WEBVTT - Ep63 "Why do brains love faces?" 0:00:04.480 --> 0:00:10.039 Brains love to look at faces, But why and what 0:00:10.200 --> 0:00:15.400 is face blindness? And what is a super recognizer? Why 0:00:15.400 --> 0:00:19.079 does your electrical plug sometimes look to you like a 0:00:19.120 --> 0:00:24.160 little face? Did aliens plant a signal for us on Mars? 0:00:24.600 --> 0:00:28.200 Or are we looking at a quirk of our own brains? 0:00:28.520 --> 0:00:30.200 What does any of this have to do with the 0:00:30.640 --> 0:00:35.360 neurologist Oliver Sachs and his inability to recognize most of 0:00:35.400 --> 0:00:38.160 the people in his life, or looking at a magazine 0:00:38.280 --> 0:00:44.480 upside down? Or what the mistakes in computerized face recognition technology? 0:00:45.080 --> 0:00:51.199 Tell us Welcome to Inner Cosmos with me David Eagleman. 0:00:51.560 --> 0:00:55.000 I'm a neuroscientist and an author at Stanford, and in 0:00:55.040 --> 0:00:59.600 these episodes we dive deeply into our three pound universe 0:01:00.080 --> 0:01:08.880 to understand some of the most surprising aspects of our lives. 0:01:12.440 --> 0:01:17.640 Today's episode is about faces. Now, let's say I posed 0:01:18.080 --> 0:01:22.679 the question to you, how does the brain recognize faces? 0:01:23.280 --> 0:01:26.600 You may reasonably think, why is that even a question? 0:01:27.120 --> 0:01:32.440 Recognizing people's faces isn't really that hard, But think about 0:01:32.480 --> 0:01:37.200 how similar faces are to one another. They're insanely similar. 0:01:37.360 --> 0:01:42.160 You have two eyes, nose, mouth, chin, ears, everyone's face 0:01:42.200 --> 0:01:46.840 looks pretty much the same people's faces are pretty much 0:01:46.959 --> 0:01:51.240 very tiny variations on a theme. The only reason you 0:01:51.320 --> 0:01:55.200 are able to distinguish faces, literally thousands of faces from 0:01:55.240 --> 0:01:59.840 one another, is because your brain devotes an enormous amount 0:01:59.880 --> 0:02:03.280 of circuitry to it. It puts so much effort in 0:02:03.920 --> 0:02:07.360 that it feels effortless to you. And we have all 0:02:07.440 --> 0:02:14.120 this circuitry because face recognition is extremely salient. It's how 0:02:14.560 --> 0:02:19.359 we recognize this package of identity from this other. Oh 0:02:19.440 --> 0:02:23.880 that's my mother, that's my neighbor, that's my wife. There's 0:02:24.440 --> 0:02:27.440 Tom Cruise and his face is similar but different from 0:02:27.800 --> 0:02:32.840 Dwayne Johnson, and that's different from Joe Biden. It's massively 0:02:32.880 --> 0:02:37.400 important for us as a social species to keep track 0:02:37.880 --> 0:02:42.200 of identities. Your brain wants to know who it is 0:02:42.280 --> 0:02:45.679 and everything that entails, like is that a friend or 0:02:45.720 --> 0:02:50.240 an enemy? What are the risks and opportunities here? And 0:02:50.280 --> 0:02:53.040 the answer to that is massively different. If you're looking 0:02:53.440 --> 0:02:57.160 at the face of Ted Bundy versus your best pal, 0:02:57.480 --> 0:03:00.040 and keep in mind that to a space alien and 0:03:00.080 --> 0:03:04.680 these faces would be essentially indistinguishable. Now, when you're trying 0:03:04.680 --> 0:03:09.040 to distinguish faces, there are lots of other cues like 0:03:09.520 --> 0:03:13.800 hairline and hairstyle and height, and details of their skin 0:03:14.040 --> 0:03:18.519 from color to smoothness, and even cues like how they 0:03:18.639 --> 0:03:23.600 walk and their clothing choice. And context also plays a 0:03:23.680 --> 0:03:28.600 large role. You've probably experienced some time when you saw 0:03:28.680 --> 0:03:31.320 somebody you know well, like a neighbor or a colleague, 0:03:31.639 --> 0:03:35.640 but it's in an unfamiliar setting and you struggle to 0:03:35.720 --> 0:03:39.720 recognize them immediately. Why does that happen Because our brains 0:03:39.880 --> 0:03:44.760 use contextual cues when identifying people, So our brains rely 0:03:45.000 --> 0:03:49.960 not just on facial features, but also on surrounding context 0:03:49.960 --> 0:03:54.760 of all sorts to identify people. But despite all these 0:03:54.800 --> 0:03:58.920 other cues that we use, just note that facial recognition 0:03:59.200 --> 0:04:02.640 is possible even over zoom, where many of those other 0:04:02.800 --> 0:04:06.160 cues are gone. And even if you've got a bunch 0:04:06.240 --> 0:04:09.400 of movie stars to put on wigs and change their 0:04:09.440 --> 0:04:12.440 hairlines and so on, you'd still be pretty good at 0:04:12.440 --> 0:04:17.840 distinguishing them just based on the details of the face itself. 0:04:18.360 --> 0:04:20.800 And I think there are at least two ways to 0:04:20.880 --> 0:04:26.479 appreciate how hard a challenge face recognition is to the brain. 0:04:26.880 --> 0:04:30.640 The first is to understand how difficult it is to 0:04:30.680 --> 0:04:35.960 get computers to recognize and distinguish human faces. Now you 0:04:36.040 --> 0:04:40.600 might think, aren't we pretty good at this with modern technology. Yes, 0:04:41.000 --> 0:04:45.400 but this has been a sixty year slog by very 0:04:45.440 --> 0:04:48.400 smart people in the computer science world. Starting in the 0:04:48.480 --> 0:04:53.000 nineteen sixties, people realized that face recognition was a really 0:04:53.040 --> 0:04:56.839 hard problem, and they've been climbing that hill ever since. 0:04:57.279 --> 0:05:01.000 So the way that researchers have gotten face recognition to 0:05:01.120 --> 0:05:05.800 work on computers nowadays is by busting the problem up 0:05:06.040 --> 0:05:12.960 into different computational stages. So first, you have algorithms just 0:05:13.000 --> 0:05:16.479 for detecting a face, like is there a face out there? 0:05:17.200 --> 0:05:20.240 Then you have a second step where you segment this 0:05:20.400 --> 0:05:23.880 face from the background and figure out how the face 0:05:24.080 --> 0:05:26.480 is aligned to tell you about the size of the 0:05:26.520 --> 0:05:29.760 face and the illumination on the face and even the 0:05:29.839 --> 0:05:32.920 pose of the face. Then you have a third step 0:05:33.160 --> 0:05:37.159 where you extract the facial features, like here's the length 0:05:37.200 --> 0:05:39.960 of the nose and the distance between the eyes, and 0:05:40.000 --> 0:05:42.520 the shape of the mouth and the position of the ears. 0:05:42.880 --> 0:05:46.160 All these things you pinpoint and you measure them, and 0:05:46.200 --> 0:05:50.279 then in the fourth step, you match those numbers against 0:05:50.320 --> 0:05:53.719 a database of faces to see if you can identify 0:05:53.800 --> 0:05:57.320 this one. So this is a really complicated algorithm and 0:05:57.400 --> 0:06:01.160 it's instructive to see how computer get things wrong. I 0:06:01.200 --> 0:06:05.400 saw a great photograph where it's a family sitting on 0:06:05.440 --> 0:06:08.960 their porch at Halloween, and the computer has identified each 0:06:09.000 --> 0:06:12.160 of their faces and put a little label next to it, 0:06:12.279 --> 0:06:16.159 like Henry and John and Sarah, and then it identifies 0:06:16.480 --> 0:06:21.040 another face and says unknown person. But that face is 0:06:21.200 --> 0:06:23.880 the Jack O lantern, the pumpkin that's sitting there on 0:06:23.920 --> 0:06:27.400 the porch. Now, a human observer would never make that 0:06:27.560 --> 0:06:30.479 kind of mistake. But this goes to show that it's 0:06:30.560 --> 0:06:36.760 a very computationally challenging problem. So some version of this 0:06:36.880 --> 0:06:41.640 very complicated algorithm, probably not exactly the same, but equally complex. 0:06:42.120 --> 0:06:44.960 This is what your brain is doing under the hood 0:06:45.600 --> 0:06:49.280 when you just glanced at some billboard and you think, oh, 0:06:49.400 --> 0:06:52.440 that's Paul Giamati. I love that guy. And we'll come 0:06:52.520 --> 0:06:54.920 back to what the brain is doing in just a moment, 0:06:55.320 --> 0:06:57.280 but I just want to say for now that quite 0:06:57.279 --> 0:07:02.520 often we find that the tasks that seem most effortless 0:07:02.560 --> 0:07:08.039 to us are the things underpinned by the most brain circuitry. 0:07:08.440 --> 0:07:10.760 And I want to look at a second way to 0:07:10.880 --> 0:07:15.360 appreciate how hard a challenge face recognition is for the brain. 0:07:16.120 --> 0:07:20.920 Just look at a bunch of Golden Retriever faces. If 0:07:20.960 --> 0:07:23.480 you were in a situation where you were hanging out 0:07:23.520 --> 0:07:27.880 with hundreds or thousands of Golden Retriever dogs, do they 0:07:27.920 --> 0:07:33.080 really look that different to you? Or consider horse faces 0:07:33.400 --> 0:07:39.560 or squirrel faces or cowfaces. Obviously we can imagine some 0:07:39.760 --> 0:07:42.200 faces at the extreme that look a little bit different, 0:07:42.200 --> 0:07:44.960 But in general, if I dropped you into the middle 0:07:44.960 --> 0:07:48.480 of a giant ranch and you looked at thousands of cows, 0:07:48.680 --> 0:07:52.840 you'd say, okay, well, each face has two eyes and 0:07:52.920 --> 0:07:55.640 a snout and a mouth, and there's just not that 0:07:55.760 --> 0:07:59.880 much difference from one face to another. You wouldn't meet 0:08:00.400 --> 0:08:03.400 recognized if one of those cows had been in a 0:08:03.440 --> 0:08:06.760 movie like oh, that's that famous cow, or if one 0:08:06.800 --> 0:08:10.600 of those cows was wanted by the legal system, like 0:08:11.120 --> 0:08:14.200 wait a minute, I recognize that cow from the poster 0:08:14.320 --> 0:08:17.000 of the post office. And of course, if one of 0:08:17.040 --> 0:08:19.960 these cows was looking at you and a bunch of 0:08:19.960 --> 0:08:23.840 other humans, she would feel exactly the same way. There's 0:08:23.920 --> 0:08:28.480 not that much difference between human faces. Now side note, 0:08:28.520 --> 0:08:31.560 which we'll return to. If you were the rancher, you 0:08:31.640 --> 0:08:35.319 could get better at distinguishing cow faces. This is because 0:08:35.360 --> 0:08:40.319 of brain plasticity and because distinguishing the cows is salient 0:08:40.640 --> 0:08:44.200 to your rancher brain. But for now, what I want 0:08:44.200 --> 0:08:49.079 to surface is just how similar faces are before you 0:08:49.160 --> 0:08:54.280 get good at them from experience. Faces are unbelievably similar. 0:08:54.640 --> 0:08:58.400 And one expression of this difficulty in distinguishing faces comes 0:08:58.400 --> 0:09:02.720 from a psychological effe fact called the other race effect, 0:09:03.200 --> 0:09:06.160 and this refers to the tendency of people to have 0:09:06.679 --> 0:09:11.640 more difficulty recognizing faces that belong to races that are 0:09:11.640 --> 0:09:14.200 not their own race. This phenomenon is also known as 0:09:14.240 --> 0:09:18.440 the cross race effect or the own race bias, and 0:09:18.520 --> 0:09:21.400 it's been widely studied and you see it across all cultures. 0:09:21.800 --> 0:09:24.440 Now people sometimes hear this and they take this as 0:09:24.440 --> 0:09:28.679 evidence for racism. It's not racism in the sense of 0:09:28.720 --> 0:09:33.040 you treating your own group better and other groups more poorly. Instead, 0:09:33.120 --> 0:09:36.440 the other race effect simply results from the fact that 0:09:36.480 --> 0:09:39.880 your brain gets trained up on the faces around you, 0:09:40.360 --> 0:09:44.640 and you are better at recognizing those faces than other faces. 0:09:45.000 --> 0:09:48.800 So if you grew up in Cambodia, you will be 0:09:48.960 --> 0:09:53.160 excellent and distinguishing Cambodian faces, and maybe not so great 0:09:53.240 --> 0:09:57.080 at distinguishing Norwegian faces. If you grew up in Norway, 0:09:57.600 --> 0:10:01.480 you're excellent and distinguishing Norwegian f but not so great 0:10:01.559 --> 0:10:06.000 at distinguishing Zimbabwean faces. And if you grew up in Zimbabwe, 0:10:06.080 --> 0:10:09.360 you're not so good at Intuit Eskimo faces and so on. 0:10:09.760 --> 0:10:14.480 So your brain becomes attuned to the type of faces 0:10:14.520 --> 0:10:18.920 that you see most often, and this heightened sensitivity makes 0:10:19.000 --> 0:10:23.679 us better at distinguishing among those familiar faces. As a 0:10:23.720 --> 0:10:27.280 side note, can you mitigate the other race effect? Sure, 0:10:27.360 --> 0:10:31.920 it's just about exposure. You can improve the ability to 0:10:32.000 --> 0:10:37.480 recognize other race faces through training and exposure. Now, Hollywood 0:10:37.520 --> 0:10:39.640 has been on the forefront of this for a long time, 0:10:39.960 --> 0:10:42.960 making sure that there's a mixture that we all get 0:10:42.960 --> 0:10:46.800 exposed to different sorts of faces, and that's presumably very useful. 0:10:47.160 --> 0:10:50.040 It could be noted that Hollywood has tended to be 0:10:50.160 --> 0:10:52.960 mostly in love with white faces and black faces, and 0:10:53.000 --> 0:10:55.080 that leaves a lot of people out of the mix. 0:10:55.480 --> 0:10:58.840 So we the viewing audience don't necessarily get much better 0:10:58.880 --> 0:11:03.320 at Cambodian face or Norwegian faces, or Inuit Eskimo faces 0:11:03.400 --> 0:11:07.680 or whatever. But it's a start to training our brains 0:11:07.760 --> 0:11:11.320 on a broader diet of faces. So the other race 0:11:11.360 --> 0:11:14.920 effect underscores how our brains can get tuned to the 0:11:14.960 --> 0:11:19.040 social world around us, and it highlights our natural tendency 0:11:19.400 --> 0:11:35.280 to recognize familiar faces more easily. Okay, so where are 0:11:35.280 --> 0:11:39.320 we so far. It's really hard to distinguish faces unless 0:11:39.360 --> 0:11:42.360 your brain has had lots and lots of practice at it. 0:11:42.840 --> 0:11:46.920 But again it seems effortless to us now. Even though 0:11:47.120 --> 0:11:50.680 we're not always great at distinguishing faces from one another, 0:11:50.880 --> 0:11:54.800 unless we've had lots of practice, we are highly programmed 0:11:54.880 --> 0:11:58.640 to identify that there is a face there and to 0:11:58.800 --> 0:12:02.640 zoom in on it. Why because faces carry so much 0:12:02.640 --> 0:12:05.840 information for us, And this is why you sometimes look 0:12:05.840 --> 0:12:09.240 at the electrical plug in your wall and you see 0:12:09.240 --> 0:12:12.400 a little face, or you see a pattern of burn 0:12:12.480 --> 0:12:14.800 marks on a piece of toast and you think, oh, 0:12:14.840 --> 0:12:17.679 that looks just like a face, or you look at 0:12:17.760 --> 0:12:20.080 the surface of the moon and you think, oh, there's 0:12:20.200 --> 0:12:23.440 a man there looking at me. In science, this is 0:12:23.480 --> 0:12:30.360 called paradolia. Paradolia refers to our tendency to perceive specific 0:12:30.520 --> 0:12:35.240 and often meaningful images in random patterns, and the most 0:12:35.280 --> 0:12:37.840 common usage of this word is in the case of 0:12:37.960 --> 0:12:42.920 perceiving the pattern of a face in random unrelated objects. 0:12:42.920 --> 0:12:45.760 And we've all been there where your brain says, oh, 0:12:46.000 --> 0:12:47.880 that looks like a face when you're looking at a 0:12:48.280 --> 0:12:53.960 cloud or a rock formation, or some arrangement of fruits 0:12:53.960 --> 0:12:55.920 on a table. And I don't know if you've ever 0:12:55.920 --> 0:12:59.480 seen this image from the surface of Mars. I'll put 0:12:59.480 --> 0:13:03.520 it on the show. But on the bumpy landscape of Mars, 0:13:03.720 --> 0:13:06.120 there are some bumps that happen to fall into the 0:13:06.160 --> 0:13:09.120 pattern that look like two y's and a nose and 0:13:09.160 --> 0:13:12.040 a mouth. And many people go nuts about this image 0:13:12.080 --> 0:13:15.520 because they suggest that it's a sculpture of a human 0:13:15.679 --> 0:13:20.080 face planted there by aliens. But probably not. This is 0:13:20.200 --> 0:13:28.199 merely our brains experiencing garden variety paradolia. We assign patterns 0:13:28.280 --> 0:13:33.560 to random inputs, and in particular, we love to see faces. Okay, 0:13:33.800 --> 0:13:36.920 but why does this happen to us so commonly. Well, 0:13:37.400 --> 0:13:42.000 it's because we are hardwired to see faces everywhere, and 0:13:42.040 --> 0:13:46.920 the reason is that faces are crucial for social interaction. 0:13:47.600 --> 0:13:51.840 Seeing a face tells us about a person's identity, their 0:13:52.000 --> 0:13:55.920 emotional state, whether they're a friend or a foe. This 0:13:56.080 --> 0:14:00.680 ability helped our ancestors to survive by recognize is saying oh, 0:14:00.720 --> 0:14:04.439 that's a tribe member. Oh, that's a potential threat. In fact, 0:14:04.520 --> 0:14:08.360 even newborn babies prefer to look at faces. They look 0:14:08.360 --> 0:14:12.640 at face like stimuli more than other stimuli. So take 0:14:13.040 --> 0:14:16.800 two little circles and a little square beneath that, and 0:14:16.840 --> 0:14:20.560 a little rectangle underneath that, they'll stare at that as 0:14:20.600 --> 0:14:24.800 opposed to the same shapes in a different configuration. And 0:14:24.840 --> 0:14:30.360 this underscores that brains are primed to focus on faces. 0:14:30.800 --> 0:14:35.120 Paridolia is just an extension of this bias to recognize faces. 0:14:35.160 --> 0:14:39.960 It's our brain's way of staying vigilant because in the wild, 0:14:40.160 --> 0:14:44.120 it is safer to mistakenly see a face where there 0:14:44.200 --> 0:14:48.720 isn't one than to miss a real face. So this 0:14:49.080 --> 0:14:53.720 hypersensitivity to faces means the difference between life and death 0:14:53.760 --> 0:14:58.480 in some situations, like spotting a lurking predator or an 0:14:58.520 --> 0:15:02.920 aggressive intruder. By the way, I'll just mention that paradolia 0:15:03.040 --> 0:15:06.720 is something that artists have always exploited. They turn random 0:15:06.760 --> 0:15:12.320 patterns or abstract shapes into recognizable faces. Just as one example, 0:15:12.440 --> 0:15:15.880 one of my favorite artists, Salvador Dhli. He very often 0:15:15.920 --> 0:15:20.800 takes advantage of paradolia to create double images that play 0:15:20.880 --> 0:15:24.280 with the viewer's perception. You can see this as a 0:15:24.520 --> 0:15:30.200 rock formation or as a face. So paradolia highlights how 0:15:30.200 --> 0:15:32.520 our brains are constantly trying to make sense of the 0:15:32.560 --> 0:15:37.200 world around us, even when there isn't anything immediately recognizable. 0:15:37.720 --> 0:15:39.640 And I want to give just one other angle on 0:15:39.760 --> 0:15:43.600 paradolia here. If you're a regular listener to this podcast, 0:15:44.040 --> 0:15:46.480 you know that I often talk about how the brain's 0:15:46.720 --> 0:15:50.600 main job is to build an internal model of the 0:15:50.680 --> 0:15:56.720 outside world, and our model influences our perception. So whatever 0:15:56.800 --> 0:16:00.920 you specialize in, you'll see lots of that world. We're 0:16:00.960 --> 0:16:05.480 all specialists at faces, and we often see those where 0:16:05.480 --> 0:16:09.480 there isn't actually one, but it generalizes beyond that. So 0:16:09.640 --> 0:16:11.680 for me personally, when I look around the world, I 0:16:11.760 --> 0:16:15.800 tend to impose shapes that look like brains. I think, oh, 0:16:15.800 --> 0:16:18.880 there's a brain, or there's a midsadgital section, or there's 0:16:18.920 --> 0:16:23.320 a cerebellum. My friends who are pilots tend to see 0:16:23.520 --> 0:16:28.440 airplane shapes and runways. My collogist friends see mushroom shapes, 0:16:28.840 --> 0:16:32.680 my dermatologists friends see melanomas, and so on. Whatever you 0:16:32.840 --> 0:16:37.360 specialize in makes you an expert in detecting that in 0:16:37.440 --> 0:16:41.840 the world, and you impose that interpretation on lots of things. 0:16:42.280 --> 0:16:48.240 It's paridolia writ large. Okay, so back to faces in particular. 0:16:48.680 --> 0:16:51.600 Now we're ready to turn to how the brain does 0:16:51.640 --> 0:16:55.600 its magic with face recognition. One of the key areas 0:16:55.640 --> 0:16:58.400 in our brain that does this is in the temporal lobe. 0:16:58.440 --> 0:17:02.760 It's a little region called the fusiform face area or 0:17:03.160 --> 0:17:07.000 FFA fusiform face area. Now, we can see in brain 0:17:07.040 --> 0:17:12.439 imaging that this area is particularly sensitive to faces. But 0:17:12.680 --> 0:17:15.239 what's interesting is it's not just any part of the 0:17:15.280 --> 0:17:19.720 face that gets attention from the FFA, it's the holistic 0:17:19.880 --> 0:17:25.080 view of the face. So the FFA responds strongly to 0:17:25.200 --> 0:17:30.200 the overall configuration of a face rather than individual features. 0:17:30.480 --> 0:17:34.720 So that allows us to quickly recognize faces and familiar faces, 0:17:34.880 --> 0:17:37.040 even if we only get a quick glimpse of them. 0:17:37.520 --> 0:17:40.639 When you look at these brain imaging studies, for example, 0:17:40.720 --> 0:17:46.400 with fMRI, you find that this fusiform face area responds 0:17:46.560 --> 0:17:51.200 more to familiar faces than to unfamiliar faces. And given 0:17:51.240 --> 0:17:54.240 what we saw a moment ago with the other race effect, 0:17:54.800 --> 0:17:57.960 the FFA is more active when you're viewing faces of 0:17:58.000 --> 0:18:01.800 your own race compared to faces of other races. So again, 0:18:02.080 --> 0:18:06.120 just like recognizing the distinction between different cows if you're 0:18:06.200 --> 0:18:10.200 the rancher, or distinguishing faces that you grew up with 0:18:10.240 --> 0:18:13.200 as opposed to other cultures that you rarely see. Your 0:18:13.320 --> 0:18:18.600 visual system is specialized to efficiently process faces that you 0:18:18.680 --> 0:18:21.960 see a lot more. Now back to this issue about 0:18:22.040 --> 0:18:27.159 recognizing the face holistically rather than detail by detail. This 0:18:27.280 --> 0:18:31.480 is what underlies the phenomenon that's called the inversion effect, 0:18:31.600 --> 0:18:34.840 which is when you see a face that is upside down, 0:18:35.480 --> 0:18:39.359 it's really hard to recognize. So try this. Just open 0:18:39.400 --> 0:18:44.280 a magazine upside down and flip through the pages and 0:18:44.280 --> 0:18:47.159 see if you can recognize the people in it. And 0:18:47.200 --> 0:18:49.240 then when you flip it right side up, it's a 0:18:49.280 --> 0:18:52.440 whole different experience. And this is because this part of 0:18:52.480 --> 0:18:57.080 the brain FFA is super specialized on faces, and it's 0:18:57.119 --> 0:19:00.960 always seen faces in a particular way, and it's recognizing 0:19:00.960 --> 0:19:04.320 the whole big picture with two eyes above the nose, 0:19:04.600 --> 0:19:08.160 above the mouth, and if you throw a ranch in that, 0:19:08.600 --> 0:19:11.280 it doesn't function so well. The key thing is that 0:19:11.359 --> 0:19:15.840 this inversion effect is much stronger for faces than for 0:19:15.960 --> 0:19:22.800 other objects, which highlights this very specialized nature of face processing. Now, 0:19:22.880 --> 0:19:25.479 what's wild is that this part of the brain is 0:19:25.600 --> 0:19:29.320 trying to see faces everywhere, as we saw in paridolia. 0:19:29.800 --> 0:19:32.520 And so if someone puts an electrode into this part 0:19:32.520 --> 0:19:35.520 of your brain, then you might look at, let's say, 0:19:35.520 --> 0:19:38.600 a soccer ball, and when the electrode turns on and 0:19:38.680 --> 0:19:42.760 stimulates this area with electricity, the hexagons of the soccer 0:19:42.800 --> 0:19:46.480 ball seem to you to become the eyes and the 0:19:46.560 --> 0:19:49.879 mouth of a face. So activity in this part of 0:19:49.880 --> 0:19:54.040 the brain is always working over time to detect faces 0:19:54.040 --> 0:19:57.280 in the world, and sometimes it does this by imposing 0:19:57.320 --> 0:20:02.240 the interpretation onto the canvas of the world. Now, consistent 0:20:02.240 --> 0:20:06.679 with what I mentioned before, the FFA isn't specialized just 0:20:06.800 --> 0:20:11.000 for faces. It also plays a role in recognizing other 0:20:11.119 --> 0:20:15.920 things that we're highly familiar with, like cars for car enthusiasts, 0:20:16.119 --> 0:20:20.480 or birds for bird watchers. So that tells us slightly 0:20:20.520 --> 0:20:24.280 more generally that this part of the brain specializes in 0:20:24.320 --> 0:20:28.159 recognizing very detailed things that are relevant and that we 0:20:28.200 --> 0:20:31.920 have a lot of experience seeing. So back to faces, 0:20:32.600 --> 0:20:35.640 the FFA, it's not just that it's sensitive to faces. 0:20:36.040 --> 0:20:38.600 It's actually critical if you want to see a face. 0:20:38.760 --> 0:20:42.240 If you get damaged to the FFA or to its connections, 0:20:42.480 --> 0:20:47.159 then you become unable to recognize faces. And will return 0:20:47.200 --> 0:20:49.440 to that in a moment, but first I just want 0:20:49.480 --> 0:20:52.879 to make one more thing clear. Like almost everything in 0:20:52.920 --> 0:20:56.840 the brain, it's not just one area that's involved. Face 0:20:56.920 --> 0:21:02.480 recognition uses a broader, specialized network. Beyond the FFA, you 0:21:02.560 --> 0:21:06.600 have other areas involved, like the occipital face area and 0:21:06.640 --> 0:21:09.919 the superior temporal sulcus. Don't worry about the details, but 0:21:09.960 --> 0:21:11.919 the thing I want to surface here is that some 0:21:12.080 --> 0:21:15.320 areas are more focused on the detailed features of the 0:21:15.359 --> 0:21:19.840 face and others are sensitive to facial expressions and movements. 0:21:20.200 --> 0:21:24.119 And the key is that together these areas form a 0:21:24.320 --> 0:21:29.360 larger network that allows us to recognize faces and interpret 0:21:29.400 --> 0:21:32.080 their expressions well. Your brain's also doing a lot of 0:21:32.119 --> 0:21:37.560 work to gauge a person's emotions and then navigating social 0:21:37.560 --> 0:21:42.680 interactions by carefully watching the reactions of someone's facial expressions. 0:21:43.200 --> 0:21:47.160 We humans are super specialists at this because our well 0:21:47.200 --> 0:21:53.040 being generally depends on our ability to recognize and understand 0:21:53.080 --> 0:21:55.439 the people around us. And just to give you a 0:21:55.480 --> 0:21:59.080 sense of how all these networks interact, there are studies 0:21:59.119 --> 0:22:04.000 showing that you're interpretation of facial expressions is influenced by 0:22:04.040 --> 0:22:08.080 the broader context. So just as an example, in one study, 0:22:08.400 --> 0:22:12.640 you're shown faces with neutral expressions, but you're given different 0:22:13.119 --> 0:22:16.639 contextual information about the person. So if you're told the 0:22:16.720 --> 0:22:21.879 person is happy, you perceive the neutral face as slightly smiling, 0:22:22.160 --> 0:22:25.520 and if you're told this person is sad, you perceive 0:22:25.600 --> 0:22:29.320 that same face as slightly frowning. So, in other words, 0:22:29.320 --> 0:22:33.720 your brain combines the facial information with the context of 0:22:33.800 --> 0:22:39.240 the larger situation whenever you're interpreting emotions and for that matter, intentions. 0:22:39.720 --> 0:22:43.480 So putting together the giant networks in the brain involved 0:22:43.480 --> 0:22:47.760 in identifying and interpreting faces, the thing that becomes clear 0:22:47.840 --> 0:22:52.840 is that we are massively finely tuned to process faces 0:22:53.320 --> 0:22:58.639 in a way that benefits are social interactions and are survival. Now, 0:22:58.960 --> 0:23:02.119 if you've been listening to other episodes of this podcast, 0:23:02.160 --> 0:23:05.560 you'll know that I'm obsessed by the differences in the 0:23:05.600 --> 0:23:09.280 internal life from person to person, from head to head. 0:23:09.800 --> 0:23:13.080 And one place this really comes up is with individual 0:23:13.200 --> 0:23:18.560 variability in face recognition. So, while most people are fairly 0:23:18.560 --> 0:23:23.080 good at recognizing faces, there's a shocking amount of variation 0:23:23.520 --> 0:23:26.359 in this ability. Some people are exceptionally good at it, 0:23:26.600 --> 0:23:29.600 and other people struggle. And this brings us to the 0:23:29.680 --> 0:23:34.919 concepts of super recognizers on the one hand, and face 0:23:35.160 --> 0:23:38.520 blindness on the other hand. So let's start with the 0:23:38.880 --> 0:23:43.440 super recognizers. These are people who have an extraordinary ability 0:23:43.760 --> 0:23:48.680 to recognize faces. They can remember and identify faces even 0:23:48.720 --> 0:23:52.760 after just brief encounters or after long periods of time. 0:23:53.040 --> 0:23:57.159 Super recognizers are terrific at picking out their acquaintances and 0:23:57.320 --> 0:24:01.240 really large crowds, and they sometimes end up getting employed 0:24:01.240 --> 0:24:05.879 in security or law enforcement to identify perpetrators. Like they'll 0:24:05.880 --> 0:24:09.040 be hired to watch the video of let's say a 0:24:09.080 --> 0:24:12.280 subway entrance and you'll see thousands of people flowing in 0:24:12.320 --> 0:24:15.040 and out, and they'll just stare at the feed for 0:24:15.160 --> 0:24:18.400 hours or days, and then they'll say, oh, there's the guy. Now. 0:24:18.440 --> 0:24:20.920 On the other end of the spectrum, we have what's 0:24:20.960 --> 0:24:25.000 known as face blindness, and this is known as prosopagnosia. 0:24:25.320 --> 0:24:28.320 This word comes from the Greek pro sopon meaning face, 0:24:28.480 --> 0:24:33.800 and agnosia meaning lack of knowledge. So people with prosopagnosia 0:24:34.359 --> 0:24:38.879 have a very hard time recognizing faces, even faces of 0:24:39.080 --> 0:24:42.120 close family and friends. Now, by the way, this isn't 0:24:42.200 --> 0:24:46.360 all or none. The condition ranges from mild to severe, 0:24:46.600 --> 0:24:52.640 like difficulty recognizing familiar faces to an inability to distinguish 0:24:52.720 --> 0:24:55.240 any faces at all. And at that far end of 0:24:55.280 --> 0:24:59.760 the spectrum, just imagine not being able to recognize your 0:24:59.760 --> 0:25:03.880 own spouse or child if you see them out of context. 0:25:04.320 --> 0:25:08.600 And this is the reality for many people with prosopagnosia. Now, 0:25:08.800 --> 0:25:13.239 how does prosopagnosia happen? Usually you're born with it. This 0:25:13.320 --> 0:25:17.160 is called congenital prosopagnosia, just meaning it's present from birth. 0:25:17.800 --> 0:25:22.760 In rarer cases, you can have acquired prosopagnosia, which just 0:25:22.840 --> 0:25:26.520 means that you get it because of a brain injury 0:25:26.760 --> 0:25:29.760 later in life, like a stroke or a traumatic brain injury. 0:25:30.000 --> 0:25:33.960 And in these acquired cases, people notice the sudden change 0:25:33.960 --> 0:25:38.000 in their ability to recognize faces, which is obviously distressing 0:25:38.040 --> 0:25:41.439 and isolating. But what's fascinating is that people born with 0:25:41.600 --> 0:25:44.720 this usually don't realize they have it because they've never 0:25:44.800 --> 0:25:48.280 known anything different, and so they just get along by 0:25:48.359 --> 0:25:52.399 using other cues like people's voices or their clothing or 0:25:52.440 --> 0:25:56.119 the way they walk, and this is how they recognize people. 0:25:56.400 --> 0:25:59.679 One example of a person who had prosopagnosia was the 0:25:59.800 --> 0:26:03.119 neurologist Oliver Sacks, and he wrote an article in The 0:26:03.119 --> 0:26:05.560 New Yorker in twenty ten on this. He talked about 0:26:05.560 --> 0:26:09.600 his personal lifelong struggle with face blindness, but he never 0:26:09.720 --> 0:26:13.680 realized there was something unusual about this until his middle age. 0:26:13.760 --> 0:26:16.520 He'd always just been really bad with recognizing faces, but 0:26:16.600 --> 0:26:19.560 that was just the way it was. But he went 0:26:19.600 --> 0:26:22.120 to visit an older brother in Australia and they got 0:26:22.160 --> 0:26:24.720 to talking and he realized that his brother had the 0:26:24.800 --> 0:26:27.919 same problems with faces that he did, and it dawned 0:26:27.920 --> 0:26:31.520 on him that this was something beyond normal variation, and 0:26:31.560 --> 0:26:35.320 that they both had this trait of face blindness, and 0:26:35.480 --> 0:26:39.040 he guessed that there was probably some distinctive genetic basis 0:26:39.040 --> 0:26:42.359 to it. And very informally, I've noticed among people that 0:26:42.440 --> 0:26:47.680 I know that those with prosopagnosia often don't enjoy movies 0:26:47.800 --> 0:26:50.720 as much as other people, because in order to follow 0:26:50.800 --> 0:26:54.520 a plot, especially if it's switching between an A story 0:26:54.520 --> 0:26:56.720 and a B story and a C story, in order 0:26:56.760 --> 0:26:59.639 to follow that plot, you really need to get it 0:27:00.040 --> 0:27:03.040 when this actor comes into the scene who you haven't 0:27:03.040 --> 0:27:06.280 seen for fifteen minutes and who's now wearing different clothes, 0:27:06.760 --> 0:27:09.679 but this is just a continuation of his plot, and 0:27:09.760 --> 0:27:13.480 everyone else in the audience recognizes him and immediately remembers 0:27:13.520 --> 0:27:15.399 what was going on with him in the last act. 0:27:16.000 --> 0:27:19.000 But just try to imagine how difficult it makes things 0:27:19.320 --> 0:27:22.840 to follow along if you don't immediately recognize the person. 0:27:23.160 --> 0:27:25.639