1 00:00:00,000 --> 00:00:01,230 Welcome to The Face Podcast. 2 00:00:01,230 --> 00:00:02,670 I'm your host, Masoud Saman. 3 00:00:03,000 --> 00:00:06,510 On this show, we explore the human face, not just as anatomy, but as 4 00:00:06,510 --> 00:00:08,640 identity, memory, and social meaning. 5 00:00:08,640 --> 00:00:10,560 The face is how we know each other. 6 00:00:10,650 --> 00:00:14,799 It's how we navigate society, and are recognized today. 7 00:00:15,069 --> 00:00:18,009 I wanna talk about something that most of us take completely for granted. 8 00:00:18,117 --> 00:00:18,297 granted. 9 00:00:18,759 --> 00:00:22,959 The effortless ability to recognize thousands of faces across a lifetime. 10 00:00:23,529 --> 00:00:27,494 By adulthood, the average person can recognize at least 5,000 faces. 11 00:00:28,033 --> 00:00:31,913 Some estimates suggest even more, and yet none of us were born 12 00:00:31,913 --> 00:00:33,893 with that ability fully formed. 13 00:00:34,133 --> 00:00:36,953 So it's something that develops slowly over years. 14 00:00:37,393 --> 00:00:40,393 What makes faces so special that the brain seems to carve out 15 00:00:40,423 --> 00:00:41,905 distinct real estate for it? 16 00:00:42,205 --> 00:00:42,325 It? 17 00:00:42,403 --> 00:00:47,563 Why does the ability mature over such a prolonged trajectory even into adulthood? 18 00:00:47,593 --> 00:00:51,943 And what happens when it fails to explore these questions and more? 19 00:00:51,973 --> 00:00:53,795 I'm speaking with Marlene Behrmann. 20 00:00:53,845 --> 00:00:54,265 Berman. 21 00:00:54,565 --> 00:00:57,018 Marlene is a cognitive neuroscientist and a professor 22 00:00:57,018 --> 00:00:59,598 of ophthalmology and psychology at the University of Pittsburgh. 23 00:00:59,838 --> 00:01:03,528 Her research has been foundational in understanding how faces are represented in 24 00:01:03,528 --> 00:01:06,198 the brain, how face recognition develops 25 00:01:06,248 --> 00:01:08,828 Why faces occupy such a privileged place. 26 00:01:09,353 --> 00:01:14,633 In the human mind, in the human brain, and how this skill unfolds across developments 27 00:01:14,903 --> 00:01:19,973 and what its breakdown reveals about the architecture of visual cognition, because 28 00:01:19,973 --> 00:01:24,653 sometimes we learn more about something's working when we study its failures. 29 00:01:25,163 --> 00:01:26,183 Marlene, welcome. 30 00:01:27,415 --> 00:01:28,435 Thank you so much. 31 00:01:28,515 --> 00:01:31,515 You make it all sounds so exciting. 32 00:01:31,515 --> 00:01:34,225 It gets me fired up all over again. 33 00:01:34,886 --> 00:01:35,606 That's wonderful. 34 00:01:35,606 --> 00:01:37,166 I'm so happy that you're here. 35 00:01:37,466 --> 00:01:39,746 I'm so excited that you accepted my invitation. 36 00:01:40,376 --> 00:01:41,576 Let's start from the beginning. 37 00:01:42,116 --> 00:01:45,536 How did you get into this amazingly interesting field of study? 38 00:01:46,794 --> 00:01:50,604 Like many people I know, I was not born to study? 39 00:01:50,604 --> 00:01:51,864 face recognition. 40 00:01:52,104 --> 00:01:56,664 In fact, I discovered it in a very circuitous trajectory. 41 00:01:57,024 --> 00:02:01,404 I'd initially been trained as a clinical speech pathologist and audiologist, 42 00:02:01,524 --> 00:02:04,569 and I worked with individuals who had communication deficits. 43 00:02:06,069 --> 00:02:10,959 I was especially interested in the fact that after stroke or tumors, 44 00:02:12,579 --> 00:02:16,779 seemed to have lost the ability to communicate a deficit known as aphasia. 45 00:02:16,899 --> 00:02:21,609 And what struck me most about this is I would go and see one patient 46 00:02:21,789 --> 00:02:26,289 who had damage, say to area X of the brain and their language impairment, 47 00:02:26,289 --> 00:02:28,089 had very particular properties. 48 00:02:28,089 --> 00:02:29,919 And then I went to see another patient. 49 00:02:30,264 --> 00:02:34,404 Who also had damage to the same area, and that patient showed the 50 00:02:34,404 --> 00:02:37,554 same breakdown of their behavior. 51 00:02:38,634 --> 00:02:42,864 occurred to me that there are clear rules in the brain that govern 52 00:02:42,864 --> 00:02:46,194 complex behavior after damage. 53 00:02:46,854 --> 00:02:51,744 Behavior breaks down in these very organized and predictable ways. 54 00:02:52,404 --> 00:02:54,554 So how did I get into faces? 55 00:02:54,654 --> 00:02:56,364 I want you to understand these rules. 56 00:02:56,544 --> 00:03:01,254 So I went back to do a PhD and acquire the skills to conduct research. 57 00:03:01,974 --> 00:03:07,664 And in the course of my graduate training, I was required to do a small side project. 58 00:03:07,784 --> 00:03:11,324 And I studied some patients who had lost the ability to recognize 59 00:03:11,324 --> 00:03:13,514 faces after brain damage. 60 00:03:13,994 --> 00:03:14,654 And this was. 61 00:03:15,284 --> 00:03:16,034 Amazing. 62 00:03:16,134 --> 00:03:17,994 There were so many questions. 63 00:03:18,714 --> 00:03:22,254 they recognize animal faces, cartoon faces? 64 00:03:22,554 --> 00:03:27,024 Some of the patients don't recognize themselves in photographs or 65 00:03:27,024 --> 00:03:28,824 even their own family members. 66 00:03:29,004 --> 00:03:33,354 And so this started me down this path of trying to understand the way 67 00:03:33,354 --> 00:03:38,034 the brain processes faces, here I am today to talk about it with you. 68 00:03:39,113 --> 00:03:39,833 That's fascinating. 69 00:03:39,833 --> 00:03:42,653 So you had a complete 180 of your career. 70 00:03:42,683 --> 00:03:45,009 You were boots on the ground, clinical field, 71 00:03:45,009 --> 00:03:48,554 diagnosis and treatment, and then you decided to study the cause. 72 00:03:48,864 --> 00:03:49,464 the cause? 73 00:03:49,704 --> 00:03:49,944 Yep. 74 00:03:50,594 --> 00:03:50,864 Exactly. 75 00:03:50,894 --> 00:03:54,494 It is very fascinating what you're telling me because prior to my, 76 00:03:54,584 --> 00:03:55,034 my, 77 00:03:55,034 --> 00:03:57,374 facial plastic surgery days, I was an 78 00:03:57,375 --> 00:03:59,024 An otolaryngologist. 79 00:03:59,325 --> 00:04:02,535 an otolaryngologist of course, we worked very closely with. 80 00:04:02,535 --> 00:04:03,523 Speech pathologist 81 00:04:03,765 --> 00:04:03,955 yes pathologists. 82 00:04:04,364 --> 00:04:07,665 I was a cancer surgeon, head and neck cancer surgeon, so I did a lot of 83 00:04:07,665 --> 00:04:12,671 laryngectomies, partial laryngectomies we worked very closely and yes, I also 84 00:04:12,671 --> 00:04:17,111 find these topics very fascinating and without having known that about you, and 85 00:04:17,111 --> 00:04:18,652 perhaps you didn't know that about me, 86 00:04:18,952 --> 00:04:19,012 Yeah. 87 00:04:19,061 --> 00:04:22,216 we came to this conversation about faces. 88 00:04:22,606 --> 00:04:27,436 Why are faces in your opinion, so important for us humans, 89 00:04:27,740 --> 00:04:36,080 I think it's pretty obvious to everybody that recognizing faces so crucial in, 90 00:04:36,080 --> 00:04:42,800 and this is so in every day of our lives, we use faces to identify friends. 91 00:04:42,890 --> 00:04:45,500 We also use faces to identify foes. 92 00:04:45,740 --> 00:04:51,205 And so in some sense it's critical for survival and obviously it's really 93 00:04:51,205 --> 00:04:53,185 important for social interactions. 94 00:04:53,702 --> 00:04:56,955 But is that something that we see across the species? 95 00:04:56,972 --> 00:05:00,362 Do we see that with other species who have preferential 96 00:05:00,362 --> 00:05:02,282 processing for faces like we do? 97 00:05:03,887 --> 00:05:05,117 it's a great question. 98 00:05:05,237 --> 00:05:09,977 And actually a number of years ago, I set out specifically to address this question 99 00:05:10,037 --> 00:05:18,547 and did a cross species, a a phylogenetic comparison of face recognition abilities. 100 00:05:19,717 --> 00:05:24,247 I'll actually tell you that I don't think that the human brain or 101 00:05:24,247 --> 00:05:29,677 other species, brains have carved out a very selective area that. 102 00:05:29,997 --> 00:05:33,412 Only is able to perform face recognition. 103 00:05:33,417 --> 00:05:34,917 So we can come to that a bit later. 104 00:05:35,247 --> 00:05:39,527 But suffice it to say that many other species use faces. 105 00:05:39,557 --> 00:05:44,657 Of course, other primates use faces in the way that humans use faces, but even. 106 00:05:45,792 --> 00:05:49,862 Chicks are born and they imprint on their parents' face. 107 00:05:49,862 --> 00:05:54,592 They the template of their parents' face, like in, in a newborn state. 108 00:05:54,892 --> 00:05:59,392 That's how important face recognition is for maternal bonding in that case. 109 00:05:59,722 --> 00:06:00,052 But 110 00:06:00,537 --> 00:06:00,827 Yeah. 111 00:06:02,182 --> 00:06:03,502 faces are important. 112 00:06:03,592 --> 00:06:03,682 And 113 00:06:03,982 --> 00:06:05,496 And we see that with infants. 114 00:06:05,512 --> 00:06:06,532 My own children. 115 00:06:06,948 --> 00:06:07,112 I 116 00:06:07,248 --> 00:06:07,518 I 117 00:06:08,042 --> 00:06:10,622 they weren't so good at recognizing my face, but maybe 118 00:06:10,622 --> 00:06:12,032 about it took about a week. 119 00:06:12,342 --> 00:06:15,342 I had to smile big and make my eyes big and, 120 00:06:15,478 --> 00:06:15,718 and 121 00:06:16,018 --> 00:06:21,312 do certain things to trigger the activity perhaps of this center of recognition 122 00:06:21,642 --> 00:06:26,172 that after about a week, they knew when it was me versus somebody completely. 123 00:06:26,817 --> 00:06:27,387 Different. 124 00:06:27,617 --> 00:06:29,807 And they would smile in, in response. 125 00:06:30,097 --> 00:06:32,347 Sometimes it wasn't my face, it was the light behind my 126 00:06:32,347 --> 00:06:33,397 head or something like this. 127 00:06:34,337 --> 00:06:37,307 And this kind of sounds like a skill, it sounds like it's 128 00:06:37,307 --> 00:06:39,107 something that we develop over time. 129 00:06:39,107 --> 00:06:41,567 It's not something that we're just born with and we 130 00:06:41,628 --> 00:06:41,898 we 131 00:06:42,407 --> 00:06:44,777 possess, and it's something perhaps that needs to be 132 00:06:44,883 --> 00:06:45,348 to be 133 00:06:45,648 --> 00:06:46,007 practiced. 134 00:06:46,187 --> 00:06:49,517 And what skill does one need to be able to recognize faces? 135 00:06:49,517 --> 00:06:50,161 So well? 136 00:06:50,177 --> 00:06:52,907 We talked about 5,000 distinct faces. 137 00:06:52,978 --> 00:06:53,267 What, 138 00:06:53,278 --> 00:06:53,638 What 139 00:06:53,807 --> 00:06:54,257 going on? 140 00:06:54,437 --> 00:06:55,187 How are we doing that? 141 00:06:55,333 --> 00:06:55,453 there? 142 00:06:55,453 --> 00:06:55,693 Yeah. 143 00:06:56,153 --> 00:06:57,443 There's no question. 144 00:06:57,443 --> 00:07:02,393 I think there's consensus in the field that it takes years for 145 00:07:02,393 --> 00:07:07,013 humans to develop this kind of expertise in recognizing faces. 146 00:07:07,013 --> 00:07:09,563 We are efficient, we are. 147 00:07:10,468 --> 00:07:13,248 It's done rapidly effortlessly. 148 00:07:14,388 --> 00:07:20,238 and one of the big challenges is that in the grand scheme of different things 149 00:07:20,238 --> 00:07:24,648 we can see in the world, faces are all remarkably similar to each other. 150 00:07:24,648 --> 00:07:29,118 We've all got two eyes above a nose that's above a mouth, and so it's 151 00:07:29,118 --> 00:07:31,248 a challenge for the visual system. 152 00:07:31,308 --> 00:07:35,178 Sometimes I think about face recognition as like a stress test. 153 00:07:35,958 --> 00:07:40,878 the visual system you have to make fine grained discriminations 154 00:07:40,878 --> 00:07:42,403 between individual faces. 155 00:07:42,708 --> 00:07:46,458 So it might be easy to discriminate between you and I now because I have 156 00:07:46,458 --> 00:07:51,198 glasses, which is just, so obvious and right upfront, dare I say. 157 00:07:52,458 --> 00:07:59,718 but really the ability to recognize faces is not very sophisticated at birth. 158 00:07:59,868 --> 00:08:02,808 So yes, infants are able to recognize. 159 00:08:04,083 --> 00:08:08,403 about faces, and particularly something about their parents' faces because 160 00:08:08,403 --> 00:08:12,753 they see them the most, almost like the chicks imprint, babies 161 00:08:12,783 --> 00:08:14,733 imprint on their parents' faces. 162 00:08:15,093 --> 00:08:19,763 And they only, we know that they can't even see very well soon after birth. 163 00:08:19,763 --> 00:08:24,203 And so they only have a very rough schematic of what your face looks like. 164 00:08:24,533 --> 00:08:27,563 And then, as you pointed out, over decades. 165 00:08:27,908 --> 00:08:32,318 The system becomes increasingly sophisticated and better configured. 166 00:08:32,618 --> 00:08:32,898 Okay. 167 00:08:32,898 --> 00:08:34,218 That that, that makes sense. 168 00:08:34,218 --> 00:08:38,358 That explains why I have to do these exaggerated things and show teeth and 169 00:08:38,358 --> 00:08:42,618 make my eyes big for, to elicit some sort of a reaction from a newborn. 170 00:08:43,101 --> 00:08:47,931 But what is the consequence of losing the ability to recognize faces? 171 00:08:47,981 --> 00:08:51,191 I'm sure if this is a skill, then perhaps. 172 00:08:51,731 --> 00:08:54,221 There was a possibility of losing it. 173 00:08:54,521 --> 00:08:54,597 it. 174 00:08:55,467 --> 00:09:02,187 So this is probably the scenario that kind of got me excited about studying 175 00:09:02,187 --> 00:09:08,307 faces is that I had this opportunity to study individuals, adults who were. 176 00:09:08,547 --> 00:09:12,927 Perfectly normal in face recognition and after some kind of brain damage 177 00:09:13,287 --> 00:09:16,257 lost that ability almost overnight. 178 00:09:16,307 --> 00:09:21,617 It's extremely dramatic, and really quite hard to understand. 179 00:09:21,617 --> 00:09:23,327 They can see just fine. 180 00:09:23,327 --> 00:09:25,037 It's not like they've gone blind. 181 00:09:25,477 --> 00:09:28,597 They can probably even say that it's a face. 182 00:09:28,777 --> 00:09:30,967 They just don't know whose face it is. 183 00:09:30,967 --> 00:09:36,337 So this is the disorder that we refer to as prosopagnosia from the Greek meaning. 184 00:09:36,377 --> 00:09:38,897 Without knowledge of the face. 185 00:09:39,437 --> 00:09:43,037 And what's been, what was interesting for me, what sort of caught me 186 00:09:43,037 --> 00:09:48,017 initially was that I didn't need to run sophisticated experiments. 187 00:09:48,017 --> 00:09:51,167 I didn't need to like brush up on my statistical knowledge. 188 00:09:51,497 --> 00:09:57,677 I just had to show these patients some pictures of their family members even. 189 00:09:58,402 --> 00:10:04,282 they were unable to recognize them or even photographs of themselves that they failed 190 00:10:04,282 --> 00:10:07,912 to recognize unless they remembered the context in which the photograph was taken. 191 00:10:08,887 --> 00:10:12,967 And they told really amazing stories like this. 192 00:10:13,067 --> 00:10:18,867 One very astute young man told me that he would go to the grocery store with his 193 00:10:18,867 --> 00:10:23,277 mother and she would go down one aisle and he would go down the other aisle. 194 00:10:23,427 --> 00:10:27,837 And if they met up at the end of the isles, he wouldn't recognize 195 00:10:27,837 --> 00:10:29,637 her unless he remembered. 196 00:10:29,907 --> 00:10:33,747 That, for example, she was wearing a red sweater that day. 197 00:10:34,167 --> 00:10:34,527 So it's 198 00:10:34,571 --> 00:10:35,411 That's fascinating. 199 00:10:36,207 --> 00:10:43,587 impairment and even reflecting on it, there's something very surreal 200 00:10:43,587 --> 00:10:50,112 or disorienting in not being able to remember even your own family members. 201 00:10:50,412 --> 00:10:53,051 Recognizing faces, it just sounds like the most 202 00:10:54,966 --> 00:10:59,346 basic, even though it's very complex from a neurocognitive and 203 00:10:59,346 --> 00:11:04,366 processing standpoint, but it sounds like how we navigate the world. 204 00:11:05,156 --> 00:11:10,406 And losing that ability can have enormous consequences and quality 205 00:11:10,406 --> 00:11:12,386 of life and how we function. 206 00:11:13,546 --> 00:11:16,351 do these people then clinging onto other visual cues. 207 00:11:16,487 --> 00:11:17,027 cues? 208 00:11:17,747 --> 00:11:18,377 Yes. 209 00:11:18,557 --> 00:11:22,817 No question about it, but also other cues like auditory cues. 210 00:11:22,997 --> 00:11:28,367 So the same young man would say as soon as his mother started speaking, it was 211 00:11:28,367 --> 00:11:30,587 so obvious to him that was his mother. 212 00:11:31,472 --> 00:11:37,662 Or other cues like hairstyle, one of the patients told me that but she and 213 00:11:37,662 --> 00:11:41,802 her husband were professors and they went to some convocation ceremony 214 00:11:41,832 --> 00:11:45,282 and they said they were going to go and put on their professorial robes 215 00:11:45,282 --> 00:11:46,962 and they'll meet back at this pillar. 216 00:11:47,412 --> 00:11:51,612 So she comes in her robes and she's standing there waiting and he is there 217 00:11:51,612 --> 00:11:55,752 as well, but she can't recognize him because he has changed his clothing. 218 00:11:56,891 --> 00:11:57,671 My goodness. 219 00:11:58,511 --> 00:11:59,591 It is dramatic. 220 00:11:59,651 --> 00:12:01,331 And how common is this? 221 00:12:01,331 --> 00:12:03,221 What is the incidence and prevalence of this? 222 00:12:03,257 --> 00:12:09,312 So this form of prosopagnosia that we've been discussing this acquired form so that 223 00:12:09,312 --> 00:12:14,082 it was normal and then something happened and they lost the ability they acquired. 224 00:12:14,082 --> 00:12:16,932 The prosopagnosia is extremely rare. 225 00:12:18,177 --> 00:12:22,602 There are, however, there is, however, a second group of Prosopagnosia 226 00:12:22,632 --> 00:12:28,962 patients whom we've come to recognize in about the last decade who seem 227 00:12:28,962 --> 00:12:31,722 to be unable to recognize faces. 228 00:12:32,202 --> 00:12:33,762 But that's always been the case. 229 00:12:33,822 --> 00:12:37,002 Like they just never learned to recognize faces. 230 00:12:37,362 --> 00:12:42,326 So I refer to this as congenital 'cause they always say it's like lifelong. 231 00:12:42,626 --> 00:12:42,836 Yep. 232 00:12:42,912 --> 00:12:51,117 people call it developmental, In my mind, they failed to master the skill 233 00:12:51,117 --> 00:12:56,397 of decoding the complex arrangement of the features of the face. 234 00:12:57,237 --> 00:13:01,297 It may be a little bit like, some kids struggle to learn to read, 235 00:13:01,537 --> 00:13:03,277 have developmental dyslexia. 236 00:13:03,997 --> 00:13:07,702 This may be the right hemisphere equivalent where they 237 00:13:07,897 --> 00:13:10,802 fail to acquire the mastery. 238 00:13:11,162 --> 00:13:14,792 Of recognizing faces, and that is not so uncommon. 239 00:13:15,122 --> 00:13:20,252 Several studies have suggested that this occurs in about two out of a hundred 240 00:13:20,252 --> 00:13:22,562 individuals, 2% of the population. 241 00:13:23,786 --> 00:13:27,146 But do you, are you suggesting that this is some sort of a skill 242 00:13:27,146 --> 00:13:33,081 acquisition failure versus anatomic or functional organic issue with the brain? 243 00:13:33,122 --> 00:13:33,482 the brain? 244 00:13:34,021 --> 00:13:34,241 Yes. 245 00:13:34,541 --> 00:13:35,592 It's a great question. 246 00:13:36,212 --> 00:13:39,512 There's a lot of debate on this particular topic. 247 00:13:40,502 --> 00:13:42,662 Will tell you because I can. 248 00:13:42,957 --> 00:13:48,177 Quote, my own work here is that I am definitely of the view that there's some 249 00:13:48,177 --> 00:13:53,997 structural anomaly that is preventing the recognition of faces, and in fact, 250 00:13:53,997 --> 00:13:56,607 in a study that we did a few years ago. 251 00:13:57,047 --> 00:14:03,167 We mapped out the circuit that the information has to flow in the 252 00:14:03,167 --> 00:14:05,627 brain in order to recognize faces. 253 00:14:05,627 --> 00:14:11,497 There's a whole little set of regions that all get activated together 254 00:14:11,497 --> 00:14:16,147 to ensure face recognition, and we showed that in these individuals. 255 00:14:16,212 --> 00:14:19,062 Who failed to acquire face recognition. 256 00:14:19,452 --> 00:14:21,492 There was like a break in the circuit. 257 00:14:21,522 --> 00:14:25,572 The white matter tracks between the two main nodes of 258 00:14:25,572 --> 00:14:28,792 the circuit were compromised. 259 00:14:29,686 --> 00:14:31,936 And I'm assuming this is a functional MRI study. 260 00:14:32,292 --> 00:14:34,032 It was a functional MRI study, and 261 00:14:34,139 --> 00:14:37,499 Was like a structural diffusion tensor imaging study. 262 00:14:38,163 --> 00:14:38,493 Let's see. 263 00:14:38,639 --> 00:14:42,479 add that these individuals the more congenital version 264 00:14:42,894 --> 00:14:44,594 it appears to run in families. 265 00:14:46,014 --> 00:14:51,174 Which again suggests that there is some genetic specification for the wiring of 266 00:14:51,174 --> 00:14:56,634 the circuit and transmitted through the genome is some architectural failure 267 00:15:01,944 --> 00:15:07,204 do we possess the ability to modify the mapping we have of that face? 268 00:15:08,550 --> 00:15:08,850 Yeah. 269 00:15:08,850 --> 00:15:09,560 This it's. 270 00:15:09,990 --> 00:15:15,210 It's such an important question because faces never remain the same 271 00:15:15,390 --> 00:15:19,800 from even, my face today and my face tomorrow will look slightly different 272 00:15:19,800 --> 00:15:21,880 for a whole host of possible reasons. 273 00:15:21,880 --> 00:15:26,080 Like I didn't sleep that well, and now my eyes are weary. 274 00:15:26,500 --> 00:15:30,460 What's really interesting here is that our ability to recognize 275 00:15:30,460 --> 00:15:33,700 faces is so robust, and this 276 00:15:33,802 --> 00:15:40,712 Of the challenge in understanding this, is that we can, we inva to these changes. 277 00:15:41,012 --> 00:15:46,772 So even if I haven't seen somebody for 15 years and they've aged 15 years. 278 00:15:47,672 --> 00:15:49,742 recognize them really well. 279 00:15:50,142 --> 00:15:52,302 Somebody's hair might have gone white. 280 00:15:52,537 --> 00:15:55,657 You still recognize them, although hair's sort of something 281 00:15:55,657 --> 00:15:57,012 of a different category. 282 00:15:57,282 --> 00:16:00,572 But, maybe somebody has a beard and then, they shave the beard. 283 00:16:00,602 --> 00:16:02,282 You can still recognize them. 284 00:16:03,632 --> 00:16:07,502 we can, of course, I can recognize you if you turn your head so I don't 285 00:16:07,502 --> 00:16:09,152 have a full frontal image of you. 286 00:16:09,487 --> 00:16:10,567 No problem. 287 00:16:10,597 --> 00:16:12,967 I know exactly who I'm looking at. 288 00:16:13,777 --> 00:16:19,007 there's something phenomenally resilient about this capacity. 289 00:16:19,357 --> 00:16:22,687 And that makes it all the more dramatic. 290 00:16:22,987 --> 00:16:25,087 When somebody can't recognize a face, 291 00:16:26,462 --> 00:16:30,902 One of the most common concerns patients have when they come to see 292 00:16:30,902 --> 00:16:36,272 me is that they wish after their aesthetic or even reconstructive surgery 293 00:16:36,392 --> 00:16:39,362 to remain looking like themselves. 294 00:16:39,662 --> 00:16:40,022 Yeah, 295 00:16:40,322 --> 00:16:45,272 Of course, it's a bit of a dichotomy because we're going to make some changes, 296 00:16:45,632 --> 00:16:48,302 yet maintaining that sense of self. 297 00:16:48,709 --> 00:16:50,672 And I think that kind of plays into this. 298 00:16:50,739 --> 00:16:57,909 And in certain patients we trigger a very prolonged time to reconcile that change 299 00:16:58,405 --> 00:16:58,705 Yeah. 300 00:16:59,139 --> 00:17:01,059 bring it back into that sense of self. 301 00:17:01,059 --> 00:17:03,879 And sometimes I tell my patients, you get a haircut and you, I know you 302 00:17:03,879 --> 00:17:06,489 mentioned the hair is a little bit different and I want to explore that, 303 00:17:06,754 --> 00:17:10,254 but you get a haircut, it takes you a week to get used to it. 304 00:17:10,270 --> 00:17:10,360 it. 305 00:17:10,477 --> 00:17:11,827 And so it's quite fascinating. 306 00:17:12,202 --> 00:17:12,472 Yeah. 307 00:17:12,522 --> 00:17:17,022 I think we do, we have a very strong sense of our own face. 308 00:17:17,887 --> 00:17:18,877 We've learned it over time. 309 00:17:18,927 --> 00:17:24,927 We've lived with this face over time and changes in, in 310 00:17:24,987 --> 00:17:27,147 really any aspect of the face. 311 00:17:27,177 --> 00:17:29,577 You have your teeth redone or whatever. 312 00:17:29,577 --> 00:17:31,827 It is really shift. 313 00:17:32,007 --> 00:17:33,537 I am, I can even imagine. 314 00:17:34,377 --> 00:17:40,407 Sort of patients having a crisis because there is this disjunct between their deep 315 00:17:40,407 --> 00:17:46,257 conceptual sense of what their face is like and also how society perceives them 316 00:17:46,977 --> 00:17:47,217 Yeah. 317 00:17:47,337 --> 00:17:51,327 that they changed face after various manipulations. 318 00:17:52,647 --> 00:17:58,017 And the time it takes for a patient to feel normal about what they 319 00:17:58,017 --> 00:18:02,997 had done on average is two months, but it's longer in older patients 320 00:18:02,997 --> 00:18:05,307 and shorter in younger patients. 321 00:18:05,607 --> 00:18:05,667 Yeah. 322 00:18:06,057 --> 00:18:10,810 I think it all is our clinical correlate of this processing system. 323 00:18:10,827 --> 00:18:13,167 This is exactly what we're talking about. 324 00:18:13,872 --> 00:18:17,562 When I tell a patient, look, it's the first six to eight weeks, you're 325 00:18:17,562 --> 00:18:22,242 not gonna necessarily recognize yourself or the way you see yourself. 326 00:18:22,572 --> 00:18:26,882 It's complex facial procedures like rhinoplasty I think they show that the 327 00:18:26,882 --> 00:18:32,252 brain's visual and self-referential system can gradually remap a kind of 328 00:18:32,252 --> 00:18:34,502 a new facial configuration itself. 329 00:18:35,007 --> 00:18:35,277 Yeah. 330 00:18:35,547 --> 00:18:38,607 So I'm pretty sure that's what happens because it's 331 00:18:38,607 --> 00:18:40,977 really one face, or two faces. 332 00:18:40,977 --> 00:18:43,617 You had one face and you've now got a slightly different face. 333 00:18:43,827 --> 00:18:48,597 So it's really just that particular face that now needs to be updated 334 00:18:48,627 --> 00:18:51,027 in your self-referential system. 335 00:18:51,387 --> 00:18:52,317 It also. 336 00:18:52,707 --> 00:18:58,347 Reminds me that I wanted to tell you how difficult it is to try and 337 00:18:58,347 --> 00:19:03,147 train individuals who are proso agnostic to learn to recognize faces. 338 00:19:03,897 --> 00:19:11,187 So most of the sort of attempts to remedy or rehabilitate the prosopagnosia has 339 00:19:11,187 --> 00:19:19,197 been done in adults, and I now, I think, have an understanding that it's as hard 340 00:19:19,587 --> 00:19:25,707 to learn to recognize faces in adulthood as it is to acquire a second language. 341 00:19:26,517 --> 00:19:28,047 It's really difficult. 342 00:19:28,287 --> 00:19:32,787 It's a little bit like not being able to teach an old dog new tricks. 343 00:19:33,222 --> 00:19:33,612 Yeah, 344 00:19:33,807 --> 00:19:35,487 the dog, the harder it 345 00:19:35,652 --> 00:19:36,192 they're harder. 346 00:19:36,312 --> 00:19:36,482 Yeah. 347 00:19:36,782 --> 00:19:36,792 Yeah. 348 00:19:36,927 --> 00:19:40,377 And so this is of course very frustrating for people who are 349 00:19:40,377 --> 00:19:41,817 impaired at face recognition. 350 00:19:41,817 --> 00:19:45,567 It's also very difficult for me because I wanna be able to 351 00:19:45,567 --> 00:19:48,447 intervene and really train this, 352 00:19:48,512 --> 00:19:48,702 Yeah. 353 00:19:49,467 --> 00:19:50,967 to be able to do this. 354 00:19:51,244 --> 00:19:56,194 . Individuals with congenital prosopagnosia, that's one category and. 355 00:19:56,240 --> 00:19:56,690 and. 356 00:19:57,274 --> 00:20:02,674 It makes me actually think of congenital deafness versus those who lose their 357 00:20:02,674 --> 00:20:08,594 hearing later on in life or later on, or proso, noia that happens later in life. 358 00:20:08,700 --> 00:20:09,240 That's right. 359 00:20:09,464 --> 00:20:15,134 What is the psychology or psychological impact to that sense of self, to 360 00:20:15,134 --> 00:20:17,264 that sense of navigating the world 361 00:20:17,754 --> 00:20:18,024 yeah. 362 00:20:18,183 --> 00:20:18,873 I have. 363 00:20:18,993 --> 00:20:25,136 I cannot think of a single study that has probed kind of phenomenological very well. 364 00:20:25,976 --> 00:20:27,356 something very obvious. 365 00:20:27,356 --> 00:20:28,061 I think we should. 366 00:20:28,781 --> 00:20:34,301 Really understand what their subjective experience is really like. 367 00:20:35,081 --> 00:20:40,151 Um, there are, however, a number of elaborated case studies in the 368 00:20:40,151 --> 00:20:46,561 literature, including the one where as Oliver Sachs the neurologist wrote The 369 00:20:46,561 --> 00:20:52,771 man who mistook his wife for a hat, he thought his wife's face was a hat and 370 00:20:52,771 --> 00:20:54,601 he tried to put it on his head, of. 371 00:20:54,601 --> 00:20:57,781 It turns out that's not actually how this disorder works at all. 372 00:20:57,781 --> 00:21:01,111 The patients almost always say it's a face. 373 00:21:01,111 --> 00:21:02,461 They don't mistake it for a head. 374 00:21:02,461 --> 00:21:05,581 They just say, I just don't know whose face it is. 375 00:21:05,611 --> 00:21:09,931 And they sort of, that's the struggle for them is they might even sack and 376 00:21:09,931 --> 00:21:12,061 see two eyes and nose and a mouth. 377 00:21:12,361 --> 00:21:17,386 I just still can't get Um, so one of the individuals in the literature, 378 00:21:17,386 --> 00:21:24,766 in this elaborate, um, case study, lament, his ability to be the same 379 00:21:24,766 --> 00:21:30,776 person that he previously and of, who he was, um, that it had limited 380 00:21:30,776 --> 00:21:33,206 him as an individual moving forward. 381 00:21:34,097 --> 00:21:34,787 I can imagine. 382 00:21:34,787 --> 00:21:39,537 I mean, this is fundamental to us as humans to be able to recognize 383 00:21:39,949 --> 00:21:43,583 at a minimum, but each other and everybody else in our society. 384 00:21:43,883 --> 00:21:44,063 Yeah. 385 00:21:44,859 --> 00:21:47,549 we kind of touched the incidents. 386 00:21:47,579 --> 00:21:49,004 Do we have, uh, specific numbers? 387 00:21:50,909 --> 00:21:56,338 several studies, maybe three, let's say were done in, one was 388 00:21:56,338 --> 00:22:01,318 done in Germany, for example, the incoming class of medical students. 389 00:22:01,528 --> 00:22:06,358 Or completed a questionnaire about whether they can recognize faces or have 390 00:22:06,568 --> 00:22:08,848 have ever experienced some difficulty. 391 00:22:08,968 --> 00:22:15,018 Another one I think was done in India and another one, um, was done to. 392 00:22:15,948 --> 00:22:17,718 I think they're Australian. 393 00:22:17,718 --> 00:22:22,578 The, the authors, um, and all of them kind of on this magic number 394 00:22:22,698 --> 00:22:26,008 of about two or in the at large. 395 00:22:26,248 --> 00:22:26,848 That's large 396 00:22:27,154 --> 00:22:27,394 rare. 397 00:22:27,484 --> 00:22:27,634 Yeah. 398 00:22:27,934 --> 00:22:29,158 So, you know, when 399 00:22:29,158 --> 00:22:30,658 I teach, uh, an 400 00:22:30,658 --> 00:22:31,498 auditorium of 401 00:22:31,498 --> 00:22:36,788 students, chances are there is somebody in that auditorium unable to recognize 402 00:22:38,339 --> 00:22:42,620 Oh I imagine that it'd be an extremely rare condition, 2% is not 403 00:22:43,136 --> 00:22:43,449 Right. 404 00:22:43,479 --> 00:22:47,139 But these are all of the congenital or developmental form. 405 00:22:47,139 --> 00:22:47,229 Mm-hmm. 406 00:22:47,469 --> 00:22:48,579 The acquired form. 407 00:22:49,599 --> 00:22:53,989 they must be, they're not even a hundred patients reported in the literature. 408 00:22:54,930 --> 00:23:00,060 And to your knowledge, are there corollaries of other pathological 409 00:23:00,060 --> 00:23:05,545 states, whether it be medical or psychiatric that, or even 410 00:23:05,845 --> 00:23:06,594 traumatic, 411 00:23:07,095 --> 00:23:10,585 that perhaps have triggered or exacerbated the condition? 412 00:23:12,834 --> 00:23:16,237 Um, I mean, definitely patients it's, even if they've had it their 413 00:23:16,237 --> 00:23:22,497 whole lives, um, it, it makes not all, but some patients report that 414 00:23:22,497 --> 00:23:26,262 it has limited their Um, be social. 415 00:23:26,352 --> 00:23:31,722 One individual told us that she was agoraphobic, like she had sort of fear 416 00:23:31,722 --> 00:23:37,282 of going because she was so anxious that she would see somebody that she 417 00:23:37,282 --> 00:23:39,322 didn't recognize she'd offend them. 418 00:23:40,717 --> 00:23:40,837 I 419 00:23:40,977 --> 00:23:41,007 Okay. 420 00:23:41,307 --> 00:23:42,697 know one of uh, one 421 00:23:42,697 --> 00:23:44,557 individual with the 422 00:23:44,557 --> 00:23:45,127 same 423 00:23:45,127 --> 00:23:49,850 congenital version who, who is interestingly a vision scientist 424 00:23:49,850 --> 00:23:54,030 himself, he can't recognize people and he's got this technique. 425 00:23:54,030 --> 00:23:59,217 goes to conferences and the first thing he does when he sees or talks to somebody, 426 00:23:59,217 --> 00:24:00,867 if she said, I just need to tell you. 427 00:24:01,572 --> 00:24:05,592 I am not going to be able to recognize your face, so I'm not 428 00:24:05,592 --> 00:24:09,702 being um, a social or a snob. 429 00:24:10,032 --> 00:24:14,452 You just need to know this is a, I have this failure, and just 430 00:24:14,452 --> 00:24:18,632 sort of is completely transparent about it, um, right up front. 431 00:24:20,082 --> 00:24:22,152 Well, I bet after saying that nobody forgets 432 00:24:22,462 --> 00:24:23,152 Yeah, for sure. 433 00:24:24,645 --> 00:24:28,912 but more into the reverse of I guess I want to explore whether or not you have 434 00:24:28,912 --> 00:24:34,965 noticed associated with this condition, meaning, I don't know uh, psychiatric, 435 00:24:34,995 --> 00:24:41,665 uh, are these patients being recognized or, um, um, diagnosed as issues? 436 00:24:42,405 --> 00:24:44,325 treat it for psychiatric problems. 437 00:24:45,414 --> 00:24:45,474 Um. 438 00:24:46,764 --> 00:24:50,124 I think probably not than in the population at large, maybe 439 00:24:50,124 --> 00:24:53,334 a little bit more because there is associated with it for 440 00:24:53,334 --> 00:24:53,604 some 441 00:24:53,715 --> 00:24:53,955 Okay? 442 00:24:55,185 --> 00:24:55,395 Of course. 443 00:24:56,094 --> 00:24:57,864 in the congenital 444 00:24:57,894 --> 00:24:59,249 individuals where there 445 00:24:59,249 --> 00:25:02,399 seems to be this failure in wiring up the system. 446 00:25:02,759 --> 00:25:06,089 There are some associated deficits, like some of the. 447 00:25:06,184 --> 00:25:12,274 Patients or individuals have difficulty do um, navigating, like finding 448 00:25:12,274 --> 00:25:14,554 their way around town, for example. 449 00:25:14,944 --> 00:25:18,067 Um, so there are some kind of comorbidities. 450 00:25:18,264 --> 00:25:21,114 Some of them definitely have difficulties reading. 451 00:25:21,214 --> 00:25:25,644 Reading is kind of some some of the same computational requirements as faces. 452 00:25:25,644 --> 00:25:28,404 There are a small number of items like the alphabet. 453 00:25:28,644 --> 00:25:31,944 You have to together to make a bigger thing, just like we have eyes and 454 00:25:31,944 --> 00:25:35,964 nose and a mouth that you have to put together to make it into a whole face. 455 00:25:36,864 --> 00:25:39,654 there are developmental comorbidities. 456 00:25:39,859 --> 00:25:45,024 Um, and these have been well reported, but much less psychiatric than 457 00:25:45,474 --> 00:25:47,304 I think you might have imagined. 458 00:25:47,825 --> 00:25:48,175 Yeah. 459 00:25:48,250 --> 00:25:48,540 Yeah. 460 00:25:48,840 --> 00:25:51,979 interesting psychiatric d you know, disorders that 461 00:25:52,009 --> 00:26:00,389 has a face component like syndrome Reduplicate These are very esoteric. 462 00:26:00,419 --> 00:26:05,519 Um, and you know, there's like a single case in in the literature where, um. 463 00:26:06,869 --> 00:26:14,309 Somebody, thinks that the, the, the, well-known face is actually calls them by 464 00:26:14,309 --> 00:26:20,549 a different name, all different kinds of historical and past experiences to them. 465 00:26:21,154 --> 00:26:26,969 sort of like the loss between the face and the memory of who that person is. 466 00:26:27,289 --> 00:26:31,774 So there are some of these, um, more, more psychiatric disorders. 467 00:26:32,150 --> 00:26:35,630 There is sometimes a phenomenon in the human brain, 468 00:26:36,620 --> 00:26:43,220 Where an input is perceived as threatening, useless, or even painful, 469 00:26:43,760 --> 00:26:46,400 and we decide to ignore that input. 470 00:26:46,829 --> 00:26:47,069 Yeah. 471 00:26:48,100 --> 00:26:52,750 Is there a possibility that these patients acquired the skill of facial 472 00:26:52,750 --> 00:26:59,860 recognition, but the input was distorted or somehow not serving them, and they 473 00:26:59,860 --> 00:27:02,260 decided I, I'm gonna put the skill aside, 474 00:27:03,814 --> 00:27:04,234 Um, 475 00:27:04,480 --> 00:27:09,030 meaning there is a cognitive, repressive, uh, not necessarily a 476 00:27:09,144 --> 00:27:09,264 a 477 00:27:09,564 --> 00:27:10,530 neurological. 478 00:27:10,973 --> 00:27:16,943 I see it in the sense that patients come in for a say, and I ask them to describe 479 00:27:16,943 --> 00:27:18,503 their current nose, and they cannot. 480 00:27:20,353 --> 00:27:20,713 Interesting. 481 00:27:20,843 --> 00:27:22,373 em too much pain, too much. 482 00:27:23,108 --> 00:27:27,101 Something they don't like and they cannot tell me if the cannot 483 00:27:27,101 --> 00:27:28,271 tell me how big the hump is, 484 00:27:29,021 --> 00:27:30,341 and it's as if it's ignored. 485 00:27:30,341 --> 00:27:32,281 It's as if they look at themselves and they don't see it. 486 00:27:32,581 --> 00:27:32,911 Yeah. 487 00:27:33,061 --> 00:27:33,871 It's interesting. 488 00:27:36,451 --> 00:27:39,211 yeah, it's a kind of like selective blindness. 489 00:27:40,421 --> 00:27:44,041 These are very, very I mean to, first of all, you're dealing with a. 490 00:27:44,491 --> 00:27:49,161 Quite a. rare condition and very difficult to set, set up the 491 00:27:49,311 --> 00:27:52,646 that give you the information to really look at every detail. 492 00:27:52,946 --> 00:27:53,336 Right. 493 00:27:54,146 --> 00:27:54,486 it does mean 494 00:27:54,626 --> 00:27:58,016 somebody like me is not going to be out of work anytime soon. 495 00:27:58,406 --> 00:28:02,156 And there are going to be people coming who also will be able to 496 00:28:02,156 --> 00:28:03,836 tackle some of these questions. 497 00:28:04,641 --> 00:28:05,721 And that's how science works. 498 00:28:05,721 --> 00:28:08,601 I mean, we are all standing on the shoulders of giants. 499 00:28:08,601 --> 00:28:12,351 We add another link to the chain and hopefully it's a meaningful 500 00:28:12,351 --> 00:28:14,391 one and, and, and we move forward. 501 00:28:14,871 --> 00:28:19,671 Marlene, when you're looking at these patients, you're looking at the studies. 502 00:28:20,211 --> 00:28:22,346 you're in, you have your scientist hat on, 503 00:28:22,646 --> 00:28:22,736 Mm-hmm. 504 00:28:23,091 --> 00:28:30,051 but sometimes you perhaps take it off and you say, wow, what, what in the hell 505 00:28:30,448 --> 00:28:30,838 Yeah, 506 00:28:31,138 --> 00:28:32,393 has that happened to you? 507 00:28:32,393 --> 00:28:34,523 And what was it about Gimme, Gimme, an 508 00:28:36,358 --> 00:28:38,278 Okay, an example. 509 00:28:38,968 --> 00:28:40,003 So, um. 510 00:28:40,925 --> 00:28:46,538 I have an individual with acquired was about just be in the, summer 511 00:28:46,538 --> 00:28:48,818 before he was due to start college. 512 00:28:49,428 --> 00:28:49,668 Yeah. 513 00:28:49,688 --> 00:28:56,308 Um, he had trained as a paramedic and worked in an ambulance, and the 514 00:28:56,338 --> 00:29:01,838 ambulance screeching out one night and they had an accident and he suffered. 515 00:29:02,128 --> 00:29:06,418 A traumatic brain injury and became prosopagnosia. 516 00:29:07,108 --> 00:29:09,508 He had some other, he broke some bones. 517 00:29:09,508 --> 00:29:10,948 Those all seem to heal. 518 00:29:11,398 --> 00:29:14,308 He has this, he's now close to 50. 519 00:29:14,698 --> 00:29:16,798 He still is prosopagnosia. 520 00:29:17,818 --> 00:29:23,848 Now, one of the odd things about this um, is that his family owns 521 00:29:23,848 --> 00:29:29,328 a photographic store and he works in the photographic store job. 522 00:29:30,628 --> 00:29:34,918 Is to check the photographs once they come off the printer. 523 00:29:35,668 --> 00:29:40,948 And so I said, I asked him, how is it possible that you can be doing this job? 524 00:29:41,758 --> 00:29:45,688 And his answer was very uh, instructive. 525 00:29:45,778 --> 00:29:51,088 He said it's easier for him to do it because he doesn't recognize the face. 526 00:29:51,178 --> 00:29:56,228 He's only looking blemishes or imperfections in the picture. 527 00:29:56,228 --> 00:30:00,278 And he's really good at picking out these small little things because 528 00:30:00,278 --> 00:30:02,138 he can basically ignore the face. 529 00:30:02,498 --> 00:30:02,858 So 530 00:30:02,918 --> 00:30:03,548 that is like. 531 00:30:05,498 --> 00:30:05,688 That's 532 00:30:05,988 --> 00:30:06,138 That's 533 00:30:06,138 --> 00:30:06,768 fascinating. 534 00:30:08,148 --> 00:30:11,126 That's amazing to take advantage of the deficit and put it 535 00:30:11,126 --> 00:30:11,363 Yeah. 536 00:30:13,088 --> 00:30:14,228 Marlene, what haven't I asked you? 537 00:30:14,308 --> 00:30:14,788 asked you? 538 00:30:17,098 --> 00:30:19,078 Has to do with artificial intelligence. 539 00:30:19,543 --> 00:30:19,833 Yeah. 540 00:30:20,368 --> 00:30:20,998 So. 541 00:30:21,913 --> 00:30:26,173 These AI models are supposed to be really good at face recognition. 542 00:30:26,263 --> 00:30:30,898 I mean, you can even use your face a, password, uh, at the bank 543 00:30:31,558 --> 00:30:33,238 you know, at the ATM, et cetera. 544 00:30:33,578 --> 00:30:33,998 Mm-hmm. 545 00:30:34,298 --> 00:30:35,248 But in fact. 546 00:30:35,325 --> 00:30:36,823 AI models are 547 00:30:36,823 --> 00:30:42,265 nowhere as good as humans are faces, for example, they fail. 548 00:30:42,355 --> 00:30:48,175 If you have changed your face in any very obvious way, they would not do so 549 00:30:48,175 --> 00:30:52,705 well with your patients after rhinoplasty or some other kind of reconstruction. 550 00:30:53,065 --> 00:30:55,150 And so there's 551 00:30:55,150 --> 00:30:56,200 There's something really 552 00:30:56,200 --> 00:30:58,060 sophisticated and. 553 00:30:59,065 --> 00:31:04,675 Almost natural about the way the human brain can compute this kind of 554 00:31:04,675 --> 00:31:09,445 information and it's not easily captured in an artificial neural network. 555 00:31:10,866 --> 00:31:16,626 One of the common almost jokes that patients tell me after surgery is that 556 00:31:16,626 --> 00:31:20,280 their phone doesn't unlock anymore and they have to redo their face, uh, 557 00:31:20,406 --> 00:31:20,856 or face 558 00:31:20,856 --> 00:31:22,446 id, and that's exactly that. 559 00:31:22,446 --> 00:31:23,976 Well, we know the brain is plastic. 560 00:31:23,976 --> 00:31:29,548 We, we know that there is, um, hardwired system can keep up with a system 561 00:31:29,558 --> 00:31:31,303 that can self-regulate and and remap. 562 00:31:31,603 --> 00:31:31,813 and 563 00:31:32,348 --> 00:31:34,207 And go that the human brain can, 564 00:31:34,258 --> 00:31:35,368 uh, and so that's good. 565 00:31:35,398 --> 00:31:35,788 That's good. 566 00:31:35,793 --> 00:31:37,258 We, we will all have jobs still. 567 00:31:37,563 --> 00:31:38,373 We'll have jobs. 568 00:31:38,954 --> 00:31:41,944 If you found this conversation meaningful, and if discussions 569 00:31:41,944 --> 00:31:45,290 like this matter to you, please share the episode and subscribe 570 00:31:45,434 --> 00:31:46,364 and we'll see you next time. 571 00:31:47,374 --> 00:31:47,974 very much.