Investigations Into Prosopagnosia

Brad Duchaine

Vision Sciences Laboratory,
Harvard University

In the last year, we have had the opportunity to assess visual perception in a number of prosopagnosics, and I will discuss a case that contributes to our understanding of the relationship between the recognition of facial identity and facial emotion, and results from a family of prosopagnosics.

Identity/Emotion: Bruce and Young’s model of face recognition posits that the recognition of facial identity and facial expression are performed by separate mechanisms, but the evidence for this distinction is not overwhelming. NM is a 40-year-old prosopagnosic with no history of head trauma who reports great difficulty recognizing facial identity yet no difficulties with facial expressions of emotion. We have tested her with six tests of facial identity recognition and her impairment was clear. In contrast, she performed normally on four tests of emotion recognition that used widely differing methods and emotions. As a result, it is clear that facial identity recognition and facial expression recognition can dissociate and this provides support for this aspect of Bruce and Young’s model.

Genetic prosopagnosia: We have assessed the face and object recognition of three members of one family: a 20-year-old man, his 56-year-old aunt, and his 52-year-old father. The father and son were both impaired on two other face discrimination tests, and the aunt was borderline on one of these tests and normal on the other. All three showed normal performance on some tests of non-face discrimination that used a method identical to that used with faces. The father and son were impaired with famous face recognition, and the aunt’s results were difficult to interpret because she had little exposure to the many celebrities that she did not recognize. It appears that their face recognition difficulties result from difficulties with holistic processing rather than featural processing. In the Jane face discrimination test, participants must decide whether two sequentially presented faces are the same or different. Faces differed either in the shape of the eyes and mouth (featural set), in the spacing of the eyes and mouth (spacing set), or in the shape of the external contour (contour set). All three performed normally with the featural set, but all three were out of the normal range on the spacing set and the contour set. Finally, all three showed normal performance on a test of gender discrimination via the face. These results indicate that genetic factors can result in impairments to particular face recognition mechanisms while not impairing other recognition mechanisms used for faces and objects.

The lecture will take place:

In Room 401, 44 Cummington St.
on Friday, November 15, 2002
at 11:00 am