Journal of Capital Medical University

Previous Articles     Next Articles

Facial features and personality traits in traditional Chinese medicine five-element typology based on 3D point-cloud data

Lin Jin1，2,  Jia Hongxiao1，2*,  Lü Hongpeng1，2*,  Dai Zhiqing1，2,  Jiang Xinyue3,   Duan Yuhang1,2,  Xu Huanshu1,2,  Zhao Ziyi1,2,  Zhang Yunhe4   

  1. 1.Beijing Anding Hospital, Capital Medical University；National Center for Mental Disorders；National Clinical Research Center for Mental Disorders；Beijing Key Laboratory of Intelligent Drug Research and Development for Mental Disorders， Beijing 100088, China; 2. Advanced Innovation Center for Human Brain Protection, Capital Medical University, Beijing 100069, China;3. Beijing University of Chinese Medicine Third Affiliated Hospital, Beijing 100029, China; 4. Chifeng Psychiatric Hospital,Chifeng 024200， Inner Mongolia Autonomous Region，China
  • Received:2026-02-02 Revised:2026-03-09 Online:2026-06-21 Published:2026-06-25
  • Supported by:
    This study was supported by National Natural Science Foundation of China(82305193，82474429), Beijing Traditional Chinese Medicine Inheritance “New 3+3” Program Demonstration Case Project（2023-ZYSF-19）,Capital Health Development Scientific Research Special Project (2026-2-2124), Capital Medical University Clinical-Basic Cooperation Platform Cultivation Project(JLPYPT2025002).

PDF

6

Abstract: Objective  To objectively quantify facial features using 3D point-cloud facial data under the traditional Chinese medicine (TCM) framework of “unity of form and spirit” (xing-shen unity), and to evaluate the predictability of personality traits from facial features in TCM five-element typology. Methods  Typical healthy participants of TCM five elements type were recruited. The 3D point-cloud facial data and the Sixteen Personality Factor Questionnaire (16PF)  were collected. A deep-learning algorithm was used to automatically localize 3D facial landmarks and extract 39 facial features. Ridge regression models were contructed for each 16PF factor as the dependent variable with facial features as predictors, adjusting for age and sex, and evaluated using 10-fold cross-validation. Factors with at least moderate predictive performance (R2≥ 0.15) were identified. Key facial features were identified based on permutation importance, and the moderating effect of sex on the associations between key facial features and personality traits was further assessed. Results  A total of 804 participants were included. Facial features showed at least moderate predictive performance for five 16PF factors: Abstractedness (M), Tension (Q4), Openness to change (Q1), Rule consciousness (G), and Perfectionism (Q3). Key predictors mainly involved indices of facial surface undulation and curvature statistics, soft-tissue morphology, and craniofacial structure as well as facial component geometry. Following the inclusion of interaction terms between key facial features and sex, the model exhibited no substantial improvement in predictive performance. Conclusion  The 3D facial features can predict several 16PF factors to a certain extent, suggesting a quantifiable correspondence between facial morphology and personality traits in TCM five-element typology, and providing modern empirical support for the TCM theory of xing-shen unity.

Key words: traditional Chinese medicine five-element typology, 3D point-cloud data, facial features, personality traits, Sixteen Personality Factor Questionnaire（16PF）, ridge regression

CLC Number: