[ Back to EurekAlert! ]

PUBLIC RELEASE DATE:
19-May-2009

[ | E-mail ] Share Share

Contact: Amy Lavoie
amy_lavoie@harvard.edu
617-496-9982
Harvard University
@HarvardResearch

'Super-recognizers,' with extraordinary face recognition ability, never forget a face

Research suggests that face recognition may vary more than previously understood

CAMBRIDGE, Mass., May 19, 2009 - Some people say they never forget a face, a claim now bolstered by psychologists at Harvard University who've discovered a group they call "super-recognizers": those who can easily recognize someone they met in passing, even many years later.

The new study suggests that skill in facial recognition might vary widely among humans. Previous research has identified as much as 2 percent of the population as having "face-blindness," or prosopagnosia, a condition characterized by great difficulty in recognizing faces. For the first time, this new research shows that others excel in face recognition, indicating that the trait could be on a spectrum, with prosopagnosics on the low end and super-recognizers at the high end.

The research is published in Psychonomic Bulletin & Review, and was led by Richard Russell, a postdoctoral researcher in the Department of Psychology at Harvard, with co-authors Ken Nakayama, Edgar Pierce Professor of Psychology at Harvard, and Brad Duchaine of the University College London.

The research involved administering standardized face recognition tests. The super-recognizers scored far above average on these tests--higher than any of the normal control subjects.

"There has been a default assumption that there is either normal face recognition, or there is disordered face recognition," says Russell. "This suggests that's not the case, that there is actually a very wide range of ability. It suggests a different model--a different way of thinking about face recognition ability, and possibly even other aspects of perception, in terms of a spectrum of abilities, rather than there being normal and disordered ability."

Super-recognizers report that they recognize other people far more often than they are recognized. For this reason, says Russell, they often compensate by pretending not to recognize someone they met in passing, so as to avoid appearing to attribute undue importance to a fleeting encounter.

"Super-recognizers have these extreme stories of recognizing people," says Russell. "They recognize a person who was shopping in the same store with them two months ago, for example, even if they didn't speak to the person. It doesn't have to be a significant interaction; they really stand out in terms of their ability to remember the people who were actually less significant."

One woman in the study said she had identified another woman on the street who served as her as a waitress five years earlier in a different city. Critically, she was able to confirm that the other woman had in fact been a waitress in the different city. Often, super-recognizers are able to recognize another person despite significant changes in appearance, such as aging or a different hair color.

If face recognition abilities do vary, testing for this may be important for assessing eyewitness testimony, or for interviewing for some jobs, such as security or those checking identification.

Russell theorizes that super-recognizers and those with face-blindness may only be distinguishable today because our communities differ from how they existed thousands of years ago.

"Until recently, most humans lived in much smaller communities, with many fewer people interacting on a regular basis within a group," says Russell. "It may be a fairly new phenomenon that there's even a need to recognize large numbers of people."

###

The research was funded by the U.S. National Eye Institute and the U.K. Economic and Social Research Council.



[ Back to EurekAlert! ] [ | E-mail Share Share ]

 


AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert! system.