Using genetic algorithms to uncover individual differences in how humans represent facial emotion

Area of Science:

• Psychology
• Cognitive Science
• Neuroscience

Background:

• Emotional facial expressions are crucial for social interaction and cognition.
• Current emotion research often assumes uniform emotional representations, neglecting individual differences.
• Standardized stimulus sets may not capture the subjective nature of emotion perception.

Purpose of the Study:

• To develop and validate an assumption-free method for generating participant-specific emotional expressions.
• To investigate individual variability in the subjective representation of basic emotions (happy, angry, fearful, sad).
• To assess the reliability and accuracy of these self-generated emotional expressions.

Main Methods:

• Utilized genetic algorithms to allow participants to generate their own emotional facial expressions.
• Collected subjective representations of happy, angry, fearful, and sad faces from 105 participants.
• Tested identification accuracy of generated faces by a separate group of 108 individuals.

Main Results:

• Observed population-level consistency for happy facial expressions.
• Found significant variability in the representation of fearful and sad faces across individuals.
• Demonstrated high test-retest reliability for all generated emotional expressions.
• Identified accurate recognition of happy and angry faces, but common misidentification of fearful and sad faces.

Conclusions:

• Individual differences significantly influence the representation of certain emotions, particularly fear and sadness.
• The developed genetic algorithm task offers a novel approach to studying subjective emotion perception.
• Findings challenge the use of standardized stimuli and pave the way for understanding atypical emotion processing.