سخنرانی عمومی در برابر یک مخاطب غیرواقعی، نیاز قدرت ضمنی و اثر تحریکی غدد درون ریز را افزایش می دهد

The present study explored the motivational characteristics of the Trier Social Stress Test (TSST; Kirschbaum, Pirke, & Hellhammer, 1993). Seventy-two participants either completed the public-speaking component of the TSST or, as a control condition, the friendly TSST (Wiemers, Schoofs, & Wolf, 2013) and wrote picture stories both before and after treatment. Stories were coded for motivational imagery related to power, achievement, and affiliation as well as for activity inhibition, a marker of functional brain lateralization during stress. The TSST had a specific arousing effect on power motivation, but not on other motivational needs, on activity inhibition, or on story length. TSST-elicited increases in power imagery, but not in achievement or affiliation imagery, were associated with a relatively greater salivary alpha-amylase response and with a relatively lesser salivary cortisol response. These findings suggest that the TSST specifically induces power-related stress.

Since its introduction more than 20 years ago (Kirschbaum et al., 1993), the Trier Social Stress Test (TSST) has become the gold standard in human stress research, with hundreds of published studies using this procedure as well as several meta-analyses and in-depth reviews about its properties and effects (e.g., Allen et al., 2014, Campbell and Ehlert, 2012 and Foley and Kirschbaum, 2010). During the original version of the TSST, research participants present, after a brief preparation period, a job talk in front of an unreceptive audience of two examiners and are then required to perform a subtraction task. Once they make a mistake, examiners ask them to start over again. This procedure, as well as a variant of the TSST that omits the subtraction task (Wiemers et al., 2013), elicits robust and reliable activation of the hypothalamic-pituitary-adrenal (HPA) stress axis, as reflected in a transient steep increase of cortisol and the adenocorticotropic hormone (ACTH) during and immediately after the TSST (Kirschbaum et al., 1993). It also elicits robust activation of the sympathetic nervous system (SNS), as indexed by transient increases in heart rate and the sympathetic catecholamines adrenaline and noradrenaline (Schommer et al., 2003) as well as in salivary alpha amylase, a biomarker of the noradrenergic component of SNS activation (Ditzen et al., 2014, Kuebler et al., 2014, Rohleder and Nater, 2009 and Wiemers et al., 2013). From a motivation science perspective, it is clear that the TSST induces a strong, aversive motivational state. However, because different types of stressors impact different motivational systems (e.g., food deprivation for energy balance; social isolation for affiliation; see also Kudielka et al., 2009 and Stroud et al., 2002), the question is what type of motivational need is challenged by the TSST. In the present research, we address this issue by examining motivational changes induced by the TSST and how they relate to endocrine changes. In so doing, we used measurement methods developed and extensively validated in the context of research on implicit motives. The implicit motive approach to human motivation is based on the assumption that people are characterized by a handful of universal motivational needs (McClelland, 1987 and Schultheiss, 2008). The most frequently studied motives are the need for power (frequently abbreviated as n Power), a concern with having impact on others; the need for achievement (n Achievement), a concern with mastering challenging tasks; and the need for affiliation (n Affiliation), a concern with establishing, maintaining, and restoring friendly relationships with others (McClelland, 1987 and Schultheiss, 2008). Measures for these needs were originally developed by experimentally arousing a given need and then studying how the content of fantasy stories changes that research participants write about pictures with ambiguous social cues (Winter, 1998). For instance, in the case of n Power, researchers examined the stories of individuals who were running for office versus those who were not (Veroff, 1957), of individuals who knew how to cheat on a card game versus those who did not (Uleman, 1972), or of individuals who listened to inspirational speeches versus of those who listened to travel descriptions (Winter, 1973). Across studies, individuals whose need for power had been aroused in these ways, but not control-condition participants, showed a similar tendency to infuse their stories with imagery related to strong, forceful action, control or regulation of others, persuasion and arguing, or impressing others (Winter, 1991). The resulting coding systems for n Power, and those for other motives derived in a similar manner, were thus sensitive to causal manipulations of motivational states (McClelland, 1958 and McClelland, 1987, chapter 6; see also Borsboom et al., 2004 Because they do not correlate substantially with self-report measures purported to assess the same motivational needs (see Köllner and Schultheiss, 2014, for meta-analytic results), picture-story measures of motives have been termed implicit by McClelland et al. (1989). Although the picture-story measurement approach was subsequently used primarily to assess stable individual differences in individuals' implicit motivational needs, its sensitivity to situational changes in motivation makes it an excellent tool for exploring which specific motivational needs are aroused by a given situational cue such as the TSST (see Schultheiss and Pang, 2007, p. 338 f.). This property of the Picture Story Exercise (PSE; McClelland et al., 1989), as the method has become known, has already been used successfully in psychoendocrinological research on the effects of movies on hormonal changes. Here, the PSE was used as a manipulation check to verify that movies intended to arouse power or affiliative concerns did, in fact, also result in the expected motivational changes (Schultheiss et al., 2004; see also Wirth and Schultheiss, 2006). So which motivational need should the TSST impact the most? We hypothesize that it is a specific stressor for n Power, because the mock job interview around which most of the TSST revolves requires a person to be persuasive and convincing, to impress others—in short: to have an impact on other people. This is the core incentive for n Power, but not for other motivational needs. If our reasoning is correct, then the TSST should lead to a specific increase in power-related imagery on the PSE, but not in other types of motivational imagery (Hypothesis 1). Some supportive evidence comes from a study by Fodor and Wick (2009), who had research participants give an impromptu speech in front of two judges acting in a negative manner. Participants with a strong dispositional n Power, measured before the task, showed greater activation of the corrugator muscle and also reported higher levels of anxiety than participants low in n Power. This difference did not emerge in a control condition in which the audience was supportive and friendly. Other supporting evidence was reported by McClelland et al. (1985), who observed that highly power-motivated individuals, but not other participants, responded with an increase in salivary noradrenaline to an exam, that is, to a situation in which an individual is subject to others' critical evaluation. Although these studies did not address whether a public-evaluation challenge actually increases power motivation in a transient manner, it is consistent with our reasoning that a situation akin to the TSST should be a relevant stressor specifically for n Power. If our hypothesis is correct, then TSST-induced changes in n Power should be associated with a specific hormonal signature of power arousal. Arousal of n Power has been linked to the release of noradrenaline (and sometimes also adrenaline) in early psychonedocrinological research by McClelland and colleagues (e.g., McClelland et al., 1985; for reviews, see McClelland, 1987 and McClelland, 1989). More recent research shows that dominance success is related to quick, transient increases in testosterone among men high in n Power, an effect that Schultheiss (2007) explained as follows, based on Sapolsky, 1985 and Sapolsky, 1986 earlier work on the interaction between stress hormones and gonadal steroid release: to the extent that a challenge activates a concern for power, it will elicit a stronger response from the SNS than from the HPA axis. In men, this results in a net increase of stimulatory action of catecholamines (relative to cortisol's inhibitory action) on the testes' Leydig cells and thus to the rapid testosterone increases observed in research on male power motivation. According to this account, power motivation arousal should lead to greater SNS activation and comparatively weaker HPA activation (although both can be activated to some extent). We thus expected variations in power motivation increases in response to the TSST to be associated with greater SNS activation and lesser HPA activation (Hypothesis 2). We tested these hypotheses in a study in which participants were either exposed to a variant of the TSST that featured the job interview task, but not the mental-arithmetic task (Wiemers et al., 2013), and thus represented a power-related incentive or to a control version of this task that explicitly lacked all power-related stressors, the friendly TSST (f-TSST; Wiemers et al, 2013). To assess changes in motivational states, we administered parallel forms of the PSE in a counterbalanced order before and after the treatment and later analyzed them for changes in motivational imagery related to power, achievement, and affiliation as well as for changes in activity inhibition, a linguistic marker of functional brain asymmetry (Schultheiss et al., 2009) that has been related to n Power and endocrine or physiological stress responses in past research (Fontana et al., 1987, McClelland, 1979 and Schultheiss and Rohde, 2002). Analyses for activity inhibition were exploratory. To measure activation of stress axes, we repeatedly sampled saliva before and after treatment and later determined levels of cortisol (HPA axis) and alpha amylase (SNS axis).

Results PSE scores We conducted repeated-measures ANOVAs for PSE motive and AI scores (square root-transformed due to deviations from a normal distribution) and word count, with time of PSE as within-subject factor (before or after stress or control condition) and condition (stress or control) as between-subject factor. We obtained a significant time × condition effect for n Power, F(1, 70) = 4.91, partial η2 = .066, p = .03. Planned comparisons showed that there were no differences between the stress and control group in n Power before the stress or control procedure (see Table 1). Afterwards, however, participants showed higher n Power after the stress condition than after the control condition, and participants in the stress condition moreover showed a significant increase in their n Power levels that was not in evidence for control-group participants (see Fig. 1). Additional analyses did not reveal conclusive evidence for a differential impact of TSST stress on the 6 specific coding categories for n Power. Repeated-measures ANOVAs for n Achievement, n Affiliation, or AI did not yield significant time × condition effects, all Fs < 0.71, all partial η2s < .010, all ps > .41. Importantly, experimental condition also had no significant effect on the length of the PSE stories, that is, the matrix in which motivational imagery was assessed (for the condition × time effect, F[1, 70] = 0.39, partial η2 = .006, p = .53). None of these findings was significantly moderated by gender. Table 1. Mean (SD) values for PSE raw motive, AI, and word count scores and within-group (df = 36 for TSST and 34 for control) and between-group (df = 70) difference significance tests. Pre Post t d p n Powera Stress 2.41 (2.40) 3.35 (2.20) 2.26 0.49 .03 Control 2.54 (1.67) 2.29 (2.41) − 0.82 − 0.17 .42 t − 0.48 2.35 d − 0.11 0.55 p .63 .02 n Achievementa Stress 4.32 (2.58) 4.11 (2.60) − 0.39 − 0.09 .70 Control 4.57 (3.12) 4.20 (2.81) − 0.50 − 0.11 .62 t − 0.20 − 0.07 d − 0.05 − 0.02 p .84 .94 n Affiliationa Stress 0.81 (1.17) 0.81 (1.35) − 0.17 − 0.04 .86 Control 0.83 (1.01) 0.51 (0.52) − 1.48 − 0.35 .15 t − 0.19 0.92 d − 0.04 0.22 p .85 .36 Activity inhibitiona Stress 2.81 (2.37) 2.51 (2.10) − 0.69 − 0.12 .50 Control 2.23 (1.48) 2.37 (2.07) 0.06 0.01 .95 t 0.93 0.30 d 0.22 0.07 p .36 .77 Word count Stress 314 (89) 304 (105) − 1.15 − 0.10 .26 Control 298 (93) 296 (108) − 0.27 − 0.02 .79 t 0.75 0.33 d 0.18 0.08 p .46 .74 a t-tests and effect size estimates are based on square-root-transformed variables. Table options Full-size image (21 K) Fig. 1. n Power imagery scores (square-root-transformed, ± SEM) before and after the stress (TSST) or the control condition (f-TSST). Figure options Cortisol Since cortisol data were not normally distributed, all data were subjected to a log-transformation after adding a constant of 1. Experimental condition had an influence on cortisol concentrations. The TSST resulted in an increase of cortisol concentration in participants while the f-TSST did not. This was reflected in the results of a repeated-measure analysis of variance (ANOVA) conducted with time of measurement (baseline, + 1, + 15, + 30) as within-subject variable and condition (stress vs. control) as between-subject variable. Results show a significant time × condition interaction effect, F(3, 171) = 24.53, partial η2 = .301, p < .0000005. Follow-up t tests show that the stress group shows significantly higher cortisol concentrations than the control group in the measurements 15 min and 30 min after the end of the stressor (see Table 2). Table 2. Mean (SD) raw values for salivary alpha-amylase and cortisol. Baseline 1 min d 15 min d 30 min d Amylase (U/ml) Stress 69.30 (49.61) 116.39 (77.94) *** 0.65 81.10 (69.92) 0.04 65.51 (47.42) − 0.13 Control 83.09 (79.85) 134.66 (116.28) 0.33 83.78 (92.39) − 0.12 70.02 (79.56) * − 0.29 t(df) − 0.45 (63) 0.30 (62) 0.25 (61) 0.32 (61) d − 0.11 0.08 0.06 0.08 p .66 .76 .80 .75 Cortisol (nmol/l) Stress 7.92 (4.75) 10.64 (5.97) *** 0.58 14.92 (10.27) *** 0.98 11.56 (7.07) *** 0.66 Control 9.19 (4.42) 8.69 (4.70) − 0.13 8.28 (4.11) − 0.21 6.56 (2.87) *** − 0.69 t(df) − 1.43 (66) 1.49 (62) 3.53 (63) 3.97 (63) d − 0.35 0.37 0.88 0.99 p .16 .14 .0008 .0002 Note. Asterisks (***p < .005, *p < .05) and horizontal ds denote differences relative to baseline. Degrees of freedom vary due to missing data. All t tests and effect size calculations were performed on log-transformed variables. Table options Alpha amylase Since amylase data were not normally distributed, all data were subjected to a log-transformation after adding a constant of 1. A repeated-measures ANOVA with time of measurement as within-subject variable and condition as between-subject variable revealed a significant main effect of time, F(3, 168) = 5.04, partial η2 = .082, p = .002, which was mainly due to participants showing an overall amylase increase 1 min after completing the TSST or the f-TSST, t(61) = 4.15, d = 0.44, p = .0001, but not 15 min later, t(61) = − 1.06, d = − 0.12, p = .30. Thirty minutes later, amylase levels were significantly lower than at baseline, t(60) = − 2.74, d = − 0.23, p = .008. The condition × time effect failed to reach significance, F(3, 168) = 1.93, partial η2 = .033, p = .13 (see Table 2). Change correlation analyses To test whether changes in n Power are differentially associated with changes in cortisol and amylase, we first created average scores across all three post-treatment assessments (log-transformed values) for each hormonal parameter and then residualized it for its respective log-transformed baseline (in this and all subsequent analyses reported in this section, we dropped one participant from the control condition whose post-treatment amylase residual score was identified as an outlier, studentized residual t = − 5.76). This yielded residualized change scores for cortisol and amylase. In the stress condition, these change scores correlated at r(32) = .47, p = .007, and in the control condition, they correlated at r(31) = .24, p = .18, suggesting somewhat tighter functional coupling of stress axes in the former condition than in the latter (for the difference, Z = 0.98, p = .33). We then entered these scores, along with square-root-transformed n Power at T1, into regressions with square-root-transformed n Power at T2 as dependent variable. We thus tested whether treatment-induced changes in hormones that were independent of initial hormone levels predicted treatment-induced changes in n Power, above and beyond differences in initial n Power. As shown in Table 3, a greater increase in amylase and a lesser increase in cortisol were both associated with an overall increase in n Power among participants in the TSST stress condition, accounting for a significant overall variance increment in n Power scores. This effect did not emerge in the f-TSST control condition. Table 3. Testing for effects of amylase and cortisol (residualized change scores) on n Power at T2 (transformed scores) in stress (TSST) and control (f-TSST) conditions. Stress Control B SE Semipartial r t p B SE Semipartial r t p n Power T1 − 0.068 0.152 − .07 − 0.45 .66 0.263 0.238 .21 1.11 .28 Amylase 0.517 0.209 .41 2.47 .02 − 0.013 0.205 − .01 − 0.06 .95 Cortisol − 0.577 0.222 − .43 − 2.59 .02 − 0.288 0.434 − .12 − 0.66 .51 R2 = .240, F(3, 28) = 2.94, p = .05 R2 = .061, F(3, 27) = 0.58, p = .63 Table options When we repeated the analyses reported in Table 3, but included either achievement or affiliation scores instead of power scores, cortisol and amylase change scores failed to predict changes in these motives in the stress condition, ps > .17. The same was also true for the control condition, ps > .32, with the exception of a specific positive association between cortisol changes on affiliation changes, B = 0.55, SE = 0.22, semipartial r = .17, t(27) = 2.55, p = .02.