Objective and perceived arousal during performance of tasks with elements of social threat: The influence of anxiety sensitivity

J. Behav. Ther. & Exp. Psychiat. 43 (2012) 967e974

Contents lists available at SciVerse ScienceDirect

Journal of Behavior Therapy and Experimental Psychiatry journal homepage: www.elsevier.com/locate/jbtep

Objective and perceived arousal during performance of tasks with elements of social threat: The influence of anxiety sensitivity Michel A. Thibodeau, Lydia Gómez-Pérez, Gordon J.G. Asmundson* Department of Psychology, University of Regina, 3737 Wascana Parkway, Regina, Saskatchewan S4S 0A2, Canada

a r t i c l e i n f o

a b s t r a c t

Article history: Received 10 November 2011 Received in revised form 27 February 2012 Accepted 7 March 2012

Background and objectives: Physiological arousal serves to maintain social anxiety disorder by skewing self-perception. Anxiety sensitivity is associated with the disorder and exaggerated perceptions of arousal, but has not been tested as explaining exaggerated perceived arousal in social contexts. The aim of this investigation was to address this issue. Methods: A total of 42 individuals participated in three tasks associated with potential social threat (i.e., a speech, typing task, hyperventilation) and completed measures of trait social anxiety and anxiety sensitivity. State anxiety, perceived arousal, and objective arousal were assessed during each task. Results: Trait social anxiety and anxiety sensitivity were correlated with state anxiety and perceived arousal, but not objective arousal, during the tasks. Anxiety sensitivity mediated the relationships between trait social anxiety and perceived arousal and between trait social anxiety and state anxiety for the typing and hyperventilation tasks. Limitations: Although the sample likely included a number of individuals with social anxiety disorder, the sample was mostly comprised of individuals without a diagnosis. The current results can be extended to clinical presentations to some extent, but future research is needed to further explore the demonstrated relationships in samples of individuals with social anxiety disorder. Conclusions: Anxiety sensitivity may play a crucial role in perceptions of arousal and state anxiety in the context of potential social threats, warranting attention from researchers and clinicians focussing on social anxiety disorder. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Anxiety sensitivity Social anxiety Physiological arousal Perceived arousal

1. Introduction Anxiety about social interactions is a ubiquitous and adaptive experience; however, when social anxiety reaches such a degree that it interferes with functioning, it may warrant a diagnosis of social anxiety disorder (American Psychiatric Association, 2000). Social anxiety disorder is a highly prevalent and disabling condition (Kessler et al., 2008; Stein et al., 2005). Cognitive-behavioural models posit that arousal (e.g., shortness of breath, sweating) serves to maintain social anxiety disorder by influencing perceptions of the self and expectations of social failure (Clark & Wells, 1995; Hofmann, 2007; Rapee & Heimberg, 1997). For example, a man in a social encounter may believe that he is perspiring profusely and that others can see his perspiration and must think he is socially inept. This perception of arousal would contribute to

* Corresponding author. Tel.: þ1 306 347 2415; fax: þ1 306 337 3275. E-mail addresses: [email protected] (M.A. Thibodeau), [email protected] (G.J.G. Asmundson). 0005-7916/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbtep.2012.03.001

perceptions of the self (e.g., “I must look like an idiot”) and serve to exacerbate symptoms of anxiety and associated consequences (e.g., avoidance). In line with this example and with contemporary theory, features that impact arousal in social situations would reasonably contribute to clinical levels of social anxiety. In recognizing the role of arousal, researchers have attempted to test if trait social anxiety (i.e., a long lasting propensity to experience social anxiety) is associated with exaggerated arousal in social contexts. Empirical findings have been incongruent, with some investigations suggesting that trait social anxiety is associated with greater objective (i.e., as measured by equipment) arousal during social interactions, and other investigations suggesting there is no such relationship (for a review, see Mauss, Wilhelm, & Groos, 2004). In contrast, investigations are consistent in showing that those with greater trait social anxiety perceive their arousal as greater (Anderson & Hope, 2009; Edelmann & Baker, 2002; Mauss et al., 2004; Mulkens, de Jong, Dobbelaar, & Bogels, 1999). Skewed perceptions of arousal are believed to also be crucial in the maintenance of other anxiety disorders, most notably panic disorder (Reiss, Peterson, Gursky, & McNally, 1986; Taylor, 1999).

968

M.A. Thibodeau et al. / J. Behav. Ther. & Exp. Psychiat. 43 (2012) 967e974

Particular attention has focused on the construct of anxiety sensitivity, which is defined as a trait-like propensity to fear anxietyrelated sensations (including arousal) based on the belief that they have harmful consequences (Reiss, 1991; Reiss & McNally, 1985). Anxiety sensitivity is believed to be a diathesis and maintenance factor for numerous anxiety-related disorders by skewing the interpretation of symptoms of anxiety (Naragon-Gainey, 2010; Reiss et al., 1986; Schmidt, Lerew, & Jackson, 1997). Specifically, anxiety sensitivity is thought to amplify anxiety responses by contributing to a self-perpetuating cycle in which symptoms of anxiety contribute to anxiety itself (Reiss et al., 1986). As with trait social anxiety, some research suggests that anxiety sensitivity is associated with exaggerated objective arousal (Conrod, 2006), whereas others research suggests that it is not (Asmundson, Norton, Wilson, & Sandler, 1994; Leen-Feldner, Feldner, Bernstein, McCormick, & Zvolensky, 2005; Stewart, Buffett-Jerrott, & Kokaram, 2001). Notwithstanding, similar to those with greater trait social anxiety, individuals with greater anxiety sensitivity have been consistently found to have exaggerated perceptions of arousal (Asmundson et al., 1994; Carter, Sbrocco, & Ayati, 2009; LeenFeldner et al., 2005; Stewart et al., 2001). Anxiety sensitivity has been empirically associated with trait social anxiety in clinical and non-clinical samples (Anderson & Hope, 2009; Carleton, Abrams, Asmundson, Antony, & McCabe, 2009; Moore, Chung, Peterson, Katzman, & Vermani, 2009), and both constructs have been associated with heightened perceptions of arousal. Therefore, it may be possible that anxiety sensitivity explains, at least in part, why individuals with greater trait social anxiety experience exaggerated perceptions of arousal and exacerbated anxiety when facing potential social threats. This has yet to be empirically tested and associated results could have implications for cognitive-behavioural models of social anxiety disorder and associated evidence-based treatments. The aim of the current investigation was to explore the relationships between trait social anxiety, anxiety sensitivity, state anxiety, and perceived and objective arousal. We hypothesized that (1) trait social anxiety and anxiety sensitivity would be significantly positively correlated and each of these constructs would be associated with greater state anxiety and perceived arousal, but not objective arousal; (2) anxiety sensitivity would mediate the relationship between trait social anxiety and perceived arousal; and (3) anxiety sensitivity would mediate the relationship between trait social anxiety and state anxiety. Support for these hypotheses would further improve our conceptual understanding of social anxiety disorder and related evidence-based treatments. 2. Materials and methods 2.1. Participants Ethical approval for this investigation was obtained from the University Research Ethics Board and all participants provided informed consent prior to participating. Exclusion criteria included: (1) medical conditions suspected to significantly influence physiological arousal patterns or that contraindicate hyperventilation (e.g., intracranial hypertension, recent subarachnoid haemorrhage, sickle cell disease, recent cardiac illness, myocardial infraction or stroke), (2) currently participating in treatment for an anxiety disorder (as treatments may include modules focussing on the interpretation of physiological arousal), (3) reported changes in psychotropic medication use in the past three months, and (4) current or past diagnosis of panic disorder or experiencing a panic attack in the previous twelve months. Participants were recruited through a university-wide email announcement that outlined the need for individuals to participate in a study exploring the

relationships between bodily arousal (e.g., heart rate, breathing), anxiety, and performance on tasks. A total of 45 individuals demonstrated interest in participating in the investigation; however, three were excluded after self-reporting medical conditions thought to be associated with differential physiological arousal or that contraindicated hyperventilation (i.e., Parkinson’s Disease, sickle cell disease, painful breathing). The final sample comprised 42 individuals (30 women, 12 men; age 18e60; Mage ¼ 24.90, SD ¼ 9.15). The majority of participants (n ¼ 32) selfreported being Caucasian, while other self-reported ethnicities included African (n ¼ 5), Asian (n ¼ 3), and mixed East Indian and Caucasian (n ¼ 2). The majority of participants endorsed being current full-time (n ¼ 32) or part-time students (n ¼ 7), and a majority endorsed being employed, either part (n ¼ 20) or fulltime (n ¼ 4). 2.2. Measures 2.2.1. Social Interaction Phobia Scale (SIPS; Carleton, Collimore, et al., 2009) The SIPS is a self-report questionnaire which includes 14-items designed to measure fear associated with social interaction and performance situations (e.g., “When mixing socially I am uncomfortable” and “I would get tense if I had to carry a tray across a crowded cafeteria”, respectively). The SIPS was derived from factor analyses of the Social Phobia Scale and the Social Interaction Anxiety Scale (Mattick & Clarke, 1998), but is much shorter and maintains acceptable psychometric properties (Carleton, Collimore, et al., 2009). Items are rated on five-point Likert scales ranging from 0 (“not at all characteristic or true of me”) to 4 (“extremely characteristic or true of me”). The SIPS has demonstrated acceptable reliability in past studies (Carleton, Peluso, Collimore, & Asmundson, 2011; Reilly, Carleton, & Weeks, 2011) as well as factorial validity and validity in discriminating between individuals with and without social anxiety disorder (Carleton, Collimore, et al., 2009; Reilly et al., 2011). 2.2.2. Anxiety Sensitivity Index ASI-3 (ASI-3; Taylor et al., 2007) The ASI-3 is an 18-item self-report questionnaire designed to measure anxiety sensitivity. Items are focused on statements relating to anxiety symptoms (e.g., “When I feel pain in my chest, I worry that I’m going to have a heart attack”), and are rated on fivepoint Likert scales ranging from 0 (“agree very little”) to 4 (“agree very much”). The ASI-3 has demonstrated reliability and validity in a number of samples across numerous countries (Taylor et al., 2007). 2.2.3. Measure of perceived arousal Participants were asked to rate the extent to which they experienced racing heart, sweaty palms, and shortness of breath during each of the periods (described below) by drawing an “X” on a 20-cm line ranging from 0 (“not at all”) to 100 (“extreme”). This measure has been demonstrated as an effective index of perceived arousal in previous investigations utilizing laboratory social threat tasks (Mauss et al., 2004). 2.2.4. Measure of state anxiety Participants were asked to rate the peak of the anxiety they experienced during each of the periods by drawing an X on a 20-cm line ranging from 1 (“not at all”) to 100 (“extremely”). Single-item visual analogue scales have been demonstrated as good measures of state anxiety and as correlating highly with other measures of state anxiety (Davey, Barratt, Butow, & Deeks, 2007). Furthermore, the visual analogue scale allows the direct assessment of state anxiety while not measuring other aspects of anxiety such as arousal.

M.A. Thibodeau et al. / J. Behav. Ther. & Exp. Psychiat. 43 (2012) 967e974

2.2.5. Measures of objective arousal Objective arousal data were recorded using the BIOPAC MP 150 Data Acquisition System (MP 150 Data Acquisition System, BIOPAC Systems Inc., Goleta, CA) calibrated using a global sampling rate of 500 samples per second. Indices of objective arousal included heart rate, skin conductance levels, and breathing rate. Heart rate was recorded using a finger pulse oximeter from BIOPAC (OXY100C). The oximeter was attached to the index finger of the non-dominant hand of participants. Skin conductance levels were recorded using a galvanic skin response amplifier from BIOPAC (GSR100C) and disposable pre-gelled Ag/AgCL electrodes (Vermed Inc., Bellow Falls, VT). The electrodes were placed onto the palmar surface of the middle phalanges of the second and third fingers of the nondominant hand (Fowles et al., 1981). Breathing rate was recorded using a respiratory amplifier from BIOPAC (RSP100C) with a transducer band wrapped around the body and centred on the sternum.1 2.3. Laboratory procedure Participants were invited to a single individual session that lasted approximately two hours. Two experimenters, one of each sex, greeted participants. Participants were informed that they would be performing a number of tasks and that their bodily reactions during these tasks would be measured. Participants provided informed consent and completed a battery of questionnaires on a computer in a waiting room prior to participating in the social threat tasks. Subsequently, participants were seated in a room regulated at 22.5  C, in which they faced a 19-inch computer monitor, a visible digital computer camera, as well as a small nonfunctional apparatus (12 by 20 by 5 cm) with a large microphone protruding from it. Physiological monitoring equipment was attached to participants by the same-sex experimenter as soon as they were seated. White noise was generated and maintained in this room at 65 decibels during the entire laboratory procedure to minimize potential auditory distractions outside of the room. Participants were asked to sit quietly and wait for 10 min to acclimatize to the laboratory environment. The last 5 min of this period were used as a baseline measurement. Following the acclimatization period, participants were asked to participate in three social threat tasksda speech task, a typing task, and a hyperventilation taskddescribed in detail below. The speech task and typing task were randomized so that each participant had an equal chance of giving the speech or of typing first. The hyperventilation task was always performed last, as it was expected to be associated with significantly greater levels of objective arousal and could have had the strongest carry-over effects. In addition, a 3-min rest period followed each task to minimize carry-over effects. Immediately following baseline and after completing each task, participants were asked to retrospectively rate the perceived arousal and state anxiety they experienced during the preceding period. 2.3.1. Speech task Speech task paradigms have been used in similar investigations and have been demonstrated to successfully elicit state anxiety and increased arousal (Feldman, Cohen, Hamrick, & Lepore, 2004; Gerlach, Wilhelm, Gruber, & Roth, 2001). The speech task adapted aspects of public speaking (i.e., speaking to an audience), but without the use of a live audience, to control for potential confounds and limitations in previous investigations (e.g.,

1 Participants also reported their perceived competence as typists on a Likert scale. Initial analyses demonstrated that perceived typing competency was not associated with any other variables of interest; consequently, methods and results associated with this measure are excluded for brevity.

969

inconsistent audience reactions (Mauss et al., 2004; Westenberg et al., 2009)). Rather than a live audience, a video camera was used to observe and record participants. Participants were notified that they had 5 min to prepare a speech that was to last 7 min, on one of three preselected topics (i.e., an autobiography, the events of their favourite vacations, their best childhood memories). They were told that their speech would be recorded and observed by three individuals in the adjacent room through a video camera, and that their performance would later be rated by a panel of judges consisting of a professor of psychology and three doctoral students. Participants were also told that the panel of judges would rate their speech on organization, vocal projection, and speaking ability for their age and academic level. Additionally, features of the Trier Social Stress Test were included to elicit state anxiety (Kirschbaum, Pirke, & Hellhammer, 1993); specifically, participants were informed that the judges were trained to monitor non-verbal behaviours and that the apparatus with the microphone was part of a voice frequency and stress analysis system This system was said to analyze the tone and pitch of their voice to identify moments of anxiety and various subtleties in their speaking patterns. After participants completed preparation of their speech, the digital camera was activated and participants were left alone in the experimental room and observed from an adjacent control room. The opposite-sex experimenter asked participants to begin the speech through a microphone by stating “please start the speech”. Participants were asked to stop after 5 min, despite the previous instruction to prepare and perform a 7-min speech. The 2-min discrepancy allowed participants to prepare more material than necessary, and increased the likelihood that they would speak for the entire 5-min period. If participants stopped speaking or appeared unsure how to proceed prior to the end of the 5-min period, they were instructed through the microphone that they may speak on another of the two assigned topics. All participants completed the entirety of the 5-min speech and only three participants required a cue to keep speaking before the end of the speech period. 2.3.2. Typing task Simple performance tasks are frequently employed by researchers in order to increase state anxiety and arousal (Shostak & Peterson, 1990; Stewart et al., 2001). Participants were asked to transcribe a neutral passage read by an experimenter as quickly as possible by means of typing with only their non-dominant hand, for a total of 5 min. Participants were instructed to make as few mistakes as possible and were told their performance would be compared to that of other participants. The social threat component of the task was increased by having the experimenter stand directly beside the participant during typing, and by informing participants that their typing performance was simultaneously being observed by experimenters on an analogous monitor in the adjacent room. The same male experimenter read to all participants to ensure a consisting reading voice and pattern. 2.3.3. Hyperventilation task The hyperventilation task was utilized to elicit state anxiety while allowing examination of the relationships of interest in the context of relatively extreme objective arousal (Carter et al., 2009). Previous findings suggest that laboratory-induced hyperventilation is associated with social anxiety in non-clinical participants when observed by an experimenter, presumably because the process of hyperventilation can be perceived as relatively embarrassing (Carter et al., 2009). The experimenter stood directly next to the participants and observed their breathing patterns, prompting participants to maintain a proper pace and depth of breaths, if necessary. Participants were asked by the experimenter to breathe through their

970

M.A. Thibodeau et al. / J. Behav. Ther. & Exp. Psychiat. 43 (2012) 967e974

mouth as deeply as possible approximately every 2 s (i.e., complete one inhalation and one exhalation) for 60 s. Participants were allowed to stop if they were unable to finish, although, only one participant ceased hyperventilation prior to the full minute. 2.4. Statistical analyses Mean heart rate, skin conductance levels, and breathing rate were calculated for the baseline period and for each of the social threat tasks. Descriptive statistics were conducted to characterize the sample. Repeated measures analyses of variance followed by pairwise comparisons were conducted in order to test the efficacy of the threat tasks in eliciting state anxiety and arousal. Specifically, state anxiety and perceived arousal scores reported immediately after each of the three tasks and objective arousal recorded during the tasks were compared to baseline levels of these measures. Correlational analyses were conducted between perceived (i.e., racing heart, sweaty palms, and shortness of breath) and objective indices of arousal (i.e., heart rate, skin conductance, breathing rate), state anxiety elicited during each of the tasks, and ASI-3 and SIPS scores to test hypothesis 1. To reduce the overall number of analyses and the likelihood of Type 1 errors produced from testing hypotheses 2 and 3, the indices of perceived arousal (i.e., racing heart, sweaty palms, and shortness of breath) were summed to create a factor representing perceived arousal reported during each of the three tasks. Bootstrapped confidence intervals were utilized to test hypotheses 2 and 3 (i.e., the role of ASI-3 scores as a mediator). Bootstrapping is a computationally intensive method that estimates relationships by repeatedly sampling from the observed relationships within the data. By repeating this process numerous times, an empirical distribution can be approximated and can be used to construct confidence intervals for effects within the data, including indirect effects. When confidence intervals for an effect do not include 0, this effect is considered as statistically significant, based on the prescribed limits of the confidence interval (e.g., 95%). Bootstrapping was performed using a macro developed by Preacher and Hayes (2008), which generated 5000 bootstrap samples to calculate confidence intervals (95%, bias-corrected) for the direct effects of SIPS and ASI-3 scores on the dependent variables (i.e., perceived arousal and state anxiety), and indirect effects of SIPS scores on the dependent variables through ASI-3 scores (i.e., the mediator; Hayes, 2009; Mackinnon, Lockwood, & Williams, 2004; Shrout & Bolger, 2002)2. Statistical p values for the correlational analyses and the test of direct effects were calculated as one-tailed, reflecting the direction of the hypotheses and the relationships reported in previous empirical literature. 3. Results Descriptive statistics for the primary variables of interest are presented in Table 1. Total ASI-3 and SIPS scores were not statistically significantly higher than scores previously reported in nonclinical Canadian samples (Carleton, Collimore, et al., 2009; Taylor et al., 2007). Repeated measures analyses of variance demonstrated that each of state anxiety, perceived arousal, and objective arousal were statistically significantly different across the different periods (all ps < .001, h2p ¼ .38e.79). Pairwise comparisons demonstrated that participants reported significantly greater state anxiety scores, greater perceived arousal scores for each index, and greater heart rate and skin conductance during each of the tasks

2 Sobel’s test (Sobel, 1982) was also calculated for each indirect effect and results were convergent with the bootstrap results; however, these findings are not reported for brevity.

Table 1 Descriptive statistics of psychological and physiological variables. Psychological and physiological variables (n ¼ 42)

Mean

SD

2e43 1e38

16.12 12.02

10.33 10.33

Baseline Perceived heart rate Perceived sweatiness Perceived shortness of breath State anxiety Heart beats per minute Skin conductance (microsiemens) Breaths per minute

0e71.50 0e72.5 0e74.5 0e68.5 47.65e91.76 0.44e19.10 10.01e19.62

18.10 8.86 14.36 19.08 68.72 5.44 15.30

19.24 15.87 17.96 18.45 8.35 5.03 2.10

Speech task Perceived heart rate Perceived sweatiness Perceived shortness of breath State anxiety Heart beats per minute Skin conductance (microsiemens) Breaths per minute

0e93.50 0e98 0e95.50 0e89 55.85e110.54 1.21e45.96 6e20

47.26 24.73 33.29 38.19 84.62 19.11 10.45

24.92 25.42 25.68 22.42 11.35 8.98 2.38

Typing task Perceived heart rate Perceived sweatiness Perceived shortness of breath State anxiety Heart beats per minute Skin conductance (microsiemens) Breaths per minute

0.1e95 0e78.50 0e69 0e83 61.6e103.09 1.26e30.71 7e19

42.04 15.45 24.52 35.13 81.84 17.80 14.70

23.22 19.39 20.03 22.48 9.53 7.48 2.19

Hyperventilation task Perceived heart rate Perceived sweatiness Perceived shortness of breath State anxiety Heart beats per minute Skin conductance (microsiemens) Breaths per minute

4.3e100 0e85 0e95 0.5e100 61.52e122.08 1.15e48.94 21e47

64.94 31.76 51.50 47.63 95.74 22.58 30.94

20.02 27.02 27.21 23.86 11.60 10.21 5.09

ASI-3a SIPSb

a b

Range

ASI-3 e Anxiety Sensitivity Index-3. SIPS e Social Interaction Phobia Scale.

compared to baseline (all ps < .001). Breathing rates were only significantly greater during the hyperventilation task compared to baseline (p < .001). Together these findings suggest that the social threat tasks successfully elicited state anxiety and perceived and objective arousal in participants, and that breathing rates do not appear to be robustly associated with anxiety in these tasks. SIPS and ASI-3 scores were significantly positively correlated. They were also positively correlated with state anxiety and the majority of perceived arousal indices (i.e., racing heart, sweaty palms, and shortness of breath) from each of the tasks (Table 2), providing support for hypothesis 1. Exceptions were that neither SIPS nor ASI-3 scores were significantly correlated with perceived sweatiness during the speech task, and SIPS scores were not significantly correlated with perceived shortness of breath during the hyperventilation task. Every index of perceived arousal was significantly positively correlated with state anxiety during each of the tasks. SIPS scores were significantly positively correlated with skin conductance levels during the typing task, but not with any other index of objective arousal. ASI-3 scores were not significantly correlated with any index of objective arousal. Bootstrapped estimations of the direct effects within the mediation models are reported in Table 3. Bootstrapped estimations of the indirect (mediated) effects of SIPS scores through ASI-3 scores in predicting the perceived arousal factors (hypothesis 2) and state anxiety (hypothesis 3) are reported in Table 4. The

M.A. Thibodeau et al. / J. Behav. Ther. & Exp. Psychiat. 43 (2012) 967e974

971

Table 2 Correlations between psychological variables. SIPSa

2

.73**

e

Perceived arousal speech 3. Heart rate 4. Sweatiness 5. Shortness of breath

.43** .21 .44**

Perceived arousal typing 6. Heart rate 7. Sweatiness 8. Shortness of breath

3

4

5

.28* .22 .36*

e .43** .71**

e .34*

e

.42** .35* .31*

.36** .45** .49**

.58** .39** .26*

.42** .70** .28*

Perceived arousal hypervent 9. Heart rate 10. Sweatiness 11. Shortness of breath

.30* .27* .09

.37** .34* .33*

.56** .30* .30*

12. State anxiety speech 13. State anxiety typing 14. State anxiety hypervent

.35* .50** .31*

.43** .60** .47**

.64** .43** .38**

2. ASI-3

b

6

7

8

9

10

11

.57** .39** .56**

e .55** .59**

e .50**

e

.27* .64** .40**

.52** .37** .15

.41** .26* .42**

.30* .69** .49**

.36** .41** .30*

e .41** .33*

e .36**

e

.44** .34* .34*

.48** .54** .45**

.59** .74** .50**

.54** .58** .45**

.38** .65** .56**

.48** .40** .70**

.27* .43** .47**

.63** .50** .38**

12

13

e .62** .59**

e .64**

Notes: *p < .05 (one-tailed); **p < .01 (one-tailed). a SIPS e Social Interaction Phobia Scale. b ASI-3 e Anxiety Sensitivity Index-3; Hypervent e hyperventilation.

indirect effect of SIPS scores on the perceived arousal factor through ASI-3 scores was statistically significant for both the typing and the hyperventilation tasks, as supported by bootstrapping confidence intervals that did not contain the value of zero. For both the typing and hyperventilation tasks, only ASI-3 scores, and not SIPS scores, were directly associated with the perceived arousal factor. SIPS scores were not significantly indirectly associated with the perceived arousal factor for the speech task; however, SIPS scores did have a direct effect on the perceived arousal factor for this task. All significant relationships were positive. These findings provide partial support for hypothesis 2. Table 3 Direct effects within mediation models. Model and direct effects

b

SIPSa to ASI-3b

SE

t

0.73

0.11

6.67**

Perceived arousal e speech task SIPS to perceived arousal ASI-3 to perceived arousal

0.48 0.07

0.25 0.25

1.94* 0.30

Perceived arousal e typing task SIPS to perceived arousal ASI-3 to perceived arousal

0.13 0.45

0.21 0.21

0.63 2.14*

Perceived arousal e hyperventilation task SIPS to perceived arousal 0.11 ASI-3 to perceived arousal 0.59

0.24 0.24

0.47 2.51*

State anxiety e speech task SIPS to state anxiety ASI-3 to state anxiety

0.10 0.10

0.42 1.72*

State anxiety e typing task SIPS to state anxiety ASI-3 to state anxiety

0.04 0.16

0.06 0.24

0.09 0.09

0.69 2.78**

State anxiety e hyperventilation task SIPS to state anxiety 0.03 ASI-3 to state anxiety 0.25

0.10 0.10

0.27 2.48**

Notes: *p < .05; **p < .01. a SIPS e Social Interaction Phobia Scale. b ASI-3 e Anxiety Sensitivity Index-3.

Similar to perceived arousal, the indirect effect of SIPS scores on state anxiety through ASI-3 scores was statistically significant for both the typing and the hyperventilation tasks; however, the indirect effect did not generalize to the speech task. For each task only ASI-3 scores had a direct effect on state anxiety. All significant relationships were positive. These findings provide partial support for hypothesis 3. 4. Discussion The aim of the present investigation was to explore the relationships between trait social anxiety, anxiety sensitivity, state anxiety, and perceived and objective arousal in the context of three tasks associated with potential social threat. The first hypothesisdthat trait social anxiety and anxiety sensitivity would be positively correlated with perceived arousal and state anxiety, but not with objective arousaldwas largely supported. Although trait social anxiety and anxiety sensitivity were positively correlated with almost every index of perceived arousal for each task, these effects did not generalize to objective arousal. These results are congruent with previous literature associating trait social anxiety (Anderson & Hope, 2009; Edelmann & Baker, 2002; Mauss et al., 2004; Mulkens et al., 1999) and anxiety sensitivity

Table 4 Testing ASI-3 scores as mediating relationships between SIPS scores and dependent variables. Dependent variable

Bootstrap M

SE

Lower limit CI

Upper limit CI

Perceived arousal Speech Typing Hyperventilation

0.06 0.43 0.32

0.14 0.20 0.17

0.17 0.15 0.10

0.30 0.82 0.72

State anxiety Speech Typing Hyperventilation

0.12 0.16 0.17

0.07 0.07 0.08

0.005 0.09 0.07

0.24 0.31 0.32

Notes: CI e 95% confidence interval; CI limits that do not include zero (i.e., are statistically significant) are bolded.

972

M.A. Thibodeau et al. / J. Behav. Ther. & Exp. Psychiat. 43 (2012) 967e974

(Asmundson et al., 1994; Carter et al., 2009; Leen-Feldner et al., 2005; Stewart et al., 2001) with exaggerated perceptions of arousal, and with previous findings suggesting no relationship between these constructs and objective arousal (Asmundson et al., 1994; Leen-Feldner et al., 2005; Mauss et al., 2004). Moreover, in line with previous investigations (Mauss et al., 2004), ratings of state anxiety elicited across the three social threat tasks were consistently associated with perceptions of arousal during those tasks. The second hypothesisdthat anxiety sensitivity would mediate the relationship between trait social anxiety and perceived arousaldwas partially supported; anxiety sensitivity was shown to mediate this relationship for the typing and hyperventilation tasks, but not for the speech task. The third hypothesisdthat anxiety sensitivity would mediate the relationship between trait social anxiety and state anxiety elicited during three social tasksdwas also partially supported; anxiety sensitivity mediated this relationship for the typing and hyperventilation tasks, but not for the speech task. For both hypothesis 2 and 3, and in each case of mediation, trait social anxiety scores were not directly associated with perceptions of arousal or state anxiety, suggesting its relationship with these dependent variables may be largely or totally mediated by anxiety sensitivity. The current findings suggest that the previously established relationship between trait social anxiety and perceived arousal, and between trait social anxiety and state anxiety (Anderson & Hope, 2009; Edelmann & Baker, 2002; Mauss et al., 2004; Mulkens et al., 1999), may be explained by the salient contributions of anxiety sensitivity to these relationships. Individuals with greater trait social anxiety typically experience heightened anxiety sensitivity (Deacon & Abramowitz, 2006; Taylor, Koch, & McNally, 1992), and this fear of anxiety-related sensations may explain their exaggerated perceptions of arousal and state anxiety, which supports the role of anxiety sensitivity as a general amplifier of fear and anxiety (Reiss, 1991). Despite these possibilities, the role of anxiety sensitivity as a potential mediator was not as salient during the speech task, suggesting its role in anxiety responses may depend on the circumstances of a threat. For example, in situations when social skills are directly under scrutiny, such as in a speech task, the role of trait social anxiety may have a more primary influence that negates the influences of anxiety sensitivity. Alternatively, the lack of an in-person audience during the speech task in the present investigation may have reduced the overall effects expected during a speech task, thereby limiting the potential role of anxiety sensitivity. In either case, trait social anxiety and anxiety sensitivity may contrast in terms of contributions to perceptions of arousal and state anxiety in differing circumstances. This possibility is consistent with previous findings that different social situations invoke heterogeneous responses (Gerlach et al., 2001). More research is needed to explore these possibilities and to determine the specific role of anxiety sensitivity as a mediator. A number of implications follow from the present findings. Contemporary cognitive-behavioural maintenance models of social anxiety disorder (Clark & Wells, 1995; Hofmann, 2007; Rapee & Heimberg, 1997) posit that arousal serves an integral part in building a mental representation of the self in social circumstances (e.g., “They must see me sweating and think I’m an idiot”). Accordingly, clarifying which constructs impact perceptions of arousal could have important implications for current understanding of how social anxiety disorder is maintained. The current findings suggest that anxiety sensitivity may be one such construct, and several mechanisms may account for the influence of anxiety sensitivity on exaggerated or catastrophic interpretations of arousal. For example, individuals with greater anxiety sensitivity may believe that their symptoms of anxiety indicate loss of cognitive control or pending insanity, signal an impending bodily

catastrophe (e.g., passing out, heart attack), or indicate an upcoming negative social outcome (e.g., others can see they are anxious; Taylor et al., 2007). Given the current findings, contemporary cognitive-behavioural models of social anxiety disorder may benefit from including anxiety sensitivity within their parameters. Specifically, the impact of anxiety sensitivity on appraisals of arousal may increase anxiety and self-focus in situations associated with potential social threat, further skewing perceptions of arousal and of the self. These negative appraisals would also serve as salient cues that a previous social interaction was a failure, and that social interactions should be avoided. Moreover, anxiety sensitivity may increase temptations to avoid potentially threatening social situations, as these would be associated with salient anxiety sensations. Resulting avoidance would relieve distress only temporarily and prevent the disproval of maladaptive thoughts regarding social interactions, perpetuating social concerns. The current findings support treatments for social anxiety disorder that highlight the importance of arousal (e.g., Stein & Walker, 2009). Given the present findings, it would seem that targeting perceptions of arousal, rather than attempting to regulate objective arousal, would be most beneficial to those with the disorder. Specifically, exposure to social situations paired with cognitive techniques targeting maladaptive thoughts focused on arousal (e.g., “I’m blushing like crazy”) may be beneficial. Additionally, the potentially salient role of anxiety sensitivity suggests that interventions that specifically target anxiety sensitivity (e.g., interoceptive exposure; Craske & Barlow, 2008; Craske, Rowe, Lewin, & Noriega-Dimitri, 1997) may be effective in modifying maladaptive beliefs about arousal, which may consequently reduce symptoms of social anxiety disorder. The current investigation has a number of limitations that provide directions for future research. First, the reliance on a relatively small sample precluded firm conclusions regarding the nature of the indirect effects demonstrated in this investigation, and if these represent forms of partial or total mediation. The lack of a direct effect of trait social anxiety on the dependent variables suggest that these relationships were totally mediated by anxiety sensitivity; however, an alternate explanation may be insufficient statistical power due to the small sample. Nevertheless, the size of the direct effects of trait social anxiety were quite small, suggesting that the relationships between trait social anxiety and the dependent variables may be largely explained by anxiety sensitivity. Moreover, a larger sample may have allowed advanced statistical modelling that could have concurrently included the four primary variables of interest (i.e., trait social anxiety, anxiety sensitivity, perceived arousal, state anxiety), albeit, a much larger sample (e.g., over 150 participants) would have been needed to adequately model the relationships in question. Such a sample size, however, would have been impractical given the nature of the data collection methods used in this investigation. Second, the exclusive use of members of a university community may limit, in part, how the current results can be generalized to clinical presentations. Nevertheless, the use of a sample with a wide range of trait social anxiety and anxiety sensitivity levels allows for relatively broad inferences, which would have been limited by the use of a clinical sample. Moreover, nine participants were likely to have social anxiety disorder based on the prescribed cutoff (i.e., >21) for the SIPS (Carleton, Collimore, et al., 2009), suggesting that the current findings may at least be partially generalizable to clinical cases. The relatively high number of potential participants with social anxiety disorder compared to the general population (Kessler et al., 2008) is likely because the investigation was advertised as examining anxiety, perhaps attracting individuals with severe anxiety. Third, participants completed the perceived arousal and state anxiety visual analogue scales consecutively, which may have artificially

M.A. Thibodeau et al. / J. Behav. Ther. & Exp. Psychiat. 43 (2012) 967e974

inflated correlations between these two measures. Fourth, although each task had intrinsically social elements, it is difficult to assess which task provoked the most social anxiety. Intuitively, a speech task should elicit the most social anxiety; however, the self-reports of participants that were disclosed during the investigation suggest that the hyperventilation task, rather than the speech task, was most embarrassing (e.g., “I must have looked so stupid breathing like that”, “that was so embarrassing!”). Future research may benefit from examining specifically what thoughts are elicited during these tasks (e.g., associated with fear of embarrassment, fear of physical sensations), how frequently and intensely they are experienced, and their associated levels of anxiety. Moreover, future research may benefit from exploring different types of social tasks, or modifications to the current tasks (e.g., using more embarrassing subject matter during the speech), that may elicit greater levels of social anxiety. Fifth, the current investigation does not allow causal inferences and the results make it impossible to determine if greater trait social anxiety or anxiety sensitivity cause perceptions of arousal or state anxiety, or perhaps if state anxiety causes perceptions of arousal, or vice-versa. Nevertheless, direct or unidirectional relationships are very unlikely. Far more likely is the possibility that the variables of interest interact complexly with each other directly and indirectly and also interact with other constructs, latent and observed, to lead to measurable responses and behaviours. The current investigation represents the first attempt at explicitly testing the role of anxiety sensitivity as mediating the relationships between trait social anxiety and state anxiety and perceived arousal during tasks that may be socially threatening. The results suggest that anxiety sensitivity may play a crucial role in perceptions of arousal and state anxiety in potentially threatening social contexts, whether these are associated with relatively extreme physiological arousal or otherwise. Cognitive-behavioural models and treatments for social anxiety disorder may benefit from considering anxiety sensitivity as a potentially critical feature in experiences of social anxiety. Declaration of interest The first author was supported by a Canadian Institutes of Health Research Doctoral award (FRN: 113434). The authors declare no conflicts of interest. References American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text revision ed.). Washington, D.C.: American Psychiatric Association. Anderson, E. R., & Hope, D. A. (2009). The relationship among social phobia, objective and perceived physiological reactivity, and anxiety sensitivity in an adolescent population. Journal of Anxiety Disorders, 23(1), 18e26. doi:10.1016/ j.janxdis.2008.03.011, S0887-6185(08)00083-2 [pii]. Asmundson, G. J. G., Norton, G. R., Wilson, K. G., & Sandler, L. S. (1994). Subjective symptoms and cardiac reactivity to brief hyperventilation in individuals with high anxiety sensitivity. Behaviour Research and Therapy, 32(2), 237e241. Carleton, R. N., Abrams, M. P., Asmundson, G. J. G., Antony, M. M., & McCabe, R. E. (2009). Pain-related anxiety and anxiety sensitivity across anxiety and depressive disorders. Journal of Anxiety Disorders, 23(6), 791e798. doi:10.1016/ j.janxdis.2009.03.003, S0887-6185(09)00078-4 [pii]. Carleton, R. N., Collimore, K. C., Asmundson, G. J., McCabe, R. E., Rowa, K., & Antony, M. M. (2009). Refining and validating the social interaction anxiety scale and the social phobia scale. Depression & Anxiety, 26(2), E71eE81. doi:10.1002/da.20480. Carleton, R. N., Peluso, D. L., Collimore, K. C., & Asmundson, G. J. (2011). Social anxiety and posttraumatic stress symptoms: the impact of distressing social events. Journal of Anxiety Disorders, 25(1), 49e57. doi:10.1016/j.janxdis.2010.08.002, S0887-6185(10)00165-9 [pii]. Carter, M. M., Sbrocco, T., & Ayati, F. (2009). Predicting anxious response to a social challenge and hyperventilation: comparison of the ASI and ASI-3. Journal of Behavior Therapy & Experimental Psychiatry, 40(3), 434e442. doi:10.1016/ j.jbtep.2009.05.001, S0005-7916(09)00026-3 [pii].

973

Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. In R. G. Heimberg, M. R. Liebowitz, & D. A. Hope (Eds.), Social phobia: Diagnosis, assessment and treatment. New York: Guildford Press. Conrod, P. J. (2006). The role of anxiety sensitivity in subjective and physiological responses to social and physical stressors. Cognitive Behaviour Therapy, 35(4), 216e225. doi:10.1080/16506070600898587, Q1G257270733773L [pii]. Craske, M. G., & Barlow, D. H. (2008). Panic disorder and agoraphobia. In D. H. Barlow (Ed.), Clinical handbook of psychological disorders. New York: Guildford Press. Craske, M. G., Rowe, M., Lewin, M., & Noriega-Dimitri, R. (1997). Interoceptive exposure versus breathing retraining within cognitive-behavioural therapy for panic disorder with agoraphobia. British Journal of Clinical Psychology, 36, 85e99. Davey, H. M., Barratt, A. L., Butow, P. N., & Deeks, J. J. (2007). A one-item question with a Likert or visual analog scale adequately measured current anxiety. Journal of Clinical Epidemiology, 60(4), 356e360. doi:10.1016/j.jclinepi.2006.07.015, S0895-4356(06)00387-8 [pii]. Deacon, B., & Abramowitz, J. (2006). Anxiety sensitivity and its dimensions across the anxiety disorders. Journal of Anxiety Disorders, 20(7), 837e857. doi:10.1016/ j.janxdis.2006.01.003. Edelmann, R. J., & Baker, S. R. (2002). Self-reported and actual physiological responses in social phobia. British Journal of Clinical Psychology, 41(Pt 1), 1e14. Feldman, P. J., Cohen, S., Hamrick, N., & Lepore, S. J. (2004). Psychological stress, appraisal, emotion and cardiovascular response in a public speaking task. Psychology & Health, 19(3), 353e368. doi:10.1080/0887044042000193497. Fowles, D. C., Christie, M. J., Edelberg, R., Grings, W. W., Lykken, D. T., & Venables, P. H. (1981). Committee report. Publication recommendations for electrodermal measurements. Psychophysiology, 18(3), 232e239. Gerlach, A. L., Wilhelm, F. H., Gruber, K., & Roth, W. T. (2001). Blushing and physiological arousability in social phobia. Journal of Abnormal Psychology, 110(2), 247e258. Hayes, A. F. (2009). Beyond Baron and Kenny: statistical mediation analysis in the new millennium. Communication Monographs, 76(4), 408e420. doi:10.1080/ 03637750903310360. Hofmann, S. G. (2007). Cognitive factors that maintain social anxiety disorder: a comprehensive model and its treatment implications. Cognitive Behaviour Therapy, 36(4), 193e209. Kessler, R. C., Berglund, P. A., Chiu, W.-T., Demler, O., Glantz, M., Lane, M. C., et al. (2008). The national comorbidity survey replication (NCS-R): cornerstone in improving mental health and mental health care in the United States. In R. C. Kessler, & T. B. Ustun (Eds.), The WHO world mental health surveys: Global perspectives on the epidemiology of mental disorders. New York: Cambridge University Press. Kirschbaum, C., Pirke, K. M., & Hellhammer, D. H. (1993). The ‘Trier social stress test’ e a tool for investigating psychobiological stress responses in a laboratory setting. Neuropsychobiology, 28(1e2), 76e81, 119004 [pii]. Leen-Feldner, E., Feldner, M., Bernstein, A., McCormick, J., & Zvolensky, M. (2005). Anxiety sensitivity and anxious responding to bodily sensations: a test among adolescents using a voluntary hyperventilation challenge. Cognitive Therapy and Research, 29(5), 593e609. doi:10.1007/s10608-005-3510-5. Mackinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Confidence limits for the indirect effect: distribution of the product and resampling methods. Multivariate Behavioral Research, 39(1), 99e128. doi:10.1207/ s15327906mbr3901_4. Mattick, R. P., & Clarke, J. C. (1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behaviour Research & Therapy, 36, 455e570, S0005-7967(97)10031-6 [pii]. Mauss, I. B., Wilhelm, F. H., & Groos, J. J. (2004). Is there less to social anxiety than meets the eye? Emotion experience, expression, and bodily responding. Cognition & Emotion, 18(5), 631e662. Moore, P. J., Chung, E., Peterson, R. A., Katzman, M. A., & Vermani, M. (2009). Information integration and emotion: how do anxiety sensitivity and expectancy combine to determine social anxiety? Cognition & Emotion, 23(1), 42e68. Mulkens, S., de Jong, P. J., Dobbelaar, A., & Bogels, S. M. (1999). Fear of blushing: fearful preoccupation irrespective of facial coloration. Behaviour Research & Therapy, 37(11), 1119e1128, S0005-7967(98)00198-3 [pii]. Naragon-Gainey, K. (2010). Meta-analysis of the relations of anxiety sensitivity to the depressive and anxiety disorders. Psychological Bulletin, 136(1), 128e150. doi:10.1037/a0018055, 2009-24669-004 [pii]. Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879e891. Rapee, R. M., & Heimberg, R. G. (1997). A cognitive-behavioral model of anxiety in social phobia. Behaviour Research & Therapy, 35(8), 741e756, S0005-7967(97) 00022-3 [pii]. Reilly, A. R., Carleton, R. N., & Weeks, J. W. (2011). Psychometric evaluation of the social interaction phobia scale. Anxiety, Stress, and Coping, Advance Online Publication, . doi:10.1080/10615806.2011.598150. Reiss, S. (1991). Expectancy model of fear, anxiety, and panic. Clinical Psychology Review, 11(2), 141e153. Reiss, S., & McNally, R. J. (1985). The expectancy model of fear. In S. Reiss, & R. R. Bootzin (Eds.), Theoretical issues in behaviour therapy (pp. 107e121). New York: Academic Press.

974

M.A. Thibodeau et al. / J. Behav. Ther. & Exp. Psychiat. 43 (2012) 967e974

Reiss, S., Peterson, R. A., Gursky, D. M., & McNally, R. J. (1986). Anxiety sensitivity, anxiety frequency and the prediction of fearfulness. Behaviour Research & Therapy, 24(1), 1e8, 0005-7967(86)90143-9 [pii]. Schmidt, N. B., Lerew, D. R., & Jackson, R. J. (1997). The role of anxiety sensitivity in the pathogenesis of panic: prospective evaluation of spontaneous panic attacks during acute stress. Journal of Abnormal Psychology, 106(3), 355e364. Shostak, B. B., & Peterson, R. A. (1990). Effects of anxiety sensitivity on emotional response to a stress task. Behaviour Research & Therapy, 28(6), 513e521, 0005-7967(90)90138-9 [pii]. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: new procedures and recommendations. Psychological Methods, 7(4), 422e445. doi:10.1037//1082-989X.7.4.422. Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhart (Ed.), Sociological methodology (pp. 290e312). San Francisco: Jossey-Bass. Stein, M. B., Roy-Byrne, P. P., Craske, M. G., Bystritsky, A., Sullivan, G., Pyne, J. M., et al. (2005). Functional impact and health utility of anxiety disorders in primary care outpatients. Medical Care, 43(12), 1164e1170, 00005650200512000-00002 [pii].

Stein, M. B., & Walker, J. R. (2009). Triumph over shyness: Conquering shyness and social anxiety. Silver Spring, MD: Anxiety Disorders Association of America. Stewart, S. H., Buffett-Jerrott, S. E., & Kokaram, R. (2001). Heartbeat awareness and heart rate reactivity in anxiety sensitivity: a further investigation. Journal of Anxiety Disorders, 15(6), 535e553. Taylor, S. (1999). Anxiety sensitivity: Theory, research, and treatment of the fear of anxiety. Mahwah, N.J.: Lawrence Erlbaum. Taylor, S., Koch, W. J., & McNally, R. J. (1992). How does anxiety sensitivity vary across the anxiety disorders? Journal of Anxiety Disorders, 6(3), 249e259. Taylor, S., Zvolensky, M. J., Cox, B. J., Deacon, B., Heimberg, R. G., Ledley, D. R., et al. (2007). Robust dimensions of anxiety sensitivity: development and initial validation of the anxiety sensitivity index-3. Psychological Assessment, 19(2), 176e188. doi:10.1037/1040-3590.19.2.176, 2007-07953-002 [pii]. Westenberg, P. M., Bokhorst, C. L., Miers, A. C., Sumter, S. R., Kallen, V. L., van Pelt, J., et al. (2009). A prepared speech in front of a pre-recorded audience: subjective, physiological, and neuroendocrine responses to the Leiden public speaking task. Biological Psychology, 82(2), 116e124. doi:10.1016/j.biopsycho.2009.06.005, S0301-0511(09)00128-8 [pii].