,272 modes of failure-- 559 00:23:40,272 --> 00:23:43,050 I fail-- one of which is that it's simply 560 00:23:43,050 --> 00:23:46,230 sensory limitations, that your sensory system can't 561 00:23:46,230 --> 00:23:50,040 handle the full amount of information that's coming in. 562 00:23:50,040 --> 00:23:53,380 So to some extent, the things that we 563 00:23:53,380 --> 00:23:56,100 miss could be chalked up to, for example, 564 00:23:56,100 --> 00:23:58,190 you have less acuity in your peripheral vision. 565 00:23:58,190 --> 00:24:01,260 You're less likely to see things. 566 00:24:01,260 --> 00:24:04,860 Subjects who are doing a visual perception task on a monitor 567 00:24:04,860 --> 00:24:07,950 are more likely to miss targets in the periphery of the monitor 568 00:24:07,950 --> 00:24:10,890 than targets that are in the center. 569 00:24:10,890 --> 00:24:14,220 But an awful lot of attentional failures 570 00:24:14,220 --> 00:24:19,370 are not just explained by sensory limitations. 571 00:24:19,370 --> 00:24:21,640 Most kind of models of attention, most of the way 572 00:24:21,640 --> 00:24:25,180 people who study attention are thinking about it, 573 00:24:25,180 --> 00:24:27,100 they're thinking of it as a bottleneck. 574 00:24:27,100 --> 00:24:30,220 Information comes in, and then there is a-- 575 00:24:30,220 --> 00:24:32,080 and then only so much of it can get 576 00:24:32,080 --> 00:24:33,830 passed on to higher levels of processing. 577 00:24:33,830 --> 00:24:35,830 So most of the information that's coming in just 578 00:24:35,830 --> 00:24:38,410 never makes it. 579 00:24:38,410 --> 00:24:40,870 And some failures in attentional tasks 580 00:24:40,870 --> 00:24:43,130 seem to be response bottlenecks. 581 00:24:43,130 --> 00:24:45,760 A response bottleneck is like if I 582 00:24:45,760 --> 00:24:50,560 tell you to push the left button if you see an F, 583 00:24:50,560 --> 00:24:52,902 and push the right button if you see a C, 584 00:24:52,902 --> 00:24:54,610 and then I show you both of them at once, 585 00:24:54,610 --> 00:24:56,568 but you'll be slower to respond to both of them 586 00:24:56,568 --> 00:24:58,840 at once than you are to either one individually, 587 00:24:58,840 --> 00:25:04,060 it probably has less to do with a sensory limitation, 588 00:25:04,060 --> 00:25:08,650 and it's possible that it's an attentional bottleneck. 589 00:25:08,650 --> 00:25:10,630 But it's also possible that coordinating 590 00:25:10,630 --> 00:25:12,820 one response and then the other response 591 00:25:12,820 --> 00:25:15,399 is where the limitation is, where 592 00:25:15,399 --> 00:25:16,690 the processing is slowing down. 593 00:25:34,500 --> 00:25:36,510 Response bottleneck might be like-- 594 00:25:36,510 --> 00:25:38,640 I was teaching my sister to drive a stick shift 595 00:25:38,640 --> 00:25:40,600 car earlier this summer. 596 00:25:40,600 --> 00:25:42,330 And so she can drive a standard. 597 00:25:42,330 --> 00:25:48,409 She has an automatic she's been driving for five yearsish. 598 00:25:48,409 --> 00:25:49,950 But she hadn't driven a stick before. 599 00:25:49,950 --> 00:25:55,230 So she can start, or stop, or shift gears. 600 00:25:55,230 --> 00:25:57,660 But ask her to do something on the streets 601 00:25:57,660 --> 00:25:59,706 where you're making a right hand turn-- so you've 602 00:25:59,706 --> 00:26:01,080 got to slow down, and then you've 603 00:26:01,080 --> 00:26:02,705 got to shift the car into another gear, 604 00:26:02,705 --> 00:26:05,195 and then you've got to flip the turn signal on, 605 00:26:05,195 --> 00:26:06,570 and you've got to turn the wheel, 606 00:26:06,570 --> 00:26:09,080 and you've got to do all of this stuff at once 607 00:26:09,080 --> 00:26:12,380 is really hard because that there's a response. 608 00:26:12,380 --> 00:26:14,850 And all of these responses to that turn 609 00:26:14,850 --> 00:26:17,490 are taking up a lot of time and a lot of attention 610 00:26:17,490 --> 00:26:20,250 that she might not otherwise have had to use. 611 00:26:20,250 --> 00:26:24,367 So bottlenecks, where attention is 612 00:26:24,367 --> 00:26:26,700 not being able to deal with everything that's happening, 613 00:26:26,700 --> 00:26:29,070 or not being able to do it as quickly and efficiently 614 00:26:29,070 --> 00:26:31,530 as it can deal with one thing happening, 615 00:26:31,530 --> 00:26:34,080 might be in the amount of information that's coming in. 616 00:26:34,080 --> 00:26:36,580 It also might be in how you're responding to that. 617 00:26:41,780 --> 00:26:43,280 All right, so we talked a little bit 618 00:26:43,280 --> 00:26:45,980 about multitasking, dual tasking. 619 00:26:45,980 --> 00:26:48,791 Dual tasking does improve with practice. 620 00:26:48,791 --> 00:26:50,540 If you want to get better at multitasking, 621 00:26:50,540 --> 00:26:52,430 this means that you should do some multitasking, 622 00:26:52,430 --> 00:26:53,930 and you'll probably see your ability 623 00:26:53,930 --> 00:26:56,540 to monitor multiple tasks go up. 624 00:27:03,590 --> 00:27:05,900 Dual task interference is just this thing, 625 00:27:05,900 --> 00:27:08,070 where doing one thing decreases your performance 626 00:27:08,070 --> 00:27:09,600 on another one. 627 00:27:09,600 --> 00:27:12,350 If you're doing two visual tasks at once-- 628 00:27:12,350 --> 00:27:14,700 you can't read two books at once, not 629 00:27:14,700 --> 00:27:16,365 at the exact same time. 630 00:27:16,365 --> 00:27:19,000 You might be able to swap back and forth between them. 631 00:27:19,000 --> 00:27:21,360 But also, if you're doing a visual task 632 00:27:21,360 --> 00:27:24,784 and an auditory task, you'll still 633 00:27:24,784 --> 00:27:26,700 see this interference, where doing one of them 634 00:27:26,700 --> 00:27:30,400 will actually disrupt your performance on the other. 635 00:27:30,400 --> 00:27:34,530 And as you get better at either of the tasks involved, 636 00:27:34,530 --> 00:27:38,880 then having something else going along with it gets easier. 637 00:27:38,880 --> 00:27:41,250 So if you're a good cook, and you cook a lot, 638 00:27:41,250 --> 00:27:43,320 then you might be able to talk on the phone 639 00:27:43,320 --> 00:27:45,430 and chop an onion at the same time, 640 00:27:45,430 --> 00:27:47,400 whereas if you don't have a lot of practice 641 00:27:47,400 --> 00:27:50,134 with those particular kinds of chef knife maneuvers, 642 00:27:50,134 --> 00:27:52,050 this may lead to a trip to the emergency room. 643 00:27:55,480 --> 00:27:58,420 AUDIENCE: But is it possible to be better when you're 644 00:27:58,420 --> 00:28:00,214 dual tasking with a certain ask than it is 645 00:28:00,214 --> 00:28:03,323 if you would do it by itself? 646 00:28:03,323 --> 00:28:05,406 ABBY NOYCE: Do you have a particular case in mind? 647 00:28:05,406 --> 00:28:06,031 AUDIENCE: Yeah. 648 00:28:06,031 --> 00:28:07,120 I type while I read. 649 00:28:07,120 --> 00:28:11,278 [INAUDIBLE] I type better when I read it because then 650 00:28:11,278 --> 00:28:14,206 I do typing when [INAUDIBLE]. 651 00:28:17,140 --> 00:28:18,600 ABBY NOYCE: Cool. 652 00:28:18,600 --> 00:28:20,940 So that's probably not-- 653 00:28:20,940 --> 00:28:22,690 so one possibility here is that that's not 654 00:28:22,690 --> 00:28:24,282 an intentional thing. 655 00:28:24,282 --> 00:28:25,990 What do you mean typing when you read it? 656 00:28:25,990 --> 00:28:27,726 You look at the screen while you type? 657 00:28:27,726 --> 00:28:28,267 AUDIENCE: No. 658 00:28:28,267 --> 00:28:28,969 [INAUDIBLE] 659 00:28:28,969 --> 00:28:30,510 ABBY NOYCE: And take notes and stuff? 660 00:28:30,510 --> 00:28:31,635 AUDIENCE: Yeah, yeah, yeah. 661 00:28:31,635 --> 00:28:34,382 And I think it's faster, [INAUDIBLE].. 662 00:28:34,382 --> 00:28:36,340 ABBY NOYCE: If you think about something that's 663 00:28:36,340 --> 00:28:38,374 like a high speed physical task like that, 664 00:28:38,374 --> 00:28:39,790 it's like sometimes something like 665 00:28:39,790 --> 00:28:42,610 that, it's like people say, my hands know how to do it. 666 00:28:42,610 --> 00:28:45,490 And then if you stop and think about it, you mess yourself up. 667 00:28:45,490 --> 00:28:46,240 Is that one those? 668 00:28:46,240 --> 00:28:47,902 I feel like one of those things. 669 00:28:47,902 --> 00:28:48,754 [INTERPOSING VOICES] 670 00:28:48,754 --> 00:28:51,170 AUDIENCE: Yeah, [INAUDIBLE],, about overanalyzing actually 671 00:28:51,170 --> 00:28:52,130 typing. 672 00:28:52,130 --> 00:28:55,360 If you really sat there and thought about A's here-- 673 00:28:55,360 --> 00:28:56,943 ABBY NOYCE: And then this means that I 674 00:28:56,943 --> 00:29:00,520 have to move my ring finger up, and my middle finger down. 675 00:29:00,520 --> 00:29:03,280 Yeah, so that's probably a different kind 676 00:29:03,280 --> 00:29:06,980 of concentration or something that's not helpful. 677 00:29:06,980 --> 00:29:07,900 I don't know. 678 00:29:07,900 --> 00:29:10,358 It's an interesting data point though, and it's a good one. 679 00:29:15,220 --> 00:29:18,160 AUDIENCE: She's experimenting. 680 00:29:18,160 --> 00:29:20,610 AUDIENCE: Yeah, how many of us [INAUDIBLE]?? 681 00:29:20,610 --> 00:29:22,570 AUDIENCE: Yeah? 682 00:29:22,570 --> 00:29:24,040 What? 683 00:29:24,040 --> 00:29:26,000 AUDIENCE: [INAUDIBLE]. 684 00:29:26,000 --> 00:29:27,740 AUDIENCE: Yes. 685 00:29:27,740 --> 00:29:29,969 AUDIENCE: [INAUDIBLE] an experiment now. 686 00:29:29,969 --> 00:29:31,260 AUDIENCE: No, it's true though. 687 00:29:31,260 --> 00:29:35,410 ABBY NOYCE: OK, how would you set up an experiment? 688 00:29:35,410 --> 00:29:37,530 Watch out, I'm going to give you homework. 689 00:29:37,530 --> 00:29:42,490 OK, so the classic dual task interference 690 00:29:42,490 --> 00:29:44,470 that is a hot political issue these days 691 00:29:44,470 --> 00:29:47,290 is talking on the cell phone while you're driving. 692 00:29:47,290 --> 00:29:49,510 Who knows somebody who does this? 693 00:29:49,510 --> 00:29:51,784 Who's been known to do it? 694 00:29:51,784 --> 00:29:52,768 [INTERPOSING VOICES] 695 00:29:52,768 --> 00:29:53,965 AUDIENCE: I do it. 696 00:29:53,965 --> 00:29:55,080 ABBY NOYCE: City kids. 697 00:29:55,080 --> 00:29:56,960 I do it. 698 00:29:56,960 --> 00:29:57,460 I do it. 699 00:29:57,460 --> 00:29:59,010 I'm a bad person. 700 00:29:59,010 --> 00:30:03,500 [INTERPOSING VOICES] 701 00:30:03,500 --> 00:30:04,193 ABBY NOYCE: So-- 702 00:30:04,193 --> 00:30:06,586 AUDIENCE: When can you get a learner's permit? 703 00:30:06,586 --> 00:30:08,394 AUDIENCE: 16. 704 00:30:08,394 --> 00:30:10,330 AUDIENCE: Unless you were born in '94 705 00:30:10,330 --> 00:30:12,670 or later, in which Massachusetts decides to increase 706 00:30:12,670 --> 00:30:13,830 [INAUDIBLE]? 707 00:30:13,830 --> 00:30:14,824 ABBY NOYCE: 17. 708 00:30:14,824 --> 00:30:15,490 They move it up. 709 00:30:15,490 --> 00:30:17,890 AUDIENCE: [INAUDIBLE]. 710 00:30:17,890 --> 00:30:19,070 ABBY NOYCE: Oh, that's late. 711 00:30:19,070 --> 00:30:20,595 It's later than New York. 712 00:30:20,595 --> 00:30:23,560 [INTERPOSING VOICES] 713 00:30:23,560 --> 00:30:25,060 AUDIENCE: No, I think '94 and later. 714 00:30:25,060 --> 00:30:26,518 ABBY NOYCE: What did it used to be? 715 00:30:26,518 --> 00:30:28,155 [INTERPOSING VOICES] 716 00:30:28,155 --> 00:30:28,780 ABBY NOYCE: OK. 717 00:30:31,820 --> 00:30:32,620 Yeah, I don't know. 718 00:30:32,620 --> 00:30:37,290 I grew up in the state of live free and kill yourself. 719 00:30:37,290 --> 00:30:40,120 Right, so we have a real life dual task interference 720 00:30:40,120 --> 00:30:41,780 situation. 721 00:30:41,780 --> 00:30:44,027 And Stayer and Johnston put together 722 00:30:44,027 --> 00:30:45,610 this study where they had participants 723 00:30:45,610 --> 00:30:49,690 do a simulated driving task, so in a little car simulator. 724 00:30:49,690 --> 00:30:51,130 And they had them either-- 725 00:30:51,130 --> 00:30:52,360 they had a course set up. 726 00:30:52,360 --> 00:30:58,030 They were asked to follow another car through a city, 727 00:30:58,030 --> 00:31:00,040 road, traffic lights, whatever. 728 00:31:00,040 --> 00:31:03,310 And they had their subjects either listen to the radio-- 729 00:31:03,310 --> 00:31:05,500 they even got to pick the radio program-- 730 00:31:05,500 --> 00:31:08,960 or talk on the phone. 731 00:31:08,960 --> 00:31:10,525 AUDIENCE: Listening to the radio, 732 00:31:10,525 --> 00:31:12,650 did they actually have to concentrate on the radio, 733 00:31:12,650 --> 00:31:13,447 or did they just-- 734 00:31:13,447 --> 00:31:16,030 ABBY NOYCE: They got to select a radio program and have it on. 735 00:31:16,030 --> 00:31:18,610 So I think what they were trying to control for was just 736 00:31:18,610 --> 00:31:23,690 having some auditory, that it wasn't merely having 737 00:31:23,690 --> 00:31:26,192 an auditory stimulus in the car. 738 00:31:26,192 --> 00:31:27,400 AUDIENCE: Oh, [INAUDIBLE] OK. 739 00:31:27,400 --> 00:31:27,610 ABBY NOYCE: Right. 740 00:31:27,610 --> 00:31:29,400 It's the control group for the talking on the phone 741 00:31:29,400 --> 00:31:30,881 so that you know-- you can't just 742 00:31:30,881 --> 00:31:32,380 say, oh, but having noise in the car 743 00:31:32,380 --> 00:31:35,247 makes people do poorly on this. 744 00:31:35,247 --> 00:31:36,830 So they had to go through this course. 745 00:31:36,830 --> 00:31:38,770 There'd be a traffic light that would 746 00:31:38,770 --> 00:31:43,480 go red or green at periodic points throughout the course. 747 00:31:43,480 --> 00:31:46,030 And participants had to push the brake when they 748 00:31:46,030 --> 00:31:48,560 saw the traffic light turn red. 749 00:31:48,560 --> 00:31:52,090 So thinking about the experiment design 750 00:31:52,090 --> 00:31:55,390 stuff we talked about in the very first week, what 751 00:31:55,390 --> 00:31:57,604 are these guys trying to study? 752 00:31:57,604 --> 00:31:59,520 What is the question they're trying to answer? 753 00:32:06,680 --> 00:32:08,305 Helen, what are they trying to answer? 754 00:32:08,305 --> 00:32:10,245 AUDIENCE: If you're talking on the phone, 755 00:32:10,245 --> 00:32:12,550 whether or not it stops you from thinking. 756 00:32:12,550 --> 00:32:15,050 ABBY NOYCE: Yeah, whether it messes up your driving ability. 757 00:32:15,050 --> 00:32:15,550 Good. 758 00:32:18,160 --> 00:32:20,060 So what's the independent variable here? 759 00:32:20,060 --> 00:32:21,851 What's the thing that the experimenters are 760 00:32:21,851 --> 00:32:25,980 changing to compare two groups? 761 00:32:25,980 --> 00:32:28,462 [? Naman. ?] 762 00:32:28,462 --> 00:32:35,185 AUDIENCE: Auditory input as well as the need to [INAUDIBLE]---- 763 00:32:35,185 --> 00:32:37,184 they have to concentrate to talk on their phone. 764 00:32:37,184 --> 00:32:37,684 So-- 765 00:32:37,684 --> 00:32:38,433 ABBY NOYCE: Right. 766 00:32:38,433 --> 00:32:40,690 So listening to the radio is just a passive task, 767 00:32:40,690 --> 00:32:42,523 whereas talking on the phone is interactive. 768 00:32:42,523 --> 00:32:46,062 AUDIENCE: [INAUDIBLE] active [? with ?] auditory. 769 00:32:46,062 --> 00:32:46,770 ABBY NOYCE: Good. 770 00:32:46,770 --> 00:32:48,394 So they've got an independent variable, 771 00:32:48,394 --> 00:32:50,820 which is what you have to do while you're driving. 772 00:32:50,820 --> 00:32:52,890 And the two levels of the independent variable 773 00:32:52,890 --> 00:32:54,090 are what, Natasha? 774 00:32:56,874 --> 00:32:58,342 AUDIENCE: The two components of it? 775 00:32:58,342 --> 00:32:59,050 ABBY NOYCE: Yeah. 776 00:32:59,050 --> 00:33:01,380 So remember, independent variables have two levels, 777 00:33:01,380 --> 00:33:02,934 they're called. 778 00:33:02,934 --> 00:33:04,350 So in our case, it's what you have 779 00:33:04,350 --> 00:33:05,520 to do while you're driving. 780 00:33:05,520 --> 00:33:07,806 And the two levels are-- 781 00:33:07,806 --> 00:33:10,435 AUDIENCE: You have to brake at the-- 782 00:33:10,435 --> 00:33:12,060 ABBY NOYCE: The thing they're studying, 783 00:33:12,060 --> 00:33:13,440 the thing that they want to find out how it 784 00:33:13,440 --> 00:33:14,820 changes people's performance. 785 00:33:14,820 --> 00:33:16,680 AUDIENCE: How does talking on the phone effect your driving? 786 00:33:16,680 --> 00:33:18,013 ABBY NOYCE: That's the question. 787 00:33:18,013 --> 00:33:20,058 AUDIENCE: [INAUDIBLE]. 788 00:33:20,058 --> 00:33:21,431 Listening and talking. 789 00:33:21,431 --> 00:33:22,180 ABBY NOYCE: Right. 790 00:33:22,180 --> 00:33:24,540 So the independent variable is what people have 791 00:33:24,540 --> 00:33:26,370 to do while they're driving. 792 00:33:26,370 --> 00:33:28,710 So either listening to the radio or talking on the phone 793 00:33:28,710 --> 00:33:29,640 are the two levels of this. 794 00:33:29,640 --> 00:33:31,150 They're breaking people into two groups. 795 00:33:31,150 --> 00:33:33,108 One group is just going to listen to the radio. 796 00:33:33,108 --> 00:33:35,010 One group's going to talk on the phone. 797 00:33:35,010 --> 00:33:36,900 And then the dependent variable, the thing 798 00:33:36,900 --> 00:33:39,060 that they're measuring is-- 799 00:33:39,060 --> 00:33:40,802 AUDIENCE: Oh, I just have a question. 800 00:33:40,802 --> 00:33:41,510 ABBY NOYCE: Sure. 801 00:33:41,510 --> 00:33:42,875 AUDIENCE: Do they talk on the phone to a person 802 00:33:42,875 --> 00:33:44,700 that was worth talking to, or-- 803 00:33:44,700 --> 00:33:46,699 ABBY NOYCE: I think you had to talk to a friend. 804 00:33:46,699 --> 00:33:48,180 And they had assigned topics. 805 00:33:48,180 --> 00:33:51,270 One topic was the 2000 Olympics, and the other topic 806 00:33:51,270 --> 00:33:52,980 was the 2000 presidential elections. 807 00:33:52,980 --> 00:33:55,500 I think they were looking for an effective heatedness 808 00:33:55,500 --> 00:33:57,312 about the conversation. 809 00:33:57,312 --> 00:33:59,520 But they didn't find any difference between these two 810 00:33:59,520 --> 00:34:00,020 groups. 811 00:34:00,020 --> 00:34:04,490 So any of the talking on the phone stuff was all the same. 812 00:34:04,490 --> 00:34:07,430 So the dependent variable then is how well they perform 813 00:34:07,430 --> 00:34:10,010 on this driving course, right? 814 00:34:10,010 --> 00:34:13,010 And they're measuring it by how well people respond 815 00:34:13,010 --> 00:34:15,440 to the red signals by breaking. 816 00:34:15,440 --> 00:34:16,940 So what did they find out? 817 00:34:16,940 --> 00:34:19,549 What do you guys think they found out? 818 00:34:19,549 --> 00:34:20,960 [INTERPOSING VOICES] 819 00:34:20,960 --> 00:34:22,880 ABBY NOYCE: This is one of the classic studies 820 00:34:22,880 --> 00:34:25,296 that people who want to get cell phones while driving band 821 00:34:25,296 --> 00:34:25,880 tend to cite. 822 00:34:28,489 --> 00:34:32,889 AUDIENCE: Well, if experienced drivers, [INAUDIBLE] 823 00:34:32,889 --> 00:34:35,850 decrease all that much. 824 00:34:35,850 --> 00:34:38,652 They'd probably get all the brakes 825 00:34:38,652 --> 00:34:40,925 because they don't have to actively-- 826 00:34:40,925 --> 00:34:42,050 ABBY NOYCE: Think about it? 827 00:34:42,050 --> 00:34:44,879 AUDIENCE: --think about it because they're experienced. 828 00:34:44,879 --> 00:34:46,875 Drivers who are experienced talking 829 00:34:46,875 --> 00:34:50,529 on the phone at the same time while driving [INAUDIBLE] I 830 00:34:50,529 --> 00:34:53,623 don't think their performance would decrease on a brake. 831 00:34:53,623 --> 00:34:56,409 AUDIENCE: It probably would decrease but not so much. 832 00:34:56,409 --> 00:35:00,390 AUDIENCE: Well, since the brakes are only at some interval, 833 00:35:00,390 --> 00:35:03,060 they can switch between-- they continue 834 00:35:03,060 --> 00:35:04,102 switching between tasks. 835 00:35:04,102 --> 00:35:05,685 ABBY NOYCE: And how to keep monitoring 836 00:35:05,685 --> 00:35:07,100 for the brake signal? 837 00:35:07,100 --> 00:35:09,370 It's an interesting hypothesis, and certainly what 838 00:35:09,370 --> 00:35:12,649 experienced drivers do in other kinds of tasks, tend to do. 839 00:35:12,649 --> 00:35:14,690 You're still going to see some amount of decrease 840 00:35:14,690 --> 00:35:16,665 in performance when people have to monitor 841 00:35:16,665 --> 00:35:17,540 two things like that. 842 00:35:20,907 --> 00:35:22,990 AUDIENCE: It's like when you do a hypothesis test, 843 00:35:22,990 --> 00:35:25,675 you measure if there's a significant difference. 844 00:35:25,675 --> 00:35:29,990 And is that average difference necessarily a lot? 845 00:35:29,990 --> 00:35:30,740 ABBY NOYCE: Right. 846 00:35:30,740 --> 00:35:33,680 So significance and effect sizes are different. 847 00:35:33,680 --> 00:35:39,860 So there's a lot of studies showing that women and men tend 848 00:35:39,860 --> 00:35:45,119 to be very subtly different in the sorts of things 849 00:35:45,119 --> 00:35:46,910 they tend to talk about with their friends, 850 00:35:46,910 --> 00:35:49,280 or how many conversations they have with their friends. 851 00:35:49,280 --> 00:35:50,030 There's good data. 852 00:35:50,030 --> 00:35:52,760 And because people run these studies on 1,000 people, 853 00:35:52,760 --> 00:35:55,462 they get differences that are statistically significant. 854 00:35:55,462 --> 00:35:57,170 But the difference between the two groups 855 00:35:57,170 --> 00:36:00,925 is still really small, or math ability. 856 00:36:00,925 --> 00:36:03,050 You'll tend to see like two overlapping curves that 857 00:36:03,050 --> 00:36:07,711 look like this, if those are both normal distributions, 858 00:36:07,711 --> 00:36:08,210 right? 859 00:36:12,710 --> 00:36:13,670 Good at math. 860 00:36:13,670 --> 00:36:15,440 This is Einstein down here. 861 00:36:15,440 --> 00:36:16,162 No, not Einstein. 862 00:36:16,162 --> 00:36:17,120 Einstein's bad at math. 863 00:36:17,120 --> 00:36:18,215 Who's good at math? 864 00:36:18,215 --> 00:36:19,085 AUDIENCE: Me. 865 00:36:19,085 --> 00:36:22,400 ABBY NOYCE: [INAUDIBLE] good at math? 866 00:36:22,400 --> 00:36:25,100 This is Einstein down here in the bad at math corner. 867 00:36:25,100 --> 00:36:29,120 This is the guy who solved the Poincare conjecture 868 00:36:29,120 --> 00:36:30,580 the other year. 869 00:36:30,580 --> 00:36:32,732 But the thing is, that here-- 870 00:36:32,732 --> 00:36:34,190 and it's really even close in this. 871 00:36:34,190 --> 00:36:35,280 I exaggerated it a bit. 872 00:36:35,280 --> 00:36:37,850 But even though guys are going to be 873 00:36:37,850 --> 00:36:39,800 slightly better at this task, and women 874 00:36:39,800 --> 00:36:46,256 are going to be worse, on average, the number of-- 875 00:36:46,256 --> 00:36:48,005 let me draw these so that they're actually 876 00:36:48,005 --> 00:36:50,267 close because they're not. 877 00:36:50,267 --> 00:36:51,100 That's more like it. 878 00:36:51,100 --> 00:36:53,060 AUDIENCE: Yeah. 879 00:36:53,060 --> 00:36:56,550 ABBY NOYCE: So the means are statistically different. 880 00:36:56,550 --> 00:37:01,961 But if you take an average woman, a lot of guys 881 00:37:01,961 --> 00:37:03,960 are still going to be worse at math than she is. 882 00:37:03,960 --> 00:37:07,800 And if you take an average guy, a lot of women 883 00:37:07,800 --> 00:37:09,800 are still going to be better at math than he is. 884 00:37:09,800 --> 00:37:12,800 So there's slight differences over the group as a whole, 885 00:37:12,800 --> 00:37:14,600 but they're not very useful in predicting 886 00:37:14,600 --> 00:37:16,550 individuals is really what I'm trying 887 00:37:16,550 --> 00:37:18,633 to go with that, that you can get differences that 888 00:37:18,633 --> 00:37:21,200 are statistically significant that aren't really 889 00:37:21,200 --> 00:37:23,660 very useful in the real world. 890 00:37:23,660 --> 00:37:25,070 You get effects kind of like this 891 00:37:25,070 --> 00:37:28,850 with things like different racial groups 892 00:37:28,850 --> 00:37:31,517 and, oh, odds of teen pregnancy, where again, 893 00:37:31,517 --> 00:37:33,350 the differences are significant because they 894 00:37:33,350 --> 00:37:36,390 run the study on 13,000 high schoolers. 895 00:37:36,390 --> 00:37:40,130 But it's not useful in any way because the difference is still 896 00:37:40,130 --> 00:37:42,050 so small. 897 00:37:42,050 --> 00:37:46,820 Back on topic, so we're studying cell phone use while driving. 898 00:37:46,820 --> 00:37:50,000 We are assigning people to either listen to the radio 899 00:37:50,000 --> 00:37:51,320 or talk on the phone. 900 00:37:51,320 --> 00:37:53,630 They have to drive through this course. 901 00:37:53,630 --> 00:37:55,890 And when the signal light flashes red, 902 00:37:55,890 --> 00:37:57,710 they have to brake. 903 00:37:57,710 --> 00:38:00,620 So what they found is that people in the phone condition 904 00:38:00,620 --> 00:38:03,020 entirely missed a break signal twice as much 905 00:38:03,020 --> 00:38:06,350 as people in the just a radio condition. 906 00:38:06,350 --> 00:38:09,290 And they also found that their response times, 907 00:38:09,290 --> 00:38:12,800 how fast the phone conditions subjects 908 00:38:12,800 --> 00:38:14,690 responded to that red light and hit 909 00:38:14,690 --> 00:38:18,595 the brake was slower than people just listening to the radio. 910 00:38:18,595 --> 00:38:20,249 AUDIENCE: About how much? 911 00:38:20,249 --> 00:38:22,040 ABBY NOYCE: I was lazy and I didn't look up 912 00:38:22,040 --> 00:38:22,831 the original study. 913 00:38:22,831 --> 00:38:24,964 I can bring it in tomorrow. 914 00:38:24,964 --> 00:38:25,880 I don't know how much. 915 00:38:25,880 --> 00:38:29,430 I didn't have-- yeah. 916 00:38:29,430 --> 00:38:31,540 AUDIENCE: Well, I think the controversy-- 917 00:38:31,540 --> 00:38:36,430 well, it's making people use bluetooth or a headset 918 00:38:36,430 --> 00:38:39,049 wouldn't really affect that attentional part. 919 00:38:39,049 --> 00:38:41,340 ABBY NOYCE: That is the big, big part of that argument. 920 00:38:41,340 --> 00:38:41,840 Yeah. 921 00:38:41,840 --> 00:38:43,683 AUDIENCE: Because [INAUDIBLE]. 922 00:38:46,300 --> 00:38:47,050 ABBY NOYCE: Right. 923 00:38:47,050 --> 00:38:47,280 Well-- 924 00:38:47,280 --> 00:38:48,863 AUDIENCE: [INAUDIBLE] attentional part 925 00:38:48,863 --> 00:38:50,240 that wouldn't really change. 926 00:38:50,240 --> 00:38:50,660 ABBY NOYCE: Yeah. 927 00:38:50,660 --> 00:38:52,550 There's something to be said for the fact that just having 928 00:38:52,550 --> 00:38:53,570 to hang on to the phone while you're 929 00:38:53,570 --> 00:38:56,410 doing everything else is eating some of your attention too. 930 00:38:56,410 --> 00:39:00,170 But probably a minuscule part of what's happening and this task 931 00:39:00,170 --> 00:39:01,040 has to do with that. 932 00:39:03,610 --> 00:39:04,660 Yeah. 933 00:39:04,660 --> 00:39:07,370 You get this kind of performance drop off with headsets too. 934 00:39:07,370 --> 00:39:11,010 Headsets as a response-- the headset requirements, 935 00:39:11,010 --> 00:39:14,510 the hands-free device requirements, as a solution 936 00:39:14,510 --> 00:39:17,570 to the fact that you see people get in accidents when 937 00:39:17,570 --> 00:39:20,760 they talk on the phone is a-- it looks good. 938 00:39:20,760 --> 00:39:21,530 Let's do it. 939 00:39:21,530 --> 00:39:24,130 But it doesn't actually solve the problem very well. 940 00:39:24,130 --> 00:39:28,230 AUDIENCE: Also, is there a law against having only one 941 00:39:28,230 --> 00:39:29,564 hand on the wheel? 942 00:39:29,564 --> 00:39:30,980 ABBY NOYCE: Not that I know about. 943 00:39:30,980 --> 00:39:33,080 I believe that you can get cited for-- 944 00:39:33,080 --> 00:39:35,099 it depends. 945 00:39:35,099 --> 00:39:36,390 And this varies state to state. 946 00:39:36,390 --> 00:39:37,848 At least in some states you can get 947 00:39:37,848 --> 00:39:41,480 cited for driving while distracted, or driving like-- 948 00:39:41,480 --> 00:39:43,660 this is one of those things where 949 00:39:43,660 --> 00:39:45,190 my friends who live in New York call 950 00:39:45,190 --> 00:39:46,940 it the driving like an idiot ticket, where 951 00:39:46,940 --> 00:39:48,814 like if a cop sees you do something it's just 952 00:39:48,814 --> 00:39:52,430 spectacularly dumb or dangerous, they can pull you over, 953 00:39:52,430 --> 00:39:55,460 or if you are proven to be distracted in some way, 954 00:39:55,460 --> 00:39:59,072 and you get into an accident, your insurance company is-- 955 00:39:59,072 --> 00:40:00,530 you're much more likely to be found 956 00:40:00,530 --> 00:40:02,840 at fault for the accident, if you're on the phone, 957 00:40:02,840 --> 00:40:06,020 or you're eating, or you're bending over 958 00:40:06,020 --> 00:40:09,260 to pick up your lapdog off the floor of the passenger seat. 959 00:40:09,260 --> 00:40:11,390 My mom is an EMT in the town where I grew up in. 960 00:40:11,390 --> 00:40:12,800 And she went on an ambulance call 961 00:40:12,800 --> 00:40:14,810 once when I was probably in middle school. 962 00:40:14,810 --> 00:40:17,390 A lady was driving down the main street 963 00:40:17,390 --> 00:40:19,790 of the town, 30 miles an hour. 964 00:40:19,790 --> 00:40:21,180 This wasn't city traffic. 965 00:40:21,180 --> 00:40:22,520 It was nice, empty road. 966 00:40:22,520 --> 00:40:24,186 And she reached over to pick up her dog, 967 00:40:24,186 --> 00:40:26,150 and she hit a phone call. 968 00:40:26,150 --> 00:40:29,960 She was OK, the dog was OK, the car was not OK. 969 00:40:29,960 --> 00:40:33,204 But take your attention off the road, and it just goes. 970 00:40:33,204 --> 00:40:34,660 AUDIENCE: [INAUDIBLE]. 971 00:40:34,660 --> 00:40:37,160 ABBY NOYCE: I thought this was a spectacularly stupid reason 972 00:40:37,160 --> 00:40:38,493 to get into an accident as well. 973 00:40:38,493 --> 00:40:39,200 But-- 974 00:40:39,200 --> 00:40:41,450 AUDIENCE: [INAUDIBLE]. 975 00:40:41,450 --> 00:40:43,280 ABBY NOYCE: Anyway, so the bigger picture 976 00:40:43,280 --> 00:40:44,779 to be learning here is that when you 977 00:40:44,779 --> 00:40:47,390 are trying to focus your attention on more than one task 978 00:40:47,390 --> 00:40:50,540 at the same time, your performance on both tasks 979 00:40:50,540 --> 00:40:52,040 is probably going to decrease. 980 00:40:55,199 --> 00:40:56,615 So we've been talking a little bit 981 00:40:56,615 --> 00:40:58,300 about two types of attention. 982 00:40:58,300 --> 00:41:01,470 There's this kind of centrally controlled top down attention. 983 00:41:01,470 --> 00:41:04,330 Another word is endogenous that some people will use. 984 00:41:04,330 --> 00:41:08,850 And there's this bottom up stimulus controlled attention 985 00:41:08,850 --> 00:41:13,280 where something that's going on in your surroundings 986 00:41:13,280 --> 00:41:16,220 causes you to remove your attention from whatever 987 00:41:16,220 --> 00:41:20,206 it was previously doing and look somewhere else. 988 00:41:20,206 --> 00:41:22,580 When David came in and dropped off the attendance sheets, 989 00:41:22,580 --> 00:41:24,705 all you all turned around to see what was going on. 990 00:41:24,705 --> 00:41:27,880 That's this exogenous stimulus-driven attention. 991 00:41:27,880 --> 00:41:29,912 AUDIENCE: Wait, when did he do that? 992 00:41:29,912 --> 00:41:31,370 ABBY NOYCE: He did it very quietly. 993 00:41:31,370 --> 00:41:32,840 Not everybody saw. 994 00:41:32,840 --> 00:41:35,630 But I know that some people over on this side of the room 995 00:41:35,630 --> 00:41:36,750 definitely saw him do it. 996 00:41:36,750 --> 00:41:42,779 [INTERPOSING VOICES] 997 00:41:42,779 --> 00:41:44,320 ABBY NOYCE: Attention is interesting. 998 00:41:44,320 --> 00:41:46,486 One of the things that's interesting about attention 999 00:41:46,486 --> 00:41:49,030 is how good people are at focusing on-- 1000 00:41:49,030 --> 00:41:51,170 like letting their top down controlled 1001 00:41:51,170 --> 00:41:54,470 attention trump bottom up stimuli 1002 00:41:54,470 --> 00:41:57,012 that are pushing in on it. 1003 00:41:57,012 --> 00:41:59,470 We probably all know people who are really good at studying 1004 00:41:59,470 --> 00:42:01,900 with lots of background noise going on, 1005 00:42:01,900 --> 00:42:04,180 and other people who just can't focus if they've 1006 00:42:04,180 --> 00:42:06,517 got that kind of chaos. 1007 00:42:06,517 --> 00:42:09,100 And that's, among other things, the difference in your ability 1008 00:42:09,100 --> 00:42:12,400 to control that centrally-based attention 1009 00:42:12,400 --> 00:42:15,040 and not let other stimuli grab onto it. 1010 00:42:18,386 --> 00:42:20,776 AUDIENCE: So when I do that [INAUDIBLE].. 1011 00:42:20,776 --> 00:42:23,644 AUDIENCE: Wait, what? 1012 00:42:23,644 --> 00:42:25,080 ABBY NOYCE: Yeah. 1013 00:42:25,080 --> 00:42:27,530 So you're creating a stimulus to get 1014 00:42:27,530 --> 00:42:29,662 people's exogenous attention. 1015 00:42:32,380 --> 00:42:35,770 And a bunch of people very carefully didn't look away. 1016 00:42:35,770 --> 00:42:38,420 Power of willpower. 1017 00:42:38,420 --> 00:42:41,110 OK. 1018 00:42:41,110 --> 00:42:42,848 What are we up to? 1019 00:42:42,848 --> 00:42:43,836 [INAUDIBLE] time. 1020 00:42:46,800 --> 00:42:48,410 So what is this attention thing? 1021 00:42:48,410 --> 00:42:50,990 So what's happening when you pay attention to something? 1022 00:42:50,990 --> 00:42:52,599 So this guy named Posner, who was 1023 00:42:52,599 --> 00:42:54,140 one of the classic guys in the field, 1024 00:42:54,140 --> 00:42:57,410 this was a model he published in the '80s. 1025 00:42:57,410 --> 00:43:00,380 And he says that in order to pay attention to something, 1026 00:43:00,380 --> 00:43:04,336 you need transfer it, to shift your attention to a new focus. 1027 00:43:04,336 --> 00:43:05,960 You need to be able to do three things. 1028 00:43:05,960 --> 00:43:08,960 You've got to disengage your attention from whatever 1029 00:43:08,960 --> 00:43:10,880 it's currently focusing on. 1030 00:43:10,880 --> 00:43:12,770 You've got to point it out whenever 1031 00:43:12,770 --> 00:43:15,020 you want to focus on next. 1032 00:43:15,020 --> 00:43:19,225 And you've got to reengage it. 1033 00:43:19,225 --> 00:43:20,600 And one of the pieces of evidence 1034 00:43:20,600 --> 00:43:22,099 that he used for the fact that there 1035 00:43:22,099 --> 00:43:25,280 are these separate mental operations involved 1036 00:43:25,280 --> 00:43:29,750 is what happens in patients with hemispatial neglect. 1037 00:43:29,750 --> 00:43:33,680 So heavy spatial neglect happens in patients usually 1038 00:43:33,680 --> 00:43:36,410 who have damage to the parietal lobe. 1039 00:43:36,410 --> 00:43:39,230 Most often, it's damage to the right parietal lobe leading 1040 00:43:39,230 --> 00:43:41,971 to neglect of the left visual field. 1041 00:43:41,971 --> 00:43:44,220 And so if you give these patients a picture of a clock 1042 00:43:44,220 --> 00:43:45,930 and ask them to copy it-- 1043 00:43:45,930 --> 00:43:48,720 these there two pretty characteristic examples 1044 00:43:48,720 --> 00:43:49,790 of a daisy-- 1045 00:43:49,790 --> 00:43:51,489 ask them to clot copy it. 1046 00:43:51,489 --> 00:43:53,780 And, of course, the original drawings for both of these 1047 00:43:53,780 --> 00:43:56,134 had features, had all the numbers 1048 00:43:56,134 --> 00:43:57,800 all the way around the clock, had petals 1049 00:43:57,800 --> 00:44:01,530 all the way around the daisy, and leaves on both sides. 1050 00:44:01,530 --> 00:44:04,100 And as you can see, these patients pretty much 1051 00:44:04,100 --> 00:44:07,224 just disregard everything in the left half, 1052 00:44:07,224 --> 00:44:08,640 on the left half of these objects. 1053 00:44:08,640 --> 00:44:10,100 They only pay attention to what's 1054 00:44:10,100 --> 00:44:12,480 on the right half of these objects. 1055 00:44:12,480 --> 00:44:15,137 So you'd ask them to copy this drawing of a clock, 1056 00:44:15,137 --> 00:44:17,720 and they'll do that, and they'll hand you one with the numbers 1057 00:44:17,720 --> 00:44:18,960 12 through five on it. 1058 00:44:18,960 --> 00:44:21,620 And you'll say, do you want to make any changes to that? 1059 00:44:21,620 --> 00:44:22,820 And they'll look at the one they drew, 1060 00:44:22,820 --> 00:44:23,960 and they'll look at the one they're copying, 1061 00:44:23,960 --> 00:44:25,160 and they'll say, no. 1062 00:44:25,160 --> 00:44:25,988 I'm good. 1063 00:44:25,988 --> 00:44:27,482 AUDIENCE: So the one on the right, 1064 00:44:27,482 --> 00:44:31,350 is that because they know a clock has 12 numbers on it, 1065 00:44:31,350 --> 00:44:32,270 so they try and-- 1066 00:44:32,270 --> 00:44:32,780 ABBY NOYCE: Maybe, or it might be 1067 00:44:32,780 --> 00:44:34,010 that their damage is a little bit less 1068 00:44:34,010 --> 00:44:35,640 severe than the other patients. 1069 00:44:35,640 --> 00:44:37,639 This is something where you'll see ranges of it. 1070 00:44:37,639 --> 00:44:39,610 You'll see really severe and milder. 1071 00:44:39,610 --> 00:44:41,060 He's got more petals on the daisy. 1072 00:44:41,060 --> 00:44:44,030 He's got a central leaf on the daisy, or she. 1073 00:44:44,030 --> 00:44:45,290 I don't know. 1074 00:44:45,290 --> 00:44:48,110 So I would imagine that this is probably just, more 1075 00:44:48,110 --> 00:44:52,757 than anything, a difference in the severity of the handicap. 1076 00:44:52,757 --> 00:44:54,967 AUDIENCE: In patient B there's two 12s. 1077 00:44:54,967 --> 00:44:56,300 ABBY NOYCE: That's a good point. 1078 00:44:56,300 --> 00:44:57,216 I hadn't noticed that. 1079 00:44:57,216 --> 00:44:58,670 Oh, is that 12, one, two? 1080 00:44:58,670 --> 00:45:00,780 AUDIENCE: It's 12, one, two. 1081 00:45:00,780 --> 00:45:03,240 AUDIENCE: Why is it-- 1082 00:45:03,240 --> 00:45:06,420 ABBY NOYCE: Look, actually, if I look at it, 1083 00:45:06,420 --> 00:45:07,982 yeah, B has all the numbers there. 1084 00:45:07,982 --> 00:45:08,690 But they're not-- 1085 00:45:08,690 --> 00:45:10,274 AUDIENCE: [INAUDIBLE] spread properly. 1086 00:45:10,274 --> 00:45:11,940 ABBY NOYCE: They're not spread properly. 1087 00:45:11,940 --> 00:45:13,100 That's interesting. 1088 00:45:13,100 --> 00:45:15,470 That might-- yeah, that's an interesting hypothesis. 1089 00:45:15,470 --> 00:45:17,960 I hadn't noticed that. 1090 00:45:17,960 --> 00:45:20,170 Cool. 1091 00:45:20,170 --> 00:45:21,760 So patients and hemispatial neglect. 1092 00:45:21,760 --> 00:45:26,200 So they won't perceive things on the left side 1093 00:45:26,200 --> 00:45:31,277 of their visual field, as demonstrated in these copies. 1094 00:45:31,277 --> 00:45:32,860 If you ask them to copy a drawing that 1095 00:45:32,860 --> 00:45:34,807 shows several objects lined up in a row, 1096 00:45:34,807 --> 00:45:36,640 they'll copy the right hand side of each one 1097 00:45:36,640 --> 00:45:37,681 and leave the left blank. 1098 00:45:40,432 --> 00:45:42,810 AUDIENCE: And it's not just that they can't see the-- 1099 00:45:42,810 --> 00:45:45,310 ABBY NOYCE: It's not just that they can't see the left side. 1100 00:45:45,310 --> 00:45:47,470 For most of these patients, if you 1101 00:45:47,470 --> 00:45:49,060 put something that is sufficiently 1102 00:45:49,060 --> 00:45:52,330 salient in their left field, like a giant flashing 1103 00:45:52,330 --> 00:45:54,550 light or something, they'll respond to that. 1104 00:45:54,550 --> 00:45:58,720 But they won't respond to it if it's just a little thing. 1105 00:45:58,720 --> 00:46:01,510 So they can see it. 1106 00:46:01,510 --> 00:46:05,440 And sometimes if you can direct their attention there, 1107 00:46:05,440 --> 00:46:08,479 then you can get them to respond to things 1108 00:46:08,479 --> 00:46:09,520 in the left visual field. 1109 00:46:09,520 --> 00:46:12,650 You can tell them to look at something over here. 1110 00:46:12,650 --> 00:46:15,850 And, again, it varies with the severity of the damage. 1111 00:46:18,530 --> 00:46:21,330 But in particular, Posner looks at this hemispatial neglect 1112 00:46:21,330 --> 00:46:25,120 thing and says that what seems to be happening with patients 1113 00:46:25,120 --> 00:46:28,540 who are in the milder form of this, where they can see things 1114 00:46:28,540 --> 00:46:32,140 in their left visual field, they don't really 1115 00:46:32,140 --> 00:46:34,442 want to direct their attention to them. 1116 00:46:34,442 --> 00:46:36,400 They have a hard time directing their attention 1117 00:46:36,400 --> 00:46:37,816 to things in the left visual field 1118 00:46:37,816 --> 00:46:40,270 unless it's explicitly directed there. 1119 00:46:40,270 --> 00:46:45,922 He says that what's happening is that there isn't-- 1120 00:46:45,922 --> 00:46:47,380 the patients are having a hard time 1121 00:46:47,380 --> 00:46:50,200 disengaging their attention from the right visual field 1122 00:46:50,200 --> 00:46:53,080 in order to move it over, that what's actually is happening 1123 00:46:53,080 --> 00:46:56,170 is that your attention is staying stuck on. 1124 00:46:56,170 --> 00:46:57,820 It can't disengage in order to move, 1125 00:46:57,820 --> 00:47:02,380 so that there isn't something about these patients left 1126 00:47:02,380 --> 00:47:05,770 visual field isn't able to send the signal 1127 00:47:05,770 --> 00:47:10,662 to disengage attention in order to start this process of moving 1128 00:47:10,662 --> 00:47:11,870 it from one place to another. 1129 00:47:15,582 --> 00:47:17,400 Oh, we're going to shift gears a little bit 1130 00:47:17,400 --> 00:47:22,580 and talk about some broad theories about how 1131 00:47:22,580 --> 00:47:25,740 the attentional bottleneck works, 1132 00:47:25,740 --> 00:47:28,620 and where in processing your systems 1133 00:47:28,620 --> 00:47:32,040 decide to, OK, 90% of what's coming in, I'm getting rid of. 1134 00:47:32,040 --> 00:47:33,960 I'm going to focus on this little bit that 1135 00:47:33,960 --> 00:47:37,140 is salient, that is relevant for some reason. 1136 00:47:37,140 --> 00:47:41,070 So one of the big questions here is the classic model 1137 00:47:41,070 --> 00:47:45,060 of attention is that selection happens very, very 1138 00:47:45,060 --> 00:47:46,710 early in the processing stages. 1139 00:47:46,710 --> 00:47:49,200 So sensory input comes in. 1140 00:47:49,200 --> 00:47:51,360 It's a fire hose of information. 1141 00:47:51,360 --> 00:47:53,340 It gets processed very basically for things 1142 00:47:53,340 --> 00:47:58,650 like in the visual system, color, and shape. 1143 00:47:58,650 --> 00:48:00,930 And then 90% of it is discarded. 1144 00:48:00,930 --> 00:48:04,020 And only a little bit makes it on to higher level processing 1145 00:48:04,020 --> 00:48:06,750 where you're actually trying to make some kind of sense 1146 00:48:06,750 --> 00:48:07,890 out of what you're hearing. 1147 00:48:11,400 --> 00:48:15,000 So in this early processing method model, 1148 00:48:15,000 --> 00:48:17,400 this is where our attentional model bottleneck is, right, 1149 00:48:17,400 --> 00:48:19,960 between this very preliminary processing 1150 00:48:19,960 --> 00:48:23,452 and then the higher order content semantic 1151 00:48:23,452 --> 00:48:25,660 processing where you actually make sense out of this. 1152 00:48:25,660 --> 00:48:29,700 So an early selection model would 1153 00:48:29,700 --> 00:48:32,420 say that if you know you're looking for your friend 1154 00:48:32,420 --> 00:48:35,160 at a crowded party who's wearing a red shirt, 1155 00:48:35,160 --> 00:48:37,350 and you're scanning the crowd of people, 1156 00:48:37,350 --> 00:48:39,000 all of the sensory input is coming in. 1157 00:48:39,000 --> 00:48:41,250 It's getting sorted by a few basic features, 1158 00:48:41,250 --> 00:48:43,090 in this case, color. 1159 00:48:43,090 --> 00:48:45,900 And you're throwing out all of the pieces of your visual field 1160 00:48:45,900 --> 00:48:48,277 that aren't attached to a red shirt. 1161 00:48:48,277 --> 00:48:49,860 And then only those pieces are getting 1162 00:48:49,860 --> 00:48:51,390 looked at more carefully to be like, 1163 00:48:51,390 --> 00:48:53,190 does this person match my friend? 1164 00:48:58,080 --> 00:49:01,950 Some of the problems with this is-- 1165 00:49:01,950 --> 00:49:03,960 the best one is what's called the cocktail party 1166 00:49:03,960 --> 00:49:07,140 effect, where if you're in a crowded room with lots 1167 00:49:07,140 --> 00:49:10,380 of conversations going on and a lot of background noise, 1168 00:49:10,380 --> 00:49:14,049 and somebody on the other side of the room mentions your name, 1169 00:49:14,049 --> 00:49:16,590 you're really likely to pick up on that piece of information, 1170 00:49:16,590 --> 00:49:18,423 even if you totally weren't paying attention 1171 00:49:18,423 --> 00:49:21,927 to this before, or if you're doing your homework 1172 00:49:21,927 --> 00:49:23,760 and your roommate in a couple of her friends 1173 00:49:23,760 --> 00:49:25,176 are gabbing in the hallway and you 1174 00:49:25,176 --> 00:49:27,770 hear your name, all of a sudden, this stream of input 1175 00:49:27,770 --> 00:49:29,520 that you haven't been paying any attention 1176 00:49:29,520 --> 00:49:32,010 to manages to [MAKES ATTENTIONAL SOUND] 1177 00:49:32,010 --> 00:49:33,917 you're talking about me. 1178 00:49:33,917 --> 00:49:36,000 AUDIENCE: Is it really called the cocktail party-- 1179 00:49:36,000 --> 00:49:38,124 ABBY NOYCE: It is called the cocktail party effect. 1180 00:49:42,210 --> 00:49:47,000 So this is a piece of bottom up stimulus driven 1181 00:49:47,000 --> 00:49:48,280 attention shifting, right? 1182 00:49:48,280 --> 00:49:52,430 You're shifting your attention to this conversation. 1183 00:49:52,430 --> 00:49:57,370 But it can't just be this preliminary basic early 1184 00:49:57,370 --> 00:49:59,120 processing that's letting it in because it 1185 00:49:59,120 --> 00:50:02,060 depends on what the content of that stimulus is. 1186 00:50:07,277 --> 00:50:08,860 Another good example of this is what's 1187 00:50:08,860 --> 00:50:13,150 called a dichroic listening study, where you have somebody 1188 00:50:13,150 --> 00:50:15,580 sit there and wear headphones. 1189 00:50:15,580 --> 00:50:19,390 And they've got two different sentences coming in, 1190 00:50:19,390 --> 00:50:20,830 like one in each ear. 1191 00:50:20,830 --> 00:50:24,400 So the left ear might be saying-- 1192 00:50:24,400 --> 00:50:27,470 I don't know-- once upon a time, a long, long time ago, 1193 00:50:27,470 --> 00:50:29,170 there was a happy princess. 1194 00:50:29,170 --> 00:50:31,360 And the right ear might be saying, 1195 00:50:31,360 --> 00:50:34,000 to repair your motorcycle, first you must-- 1196 00:50:34,000 --> 00:50:37,884 two very two streams of text with very different context. 1197 00:50:37,884 --> 00:50:39,550 And if you tell people to listen to just 1198 00:50:39,550 --> 00:50:41,924 what's in their left ear or just what's in the right ear, 1199 00:50:41,924 --> 00:50:43,360 they can do that. 1200 00:50:43,360 --> 00:50:46,240 And they're actually reasonably good at-- 1201 00:50:48,850 --> 00:50:51,510 they're actually really bad at remembering 1202 00:50:51,510 --> 00:50:52,510 what's in the other ear. 1203 00:50:52,510 --> 00:50:54,760 If you ask them what was the other steam? 1204 00:50:54,760 --> 00:50:56,320 They'll be like, I have no clue. 1205 00:50:56,320 --> 00:50:58,360 I was only focusing on this one. 1206 00:50:58,360 --> 00:51:03,760 But if you switch them halfway through that the left ear goes, 1207 00:51:03,760 --> 00:51:06,910 once upon a time, a long, long time ago-- 1208 00:51:06,910 --> 00:51:09,355 first you must disassemble the engine and-- 1209 00:51:09,355 --> 00:51:14,380 and the left ear goes, to repair your motorcycle-- first, 1210 00:51:14,380 --> 00:51:17,800 there was a happy princess who lived-- 1211 00:51:17,800 --> 00:51:19,960 and you tell them to attend to the left ear, 1212 00:51:19,960 --> 00:51:23,260 they won't hear the ear. 1213 00:51:23,260 --> 00:51:26,110 They will continue following one of the threads of this, 1214 00:51:26,110 --> 00:51:30,380 even though which ear it's coming from has switched. 1215 00:51:30,380 --> 00:51:33,310 So they'll hear only the princess story, or only 1216 00:51:33,310 --> 00:51:36,418 the motorcycle repair guide despite-- 1217 00:51:36,418 --> 00:51:38,251 AUDIENCE: Is that even though they're trying 1218 00:51:38,251 --> 00:51:39,750 to listen to a specific ear? 1219 00:51:39,750 --> 00:51:41,207 ABBY NOYCE: No, whichever ear you 1220 00:51:41,207 --> 00:51:42,790 tell them to listen to, they'll follow 1221 00:51:42,790 --> 00:51:46,060 the thread of that story, even if you switch it. 1222 00:51:46,060 --> 00:51:49,420 Even if you switch it back-- 1223 00:51:49,420 --> 00:51:51,830 and they won't know that it switched. 1224 00:51:51,830 --> 00:51:54,459 You then said, did you notice anything odd about that? 1225 00:51:54,459 --> 00:51:56,000 Did you notice that it switched ears? 1226 00:51:56,000 --> 00:51:57,616 Nope. 1227 00:51:57,616 --> 00:51:59,770 And so this is another case where 1228 00:51:59,770 --> 00:52:02,620 which stimulus is being attended to 1229 00:52:02,620 --> 00:52:06,860 can't just be explained by basic early processing features. 1230 00:52:06,860 --> 00:52:09,460 Again, it's really dependent on the content 1231 00:52:09,460 --> 00:52:11,590 of what's coming in, that you're following 1232 00:52:11,590 --> 00:52:15,859 a piece of content despite one of its biggest early processing 1233 00:52:15,859 --> 00:52:17,900 features, which is which ear that noise is coming 1234 00:52:17,900 --> 00:52:21,040 into just changed entirely. 1235 00:52:21,040 --> 00:52:24,970 So this whole like early selection attentional model 1236 00:52:24,970 --> 00:52:27,930 has some big holes in it. 1237 00:52:27,930 --> 00:52:30,017 There's a couple of later selection models 1238 00:52:30,017 --> 00:52:32,100 that we're going to talk about, the first of which 1239 00:52:32,100 --> 00:52:34,620 is what's called a spotlight theory. 1240 00:52:34,620 --> 00:52:37,770 And the spotlight theory is a theory of visual attention. 1241 00:52:37,770 --> 00:52:42,320 Most of these, most of the really well-defined theories 1242 00:52:42,320 --> 00:52:44,310 are theories of visual attention, 1243 00:52:44,310 --> 00:52:46,650 partly because more research gets done on vision 1244 00:52:46,650 --> 00:52:48,487 than on any other modality. 1245 00:52:51,900 --> 00:52:54,620 And the spotlight theory basically-- 1246 00:52:54,620 --> 00:52:57,480 this model's attention is being like a spotlight. 1247 00:52:57,480 --> 00:53:02,310 Your attention flits around the visual scene and shines 1248 00:53:02,310 --> 00:53:04,680 on a particular spot. 1249 00:53:04,680 --> 00:53:07,634 And whatever's in that spot is what you're attending to. 1250 00:53:11,010 --> 00:53:14,480 It's kind like you being in a dark room with a flashlight. 1251 00:53:14,480 --> 00:53:18,440 You can only see a very small portion of it at that time. 1252 00:53:21,280 --> 00:53:25,051 And spotlight theorists say that when you attend to a spot, 1253 00:53:25,051 --> 00:53:25,550 that's all. 1254 00:53:25,550 --> 00:53:27,008 You're just focusing your attention 1255 00:53:27,008 --> 00:53:29,450 on a particular location, and everything that's there gets 1256 00:53:29,450 --> 00:53:30,020 attended to. 1257 00:53:34,550 --> 00:53:37,880 One little flaw with this theory that the reading makes 1258 00:53:37,880 --> 00:53:40,540 a big deal out of that I don't think it's that big a deal, it 1259 00:53:40,540 --> 00:53:46,310 says if I have multiple objects, two objects, 1260 00:53:46,310 --> 00:53:49,760 and I ask you to attend to the water bottle, 1261 00:53:49,760 --> 00:53:52,794 and then switch your attention to the pile of paper, 1262 00:53:52,794 --> 00:53:54,710 how long it takes you to switch your attention 1263 00:53:54,710 --> 00:53:57,980 between these two is the same whether or not 1264 00:53:57,980 --> 00:54:00,860 there's a third object in the middle 1265 00:54:00,860 --> 00:54:03,740 that an honest to goodness smoothly moving 1266 00:54:03,740 --> 00:54:06,771 spotlight would have to hop over. 1267 00:54:06,771 --> 00:54:08,270 But if we don't think of a spotlight 1268 00:54:08,270 --> 00:54:12,050 as having to move smoothly, this works pretty well. 1269 00:54:12,050 --> 00:54:17,060 A bigger problem is if you ask people to do something, 1270 00:54:17,060 --> 00:54:23,780 to look at items like this where there is a box and a line, 1271 00:54:23,780 --> 00:54:25,940 and ask them to do-- it you ask them 1272 00:54:25,940 --> 00:54:30,050 to do two tasks about the boxes, the box, big or little, 1273 00:54:30,050 --> 00:54:33,470 and is the gap in the box on the left side or the right side, 1274 00:54:33,470 --> 00:54:36,530 they can do two-box tasks reasonably well. 1275 00:54:36,530 --> 00:54:40,190 If you ask them is the line diagonal or upright, 1276 00:54:40,190 --> 00:54:43,490 and is it solid or dotted, and to respond to both of those, 1277 00:54:43,490 --> 00:54:46,610 they can do two line tasks reasonably well. 1278 00:54:46,610 --> 00:54:50,780 But if you ask them, is the gap on the left side of the box? 1279 00:54:50,780 --> 00:54:53,160 And is the line diagonal or upright? 1280 00:54:53,160 --> 00:54:56,150 Then their performance is worse. 1281 00:54:56,150 --> 00:55:00,680 Even though both the box and the line are in the same location, 1282 00:55:00,680 --> 00:55:04,220 the fact that they're having to think about two objects, two 1283 00:55:04,220 --> 00:55:07,880 things that are being registered as separate objects, 1284 00:55:07,880 --> 00:55:11,830 reduces performance. 1285 00:55:11,830 --> 00:55:16,130 And so this is evidence that attention 1286 00:55:16,130 --> 00:55:19,430 seems to be focused not just on locations, but actually 1287 00:55:19,430 --> 00:55:21,200 on objects. 1288 00:55:21,200 --> 00:55:23,270 And frequently, objects art in locations, 1289 00:55:23,270 --> 00:55:25,647 but biding your attention onto an object 1290 00:55:25,647 --> 00:55:27,980 is different from biding your attention onto a location. 1291 00:55:48,390 --> 00:55:52,410 OK, third theory, it's what's called feature integration 1292 00:55:52,410 --> 00:55:53,300 theory. 1293 00:55:53,300 --> 00:55:54,970 So I'm going to show you two images. 1294 00:55:54,970 --> 00:55:57,750 Your mission is to find the black circle. 1295 00:55:57,750 --> 00:55:59,706 Who sees the black circle in this image? 1296 00:55:59,706 --> 00:56:01,080 You see it just like that, right? 1297 00:56:01,080 --> 00:56:03,429 It pops out at you. 1298 00:56:03,429 --> 00:56:05,655 AUDIENCE: Technically, that's-- 1299 00:56:05,655 --> 00:56:07,530 ABBY NOYCE: Not quite circular, is that where 1300 00:56:07,530 --> 00:56:08,220 you're going with that? 1301 00:56:08,220 --> 00:56:08,761 AUDIENCE: No. 1302 00:56:08,761 --> 00:56:12,100 The circle [INAUDIBLE]. 1303 00:56:12,100 --> 00:56:14,525 That's the inside of a circle. 1304 00:56:17,359 --> 00:56:18,650 ABBY NOYCE: Here's another one. 1305 00:56:18,650 --> 00:56:19,820 Find the black circle. 1306 00:56:19,820 --> 00:56:22,970 This one is probably at least a little bit harder. 1307 00:56:22,970 --> 00:56:24,620 It should take a little bit longer. 1308 00:56:24,620 --> 00:56:27,775 Not a lot, but if we're having you hit a key when 1309 00:56:27,775 --> 00:56:29,150 you find it-- and we're measuring 1310 00:56:29,150 --> 00:56:32,810 your responses in milliseconds-- you will see the difference. 1311 00:56:32,810 --> 00:56:37,277 What's different about these two [? stimulin. ?] 1312 00:56:37,277 --> 00:56:39,860 AUDIENCE: Before, we were just looking for a black, basically, 1313 00:56:39,860 --> 00:56:44,080 just solid black [INAUDIBLE]. 1314 00:56:44,080 --> 00:56:44,830 ABBY NOYCE: Right. 1315 00:56:44,830 --> 00:56:46,990 So a feature identity integration theorist 1316 00:56:46,990 --> 00:56:48,800 would say, in the first task, there 1317 00:56:48,800 --> 00:56:51,230 was one feature that differentiated the target 1318 00:56:51,230 --> 00:56:52,130 from everything else. 1319 00:56:52,130 --> 00:56:54,185 And that feature was color. 1320 00:56:54,185 --> 00:56:55,685 And so you could just be like, done. 1321 00:56:55,685 --> 00:56:56,315 There it is. 1322 00:56:56,315 --> 00:56:58,160 In this one, there's two features 1323 00:56:58,160 --> 00:57:01,160 that are distinguishing it. it's color and shape. 1324 00:57:01,160 --> 00:57:03,347 And so feature integration theory 1325 00:57:03,347 --> 00:57:05,180 says that these are two very different kinds 1326 00:57:05,180 --> 00:57:06,240 of searching tasks. 1327 00:57:06,240 --> 00:57:09,422 This is a visual search task. 1328 00:57:09,422 --> 00:57:14,603 AUDIENCE: Could that imply that distinguishing colors is easier 1329 00:57:14,603 --> 00:57:16,089 than distinguishing shapes? 1330 00:57:16,089 --> 00:57:17,630 ABBY NOYCE: No, because if I give you 1331 00:57:17,630 --> 00:57:20,480 a circle among squares, and they're all the same color, 1332 00:57:20,480 --> 00:57:22,122 you get those fast too. 1333 00:57:22,122 --> 00:57:23,540 AUDIENCE: Oh, OK. 1334 00:57:23,540 --> 00:57:25,450 ABBY NOYCE: And the classic example of this 1335 00:57:25,450 --> 00:57:30,850 is a red X among red O's, or a red X among green X's are 1336 00:57:30,850 --> 00:57:32,740 both very fast to find. 1337 00:57:32,740 --> 00:57:38,110 But a red X among red O's and green X's is a lot slower 1338 00:57:38,110 --> 00:57:40,690 to find because it's being distinguished 1339 00:57:40,690 --> 00:57:43,120 by two features rather than one. 1340 00:57:43,120 --> 00:57:46,210 So terminology, the first of those, 1341 00:57:46,210 --> 00:57:49,360 where we had just the black circle among the white circles, 1342 00:57:49,360 --> 00:57:51,010 is a disjunctive trial. 1343 00:57:51,010 --> 00:57:53,080 There's one feature that distinguishes the target 1344 00:57:53,080 --> 00:57:54,340 from the distracters. 1345 00:57:54,340 --> 00:57:56,050 The distracters are all the things 1346 00:57:56,050 --> 00:57:59,680 in the image that are not targets. 1347 00:57:59,680 --> 00:58:03,850 The second one is a conjunctive trial, 1348 00:58:03,850 --> 00:58:05,980 where there is more than one feature that 1349 00:58:05,980 --> 00:58:07,007 defines the target. 1350 00:58:12,150 --> 00:58:16,560 And feature integration notes that one of the things that's 1351 00:58:16,560 --> 00:58:20,160 interesting about this, is with disjunctive trials, where 1352 00:58:20,160 --> 00:58:22,350 you're trying to find a target that differs 1353 00:58:22,350 --> 00:58:25,075 from its distracters by-- 1354 00:58:25,075 --> 00:58:26,700 they're are all the same color and then 1355 00:58:26,700 --> 00:58:28,290 one's a different shape, or they're all the same shape, 1356 00:58:28,290 --> 00:58:30,180 and one's a different color, these kinds 1357 00:58:30,180 --> 00:58:33,630 of single-feature differences, if I 1358 00:58:33,630 --> 00:58:38,100 ask you to find a target with four distracters and a target 1359 00:58:38,100 --> 00:58:39,540 with 25 distracters, you're going 1360 00:58:39,540 --> 00:58:42,790 to be equally fast on both of those tasks. 1361 00:58:42,790 --> 00:58:45,780 So people say that disjunctive search is pre-attentive. 1362 00:58:45,780 --> 00:58:49,920 You don't have to attend to the individual items in the field 1363 00:58:49,920 --> 00:58:53,202 in order to distinguish the target from the distracters. 1364 00:58:58,800 --> 00:59:00,960 In a conjunctive trial, where you've 1365 00:59:00,960 --> 00:59:03,940 got the black squares, and the white circles, 1366 00:59:03,940 --> 00:59:08,010 and you're looking for the black circle, 1367 00:59:08,010 --> 00:59:10,476 as I add items to the number of distracters, 1368 00:59:10,476 --> 00:59:11,850 if I start with three distracters 1369 00:59:11,850 --> 00:59:14,490 and then test you with four, and then five, and then six, 1370 00:59:14,490 --> 00:59:16,470 and on up, your performance gets-- 1371 00:59:16,470 --> 00:59:20,030 your average performance, your average time to find the target 1372 00:59:20,030 --> 00:59:22,040 gets steadily slower with each one, 1373 00:59:22,040 --> 00:59:25,860 with each distracter that you add. 1374 00:59:25,860 --> 00:59:28,320 So this is an attention required task, 1375 00:59:28,320 --> 00:59:30,750 where you have to actually attend to individual items 1376 00:59:30,750 --> 00:59:32,190 until you find the target. 1377 00:59:35,570 --> 00:59:40,280 So if you were graphing this, and it was like number 1378 00:59:40,280 --> 00:59:48,825 of distracters, response time-- 1379 00:59:48,825 --> 00:59:51,530 rt is response time-- 1380 00:59:51,530 --> 00:59:55,110 so we have a disjunctive trial. 1381 00:59:55,110 --> 00:59:58,730 That's the black circle among white circles. 1382 00:59:58,730 --> 01:00:03,140 And your response is going to stay pretty much even, 1383 01:00:03,140 --> 01:00:06,560 whereas for conjunctive trial, it's usually harder. 1384 01:00:06,560 --> 01:00:08,690 You're going to start out slower. 1385 01:00:08,690 --> 01:00:10,694 And as the number of distracters-- 1386 01:00:10,694 --> 01:00:12,610 this is not the order I wanted to do these in. 1387 01:00:12,610 --> 01:00:13,109 I'm going to 1388 01:00:13,109 --> 01:00:14,260 do this in this order. 1389 01:00:14,260 --> 01:00:16,790 That's the graph I'm drawing. 1390 01:00:16,790 --> 01:00:19,790 Number of distracters-- there we go. 1391 01:00:19,790 --> 01:00:20,707 That makes more sense. 1392 01:00:20,707 --> 01:00:22,539 AUDIENCE: Variable goes on the [INAUDIBLE].. 1393 01:00:22,539 --> 01:00:24,840 ABBY NOYCE: Yes, number of distracters, reaction time. 1394 01:00:24,840 --> 01:00:27,170 And as the number of distracters increases 1395 01:00:27,170 --> 01:00:28,700 with a conjunctivitis task-- 1396 01:00:28,700 --> 01:00:32,450 AUDIENCE: Is that 0 right there? 1397 01:00:32,450 --> 01:00:37,790 [INAUDIBLE] 0, would there be a point at 0 for zero distrators? 1398 01:00:37,790 --> 01:00:40,634 Because you'd automatically know. 1399 01:00:40,634 --> 01:00:42,130 So it would be a little hole there. 1400 01:00:42,130 --> 01:00:43,180 ABBY NOYCE: Well, there's still some reaction time 1401 01:00:43,180 --> 01:00:43,540 involved, right? 1402 01:00:43,540 --> 01:00:44,400 We're measuring how long it takes 1403 01:00:44,400 --> 01:00:46,060 you to press the spacebar that says, 1404 01:00:46,060 --> 01:00:48,612 yes, my target is in this field. 1405 01:00:48,612 --> 01:00:50,320 There's no such thing as a zero response. 1406 01:00:50,320 --> 01:00:52,900 This might be the 75-millisecond line though. 1407 01:00:55,299 --> 01:00:55,840 I don't know. 1408 01:00:55,840 --> 01:00:57,970 I don't have good numbers for exactly how long would take. 1409 01:00:57,970 --> 01:00:58,810 But it's fast. 1410 01:00:58,810 --> 01:01:02,200 And the thing that's relevant is that it's consistently fast 1411 01:01:02,200 --> 01:01:04,380 no matter how many distracters I give you. 1412 01:01:04,380 --> 01:01:06,520 If I give you 100 white circles and one black one, 1413 01:01:06,520 --> 01:01:08,980 you're just as fast at finding the black one 1414 01:01:08,980 --> 01:01:11,980 as you were with 15 others. 1415 01:01:11,980 --> 01:01:14,350 With conjunctive tasks on the other hand, 1416 01:01:14,350 --> 01:01:18,540 reaction time goes up, pretty much probably not quite 1417 01:01:18,540 --> 01:01:20,184 that steeply, but pretty much linearly. 1418 01:01:20,184 --> 01:01:22,100 AUDIENCE: Shouldn't this guard the same points 1419 01:01:22,100 --> 01:01:25,192 since they'd only-- if there are zero distractors, 1420 01:01:25,192 --> 01:01:26,650 would take the same amount of time. 1421 01:01:26,650 --> 01:01:27,070 ABBY NOYCE: Yeah. 1422 01:01:27,070 --> 01:01:28,240 We're probably not starting at zero. 1423 01:01:28,240 --> 01:01:30,970 We're probably starting at five or something, something small, 1424 01:01:30,970 --> 01:01:32,396 three. 1425 01:01:32,396 --> 01:01:33,770 Reaction time will start at zero. 1426 01:01:33,770 --> 01:01:35,770 We won't do a cheater graph where we skip 1427 01:01:35,770 --> 01:01:37,112 the first 300 or something. 1428 01:01:37,112 --> 01:01:38,570 But the thing that's interesting is 1429 01:01:38,570 --> 01:01:42,040 conjunctive trials tend to get linear. conjunctive trials, 1430 01:01:42,040 --> 01:01:44,510 every time you add a distracter, the amount of extra work 1431 01:01:44,510 --> 01:01:48,010 you have to do to find that target goes up. 1432 01:01:48,010 --> 01:01:50,260 And so what people conclude from this 1433 01:01:50,260 --> 01:01:52,000 is that you have to focus your attention 1434 01:01:52,000 --> 01:01:55,810 on each item individually in a conjunctive search 1435 01:01:55,810 --> 01:01:58,850 in order to confirm that it is or is not your target. 1436 01:02:01,810 --> 01:02:04,350 So feature integration theorists say that how 1437 01:02:04,350 --> 01:02:06,990 you do this is your-- 1438 01:02:06,990 --> 01:02:10,230 we talked last week about how the visual system is modular. 1439 01:02:10,230 --> 01:02:12,540 Different parts of it specialize in different features. 1440 01:02:12,540 --> 01:02:15,870 That's not a hugely new idea. 1441 01:02:15,870 --> 01:02:17,250 And feature integration theorists 1442 01:02:17,250 --> 01:02:20,430 say that for each feature, you have information 1443 01:02:20,430 --> 01:02:22,110 about the location of it. 1444 01:02:22,110 --> 01:02:25,620 So you have information about color and location 1445 01:02:25,620 --> 01:02:27,300 over in the color module. 1446 01:02:27,300 --> 01:02:29,220 You have information about shape and location 1447 01:02:29,220 --> 01:02:31,110 over in the shape module. 1448 01:02:31,110 --> 01:02:34,260 You have information about direction 1449 01:02:34,260 --> 01:02:36,990 of movement and location over in the direction 1450 01:02:36,990 --> 01:02:39,180 of movement module. 1451 01:02:39,180 --> 01:02:41,850 And this junctive search is required that you only 1452 01:02:41,850 --> 01:02:43,110 consult one of these maps. 1453 01:02:43,110 --> 01:02:44,984 You just say, hey, look, color map. 1454 01:02:44,984 --> 01:02:45,900 There's the black one. 1455 01:02:45,900 --> 01:02:47,210 And it goes. 1456 01:02:47,210 --> 01:02:50,160 And in a conjunctive search, you have 1457 01:02:50,160 --> 01:02:54,120 to line the color map and the shape map up against each other 1458 01:02:54,120 --> 01:02:57,930 and look for the point where you've got both your targets, 1459 01:02:57,930 --> 01:03:00,960 both your target features, where you've got circle and black 1460 01:03:00,960 --> 01:03:01,875 in the same location. 1461 01:03:05,640 --> 01:03:07,900 So you'll notice that like feature integration theory, 1462 01:03:07,900 --> 01:03:12,310 like spotlight theory, feature integration theory is visual, 1463 01:03:12,310 --> 01:03:17,770 and it's about attention bound to location in a big way. 1464 01:03:17,770 --> 01:03:19,669 This whole attention and location thing 1465 01:03:19,669 --> 01:03:22,210 works really well when you're talking about visual attention. 1466 01:03:22,210 --> 01:03:24,280 It works a lot less well when you're 1467 01:03:24,280 --> 01:03:29,500 attending to an auditory stream, or to some kind 1468 01:03:29,500 --> 01:03:31,490 of internal imagery. 1469 01:03:31,490 --> 01:03:34,480 Think about your first grade birthday party. 1470 01:03:34,480 --> 01:03:36,340 It's not a location bound task. 1471 01:03:36,340 --> 01:03:39,720 It's an imagery task. 1472 01:03:39,720 --> 01:03:43,030 So these are some of the big gaps in how 1473 01:03:43,030 --> 01:03:46,595 people think about this stuff. 1474 01:03:46,595 --> 01:03:49,630 AUDIENCE: When it says two, does it mean two or more or-- 1475 01:03:49,630 --> 01:03:51,393 ABBY NOYCE: Yes, two or more. 1476 01:03:55,337 --> 01:03:59,172 AUDIENCE: --or do we only manage to compare two maps at a time. 1477 01:03:59,172 --> 01:04:01,130 ABBY NOYCE: No, you can do conjunctive searches 1478 01:04:01,130 --> 01:04:01,921 that require three. 1479 01:04:01,921 --> 01:04:07,010 If we take our white circles and black squares, and we add-- 1480 01:04:07,010 --> 01:04:08,582 I don't know. 1481 01:04:08,582 --> 01:04:10,790 And if we have targets that are both drifting upwards 1482 01:04:10,790 --> 01:04:12,322 and drifting downwards on a screen, 1483 01:04:12,322 --> 01:04:14,030 and we're looking for a black circle that 1484 01:04:14,030 --> 01:04:16,410 is drifting downwards that's a third feature, 1485 01:04:16,410 --> 01:04:18,200 and these theorists would say you've 1486 01:04:18,200 --> 01:04:21,440 got a third map that you're comparing. 1487 01:04:21,440 --> 01:04:23,360 AUDIENCE: Do you pair [INAUDIBLE] 1488 01:04:23,360 --> 01:04:26,250 at the same time [INAUDIBLE]? 1489 01:04:26,250 --> 01:04:27,720 ABBY NOYCE: Well, even if you only 1490 01:04:27,720 --> 01:04:29,290 compare-- in a conjunctive search, 1491 01:04:29,290 --> 01:04:31,050 even if you compare two of the three, 1492 01:04:31,050 --> 01:04:33,544 you're going to have multiple potential targets. 1493 01:04:33,544 --> 01:04:34,710 So you're going to have to-- 1494 01:04:34,710 --> 01:04:35,667 AUDIENCE: [INAUDIBLE]. 1495 01:04:35,667 --> 01:04:36,750 ABBY NOYCE: For the third? 1496 01:04:36,750 --> 01:04:37,110 AUDIENCE: For the third. 1497 01:04:37,110 --> 01:04:38,250 ABBY NOYCE: Yeah. 1498 01:04:38,250 --> 01:04:41,550 I don't know if anybody has come up 1499 01:04:41,550 --> 01:04:43,650 with a good theory about what happens 1500 01:04:43,650 --> 01:04:47,820 in that case about the details of how it happens, 1501 01:04:47,820 --> 01:04:50,160 or how you would get experimental evidence 1502 01:04:50,160 --> 01:04:52,520 to confirm one or the other. 1503 01:04:55,887 --> 01:04:57,280 It's an interesting question. 1504 01:04:57,280 --> 01:04:58,154 It's a good question. 1505 01:05:01,460 --> 01:05:06,970 All right, most neuroscientists who look at attention 1506 01:05:06,970 --> 01:05:08,710 don't think about attention quite the way 1507 01:05:08,710 --> 01:05:10,168 that the cognitive psych people do. 1508 01:05:10,168 --> 01:05:11,980 They say that attention is what happens. 1509 01:05:11,980 --> 01:05:13,240 You've got different inputs. 1510 01:05:13,240 --> 01:05:16,540 The inputs compete with each other. 1511 01:05:16,540 --> 01:05:20,920 And the strongest input gets bumped up, 1512 01:05:20,920 --> 01:05:24,947 and it gets brought into conscious awareness. 1513 01:05:24,947 --> 01:05:27,280 Different inputs can compete at any stage of processing. 1514 01:05:27,280 --> 01:05:30,010 You can have very early processing competing 1515 01:05:30,010 --> 01:05:32,380 with very late processing. 1516 01:05:32,380 --> 01:05:36,220 And exactly what makes an input stronger 1517 01:05:36,220 --> 01:05:37,660 has a lot of variation. 1518 01:05:37,660 --> 01:05:40,330 It can depend on what kind of task you're trying to do. 1519 01:05:40,330 --> 01:05:44,590 It can depend on characteristics of the inputs. 1520 01:05:44,590 --> 01:05:46,090 A loud bang is probably always going 1521 01:05:46,090 --> 01:05:48,835 to be a reasonably salient piece of stimulus. 1522 01:05:51,520 --> 01:05:53,920 And yeah, this competition theory 1523 01:05:53,920 --> 01:05:56,302 is how a lot of neural people who look at the stuff 1524 01:05:56,302 --> 01:05:57,260 like to think about it. 1525 01:06:02,595 --> 01:06:03,800 Oops, sorry. 1526 01:06:03,800 --> 01:06:05,156 Go back. 1527 01:06:05,156 --> 01:06:06,635 Give me a thing. 1528 01:06:10,090 --> 01:06:12,269 Come on, little laptop. 1529 01:06:12,269 --> 01:06:13,741 There we go. 1530 01:06:13,741 --> 01:06:14,241 No. 1531 01:06:19,171 --> 01:06:24,960 OK, questions about attention, the theories of attention, 1532 01:06:24,960 --> 01:06:28,555 models of attention? 1533 01:06:28,555 --> 01:06:31,537 All right. 1534 01:06:31,537 --> 01:06:34,519 AUDIENCE: Can you go back to the slide that introduced 1535 01:06:34,519 --> 01:06:36,368 the spotlight theory? 1536 01:06:36,368 --> 01:06:37,472 ABBY NOYCE: I can. 1537 01:06:37,472 --> 01:06:39,430 AUDIENCE: About half way through copying down I 1538 01:06:39,430 --> 01:06:40,679 started listening to you and-- 1539 01:06:40,679 --> 01:06:41,645 ABBY NOYCE: That's OK. 1540 01:06:41,645 --> 01:06:44,230 So spotlight theory, again, is this idea 1541 01:06:44,230 --> 01:06:47,920 that attention focuses on a location in space, 1542 01:06:47,920 --> 01:06:50,224 that when you attend to a location in space, 1543 01:06:50,224 --> 01:06:52,265 you tend to everything within this little circle. 1544 01:06:55,680 --> 01:07:01,500 It's not a great theory, but it's a longstanding one. 1545 01:07:01,500 --> 01:07:06,840 OK, the paper we are reading this week is "Covert-Orienting 1546 01:07:06,840 --> 01:07:10,620 of Attention and Over Eye Movements Activate Identical 1547 01:07:10,620 --> 01:07:11,460 Brain Regions." 1548 01:07:11,460 --> 01:07:15,780 I figured we'd read an FMRI study because they're fun. 1549 01:07:15,780 --> 01:07:19,800 So I'm going to take one, pass them. 1550 01:07:19,800 --> 01:07:20,820 I'm going to apologize. 1551 01:07:20,820 --> 01:07:23,130 My printer decided that we need to print stuff 1552 01:07:23,130 --> 01:07:26,885 so that it flips top to bottom and not left to right because I 1553 01:07:26,885 --> 01:07:27,510 don't know why. 1554 01:07:27,510 --> 01:07:30,600 Dan did something to it, and it wouldn't print at all for me 1555 01:07:30,600 --> 01:07:31,350 for several hours. 1556 01:07:31,350 --> 01:07:33,360 And then when it did, it did this. 1557 01:07:33,360 --> 01:07:35,080 AUDIENCE: [INAUDIBLE]. 1558 01:07:35,080 --> 01:07:36,780 ABBY NOYCE: Yes. 1559 01:07:36,780 --> 01:07:40,290 And I'm good with computers, but he's 1560 01:07:40,290 --> 01:07:42,180 a software engineer, and a network geek, 1561 01:07:42,180 --> 01:07:43,360 and does this for a living. 1562 01:07:43,360 --> 01:07:45,600 So he has ownership of our home network. 1563 01:07:45,600 --> 01:07:47,019 So then he changes things. 1564 01:07:47,019 --> 01:07:48,060 And then he goes to work. 1565 01:07:48,060 --> 01:07:53,160 And we were not happy this afternoon. 1566 01:07:53,160 --> 01:07:55,040 AUDIENCE: [INAUDIBLE]. 1567 01:07:55,040 --> 01:07:55,690 ABBY NOYCE: No. 1568 01:07:55,690 --> 01:07:56,898 My printer's name is Papyrus. 1569 01:07:56,898 --> 01:07:58,830 It is Papyrus.local. 1570 01:07:58,830 --> 01:07:59,556 AUDIENCE: Really? 1571 01:07:59,556 --> 01:08:00,222 ABBY NOYCE: Yes. 1572 01:08:03,470 --> 01:08:05,140 It's a really nice printer. 1573 01:08:05,140 --> 01:08:06,490 It actually has a web interface. 1574 01:08:06,490 --> 01:08:09,280 If you go to just type Papyrus into your web browser, 1575 01:08:09,280 --> 01:08:11,120 then it gives you the controls for it. 1576 01:08:18,010 --> 01:08:34,950 All right, so this is a first really neuro paper 1577 01:08:34,950 --> 01:08:35,850 that we've looked at. 1578 01:08:35,850 --> 01:08:36,700 Oh, come on. 1579 01:08:36,700 --> 01:08:38,039 None of that. 1580 01:08:38,039 --> 01:08:38,830 What are you doing? 1581 01:08:38,830 --> 01:08:40,044 What are you doing? 1582 01:08:43,529 --> 01:08:48,390 There we go, the first really neural paper that we looked at. 1583 01:08:48,390 --> 01:08:51,240 I think these can be a little bit more intense to read. 1584 01:08:51,240 --> 01:08:53,010 There's usually more new terminology 1585 01:08:53,010 --> 01:08:56,020 talking about particular brain regions and stuff. 1586 01:08:56,020 --> 01:08:57,660 Let's take a minute to just-- 1587 01:08:57,660 --> 01:09:00,330 let's try and read the abstract and make sure we can make sense 1588 01:09:00,330 --> 01:09:01,244 out of that. 1589 01:09:01,244 --> 01:09:02,910 And then I'm going to do like last week. 1590 01:09:02,910 --> 01:09:05,460 I'm going to break you guys up, and have 1591 01:09:05,460 --> 01:09:08,569 each group have a section of the thing 1592 01:09:08,569 --> 01:09:09,819 to report on to the group. 1593 01:09:22,312 --> 01:09:23,770 AUDIENCE: How long is the abstract? 1594 01:09:23,770 --> 01:09:25,090 ABBY NOYCE: The abstract is just the one paragraph 1595 01:09:25,090 --> 01:09:26,090 right on the first page. 1596 01:09:26,090 --> 01:09:26,840 AUDIENCE: Oh, OK. 1597 01:09:26,840 --> 01:09:29,340 ABBY NOYCE: So an abstract is usually a couple hundred words 1598 01:09:29,340 --> 01:09:30,215 summary of the paper. 1599 01:09:30,215 --> 01:09:32,298 So this is what you read to decide if you actually 1600 01:09:32,298 --> 01:09:33,570 want to read the whole thing. 1601 01:09:33,570 --> 01:09:34,986 AUDIENCE: [INAUDIBLE] investigated 1602 01:09:34,986 --> 01:09:40,630 the anatomical relationship between [INAUDIBLE].. 1603 01:09:40,630 --> 01:09:43,430 ABBY NOYCE: OK, so anatomical relationship, 1604 01:09:43,430 --> 01:09:45,800 so they mean they're looking at what kind of-- when they 1605 01:09:45,800 --> 01:09:47,383 say anatomy, we're presuming they mean 1606 01:09:47,383 --> 01:09:49,109 brain anatomy in this case. 1607 01:09:49,109 --> 01:09:51,109 What parts of your neuro anatomy are 1608 01:09:51,109 --> 01:09:55,860 involved when you covertly or overtly shift your attention? 1609 01:09:55,860 --> 01:09:57,860 We think we've got a pretty good grasp of what's 1610 01:09:57,860 --> 01:09:59,820 involved in shifting attention. 1611 01:09:59,820 --> 01:10:03,040 What's this covert and overt thing? 1612 01:10:03,040 --> 01:10:05,870 So covert, overt attention shifting 1613 01:10:05,870 --> 01:10:08,270 is like actually moving your gauge, 1614 01:10:08,270 --> 01:10:10,190 directing your gaze to actually look 1615 01:10:10,190 --> 01:10:11,960 at a different object in the room. 1616 01:10:11,960 --> 01:10:13,910 So you could shift your gaze to the screen, 1617 01:10:13,910 --> 01:10:15,890 or to the graph on the board, and all of these 1618 01:10:15,890 --> 01:10:20,216 would be overt shifts of attention to new stimuli. 1619 01:10:20,216 --> 01:10:23,750 A covert shift of attention it would be you 1620 01:10:23,750 --> 01:10:26,380 keep looking at the same point. 1621 01:10:26,380 --> 01:10:27,920 But while keeping your eyes fixated, 1622 01:10:27,920 --> 01:10:30,740 say, on me, you nonetheless attended 1623 01:10:30,740 --> 01:10:34,970 to what was happening on the screen, which is hard. 1624 01:10:34,970 --> 01:10:37,670 We're pretty strongly wired to want to look directly 1625 01:10:37,670 --> 01:10:40,040 at whatever it is that we're attending to, 1626 01:10:40,040 --> 01:10:42,490 but we certainly can do it. 1627 01:10:42,490 --> 01:10:44,240 So you can look directly at me and think 1628 01:10:44,240 --> 01:10:46,531 about the laptop, or the screen, or something that's 1629 01:10:46,531 --> 01:10:48,530 a little ways away from wherever you're actually 1630 01:10:48,530 --> 01:10:49,682 fixating your eyes. 1631 01:10:49,682 --> 01:10:51,140 So that's a shift of attention that 1632 01:10:51,140 --> 01:10:56,602 doesn't have a physical moving of gaze that goes with it. 1633 01:10:56,602 --> 01:10:58,810 AUDIENCE: What if you keep looking at the same thing, 1634 01:10:58,810 --> 01:11:00,862 but you're actually concentrating 1635 01:11:00,862 --> 01:11:03,299 on a different aspect of that? 1636 01:11:03,299 --> 01:11:05,340 ABBY NOYCE: A different aspect of the same thing? 1637 01:11:05,340 --> 01:11:06,506 AUDIENCE: Of the same thing. 1638 01:11:06,506 --> 01:11:09,910 ABBY NOYCE: So that's a shift in attention 1639 01:11:09,910 --> 01:11:12,120 within the same object. 1640 01:11:12,120 --> 01:11:16,076 I don't know which of those categories I would put that in. 1641 01:11:16,076 --> 01:11:17,700 It's not quite-- they're looking at two 1642 01:11:17,700 --> 01:11:20,599 that are a little bit better distinguished than that. 1643 01:11:20,599 --> 01:11:22,140 I'm going to talk tomorrow, actually, 1644 01:11:22,140 --> 01:11:25,100 about some research that the lab I used to work in was doing. 1645 01:11:25,100 --> 01:11:26,558 And one of the things they're doing 1646 01:11:26,558 --> 01:11:29,369 is have people attend to different-- not people-- that 1647 01:11:29,369 --> 01:11:31,410 attend to different aspects of the same stimulus. 1648 01:11:31,410 --> 01:11:34,706 AUDIENCE: [INAUDIBLE]. 1649 01:11:34,706 --> 01:11:36,080 ABBY NOYCE: Overt is the actually 1650 01:11:36,080 --> 01:11:37,079 moving your gaze around. 1651 01:11:37,079 --> 01:11:39,230 Covert is the being sneaky about it. 1652 01:11:39,230 --> 01:11:42,390 Covert operations is spy operations. 1653 01:11:42,390 --> 01:11:48,120 So it's being subtle about how you're shifting attention. 1654 01:11:48,120 --> 01:11:56,050 OK, second sentence, who wants to read? 1655 01:11:56,050 --> 01:11:57,606 [? Zechariah. ?] 1656 01:11:57,606 --> 01:11:59,186 AUDIENCE: Previous studies have found 1657 01:11:59,186 --> 01:12:01,180 that the areas of the brain activated 1658 01:12:01,180 --> 01:12:03,862 by the covert and overt shifts are 1659 01:12:03,862 --> 01:12:05,320 [INAUDIBLE] are very similar. 1660 01:12:05,320 --> 01:12:06,778 ABBY NOYCE: OK, questions? 1661 01:12:10,400 --> 01:12:11,816 Third sentence, who wants to read? 1662 01:12:15,220 --> 01:12:17,586 Natasha? 1663 01:12:17,586 --> 01:12:21,418 AUDIENCE: [INAUDIBLE]. 1664 01:12:21,418 --> 01:12:23,334 There are few issues [INAUDIBLE].. 1665 01:12:26,700 --> 01:12:27,844 ABBY NOYCE: Jenny? 1666 01:12:27,844 --> 01:12:32,602 AUDIENCE: Primarily, [INAUDIBLE] to produce covert [INAUDIBLE] 1667 01:12:32,602 --> 01:12:33,870 were not always comparable. 1668 01:12:33,870 --> 01:12:36,120 ABBY NOYCE: All right, so they're listing off problems 1669 01:12:36,120 --> 01:12:36,790 with-- 1670 01:12:36,790 --> 01:12:38,980 other people have tried to look at covert and overt attention 1671 01:12:38,980 --> 01:12:41,320 to see if they're using the same parts of your brain. 1672 01:12:41,320 --> 01:12:43,990 And these guys are like, OK, we think 1673 01:12:43,990 --> 01:12:45,490 there's some good evidence for this, 1674 01:12:45,490 --> 01:12:46,990 but there's some problems in the previous work. 1675 01:12:46,990 --> 01:12:48,685 So we're going to try and put together 1676 01:12:48,685 --> 01:12:50,140 a study that fixes these problems, 1677 01:12:50,140 --> 01:12:54,990 and tries to prove once and for all that this is really true. 1678 01:12:54,990 --> 01:12:56,430 And this is fairly typical. 1679 01:12:56,430 --> 01:12:58,970 This is trying to reproduce somebody else's results 1680 01:12:58,970 --> 01:13:02,320 sort of thing that is a really important part of what 1681 01:13:02,320 --> 01:13:05,324 makes science go, is having results that you can have. 1682 01:13:05,324 --> 01:13:07,240 You can do the same experiment more than once, 1683 01:13:07,240 --> 01:13:08,492 get the same results. 1684 01:13:08,492 --> 01:13:10,450 But on the other hand, it's hard to get funding 1685 01:13:10,450 --> 01:13:12,908 for an experiment if you're just reproducing someone else's 1686 01:13:12,908 --> 01:13:13,829 results. 1687 01:13:13,829 --> 01:13:15,620 So really frequently you'll see people say, 1688 01:13:15,620 --> 01:13:17,494 we're going to try to reproduce their results 1689 01:13:17,494 --> 01:13:20,340 and we're going to fix some flaws in the original study. 1690 01:13:20,340 --> 01:13:24,740 All right, so not always comparable. 1691 01:13:24,740 --> 01:13:27,880 So if you're asking people to do different tasks, tasks that 1692 01:13:27,880 --> 01:13:30,220 are not sufficiently similar, and then you're 1693 01:13:30,220 --> 01:13:35,632 trying to compare how brain activity is changing, 1694 01:13:35,632 --> 01:13:37,840 then it's possible that the differences you're seeing 1695 01:13:37,840 --> 01:13:40,090 are not differences in overt and covert attention 1696 01:13:40,090 --> 01:13:41,140 shifting per se. 1697 01:13:41,140 --> 01:13:42,790 They're actually differences that have 1698 01:13:42,790 --> 01:13:46,110 to do with the original task. 1699 01:13:46,110 --> 01:13:46,910 What are you up to? 1700 01:13:46,910 --> 01:13:49,810 Secondly, [? Naman, ?] can you read? 1701 01:13:49,810 --> 01:13:51,280 AUDIENCE: [INAUDIBLE]. 1702 01:13:55,690 --> 01:13:58,900 ABBY NOYCE: OK, so another change they're trying to-- 1703 01:13:58,900 --> 01:14:01,210 another thing they're trying to clarify in this study. 1704 01:14:04,090 --> 01:14:07,355 Wayne, thirdly? 1705 01:14:07,355 --> 01:14:09,615 AUDIENCE: Thirdly, the previous studies differed 1706 01:14:09,615 --> 01:14:14,282 on what are shifts of attention were based on exogenous 1707 01:14:14,282 --> 01:14:17,120 or endogenous [INAUDIBLE]. 1708 01:14:17,120 --> 01:14:19,445 ABBY NOYCE: Right, that's that exogenous and endogenous 1709 01:14:19,445 --> 01:14:21,570 attention that we talked about, the stimulus driven 1710 01:14:21,570 --> 01:14:26,540 versus centrally controlled attentional shifts. 1711 01:14:26,540 --> 01:14:30,150 Exogenous, exo, outside, that's a stimulus 1712 01:14:30,150 --> 01:14:31,840 controlled attention shift. 1713 01:14:31,840 --> 01:14:34,000 Endogenous is coming from within. 1714 01:14:34,000 --> 01:14:38,430 It's the internal centrally controlled attention shift. 1715 01:14:38,430 --> 01:14:40,028 Finally, Jen. 1716 01:14:40,028 --> 01:14:42,890 AUDIENCE: Finally, the statistical analyses 1717 01:14:42,890 --> 01:14:45,812 used by all previous studies failed 1718 01:14:45,812 --> 01:14:49,708 to account for between subject variability. 1719 01:14:49,708 --> 01:14:51,656 ABBY NOYCE: Mm-hm. 1720 01:14:51,656 --> 01:14:53,604 Do you know what that means? 1721 01:14:57,030 --> 01:15:00,324 So we didn't talk about this too much when 1722 01:15:00,324 --> 01:15:01,740 we talked about experiment design, 1723 01:15:01,740 --> 01:15:04,890 but there's two big ways of setting up an experiment where 1724 01:15:04,890 --> 01:15:06,760 you want to compare two tasks. 1725 01:15:06,760 --> 01:15:08,942 And one is a between subjects design, 1726 01:15:08,942 --> 01:15:10,650 where you've got a group A and a group B, 1727 01:15:10,650 --> 01:15:14,520 and you measure group A, and you measure Group B, 1728 01:15:14,520 --> 01:15:16,170 and you see how they are different. 1729 01:15:16,170 --> 01:15:17,754 The other is a within subjects design. 1730 01:15:17,754 --> 01:15:19,170 If you want to know whether people 1731 01:15:19,170 --> 01:15:21,390 are better at driving that course while on the phone 1732 01:15:21,390 --> 01:15:24,450 or listening to the radio, you could have, for each person, 1733 01:15:24,450 --> 01:15:27,270 have them drive the course twice, 1734 01:15:27,270 --> 01:15:29,940 once listening to the radio, once talking on the phone, 1735 01:15:29,940 --> 01:15:32,890 and see how each person's performance changes. 1736 01:15:32,890 --> 01:15:34,842 So that's a within subject study. 1737 01:15:34,842 --> 01:15:36,300 It's probably a bad example for it. 1738 01:15:36,300 --> 01:15:37,383 There's a lot of confines. 1739 01:15:37,383 --> 01:15:38,370 AUDIENCE: [INAUDIBLE]? 1740 01:15:38,370 --> 01:15:41,710 ABBY NOYCE: Matched pair is a between subjects study. 1741 01:15:41,710 --> 01:15:47,020 But it's a form of counter-balancing to make sure 1742 01:15:47,020 --> 01:15:48,946 that you don't have a confound where-- 1743 01:15:48,946 --> 01:15:50,820 I don't know-- all the people in group A come 1744 01:15:50,820 --> 01:15:52,800 from rich households, and all the poor people in group B 1745 01:15:52,800 --> 01:15:54,060 come from poor households. 1746 01:15:54,060 --> 01:15:56,730 And then you have a big socioeconomic factor 1747 01:15:56,730 --> 01:15:59,220 that's going to be a big problem for whatever, that 1748 01:15:59,220 --> 01:16:01,303 could be accounting for whatever other differences 1749 01:16:01,303 --> 01:16:02,174 you're finding. 1750 01:16:02,174 --> 01:16:03,840 And if what you're really trying to find 1751 01:16:03,840 --> 01:16:07,320 is the effects of caffeine on a memorization task, 1752 01:16:07,320 --> 01:16:09,900 having this total other issue that might be affecting 1753 01:16:09,900 --> 01:16:13,500 people's performance is bad. 1754 01:16:13,500 --> 01:16:15,654 So a match pair would say, OK, you take-- 1755 01:16:15,654 --> 01:16:17,070 if we're looking at socioeconomic, 1756 01:16:17,070 --> 01:16:19,581 you take two people with even socioeconomic status. 1757 01:16:19,581 --> 01:16:21,330 You put one in group A and one in group B, 1758 01:16:21,330 --> 01:16:23,280 and you do the next same thing, and you 1759 01:16:23,280 --> 01:16:25,290 make sure that they're evenly spread. 1760 01:16:25,290 --> 01:16:30,150 So it's more it's more careful than random assignment. 1761 01:16:30,150 --> 01:16:32,240 With smaller groups, where there's literally 1762 01:16:32,240 --> 01:16:33,740 a big spread on something like that, 1763 01:16:33,740 --> 01:16:37,864 is where you're likely to see a match pair paradigm. 1764 01:16:37,864 --> 01:16:40,350 All right, so previous studies failed 1765 01:16:40,350 --> 01:16:42,850 to count for between subject variability. 1766 01:16:42,850 --> 01:16:46,740 So where if person A's pattern of activation 1767 01:16:46,740 --> 01:16:49,284 is different than person B's is different than person C's, 1768 01:16:49,284 --> 01:16:50,700 these guys are trying to take that 1769 01:16:50,700 --> 01:16:52,824 into account in their analysis, and try and come up 1770 01:16:52,824 --> 01:16:55,340 with a broader, good picture. 1771 01:16:55,340 --> 01:16:57,360 All right, the present study was designed 1772 01:16:57,360 --> 01:16:58,710 to address all these issues. 1773 01:17:01,510 --> 01:17:04,420 12 healthy, Helen, do you want to read? 1774 01:17:04,420 --> 01:17:06,520 AUDIENCE: Oh, 12 healthy subjects 1775 01:17:06,520 --> 01:17:08,720 performed either covert or overt shifts of attention 1776 01:17:08,720 --> 01:17:12,943 are the function of magnetic resonance imaging, MRI-- 1777 01:17:12,943 --> 01:17:17,900 fMRI [INAUDIBLE] was continuously measured. 1778 01:17:17,900 --> 01:17:19,650 ABBY NOYCE: OK, so they had 12 subjects, 1779 01:17:19,650 --> 01:17:22,460 had them do these shifts of attention, 1780 01:17:22,460 --> 01:17:24,586 measured the FMRI signal. 1781 01:17:24,586 --> 01:17:26,460 We talked about FMRI as an imaging mechanism. 1782 01:17:26,460 --> 01:17:29,520 FMRI tracks changes in blood flow 1783 01:17:29,520 --> 01:17:34,290 in the brain, which is believed to be good, 1784 01:17:34,290 --> 01:17:37,150 to correlate really well with changes in neural activity. 1785 01:17:37,150 --> 01:17:41,850 So it's an indirect measure of neural activity in the brain. 1786 01:17:41,850 --> 01:17:42,600 What are we up to? 1787 01:17:42,600 --> 01:17:43,100 In line. 1788 01:17:43,100 --> 01:17:44,516 [INAUDIBLE],, do you want to read? 1789 01:17:44,516 --> 01:17:46,464 AUDIENCE: In line with the previous studies, 1790 01:17:46,464 --> 01:17:49,440 the previous study showed [INAUDIBLE].. 1791 01:18:03,328 --> 01:18:05,940 ABBY NOYCE: OK, so pretty much the same brain 1792 01:18:05,940 --> 01:18:08,670 areas are involved in both. 1793 01:18:08,670 --> 01:18:09,610 Sarah, furthermore? 1794 01:18:09,610 --> 01:18:13,002 AUDIENCE: Furthermore, [INAUDIBLE] endogenous 1795 01:18:13,002 --> 01:18:16,898 [INAUDIBLE] shifts of attention, the study 1796 01:18:16,898 --> 01:18:19,820 found that the covert shifts of attention 1797 01:18:19,820 --> 01:18:22,255 resulted in high level activations of the brain 1798 01:18:22,255 --> 01:18:26,180 than overt shifts of attention. 1799 01:18:26,180 --> 01:18:27,020 ABBY NOYCE: Good. 1800 01:18:27,020 --> 01:18:29,870 So remember, endogenous driven shifts 1801 01:18:29,870 --> 01:18:33,890 are driven by centrally controlled shifts. 1802 01:18:33,890 --> 01:18:38,810 You're choosing to attend to one stimulus or another. 1803 01:18:38,810 --> 01:18:41,780 And those resulted in higher levels of activation. 1804 01:18:41,780 --> 01:18:43,600 Good. 1805 01:18:43,600 --> 01:18:45,420 Last sentence. 1806 01:18:45,420 --> 01:18:46,982 [? Genti, ?] you want to read? 1807 01:18:46,982 --> 01:18:48,440 AUDIENCE: The results of this study 1808 01:18:48,440 --> 01:18:51,254 provide support for the [INAUDIBLE] theory 1809 01:18:51,254 --> 01:18:52,192 of attention. 1810 01:18:52,192 --> 01:18:55,160 That posits that the attentional motor systems 1811 01:18:55,160 --> 01:18:57,690 are the result of neural activation 1812 01:18:57,690 --> 01:18:59,540 in the same areas of the brain. 1813 01:18:59,540 --> 01:19:00,740 ABBY NOYCE: Good. 1814 01:19:00,740 --> 01:19:03,140 Cool. 1815 01:19:03,140 --> 01:19:04,690 OK.