Abstract
What aspects of face recognition, if any, are spared in a prosopagnosic (DC) whose object recognition and reading are intact? We compared DC's performance with that of a person with object agnosia and pure alexia (CK, see Moscovitch et al., 1997), and with that of a person with only pure alexia (WK), because reading, like object recognition, is presumed to be mediated by a part-based processing system which may also be involved in some aspects of face-recognition, such as recognition of inverted and fractured faces. We also included a control, DCB, matched for social and educational background to DC (his 2.5-years-older brother). DC could recognize only about 37% of all upright full-viewed faces, whereas CK, WK and DCB recognized about 80% of them. When these faces were inverted, DC and CK were impaired (9% and 14%) compared to DCB and WK (50% and 70%). For DC, this inversion weakness was true whether the internal or external facial features were inverted, but for CK, WK, and DCB, it was only true when the internal parts were inverted. When the faces were disguised and when parts were missing, DC only recognized about 27% of them, whereas CK, WK, and DCB had no difficulty. Only CK and DC were impaired on recognizing fractured faces. We conclude that (1) only the part-based object system contributes to some aspects of face recognition; (2) it does so by interacting with an intact face-system; and, (3) damage to the face system leads to global face recognition deficits.
Supported by NSERC grant 0155930 (JR) and NSERC grant A8347 (MM).