The Impact of Affect and Age on Facial Expression Predictions

Poster Session D - Monday, March 31, 2025, 8:00 – 10:00 am EDT, Back Bay Ballroom/Republic Ballroom

Eva Cortés Velasco¹, Lorena Chanes¹; ¹Universitat Autònoma de Barcelona

Prior research investigated predictive processing in individuals with depression, revealing decreased predictability of facial expressions, roughly reflecting increased prediction errors, particularly in negative emotion contexts. We here extend such investigation to healthy individuals, exploring the relationship between predictability and positive (PA) and negative (NA) affect. We hypothesized that lower PA and higher NA scores would be associated to decreased predictability ratings (i.e., increased prediction errors), notably for negative-valence emotion contexts. Additionally, exploratory analyses were performed to investigate the potential influence of age on predictability ratings. A database of 288 participants having completed a predictive processing and social perception task and the Positive and Negative Affect Schedule (PANAS) was used. The task involved scenarios (texts) designed to evoke emotions (fear, sadness, or happiness) followed by a facial expression that could either match or mismatch the evoked emotion. Participants were asked to predict the facial expression (predictability rating). The data was analyzed using hierarchical linear modeling, with predictability ratings as the outcome variable and match condition or evoked emotion and affect or age as predictors. Lower PA scores were associated with decreased predictability ratings, the effect being stronger for negative- than positive-valence emotions, in line with what was observed for depression. NA scores did not show significant effects. The exploratory analysis for age suggested an association with predictability ratings, possibly in interaction with match condition, pointing at decreased predictability or stereotypicality of predictions with age. These findings extend the observations made for depression and suggest potential modulations of predictions across the lifespan.

Topic Area: EMOTION & SOCIAL: Emotion-cognition interactions