Holistic processing of face gender in developmental prosopagnosia

Abstract

Normal face identity recognition and gender recognition has been shown to rely on holistic and configural processing, as evidenced by recognition difficulties with face inversion, viewing scrambled face parts, and when recombining the top of one face with the bottom of another (composite effect). Face gender recognition, however, can also be successfully performed using parts-based analysis, such as using eyebrow thickness or pigmentation cues. In developmental prosopagnosia (DP), face recognition is severely compromised and evidence suggests that impaired holistic face identity processing plays a predominant role. However, DPs typically demonstrate normal face gender recognition. It may be that DPs' holistic processing deficits are more specific to face identity and that DPs engage holistic processing mechanisms to successfully recognize gender. Alternatively, DPs may have general holistic processing deficits with faces and recognize gender using more parts-based analyses. To test these alternatives, we assessed 9 DPs on face gender recognition ability, and tested their holistic processing of gender: comparing upright gender recognition to inversion and viewing scrambled face parts, as well examining the gender composite effect. Our results show gender recognition in DPs is not significantly different than normal controls. Both groups were significantly impaired at gender recognition by inversion and displaying scrambled parts, and these decrements in performance are similar in magnitude. Like controls, DPs also showed a significant composite effect for gender. This was reduced compared to the effect in healthy controls. Together, these results suggest that face gender is processed holistically in DPs and calls into question the assertion that DP is a generalized holistic processing deficiency.

NEI-NIH grant to KN.