Research highlight

Social knowledge alters face perception

Nature Neuroscience

May 3, 2016

The neural representations of race, sex and emotion become linked in the brain according to a person’s stereotypical bias, reports a study published online this week in Nature Neuroscience. The study suggests that an individual’s personal links between these broad social categories can influence how faces are processed in the brain and socially perceived.

Computer modeling suggests that distinct social categories, such as sex, race and emotion, can become interlinked or entangled in the mind as a result of conceptual knowledge or stereotypes. Although behavioral research suggests that these stereotypical links between sex, race and emotional traits influence perceptual judgments about a person, the brain regions involved in these links have not been identified.

Ryan Stolier and Jonathan Freeman examined whether two brain regions involved in processing faces and social categories - the fusiform gyrus and the orbitofrontal cortex - could mediate the influence of social categories on face perception. A total of 43 healthy young adults passively viewed computer-generated faces that varied in sex (Male, Female), race (Black, White, Asian), and emotion (Angry, Happy) while their brain activity was measured in two functional magnetic resonance imaging studies. Outside the scanner, participants categorized faces and traits such as intelligence and aggression as they related to sex, emotion, and race. They found that a person’s categorization of traits (reflecting personal stereotypes or biases) was related to how they categorized faces in terms of race + emotion, race + sex, and sex + emotion pairings. In turn, these face categorizations were related to the similarity of neural representations across sex and emotion, race and sex, or race and emotion specifically in the fusiform gyrus and orbitofrontal cortex. The behavioral and neural links between these categories were not due to similarities in the visual features of the faces.

doi: 10.1038/nn.4296

Return to research highlights

PrivacyMark System