ABSTRACT

Background & objectives
Existing research indicates that not only own stress leads to physiological stress reactions, but also observing stress in others. So far, a standardized paradigm to reliably induce physiological stress contagion based on direct face-to-face stress observation compared to an active non-stress observing control group is lacking. Here, we tested a standardized randomized placebo-controlled experimental paradigm to investigate physiological reactivity to direct stress observation and characterized the stress contagion response of major endocrine stress systems including full reactivity kinetics.

Methods
Healthy young male participants were randomly assigned to (1) undergo an adapted version of the Trier-Social-Stress-Test (“TSST participants”, n=20), (2) observe it (“stress observers”, n=36), or (3) observe a corresponding placebo-stress control condition (“non-stress observers”, n=30). We repeatedly assessed heart rate, salivary alpha-amylase, salivary cortisol, as well as salivary aldosterone.

Results
Stress observers exhibited greater physiological reactivity to stress observation as compared to non-stress observers to placebo-stress observation in heart rate, salivary alpha-amylase, and cortisol (p’s≤.۰۲۷), but not in aldosterone. We observed similar reactivity kinetics in TSST participants and stress observers, but less pronounced in stress observers.

Discussion
Extending previous literature, our findings indicate that independent of secondary effects of the observation setting, direct observation of stress in other individuals induces activation of the hypothalamus-pituitary-adrenal axis and the sympathetic-adrenal-medullary axis. Moreover, the physiological stress contagion response resembles the physiological reactivity to first-hand stress but is less pronounced. Potential implications of physiological stress contagion regarding health, cognition, or behavior, as well as modulating factors need to be further elucidated.

INTRODUCTION

While the individual experience of and reactivity to direct stress exposure has been extensively studied over decades, comparably little is known about interindividual contagion effects of such stress exposure. Given the increasing stress burden on the one hand (Techniker Krankenkasse, 2021) and increasing population numbers especially in metropolitan areas on the other hand (Moreno-Monroy et al., 2021, United Nations, 2022), a better understanding of the contagion of stress is of importance.

In research, the interindividual contagion of stress, or stress contagion, respectively (Dimitroff et al., 2017, Erkens et al., 2019, Schury et al., 2020, Waters et al., 2020, Waters et al., 2017, Waters et al., 2014) is captured by additional different terms such as crossover effects of stress (for review see: (Wethington, 2000), autonomic contagion (Ebisch et al., 2012), vicarious autonomic response (Manini et al., 2013), empathic physiological resonance (Blons et al., 2021, Buchanan et al., 2012), and vicarious or resonant empathic stress (Blons et al., 2021, Engert et al., 2019, Engert et al., 2014, Engert et al., 2018, Park et al., 2021) that in part differ slightly with respect to implications. Stress contagion research started in the 1990s, with a focus on stress contagion at the cognitive and emotional level between close individuals (Wethington, 2000). Within the last 10 years, investigation of stress contagion has been extended to the physiological level with studies assessing physiological responses in reaction to the observation of stressed individuals.

So far, studies investigated physiological stress contagion in terms of significant physiological increases following the observation of other stressed individuals and/or of covariation of physiological parameters between observers and stressed individuals (for review see: (Engert et al., 2019, White and Buchanan, 2016). Hitherto, the physiological contagion of stress was examined between parents (mostly mothers) or stranger women and children (Ebisch et al., 2012, Manini et al., 2013, Waters et al., 2020, Waters et al., 2017, Waters et al., 2014), couples (Engert et al., 2014, Engert et al., 2018), or strangers (Blons et al., 2021, Buchanan et al., 2012, Dimitroff et al., 2017, Engert et al., 2014, Erkens et al., 2019, Park et al., 2021, Schury et al., 2020) in various settings.

RESULTS

۳٫۱٫ Participant characteristics
Our final sample comprised 20 TSST participants, 36 stress observers, and 30 non-stress observers. Participant characteristics are depicted in Table 1. As expected, participants did not significantly differ in terms of age and BMI as well as in terms of physiological measures at baseline (p’s ≥.۰۷۴). Moreover, panel size did not differ between stress observation (2.85 ±.۱۷) and placebo stress observation condition (2.67 ±.۲۶) (t(30) =.63, p =.54).

۳٫۲٫ Intraclass-correlation coefficients
Table 2 shows the ICCs for all dependent variables at every measurement timepoint. With the exception of aldosterone and changes from baseline in aldosterone and HR, the proportion of variance that can be explained by study group was generally rather low (both for stress observers and non-stress observers) which suggests that ANOVAs and t-tests should be rather robust. Nevertheless, alternatives for the inferential-statistical procedures are reported in Appendix A. These are SEM-analyses that consider potential between-study group differences in accordance with the ICCs. No substantially different conclusions were yielded with the alternative analyses.