Representation of facial identity includes expression variability.

Area of Science:

• Cognitive Psychology
• Neuroscience
• Computer Vision

Background:

• Understanding how the human brain forms and retrieves facial representations is crucial for cognitive science.
• Existing models like exemplar and prototype theories offer partial explanations for facial recognition.
• The role of dynamic facial expressions in forming stable identity representations remains an active area of research.

Purpose of the Study:

• To investigate the contribution of expression variability to the formation of face representations.
• To determine if learning faces with varying expressiveness impacts subsequent recognition performance.
• To propose a refined model of face representation that incorporates expression dynamics.

Main Methods:

• Participants learned novel identities from face images characterized by either low or high expressiveness.
• Recognition performance was assessed using a test set of novel facial images.
• Response times and accuracy were recorded to measure recognition efficiency and effectiveness.

Main Results:

• Recognition accuracy after low expressiveness training was significantly modulated by image expressiveness, with more expressive images leading to slower responses.
• Recognition after high expressiveness training showed minimal dependence on image expressiveness.
• These findings challenge existing exemplar and prototype theories of face representation.

Conclusions:

• Facial representations incorporate variability in expressions, suggesting a dynamic rather than static model of face perception.
• Learning to recognize an individual involves understanding the range of expressions they exhibit.
• A combined model of average and exemplar representations, preserving within-person variability, best explains the observed results.