Facial contraction patterns and their association with psychological well-being: a cross-sectional study
摘要
Facial expressions are known to play a key role in emotional communication, but their association with psychological well-being is less understood. This study explores whether facial contraction patterns, specifically in the Glabella, Frontalis, and Smile regions, correlate with psychological well-being, as measured by the WHO-5 Well-being Index.
MethodsA cross-sectional observational study was conducted with 300 participants, aged 18 to 80 years, to assess the relationship between facial contraction patterns and psychological well-being. Participants’ facial contractions were photographed under standard conditions, and psychological well-being was measured using the WHO-5 Well-being Index. The contraction patterns were categorized for Glabella (e.g., U-pattern, V-pattern), Frontalis, and Smile (Duchenne vs. Non-Duchenne), and the results were analysed using ANOVA, multiple linear regression, logistic regression for depression risk, and machine learning predictive models.
ResultsThe Random Forest model emerged as the best predictive model, with an accuracy of 79.5%, AUC of 0.85, and identifying Non-Duchenne (PanAm) smile and Inverted Omega pattern in the Glabella region as the strongest predictors of poor well-being. The study found that participants with non-Duchenne smiles had lower WHO-5 scores (β = -6.52, p < 0.001), while those with the Inverted Omega pattern also exhibited lower well-being (β = -3.45, p = 0.012). Logistic regression revealed that non-Duchenne smiles (OR = 4.23) and Inverted Omega patterns (OR = 5.12) were strongly associated with depressive symptoms.
ConclusionsThe study supports the Facial Feedback Hypothesis, demonstrating that specific facial contraction patterns, particularly the non-Duchenne smile and Inverted Omega pattern, are indicative of poor psychological well-being and an increased likelihood of depressive symptoms. These findings suggest that facial expressions could serve as non-invasive markers for mental health screening and intervention strategies.
Level of EvidenceNot gradable.