Audio: Hear more on this feature by downloading .
It was the ultimate cold shoulder. Having spent a perfectly nice evening on a date with a girl from college, Jacob Hodes then spent the rest of the year ignoring her. “I never saw her again,” he says. “Well, I’m sure I walked past her plenty of times, but I just didn’t see łó±đ°ů.”
Hodes didn’t set out to play it cool. He just couldn’t for the life of him remember what his date looked like. He had had the same trouble all his life: people would say “Hi”, and he wouldn’t have a clue who they were, and he’d feel shy even with people he’d met several times. He knew the names of celebs but he could have walked past any one of them in the street without batting an eyelid.
Advertisement
It wasn’t until five years ago that it all made sense. That was when Hodes was diagnosed with prosopagnosia, a condition that means he is unable to recognise faces. He is far from alone. In fact, the condition is so common that if you’re not prosopagnosic yourself, you almost certainly know someone who is. Strange as it might sound, until they’re tested most sufferers don’t even realise they have a problem with face recognition.
Research into prosopagnosia, or face blindness, may do more than just explain why some people cannot recognise their nearest and dearest. It might also settle an ongoing debate about how the rest of us recognise faces, which happens to be one of the hottest topics in brain research. Do our brains have specialised modules for the job, or are faces dealt with by generic recognition modules that also recognise other objects, be they makes of car or breeds of dog? This, in turn, is part of a wider debate about how the brain deals with information: is it like a Swiss army knife with a separate tool for each task or is it more of a general-purpose information processor? With more cases of prosopagnosia being diagnosed every year, and a glut of papers on the subject, we could soon have some answers.
Prosopagnosia was first recognised in 1947 when a German neurologist, Joachim Bodamer, described the condition in a 24-year-old man who’d lost the ability to recognise his friends and family, and even his own reflection, after a bullet wound to the head. Until a few years ago only a few dozen cases had ever been described, all caused by brain injury, and the condition was considered extremely rare. Recently, though, researchers identified a second form of face blindness, “developmental prosopagnosia”, which is either present from birth or develops very early in life.
Developmental prosopagnosia is surprisingly common. In May a team from Harvard University and University College London announced the results of a web survey of 1600 people suggesting that up to 2 per cent of people have some degree of face blindness (about 2.5 per cent of people are colour-blind and 5 per cent are dyslexic). Then in August Martina GrĂĽter and colleagues at the Institute for Human Genetics in MĂĽnster, Germany, similarly reported that 2.5 per cent of 700 secondary-school pupils they had tested had trouble recognising faces (American Journal of Medical Genetics, vol 140A, p 1617).
The results took everyone by surprise, not least many of the subjects. It seems that if you have never known what it is to recognise a face, you don’t necessarily know that you’re supposed to be able to. This probably explains why developmental prosopagnosia was unknown until recently. “Prosopagnosics almost always know that they have trouble recognising people, but they often don’t realise that other people are better at it,” says Brad Duchaine, a cognitive neuroscientist at University College London. “Sometimes they don’t realise that it is specifically a problem with the face.”
“Prosopagnosics almost always know that they have trouble recognising people, but they often dont realise that other people are better at it, says Brad Duchaine, a cognitive neuroscientist at University College London. Sometimes they dont realise that it is specifically a problem with the face.”
“It’s like asking someone who’s colour blind what’s it like not to see colours,” adds Thomas Grüter, an independent scientist and prosopagnosic who works with his wife Martina’s team in Münster.
Like most prosopagnosics, Grüter has spent his whole life inventing strategies to compensate for his face blindness. He says he recognises people by the way they walk, the clothes they wear or the way they speak. Hodes finds hairstyles useful, although he admits this strategy isn’t foolproof. He describes watching a colleague pull her hair into a ponytail and “disappear” before his eyes. “Her identity dissolved in front of me,” he says. “It was very disconcerting.”
“He describes watching a colleague pull her hair into a ponytail and disappear before his eyes. Her identity dissolved in front of me, he says. It was very disconcerting.”
Despite these difficulties most developmental prosopagnosics are so good at getting around their problem that, unless they see a familiar person out of context, with a new hairstyle or in different clothes to normal, they can recognise people just fine.
Even so, the discovery that prosopagnosia may be much more common than once thought has attracted attention from neuroscientists and cognitive psychologists keen to nail down the face-recognition debate. Discovering what is different about the brains of face-blind people could, they believe, answer the question of whether we have specialised face-recognition modules in our brains.
Macaques, apples and faces
Duchaine, who has spent the past eight years researching prosopagnosia, is firmly in the modular camp. He points to evidence from brain-imaging studies in people without prosopagnosia showing that several regions of the brain respond selectively to faces.
One such region is the fusiform gyrus, also called the fusiform face area or FFA. Neuroscientists have known for many years that this area lights up in response to faces, and also to patterns suggestive of a face. But the FFA also responds to other objects, albeit less strongly, casting doubt on the idea that it is specifically tuned in to faces.
A paper published in Science in February (vol 311, p 670) gave the strongest evidence yet that the FFA is indeed a specialised face centre. In an experiment on macaques, Doris Tsao of Harvard University used fine electrodes to record the output of individual neurons in the FFA. She found that 97 per cent of cells in the FFA responded almost exclusively to faces; they also responded weakly to apples, clock faces and other round, face-like objects but other objects suppressed activity in these cells. “I think it settles the debate about whether there are areas in the brain dedicated to face processing,” says Tsao. Duchaine agrees. “Doris’s evidence is overwhelming,” he says. “It’s hard to argue with that.”
Isabel Gauthier, a cognitive neuroscientist at Vanderbilt University in Nashville, Tennessee, begs to differ. She is one of the main supporters of the “expertise” hypothesis, which holds that we use the same brain circuits to distinguish between all objects, including faces, and that we are only good at recognising faces because we have been studying them all our lives.
“We’ve known for a long time that there are face cells in the FFA,” she says. “But they didn’t do the same experiments after training the monkeys for something other than faces, and show that another part of the brain is picking it up.” If the FFA is just for faces, she argues, then other familiar objects must be handled elsewhere. But that’s not what happens. “The FFA responds when we show people objects with which they’re expert.” The FFA of a birdwatcher, for example, will activate for faces and birds, but not some other kind of object – cars, say – in which they’re not interested, she says.
Critics point out that in these experiments the FFA response was half as strong for birds or cars as it was for faces. Gauthier thinks that’s hardly surprising. “Even if you love cars you spend more time with people and have done over your whole life,” she says. That, however, doesn’t explain why the macaques’ “face cells” did not respond to images of macaque bodies, something they have a lot of experience of perceiving.
So what does the study of prosopagnosics bring to the table? Surprisingly, initial studies were a blow for the Swiss-army-knife crowd: sufferers often show no deficit in their FFA response. However, unpublished work from a team led by Galit Yovel of Tel Aviv University, Israel, has found evidence that while the FFA does respond to faces in prosopagnosics, it doesn’t always do so normally.
When unaffected people are shown the same face twice in succession, the FFA responds less strongly on the second showing. This is called “adaptation”: once the brain has clocked a stimulus once, it doesn’t bother to do it again. When Yovel’s team did the same experiment with seven prosopagnosics, five of them failed to show adaptation. “You show them the same face twice, they show the same response as if you’d showed them different faces,” says Duchaine. That doesn’t mean adaptation is the underlying problem, but it does at least show that the FFA isn’t working as normal. Yovel is now doing tests to see if lack of adaptation is a general problem among prosopagnosics.
Other lines of evidence from prosopagnosics also support the idea that our brains have specialist equipment for recognising human faces. Duchaine cites a case study where a man who acquired prosopagnosia after a brain inury gave up his city-based job to become a sheep farmer and soon discovered that, though he still couldn’t recognise the faces of his wife and family, he was able to identify individual sheep. The expertise hypothesis would suggest that he shouldn’t have been able to develop that skill, argues Duchaine.
There is also at least one example of the opposite case – a man known as Mr C. K. who, as the result of a brain injury, lost the ability to tell one object from another, though he has no problem with faces. Intriguingly, Mr C. K. used to be an expert plane-spotter but can no longer tell one plane from another. “He lost expert abilities yet didn’t lose his face abilities,” says Duchaine.
Gauthier is equally convinced that it is impossible to use this evidence to settle the debate. “Prosopagnosia is fascinating but it’s not going to solve the question of face recognition in the brain,” she says. She points out that some people have brain damage where they specifically lose the ability to read. “That doesn’t mean there’s a special part of the brain for reading,” she says. Perhaps faces will turn out to be the same.
Whatever the outcome, few researchers think that the FFA is the full story. Studies of people with brain damage have shown that another area, the occipital face area, also seems to be necessary for face recognition. Damage to either the FFA or the OFA causes problems. How these elements fit together, whether they work in tandem or along the same pathway, and how they tie in with our memory of people we know, are still mysteries. What is more, says Tsao, there may be other brain regions outside of these that also contribute to face recognition.
Genetic glitches
The waters are muddied further by the fact that no two prosopagnosics are the same. Some have problems only with faces, while others have trouble with objects or animals as well. Some cannot recognise faces but have no trouble perceiving whether a face is attractive, or what emotion it might be conveying, while others cannot do this. Some can train themselves to recognise specific faces, others can’t even recognise their own in the mirror. Because of this diversity, working out the cause of prosopagnosia will not be easy. Individual cases may have slightly different glitches in a complicated system.
Some of these glitches may turn out to be genetic. In Grüter’s prevalence study, all of the prosopagnosics who agreed to have their families tested reported a parent with the condition. Having looked at 38 cases in seven families, including Thomas Grüter’s, the team believe they have good evidence that a single gene could be responsible.
Duchaine also has some evidence that face blindness can be hereditary but cautions against putting too much emphasis on a single gene. Other factors, he says, may be important. For example, studies of babies with congenital cataracts have shown that being unable to see with the left eye during the first two months of life is a major risk factor for prosopagnosia. This raises the possibility that there is a developmental window for learning face recognition. Similarly, many people who contact Duchaine after hearing about his work turn out to have a condition called amblyopia, where the eyes cannot focus on one point so the brain ignores the input from one eye. “It looks like you need input from the left eye to stimulate the brain in the first months of life to activate the mechanisms in face processing,” he says.
Whatever the cause, what most prosopagnosics want to know is whether they can do anything to improve their face-recognition skills. The bad news is that as yet there is no treatment beyond the coping strategies people like Hodes already develop for themselves. That doesn’t mean nothing can be done, however. Joseph DeGutis, a graduate student at the University of California, Berkeley, recently reported successfully training a severe developmental prosopagnosic to recognise faces. Not only did the patient find it easier to pick out faces in laboratory tests, she also reported that recognising faces in everyday life became a little easier. Brain-imaging studies also revealed more activity in her FFA after training, and better connectivity between the parts of the brain involved in face processing.
Duchaine now plans to attempt to train sufferers to recognise the five people that they most need to know – maybe their close family and colleagues, for example.
Thomas Grüter, however, is not convinced it will work. “I don’t know how you can have more training than you have already had,” he says. “Humans already spend all day looking at faces.” He also points out that it’s sometimes easy to cheat the tests. “One person we studied said that when she was doing the face-recognition test she memorised the distance between nose and upper lip. So you can perform well on the test and not do well in real life and vice versa.”
Hodes is living proof of this. Laboratory tests revealed he in the bottom 2 per cent in terms of face recognition, but he is keen to point out that, blind dates aside, it doesn’t cause him too many problems. “Day to day it’s been a fascinating process of understanding how my mind works.”
