Race-specific norms for coding face identity and a functional role for norms

Area of Science:

• Cognitive Psychology
• Neuroscience
• Computer Vision

Background:

• Face perception models often use a multidimensional space with an average face as a norm.
• Similar visual properties of faces suggest encoding in a common space, but subclassification may occur.
• Race-specific norms might influence how faces are processed and categorized.

Purpose of the Study:

• To investigate whether race-specific norms are utilized in human face perception.
• To examine the impact of race on identity aftereffects and discrimination thresholds.
• To determine if distinct average faces (norms) exist for different racial groups.

Main Methods:

• Utilized high-level aftereffects paradigm in face perception studies.
• Measured identity aftereffects using adapt-test pairs relative to race-specific (Asian, Caucasian) and mixed-race averages.
• Assessed identification thresholds for targets presented around race-specific and mixed-race averages.

Main Results:

• Identity aftereffects were significantly larger for race-specific compared to mixed-race adapt-test pairs.
• Aftereffects were greater for pairs opposite relative to a norm, supporting norm-based coding.
• Identification thresholds were lower when targets were centered around race-specific averages.

Conclusions:

• Evidence suggests that the human visual system employs race-specific norms for encoding facial identity.
• These race-specific norms appear to facilitate more efficient identity discrimination within racial groups.
• Findings support the existence of specialized neural mechanisms for processing faces of different races.