Ken Nakayama

时间： 2010-06-01 15:00 - 17:00

地点： 哲学楼103房间

Range and scope of face recognition abilities in humans

Because so much of our life is social and because noticing subtle changes in faces is so important in navigating through our extraordinarily complex social life, we humans are face recognition experts. We rely on the many differences that exist between one face and another. Configural processing is present but small local features are also important. Very recently, and to our surprise, we have found that the ability to learn new faces does not peak in late adolescence as previously assumed but occurs later, in our early thirties. Furthermore, our abilities decline rather slowly after this point. Some people (developmental prosopagnosics) are very bad at face recognition, so much so that they are severely handicapped in real life. Some of these deficits are restricted to faces but many others have a more general visual memory loss. To our surprise, the ability to discriminate other subtle facial differences (emotion, gender, age, attractiveness) is mostly intact despite major problems in recognizing people. In addition, preliminary tests indicate that episodic memory is normal at least for verbal materials. We have also identified others at the opposite end of the spectrum, whom we call super-recognizers. Quantitatively, these individuals appear to be as extraordinarily good as prosopagnosics are regretfully bad, suggesting a wide range of face recognition capacities in the population. It’s likely that face recognition abilities have a genetic basis independent of other mental faculties because face recognition problems run in families and face recognition scores in MZ twins are highly correlated.  We discuss these results in terms of contemporary accounts of face processing as well as their possible
relation to more general concepts of perception, memory and learning.

2010-06-01

2010-06-01