استفاده از ربات برای القای واکنش استرس
Abstract
1. Introduction
2. Materials and methods
3. Results
4. Discussion
5. Conclusions
Funding
Declarations of interest
References
Abstract
This study examined whether a non-human robot audience can elicit a stress response in human participants. A 90-min experimental laboratory session based on the Trier Social Stress Test (TSST) using a pre-recorded robot audience, was presented as a live on-screen simulation. Nineteen participants (female = 16) aged 21–۵۷ years (M = 29.74) underwent a 10-min mock interview and mathematics task in front of the robot audience. Salivary cortisol was assessed at 10-min before and immediately prior to the start of the stress test, and +10-, +30- and +40-min after the start of the test. Heart rate was assessed 20 min before, at 5 min into and 40-min after the test. Perceived stress and trait coping responses were provided at entry and participants were interviewed post task about their subjective experience. Significant increases in salivary cortisol and heart rate were observed over time with no significant interactions by participant subjective report. Coping responses including active coping and planning showed significant relationships with cortisol and heart rate reactivity and recovery. Until now, a non-human robot audience has not been used in a social stress testing paradigm. This methodology offers an innovative application with potential for further in-depth evaluation of stress reactivity and adaptation.
Introduction
It is uncontested that the Trier Social Stress Test (TSST) (Kirschbaum, Pirke, & Hellhammer, 1993) offers a well validated and established acute stress experimental paradigm. A recent review cites it as the “gold standard in human experimental stress research” having been applied across a range of settings, different populations and age groups (Allen et al., 2017, p. 115). Various adaptations of this test include those that have changed the setting or other relevant parameters including the TSST-G designed for group testing (von Dawans, Kirschbaum, & Heinrichs, 2011), the inclusion of a placebo (Het, Rohleder, Schoofs, Kirschbaum, & Wolf, 2009) or friendly version (Het et al., 2009) as a control comparison and a growing number of virtual reality (VR) versions (Jonsson et al., 2010; Kotlyar et al., 2008; Shiban et al., 2016; Wallergard, Jonsson, Osterberg, Johansson, & Karlson, 2011; Kothgassner et al., 2016). Throughout these adaptations the audience has typically involved a panel or audience of two or three adults, whether in-person (termed “in vivo” by some authors e.g. Shiban et al. (2016) as in the original TSST (Kirschbaum et al., 1993) or delivered via VR. Some exceptions have also experimented with video delivery using a pre-recorded audience such as in the Leiden public speaking task (Westenberg et al., 2009). The presence or absence of an in-person audience and the degree of human authenticity differentiates within social stress testing paradigms, highlighting the importance of a key stressor characteristic, that of social evaluative threat (SET). Described by Dickerson and Kemeny (2004) in the context of their social self-preservation theory, the selfpreservation system is attuned to detect esteem or status threats to the social self, initiating psychological and physiological responses to protect against these experiences.