Citation

Lau WKW, Tai APL, Chan JNM, Lau BWM, Geng X. Psychoneuroendocrinology 2021; 129: 105267.

Copyright

(Copyright © 2021, Elsevier Publishing)

DOI

10.1016/j.psyneuen.2021.105267

PMID

unavailable

Abstract

The hypothalamic-pituitary-adrenal (HPA) and parasympathetic nervous systems have been reported to play important roles in emotion regulation and stress coping. Yet, their direct relationship with psychological resilience remains unclear. These biophysiological features should be considered together with the traditional psychometric properties in studying resilience more comprehensively. The current study aimed to examine the role of these systems during a laboratory stress task and to determine the prediction power of resilience by combining psychological and biophysiological features. One hundred and seven (52 females) university students without psychiatric disorders underwent the Trier Social Stress Task (TSST). Psychometric properties of resilience were measured at rest; vagal heart rate variability (HRV), salivary cortisol, and dehydroepiandrosterone (DHEA) levels were captured at baseline, during, and after TSST. Multivariate linear regression as well as support vector regression machine-learning analyses were performed to investigate significant predictors and the prediction power of resilience.

RESULTS showed that positive and negative affects, HRV during the anticipatory phase of stress, and the ratio of cortisol/DHEA at the first recovery time point were significant predictors of resilience. The addition of biophysiological features increased the prediction power of resilience by 1.2-fold compared to psychological features alone.

RESULTS from machine learning analyses further demonstrated that the increased prediction power of resilience by adding the ratio of cortisol/DHEA was significant in "cortisol responders"; whereas a trend level was observed in "cortisol non-responders". Our findings extend the knowledge from the literature that high vagal activity during the anticipating phase of stress and the ability to restore the balance between cortisol and DHEA after a stress event could be an important feature in predicting resilience. Our findings also further support the need of combining psychological and biophysiological features in studying/predicting resilience.

Language: en

Keywords

Machine learning; Resilience; Affective styles; Cortisol; Dehydroepiandrosterone; Heart rate variability