Sage Journals: Discover world-class research

Abstract

Previous studies have shown that social anxiety was associated with enhanced acquisition of anxiety as measured by subjective ratings in conditioning tasks using faces as the conditioned stimulus and negative evaluation as the unconditioned stimulus. However, a recent study failed to replicate the effect. The current study re-examined the enhanced acquisition effect with a larger sample, explored whether differences in expectancy of negative evaluation was a potential mechanism, and compared the contribution of social anxiety to that of depression on enhanced acquisition. Two hundred and sixty-three unselected participants took part in a differential conditioning task in which three faces each were paired with hostile, neutral, and friendly reaction during acquisition, and all three were paired with neutral reaction during extinction. Results replicated earlier findings that participant social anxiety was associated with enhanced acquisition of anxiety. Socially anxious participants did not show higher expectancy of hostile reaction during acquisition, which suggested the need to consider alternative mechanisms underlying enhanced acquisition. Depression was also associated with enhanced acquisition; however, that association was accounted for by social anxiety. The effect of social anxiety was significant over and above depression, which supported its diagnostic validity.

Keywords

Conditioning depression expectancy extinction social anxiety

Introduction

Chronic social anxiety, characterized by anxiety in and avoidance of social situations, is a prevalent, debilitating condition (American Psychiatric Association, 2013). Social events characterized by rejection and hostile reaction are linked to the development of social anxiety (Spence & Rapee, 2016). Exposure to social situations in treatment contexts and in naturalistic settings predicts lower subsequent social anxiety (e.g., Deacon & Abramowitz, 2004; Miers, Blöte, Heyne, & Westenberg, 2014).

Researchers increasingly draw on experimental learning models to understand the development and maintenance of anxiety (e.g., Vervliet, Craske, & Hermans, 2013). Acquisition refers to the pairing of a neutral conditioned stimulus with a threatening unconditioned stimulus that results in increased anxiety toward the conditioned stimulus. Extinction refers the pairing of an anxiety-provoking conditioned stimulus with no unconditioned stimulus that leads to a reduction in anxiety toward the conditioned stimulus. Conventional models posit that learning depends, in part, on changes in the expectation of the unconditioned stimulus, given the conditioned stimulus, and the intensity of the unconditioned stimulus (e.g., Davey, 1989; Pearce & Hall, 1980; Rescorla & Wagner, 1972).

The observation that only a portion of individuals who experienced negative events develop anxiety challenged learning models. One hypothesis put forward is that learning processes are different for the anxious individual, specifically, that individuals who develop anxiety disorders may more readily acquire fear and anxiety from the same negative experience (e.g., Mineka & Zinbarg, 2006). However, a recent meta-analysis found that the effects of acquisition and extinction were not significantly different for participants with and without anxiety disorders (Duits et al., 2015). One limitation of those studies is that most of the studies used general stimuli, such as shapes and shocks, whereas different types of anxiety are proposed to be characterized by idiosyncratic conditioned and unconditioned stimulus (Foa & Kozak, 1986; Mineka & Zinbarg, 2006). Thus, although the general tendency for acquisition and extinction was comparable for those with and without anxiety disorders, there is a dearth of research examining whether different patterns of learning may be observed when stimuli relevant to the type of anxiety are used, for example, socially relevant stimuli presented to socially anxious individuals.

Early studies on social anxiety examined acquisition and extinction using faces as the conditioned stimulus and nonsocial threats, such as unpleasant odors and painful pressure, as the unconditioned stimulus (Hermann, Ziegler, Birbaumer, & Flor, 2002; Schneider et al., 1999; Veit et al., 2002). Although main effects of acquisition and extinction were observed, they were generally comparable for participants across social anxiety levels. Later studies, which used negative evaluation as the unconditioned stimulus instead, found main effects of acquisition and extinction, and additionally, an enhanced acquisition effect for both participants with social anxiety and social anxiety disorder (Lissek et al., 2008; Pejic, Hermann, Vaitl, & Stark, 2013). Potential mechanisms were investigated. Results suggested that social anxiety was correlated with heightened activation in the left amygdala and hippocampus (Pejic et al., 2013) but not with reactivity to negative evaluation (Lissek et al., 2008; Pejic et al., 2013). Taken together, this line of research suggests that social anxiety is characterized not by the general tendency to acquire anxiety but by the tendency to acquire anxiety when socially relevant stimuli are used. It was unclear whether enhanced acquisition was specific to social anxiety, because the effect of other symptoms of psychopathology was not measured. Also, although neural underpinnings were identified, psychological mechanisms have not been established, although reactivity to negative evaluation is unlikely to be one.

A recent study (Tinoco-González et al., 2015) examined conditioning using socially relevant stimuli in individuals with social anxiety, social anxiety disorder, panic disorder, and healthy controls. The results suggested that the effect of acquisition was comparable for all groups. The failure to replicate raises doubts about the effect of social anxiety on enhanced acquisition, in addition to failing to find support for the specificity of enhanced acquisition to social anxiety. One possibility is that the small sample sizes used in past studies (i.e., Ns < 50) precluded the detection of a reliable effect. Other studies that involved a differential conditioning task using socially relevant stimuli but did not examine the effect of participant social anxiety are not presented here, because they cannot provide evidence for or against diagnostic validity (Davis, Johnstone, Mazzulla, Oler, & Whalen, 2010; Haddad, Lissek, Pine, & Lau, 2011; Wiggert, Wilhelm, Reichenberger, & Blechert, 2015).

The purpose of the current study was threefold. First, we examined whether the effect of social anxiety on enhanced acquisition could be replicated. Second, we examined whether expectancy of negative evaluation for the conditioned stimulus was a mechanism that partially explains the association between social anxiety and enhanced acquisition, given that expectancy is an important mechanism according to learning models (e.g., Pearce & Hall, 1980; Rescorla & Wagner, 1972). Third, the diagnostic validity (Boddez, Davey, & Vervliet, 2018; Davey, 2017) of social anxiety as a moderator of acquisition and extinction was examined. Depression was chosen as a comparison, because it tends to correlate with social anxiety and shares common patterns of neural activity, as well as symptomatic features, biased-information processing of social threats, social rejection as a contributor, and response to interpersonally focused treatments (e.g., Adams, Balbuena, Meng, & Asmundson, 2016; Cuijpers et al., 2011; Gotlib et al., 2004; Hamilton, Chen, Waugh, Joormann, & Gotlib, 2013; Langer et al., 2019; Slavich, O’Donovan, Epel, & Kemeny, 2010). Thus, the effect of social anxiety on enhanced acquisition may be due to the influence of depression. The current study aimed to examine this possibility, which would, in turn, inform the diagnostic validity of social anxiety on enhanced acquisition.

A sample of over 200 participants, which was larger than the combined total of previous studies (Lissek et al., 2008; Pejic et al., 2013; Tinoco-González et al., 2015), was recruited to increase power to detect differences. Given the dimensional nature of social anxiety and related dysfunction, participants were not selected on the basis of clinical diagnosis (Stein, Walker, & Forde, 1994). The advantage of examining social anxiety and depression on a continuum is the greater sensitivity of detecting effects. Given that the effect of social anxiety and depression on learning was expected to be similar because of the strong association between the two (Adams et al., 2016), the current design allowed us to examine the effect of social anxiety accounting for the depression and vice versa.

In line with past studies, participants completed a differential conditioning task using socially relevant conditioned and unconditioned stimulus. The task involved acquisition, in which a face was repeatedly paired with hostile reaction, using two other faces paired with neutral and friendly reactions as control and comparison, followed by extinction, in which all three faces were paired with neutral reactions. Subjective ratings of anxiety were measured and considered the primary dependent variable (Ledoux & Hofmann, 2018). Expectancy was measured through ratings of probability of subsequent hostile reaction for each conditioned stimulus (Lovibond, 2007) before and after acquisition and extinction trials. Also, during each conditioning trial, participants were instructed to indicate the valence of the conditioned stimulus, that is, whether it was hostile, neutral, or friendly. The reaction time required to indicate hostile reactions was used as a parallel measure of expectancy, with lower reaction time indicating greater expectancy (Craddock, Molet, & Miller, 2012). After the conditioning task, participants completed measures of social anxiety and depression.

Method

Participants and procedure

The research methods were approved by the university behavioral research committee. The sample consisted of 263 undergraduate students, who provided informed consent and subsequently participated in the study for partial course credit. The mean age of participants was 20.67 years (SD = 3.33 years) and 76% were female. During a laboratory session, participants first reported on demographics and then on questionnaires assessing depression and social anxiety. They then completed the conditioning task using the software E-PRIME 2.0 (Psychology Software Tools, Pittsburgh, Pennsylvania, USA).

Materials

Depression questionnaire

Depression was assessed using the depression subscale of the Depression, Anxiety, and Stress Scale depression subscale, which was shown to strongly correlate with other measures of depression, and may better differentiate depression from other affective states compared to other questionnaires (Lovibond & Lovibond, 1995) The questionnaire consisted of 14-items each rated on a 4-point scale. Scores were summed. In the current sample, participant scores ranged from 0 to 38 with a mean of 8.83 and standard deviation of 8.29. Cronbach’s α for the scale was .94.

Social anxiety questionnaire

Social anxiety was assessed using the straightforward items (Rodebaugh et al., 2011) of the Social Interaction Anxiety Scale (Mattick & Clarke, 1998). The questionnaire consisted of 17 items each rated on a 5-point scale. Scores were summed. In the current sample, scores ranged from 0 to 59 with a mean of 22.96 and standard deviation of 14.22. Cronbach’s α for the scale was .94. Social anxiety was correlated with depression, r = .48, p < .001.

Conditioning task

The basic building block of the conditioning task was one trial. The procedure is presented in Figure 1. For each trial, a fixation cross appeared in the middle of the screen for 1 s. Then, the cross was replaced by the neutral expression of a face (i.e., the conditioned stimulus), which was presented for 1 s, followed by the same face showing either a hostile, neutral, or friendly facial expression, as well as various valence–congruent verbal statements presented as written text beneath the face (i.e., the conditioned and unconditioned stimulus). Verbal statements were randomly selected from a database of 25 hostile (e.g., “What’s wrong with you?!”), 45 neutral (e.g., “Hello.”), and 25 friendly (e.g., “Good job!”) statements created for the purpose of the experiment (Table 1). Three research assistants, who were not involved in data collection and were blind to the purpose of the study, were asked to rate the statements in randomized order on how friendly, neutral, and hostile they were on a scale of 0 (disagree) to 10 (agree). Friendly statements were rated significantly more friendly (M = 8.76, SD = 1.63) than neutral (M = 2.37, SD = 2.77) and hostile statements (M = 0.29, SD = 0.94), neutral statements were rated significantly more neutral (M = 8.80, SD = 1.91) than friendly (M = 1.88, SD = 2.31) and hostile statements (M = 1.11, SD = 2.06), and hostile statements were rated significantly more hostile (M = 8.48, SD = 2.03) than friendly (M = 0.47, SD = 1.44) and neutral statements (M = 1.13, SD = 1.88) according to independent t-tests, ps < .001. The conditioned and unconditioned stimulus presentation remained until participants indicated whether the face was hostile, neutral, or friendly by pressing keyboard buttons “d,” “s,” and “a” labeled with “red,” “yellow,” and “green” colored stickers. Reaction time was measured by the computer software. Expectation of a valanced reaction (i.e., hostile, neutral, or friendly) for a conditioned stimulus was operationalized as the reaction time needed to correctly indicate the valence for that conditioned stimulus, with a lower reaction time indicating stronger expectation.

Figure 1.

Procedure for one conditioning trial. Note. CS = conditioned stimulus; US = unconditioned stimulus.

Table 1.

Hostile, neutral, and friendly statements.

Hostile	Neutral		Friendly
I don’t like you.	Okay.	See you later.	Cool!
What’s wrong with you?	Pardon me.	Say that again?	You’re so interesting.
Stay away from me.	What’s going on?	Got it.	Awesome!
Loser.	Hold on.	I understand.	You rock!
Jerk.	Hi.	I am in second year.	Good job!
Go away.	Good morning.	I have no idea.	Always good to see you.
Leave me alone.	What’s your name?	Right?	You’re looking great!
Not interested.	What happened?	Really?	You’re so smart!
Shut up.	When is the next bus?	Is this seat taken?	Nice talking to you!
Don’t talk to me.	I am a student.	There we go.	Brilliant!
What’s your problem?	I don’t know.	I am hungry.	You can do it!
Not you again.	The library closes at 10.	Can I borrow this?	You’re the best!
I don’t care.	What time is it?	How much is it?	Great idea!
You’re hopeless.	It will rain later.	What?	Take care!
Buzz off.	Where is the washroom?	I’ll be back.	That’s very thoughtful.
You’re useless.	It’s lunchtime.	I drink coffee.	I like you.
Please stop.	I came from class.	Huh?	You’re really good at this.
You’re so awkward.	It will be sunny.		You’re right.
You suck.	Where’s the washroom?		Well done!
You’re so boring.	How’s the weather?		Fantastic!
You’re unbelievable.	I have a question.		Nice outfit.
Stupid.	Go ahead.		That’s very nice of you.
That’s enough!	Any questions?		Terrific!
No, you may not.	Is that clear?		Way to go!
You’re ugly.	Anything else?		Love it!

The overall procedure is summarized in Table 2 and described below. First, participants were instructed that they would be shown a person followed by a facial expression and statement the person says and that participants were to indicate the valence of the person by pressing the respective keyboard button as quickly as they can with their dominant hand. Participants completed a practice sequence in which a practice conditioned stimulus (i.e., CS_practice) was paired with two trials of hostile, neutral, and friendly reactions each. During this phase, participants were given feedback on whether they correctly indicated the valence of the reaction.

Table 2.

Conditioning task procedure.

Note. CS = conditioned stimulus. Conditioning trials are shaded in grey. Trials were randomized within each sequence.

The main conditioning procedure consisted of an acquisition and an extinction sequence. During acquisition, three female conditioned stimuli were paired with either 10 trials of hostile, neutral, or friendly reaction. Conditioned stimuli paired with hostile, neutral, and friendly reaction during acquisition are referred to as CS_hostile, CS_neutral, and CS_friendly, respectively. The roles of the conditioned stimuli were counterbalanced across participants. During extinction, each conditioned stimulus was paired with 10 trials of faces with neutral reactions. Before and after acquisition and extinction, participants underwent a sequence of reaction time test trials to measure their expectation of each conditioned stimulus with each type of valanced reaction. Each test trial sequence consisted of nine trials, in which each of the three conditioned stimulus was paired with one hostile, one neutral, and one friendly reaction. The presentation of trials was randomized within sequences.

Subjective ratings were measured for each conditioned stimulus at three points, before pre-acquisition test trials, after post-acquisition test trials, and after post-extinction test trials. Conditioned stimuli displayed a neutral expression and no valanced reaction followed. For each conditioned stimulus, participants rated the extent to which they felt anxious (0 = not anxious at all; 100 = extremely anxious), probability that the conditioned stimulus would be hostile on the next trial (0 = certainly not; 100 = certainly), pleasantness (0 = extremely unpleasant; 100 = extremely pleasant), and probability that the conditioned stimulus would be hostile on the next trial (0 = certainly not; 100 = certainly) in the listed order for all participants.

Results

The computer software “R” version 3.3.1 (R Core Team, 2019) package “lmer” (Bates, Maechler, Bolker, & Walker, 2015) was used to analyze data. Compared to analysis of variance, multilevel modeling is a more flexible alternative for analyzing repeated measures data (Quené & Van Den Bergh, 2004) and was used for the current analyses. In the current analyses, using multilevel modeling has the notable advantages of being able to examine social anxiety and depression as continuous predictors, which increased power (Royston, Altman, & Sauerbrei, 2006), and to handle missing data. For all analyses, maximum likelihood was used for estimation.

Only dependent variables that were of immediate relevance to the research question, that is, subjective ratings of anxiety, subjective ratings of probability of hostile reaction, and reaction time for hostile reactions (RT_hostile), were analyzed and reported. Four separate multilevel regression analyses were run to examine the effect of social anxiety and depression on subjective ratings of anxiety across acquisition and extinction, the effect of social anxiety and depression on subjective ratings probability of hostile reaction across acquisition and extinction, the effect of social anxiety and depression RT_hostile across acquisition, and the effect of social anxiety and depression for RT_hostile across extinction.

Based on past findings, eight main effects of acquisition and extinction for CS_hostile and CS_friendly on the dependent variables of subjective ratings of anxiety and probability of hostile reaction were expected, as well as two main effects of acquisition and extinction for CS_hostile on RT_hostile. Thus, statistical significance for these effects were compared to a two-tailed level of α = .05. The moderation of social anxiety and depression on those effects, that is, 20 effects, were explored, and hence, statistical significance was compared to an adjusted, two-tailed level of α = .05/20 = .0025.

Follow-up analyses were conducted for any effects that were significantly moderated by social anxiety or depression or both. To test whether the social anxiety or depression interaction accounted for the observed data over and above that of the other moderator, the full model, which consisted of social anxiety and depression simultaneously entered as moderators, was compared to two reduced models using the deviance test, the first one without social anxiety interaction term and the second one without the depression interaction term.

Subjective ratings of anxiety as the dependent variable

Out-of-range and nonsensical values for subjective ratings, that is, nonnumerical values and numerical values less than 0 or greater than 100, were removed, and 2,349 of the 2,367 original responses were analyzed. Although not systematically assessed, anecdotal reports from participants suggest that the values were because they were confused with E-PRIME’s text display layout, not knowing that the backspace was functional, and typographic errors. The exclusion criteria were not set a priori, because these types of responses were not foreseen. Subsequently, the following models were used to analyze the data:

An x_{i j} = γ_{00} + γ_{10} \times C S_{hostile} + γ_{20} \times C S_{friendly} + γ_{30} \times Acq + γ_{40} \times Ext + γ_{50} \times C S_{hostile} \times Acq + γ_{60} \times C S_{hostile} \times Ext + γ_{70} \times C S_{friendly} \times Acq + γ_{80} \times C S_{friendly} \times Ext + γ_{01} \times Moderator + γ_{11} \times C S_{hostile} \times Moderator + γ_{21} \times C S_{friendly} \times Moderator + γ_{31} \times Acq \times Moderator + γ_{41} \times Ext \times Moderator + γ_{51} \times C S_{hostile} \times Acq \times Moderator + γ_{61} \times C S_{hostile} \times Ext \times Moderator + γ_{71} \times C S_{friendly} \times Acq \times Moderator + γ_{81} \times C S_{friendly} \times Ext \times Moderator + u_{0 j} + e_{i j}

Subjective ratings of anxiety or probability of hostile reaction (subscripted i) were nested within participants (subscripted j). The following dummy codes were used for CS_hostile (CS_neutral = 0, CS_hostile = 1), CS_friendly (CS_neutral = 0, CS_friendly = 1), acquisition (i.e., pre-acquisition = −1, post-acquisition = 0), and extinction (i.e., post-acquisition = 0, post-extinction = 1). Hence, γ₀₀ is the mean for CS_neutral at post-acquisition; γ₁₀ and γ₂₀ are the differences in ratings between CS_hostile and CS_neutral, and between CS_friendly and CS_neutral, at post-acquisition; γ₃₀ and γ₄₀ are the changes in ratings for CS_neutral across acquisition and extinction; γ₅₀, γ₆₀, γ₇₀, and γ₈₀ are interaction representing change in ratings for CS_hostile compared to CS_neutral across acquisition, CS_friendly compared to CS_neutral across acquisition, CS_hostile compared to CS_neutral across extinction, and CS_friendly compared to CS_neutral across extinction; γ₀₁, γ₁₁, γ₂₁, γ₃₁, γ₄₁, γ₅₁, γ₆₁, γ₇₁, and γ₈₁ are the interactions between the moderator, that is, participant social anxiety or depression, and the respective effects in the order mentioned above; u _0j is between-participant variance; and e_ij is residual variance. γ₅₁, γ₆₁, γ₇₁, and γ₈₁ were the effects tested for statistical significance.

The model entering social anxiety as the moderator was first examined (Table 3, column A; Figure 2). Social anxiety significantly moderated acquisition for CS_hostile (i.e., CS_hostile × Acq × SocialAnxiety), γ₅₁ = .54, SE = .15, t(2,078.80) = 3.60, p < .001. Probing the interaction suggested that, for participants with social anxiety one standard deviation above the mean, ratings for CS_hostile increased 19.32, SE = 2.11, t(2,078.80) = 9.17, p < .001, and ratings for CS_neutral decreased 3.01, SE = 2.11, t(2,078.80) = −1.43, p = .076, across acquisition. For participants with mean social anxiety, ratings for CS_hostile increased 14.92, SE = 1.50, t(2,078.80) = 4.98, p < .001, and ratings for CS_neutral increased .19, SE = 1.49, t(2,078.80) = 0.13, p = .55. For participants with social anxiety one standard deviation below the mean, ratings for CS_hostile increased 10.52, SE = 2.12, t(2,078.80) = 4.95, p < .001, and ratings for CS_neutral increased 3.40, SE = 2.12, t(2,078.80) = 1.61, p = .054.

Figure 2.

Subjective ratings of anxiety across acquisition and extinction as predicted by the multivariate multilevel model. Note. CS = conditioned stimulus. ^aSocial anxiety was significantly associated, and depression was marginally significantly associated with greater increase in anxiety for CS_hostile compared to CS_neutral (i.e., enhanced acquisition) when entered as moderators separately.

Table 3.

Results for multilevel regression analyses.

Model	DV: subjective ratings of anxiety					DV: subjective ratings of probability of hostile reaction		DV: RT_hostile
Model	A	B	C	D	E	F	G	H	I	J	K
Fixed
Intercept	27.42 (1.42)***	27.44 (1.48)***	27.41 (1.41)***	27.41 (1.42)***	27.41 (1.41)***	30.88 (1.30)***	30.78 (1.31)***	7.03 (0.03)***	7.04 (0.03)***	6.84 (0.03)***	6.84 (0.03)***
CS_hostile	14.39 (1.49)***	14.4 (1.49)***	14.39 (1.48)***	14.39 (1.48)***	14.39 (1.48)***	19.66 (1.48)***	19.86 (1.48)***	−0.04 (0.03)	−0.04 (0.03)	−0.11 (0.03)***	−0.11 (0.03)***
CS_friendly	−8.77 (1.49)***	−8.80 (1.49)***	−8.76 (1.48)***	−8.77 (1.48)***	−8.77 (1.48)***	−8.96 (1.48)***	−8.93 (1.48)***	−0.05 (0.03)	−0.04 (0.03)	0.04 (0.03)	0.04 (0.03)
Acq	0.19 (1.49)	0.18 (1.49)	0.19 (1.49)	0.17 (1.49)	0.19 (1.49)	0.61 (1.48)	0.57 (1.48)	−0.06 (0.03)*	−0.06 (0.03)*
Ext	−0.70 (1.49)	−0.68 (1.49)	−0.69 (1.48)	−0.69 (1.49)	−0.69 (1.48)	0.45 (1.48)	0.47 (1.48)			0.02 (0.03)	0.02 (0.03)
CS_hostile × Acq	14.73 (2.11)***	14.83 (2.12)***	14.72 (2.11)***	14.77 (2.11)***	14.71 (2.11)***	19.34 (2.10)***	19.52 (2.10)***	−0.25 (0.04)***	−0.25 (0.04)***
CS_hostile × Ext	−10.02 (2.1)***	−10.09 (2.10)***	−10.02 (2.10)***	−.10.00 (2.10)***	−10.02 (2.10)***	−15.07 (2.10)***	−15.18 (2.10)***			0.06 (0.04)	0.06 (0.04)
CS_friendly × Acq	−8.99 (2.11)***	−8.96 (2.11)***	−8.86 (2.11)***	−8.95 (2.11)***	−8.97 (2.11)***	−9.72 (2.10)***	−9.71 (2.10)***	0.02 (0.04)	0.01 (0.04)
CS_friendly × Ext	8.12 (2.1)***	8.09 (2.10)***	8.11 (2.10)***	8.11 (2.10)***	8.11 (2.10)***	6.32 (2.10)*	6.29 (2.10)*			−0.05 (0.04)	−0.06 (0.04)
SocialAnxiety	0.57 (0.10)***		0.65 (.11)*	0.76 (.11)***	0.63 (.11)***	0.36 (0.09)***		0.001 (0.002)		−0.0001 (0.002)
CS_hostile × SocialAnxiety	0.05 (0.1)		−0.06 (.12)	−0.28 (.08)***	−0.01 (.11)	−0.19 (0.10)		0.002 (0.002)		0.004 (0.002)
CS_friendly × SocialAnxiety	−0.04 (0.1)		−0.15 (.12)	−0.26 (.11)*	−0.12 (.12)	−0.11 (0.10)		−0.002 (0.002)		0.002 (0.002)
Acq × SocialAnxiety	−0.23 (0.11)*		−0.11 (.12)	0.11 (8.47)	−0.16 (.11)	−0.02 (0.10)		−0.002 (0.002)
Ext × SocialAnxiety	0.11 (0.1)		0.09 (.12)	0.02 (.11)	0.11 (.11)	−0.07 (0.10)				−0.003 (0.002)
CS_hostile × Acq × SocialAnxiety	0.54 (0.15)***		0.45 (.17)**		0.22 (.14)***	0.04 (0.15)		−0.002 (0.003)
CS_hostile × Ext × SocialAnxiety	−0.09 (0.15)		−0.05 (.17)	0.17 (0.15)	0.31 (0.19)	0.16 (0.15)				−0.002 (0.003)
CS_friendly × Acq × SocialAnxiety	0.34* (0.15)*		0.27 (0.17)	0.04 (0.15)	−0.25 (0.14)	0.19 (0.15)		−0.002 (0.003)
CS_friendly × Ext × SocialAnxiety	−0.19 (0.15)		−0.10 (0.17)	0.01 (0.16)	−0.02 (0.19)	0.03 (0.15)				0.005 (0.003)
Depression		0.24 (0.18)	−0.29 (0.19)	−0.38 (0.19)*	−0.21 (0.18)		0.06 (0.16)		−0.001 (0.003)		−0.001 (0.003)
CS_hostile × Depression		0.33 (0.18)	0.37 (0.20)	0.55 (0.19)**	0.54 (0.15)***		0.32 (0.18)		−0.002 (0.003)		0.0004 (0.003)
CS_friendly × Depression		0.27 (0.18)	0.39 (0.20)	0.48 (0.19)*	−0.10 (0.16)		0.13 (0.18)		−0.005 (0.003)		−0.002 (0.003)
Acq × Depression		−0.49 (0.18)*	−0.40 (0.20)*	−0.58 (0.19)**	0.31 (0.16)		−0.06 (0.18)		−0.001 (0.004)
Ext × Depression		0.13 (0.18)	0.06 (0.20)	0.15 (0.20)	−0.12 (0.17)		0.22 (0.18)				−0.01 (0.004)
CS_hostile × Acq × Depression		0.67 (0.25)*	0.31 (0.28)	0.67 (0.25)**			0.33 (0.25)		0.003 (0.005)
CS_hostile × Ext × Depression		−0.18 (0.25)	−0.13 (0.28)	−0.31 (0.28)	0.02 (0.25)		−0.36 (0.25)				0.004 (0.01)
CS_friendly × Acq × Depression		0.45 (0.25)	0.24 (0.28)	0.42 (0.28)	0.09 (0.25)		0.03 (0.25)		0.001 (0.005)
CS_friendly × Ext × Depression		−0.39 (0.25)	−0.31 (0.28)	−0.40 (0.28)	−0.23 (0.28)		−0.33 (0.25)				0.01 (0.01)
Random
Between-participant variance	236.2	282.8	236.5	236.5	236.5	151.1	159.6	0.08	0.08	0.07	0.07
Residual variance	289.1	290.7	287.2	288.1	287.3	286.6	287.4	0.10	0.10	0.10	0.10
Deviance	20,455.5	20,587.4	20,441.3	20,448.4	20,442.5	20,414.2	20,510.3	1,185.4	1,219.2	1,222.4	1,244.1
AIC	20,495.5	20,627.4	20,499.3	20,504.4	20,498.5	20,454.2	20,550.3	1,213.4	1,247.2	1,250.4	1,272.1
BIC	20,610.6	20,742.6	20,666.3	20,665.6	20,659.7	20,569.5	20,665.6	1,287.0	1,320.8	1,323.9	1,345.7

Note. RT = reaction time for hostile reactions; CS = conditioned stimulus; DV = Dependent Variable; AIC = Akaike information criterion; BIC = Bayesian information criterion. γ₅₀, γ₆₀, γ₇₀, and γ₈₀ are compared to α = .05. γ₅₁, γ₆₁, γ₇₁, and γ₈₁ are compared to an adjusted α = .002.

*p < .05; **p < .002; ***p < .001.

In the model entering depression as the moderator (Table 3, column B), the moderation of depression on acquisition for CS_hostile compared to CS_neutral (i.e., CS_hostile × Acq × Depression) was marginally significant, γ₅₁ = .67, SE = .25, t(2,086.40) = 2.67, p = .008. Follow-up analyses were conducted to examine whether social anxiety better accounted for greater increases in ratings toward CS_hostile compared to CS_neutral after acquisition over and above the effects of depression, and vice versa. First, the full model simultaneously entering interactions with social anxiety and depression as moderators was examined (Table 3, column C). The effects of social anxiety on acquisition for CS_hostile compared to CS_neutral (i.e., CS_hostile × Acq × SocialAnxiety) remained significant when adjusted for depression in the full model. The full model then was compared a reduced model without the CS_hostile × Acq × SocialAnxiety interaction (Table 3, column D), and results suggested that the full model provided significantly better model fit, χ²(1) = 7.04, p = .008. The full model was compared to another reduced model without the CS_hostile × Acq × Depression interaction (Table 3, column E), and results suggested that the full model did not significantly improve fit, χ²(1) = 1.15, p = .28.

As expected, for the sample as a whole, there were significant effects of extinction for CS_hostile (i.e., CS_hostile × Ext), as well as acquisition and extinction for CS_friendly (i.e., CS_friendly × Acq, CS_friendly × Ext). Social anxiety and depression did not moderate any of the effects in their respective models, ps > .024 (Table 3, columns A and B).

Subjective ratings of probability of hostile reaction as the dependent variable

The same exclusion criteria for out-of-range and nonsensical values for subjective ratings was applied based on the same rationale, and 2,358 of the 2,367 original responses were analyzed. The same models as above were used to analyze the data except that the dependent variable is subjective ratings of probability of reaction and the same effects were tested for significance.

As expected and shown in Figure 3, there were significant effects of acquisition and extinction for CS_hostile (i.e., CS_hostile × Acq, CS_hostile × Ext) and CS_friendly (i.e., CS_friendly × Acq, CS_friendly × Ext). Social anxiety and depression did not moderate any of the effects in their respective models, ps > .15 (Table 3, columns F and G).

Figure 3.

Subjective ratings of probability of hostile reaction across acquisition and extinction as predicted by the multivariate multilevel model.

Reaction time for hostile reaction (RT_hostile) as the dependent variable

Incorrect responses, that is, responses indicating the hostile reaction as neutral or friendly, were removed. Of 3,156 responses, 2,835, that is, 90%, were accurate. The effects of acquisition and extinction were analyzed in two separate models. First, the following model was used to analyze acquisition:

R T_{hostil e_{i j}} = γ_{00} + γ_{10} \times C S_{hostile} + γ_{20} \times C S_{friendly} + γ_{30} \times Acq + γ_{50} \times C S_{hostile} \times Acq + γ_{70} \times C S_{friendly} \times Acq + γ_{01} \times Moderator + γ_{11} \times C S_{hostile} \times Moderator + γ_{21} \times C S_{friendly} \times Moderator + γ_{31} \times Acq \times Moderator + γ_{51} \times C S_{hostile} \times Acq \times Moderator + γ_{71} \times C S_{friendly} \times Acq \times Moderator + u_{0 j} + e_{i j}

The following model was used to analyze extinction:

R T_{hostil e_{i j}} = γ_{00} + γ_{10} \times C S_{hostile} + γ_{20} \times C S_{friendly} + γ_{40} \times Ext + γ_{60} \times C S_{hostile} \times Ext + γ_{80} \times C S_{friendly} \times Ext + γ_{01} \times Moderator + γ_{11} \times C S_{hostile} \times Moderator + γ_{21} \times C S_{friendly} \times Moderator + γ_{41} \times Ext \times Moderator + γ_{61} \times C S_{hostile} \times Ext \times Moderator + γ_{81} \times C S_{friendly} \times Ext \times Moderator + u_{0 j} + e_{i j}

Log transformed RT_hostile (subscripted i) were nested within participants (subscripted j). Raw RT_hostile was not used, because it was positively skewed and the resulting model would violate the assumption of homoscedasticity. The following dummy codes were used for CS_hostile (CS_neutral = 0, CS_hostile = 1), CS_friendly (CS_neutral = 0, CS_friendly = 1), acquisition (i.e., pre-acquisition = 0, post-acquisition = 1), and extinction (i.e., pre-extinction = 0, post-extinction = 1). Hence, γ₀₀ is the mean for log RT_hostile at pre-acquisition or pre-extinction; γ₁₀ and γ₂₀ are the differences in ratings between CS_hostile and CS_neutral, and between CS_friendly and CS_neutral, at pre-acquisition or pre-extinction; γ₃₀ and γ₄₀ are the changes in ratings for CS_neutral across acquisition and extinction; γ₅₀, γ₆₀, γ₇₀, and γ₈₀ are interactions representing change in ratings for CS_hostile compared to CS_neutral across acquisition, CS_friendly compared to CS_neutral across acquisition, CS_hostile compared to CS_neutral across extinction, and CS_friendly compared to CS_neutral across extinction; γ₀₁, γ₁₁, γ₂₁, γ₃₁, γ₄₁, γ₅₁, γ₆₁, γ₇₁, and γ₈₁ are the interactions between the moderator, that is, participant social anxiety or depression, and the respective effects in the order mentioned above; u _0j is between-participant variance; and e_ij is residual variance. γ₅₁, γ₆₁, γ₇₁, and γ₈₁ were the effects tested for statistical significance.

As shown in Figure 4, there was a significant effect of acquisition (CShostile × Acq) but insignificant trend of extinction (i.e., CShostile × Ext) for CShostile. Social anxiety and depression did not moderate any of the effects in their respective models, ps > .13 (Table 3, columns H, I, J, and K).

Figure 4.

log RT_hostile across acquisition and extinction predicted by the multilevel model. RT = reaction time.

Post hoc power analysis

Much of the literature suggests that power for a multilevel model interaction is difficult to determine, because it requires the estimation of the interaction coefficient’s standard error, which in turn depends on several other parameters unknown to the researchers before the study (e.g., Hox, 2010). Hence, sample size was determined based on a coarse analysis before the study. The analyses were treated as the correlational effect of social anxiety and depression on acquisition and extinction. Analyses suggested that a sample of 200 had power of .92 to detect moderate effects (i.e., r = .30) after adjusting for the 20 comparisons. Data for 200 participants were collected, but data collection continued after the goal was reached until the end of the university term, which resulted in the current sample.

Post hoc power analyses were conducted following recommendations by Hox (2010, equation 12.3) using standard errors for the interaction estimated from the models above, that is, those for CS_hostile × Acq × SocialAnxiety, CS_hostile × Ext × SocialAnxiety, CS_friendly × Acq × SocialAnxiety, CS_friendly × Ext × SocialAnxiety, CS_hostile × Acq × Depression, CS_hostile × Ext × Depression, CS_friendly × Acq × Depression, and CS_friendly × Ext × Depression and adjusting for multiple comparisons. Results suggested that the current study had .80 power to detect interactions with sizes of .54 for social anxiety moderation on subjective ratings of anxiety and probability of hostile reaction, .91 for depression moderation on subjective ratings of anxiety and probability of hostile reaction, .01 for social anxiety moderation on log RT_hostile, and .02 for depression moderation on log RT_hostile, which converted to standardized coefficients in the range of .31–.35.

Discussion

These results replicated earlier findings that social anxiety was associated with enhanced acquisition, that is, social anxiety was associated with greater increases in anxiety for the conditioned stimulus paired with hostile reaction relative to changes paired with neutral reaction (Lissek et al., 2008; Pejic et al., 2013; cf. Tinoco-González et al., 2015). Notably, the current study used a sample size greater than the combined total used in past studies (Lissek et al., 2008; Pejic et al., 2013; Tinoco-González et al., 2015), and had adequate power (i.e., .80) to detect moderate effects even after adjusting for multiple comparisons. Effects that were marginally significant in the current study, for example, that social anxiety was associated with smaller decreases in anxiety toward the conditioned stimulus paired with friendly reaction (Table 3, column A), suggest that there may be smaller effects that did not achieve significance. These effects can be pursued through future studies involving an even larger sample. One question regarding the conditioning framework of anxiety is that only a subset of individuals who experience acquisition-like events involving negative evaluation develop social anxiety. The current findings provide preliminary support for the explanation that individuals who acquire anxiety more readily from negative evaluation may be at risk for developing social anxiety (Mineka & Zinbarg, 2006). However, only future studies using longitudinal designs would provide more definitive answers (e.g., Lonsdorf & Merz, 2017).

Expectancy of negative evaluation was evaluated as a potential mechanism of the acquisition effect. Subjective ratings and reaction time toward negative evaluation were measured. Results suggested that acquisition and extinction, respectively, increased and decreased expectancy, although the latter effect was insignificant for the reaction time measure. It is worth noting that other studies have also found that extinction was not observed on measures, such as fear-potentiated startle (e.g., Lissek et al., 2008). Social anxiety was not associated with increased expectancy for negative evaluation on either measure. Thus, the results tentatively suggest that, although expectancy of the unconditioned stimulus may underlie anxiety toward the conditioned stimulus (e.g., Pearce & Hall, 1980; Rescorla & Wagner, 1972), the enhanced acquisition effect for social anxiety was not likely due to differences in learning expectancy of negative evaluation. Combining these findings with past work, the results tentatively suggest that the effect of social anxiety on enhanced acquisition is not due to reactivity to the negative evaluation (Lissek et al., 2008; Pejic et al., 2013) or expectancy of negative evaluation from the conditioned stimulus. Future studies could investigate other mechanisms, for example, the salience of the conditioned stimulus (Pearce & Hall, 1980; Rescorla & Wagner, 1972), or the input of higher order cognitive processes (Thorwart & Livesey, 2016).

Depression displayed a marginal significant moderation of enhanced acquisition, which was not surprising given its correlation and shared neural patterns, symptoms, biased-information processing of social threats, and contributor of social rejection. The follow-up analyses resulted in an important finding that depression was no longer a significant moderator of acquisition when social anxiety was controlled, but that the effect of social anxiety on enhanced acquisition remained significant even accounting for depression. These results suggested that the effect of social anxiety on enhanced acquisition was not due to depression, but vice versa. The findings support the diagnostic validity of social anxiety on enhanced acquisition and may also contribute to understanding the relation between social anxiety and depression in future research.

The desire to make a good impression on others has been proposed to be a crucial factor underlying social anxiety (Schlenker & Leary, 1982). Hence, the absence of an actual or perceived audience in the current study presents a potential limitation to generalizing the results to naturalistic settings. Another difference between social conditioning tasks and naturalistic environments is the intensity of negative evaluation. In laboratory studies, negative evaluation is confined to hostile facial expressions and statements from pictures of strangers because of consideration for experimental control and participant well-being, whereas negative evaluation in naturalistic settings can involve more severe embarrassment or humiliation (e.g., McCabe, Antony, Summerfeldt, Liss, & Swinson, 2003). Be that as it may, the operationalization of negative evaluation in conditioning tasks is certainly not trivial, considering its impact on neurophysiological indices of anxiety (Lissek et al., 2008; Pejic et al., 2013) and its long-lasting effects (Wiggert et al., 2017).

In sum, the current study replicated the association between social anxiety and enhanced acquisition of anxiety toward socially threatening stimuli and showed that social anxiety accounted for the effect over and above the effect of depression. Socially anxious participants did not show differences in learning about the expectancy of negative evaluation, and the results tentatively suggest that expectations are not likely to be a mechanism underlying enhanced acquisition.

Footnotes

Author’s note

Detailed information not reported in the main text is available upon request from the first author.

Acknowledgments

The authors would like to thank Cindy Choi, Sarah Lillemo, and Jasmine Veark for their help in data collection, and Rachel Hueller, Taylyn Jameson, and Priyam Joshi for their help in creating the stimuli.

Data accessibility statement

The stimulus presentation file, data, and analysis code can be accessed through .

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the University of British Columbia Faculty of Graduate Studies Doctoral Four Year Fellowship, and Canadian Psychological Association Ken Bowers Award awarded to Klint Fung, and the Social Sciences and Humanities Research Council of Canada (SSHRC) 435-2012-0350 awarded to Lynn E. Alden.

ORCID iD

Klint Fung

References

Adams

G. C.

Balbuena

Meng

X. F.

Asmundson

G. J. G.

(2016). When social anxiety and depression go together: A population study of comorbidity and associated consequences. Journal of Affective Disorders, 206, 48–54. doi:10.1016/j.jad.2016.07.031

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: Author.

Bates

Maechler

Bolker

Walker

(2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67, 1–48. doi:10.18637/jss.v067.i01

Boddez

Davey

Vervliet

(2018). Editorial: Experimental psychopathology: Defining the field. Psychopathology Review, a4, 109–111. doi:10.1177/20438087170a400201

Craddock

Molet

Miller

R. R.

(2012). Reaction time as a measure of human associative learning. Behavioural Processes, 90, 189–197. doi:10.1016/j.beproc.2012.01.006

Cuijpers

Geraedts

A. S.

van Oppen

Andersson

Markowitz

J. C.

van Straten

(2011). Interpersonal psychotherapy for depression: A meta-analysis. American Journal of Psychiatry, 168, 581–592. doi:10.1176/appi.ajp.2010.10101411

Davey

G. C. L.

(1989). UCS revaluation and conditioning models of acquired fears. Behaviour Research and Therapy, 27, 521–528. doi:10.1016/0005-7967(89)90086-7

Davey

G. C. L.

(2017). A research pathway for experimental psychopathology: The role of external validity criteria. Psychopathology Review, a4, 129–140. doi:10.5127/pr.045015

Davis

F. C.

Johnstone

Mazzulla

E. C.

Oler

J. A.

Whalen

P. J.

(2010). Regional response differences across the human amygdaloid complex during social conditioning. Cerebral Cortex, 20, 612–621. doi:10.1093/cercor/bhp126

10.

Deacon

B. J.

Abramowitz

J. S.

(2004). Cognitive and behavioral treatments for anxiety disorders: A review of meta-analytic findings. Journal of Clinical Psychology, 60, 429–441. doi:10.1002/jclp.10255

11.

Duits

Cath

D. C.

Lissek

Hox

J. J.

Hamm

A. O.

Engelhard

I. M.

… Baas

J. M. P.

(2015). Updated meta-analysis of classical fear conditioning in the anxiety disorders. Depression and Anxiety, 32, 239–253. doi:10.1002/da.22353

12.

Foa

E. B.

Kozak

M. J.

(1986). Emotional processing of fear: Exposure to correct information. Psychological Bulletin, 99, 20–35.

13.

Gotlib

I. H.

Kasch

K. L.

Traill

Arnow

B. A.

Joormann

Johnson

S. L.

(2004). Coherence and specificity of information-processing biases in depression and social phobia. Journal of Abnormal Psychology, 113, 386–398. doi:10.1037/0021-843X.113.3.386

14.

Haddad

A. D. M.

Lissek

Pine

D. S.

Lau

J. Y. F.

(2011). How do social fears in adolescence develop? Fear conditioning shapes attention orienting to social threat cues. Cognition and Emotion, 25, 1139–1147. doi:10.1080/02699931.2010.524193

15.

Hamilton

J. P.

Chen

M. C.

Waugh

C. E.

Joormann

Gotlib

I. H.

(2013). Distinctive and common neural underpinnings of major depression, social anxiety, and their comorbidity. Social Cognitive and Affective Neuroscience, 10, 552–560. doi:10.1093/scan/nsu084

16.

Hermann

Ziegler

Birbaumer

Flor

(2002). Psychophysiological and subjective indicators of aversive Pavlovian conditioning in generalized social phobia. Biological Psychiatry, 52, 328–337. doi:10.1016/S0006-3223(02)01385-9

17.

Hox

J. J.

(2010). Multilevel Analysis: Techniques and Applications (2nd ed.). New York: Routledge.

18.

Langer

J. K.

Tonge

N. A.

Piccirillo

Rodebaugh

T. L.

Thompson

R. J.

Gotlib

I. H.

(2019). Symptoms of social anxiety disorder and major depressive disorder: A network perspective. Journal of Affective Disorders, 243, 531–538. doi:10.1016/j.jad.2018.09.078

19.

Ledoux

J. E.

Hofmann

S. G.

(2018). The subjective experience of emotion: A fearful view. Current Opinion in Behavioral Sciences, 19, 67–72. doi:10.1016/j.cobeha.2017.09.011

20.

Lissek

Levenson

Biggs

A. L.

Johnson

L. L.

Ameli

Pine

D. S.

Grillon

(2008). Elevated fear conditioning to socially relevant unconditioned stimuli in social anxiety disorder. American Journal of Psychiatry, 165, 124–132.

21.

Lonsdorf

T. B.

Merz

C. J.

(2017). More than just noise: Inter-individual differences in fear acquisition, extinction and return of fear in humans—Biological, experiential, temperamental factors, and methodological pitfalls. Neuroscience and Biobehavioral Reviews, 80, 703–728. doi:10.1016/j.neubiorev.2017.07.007

22.

Lovibond

(2007). Fear and avoidance: An integrated expectancy model. In Craske

M. G.

Hermans

Vansteenwegen

(Eds.), Fear and learning: From basic processes to clinical implications (pp. 117–132). Washington, DC: American Psychological Association. doi:10.1037/11474-006

23.

Lovibond

P. F.

Lovibond

S. H.

(1995). The structure of negative emotional states: Comparison of the Depression Anxiety Stress Scales (DASS) with the beck depression and anxiety inventories. Behaviour Research and Therapy, 33, 335–343. doi:10.1016/0005-7967(94)00075-U

24.

Mattick

Clarke

(1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behavior Research and Therapy, 36, 70. doi:10.1016/S0005-7967(97)10031-6

25.

McCabe

R. E.

Antony

M. M.

Summerfeldt

L. J.

Liss

Swinson

R. P.

(2003). Preliminary examination of the relationship between anxiety disorders in adults and self-reported history of teasing or bullying experiences. Cognitive Behaviour Therapy, 32, 187–193. doi:10.1080/16506070310005051

26.

Miers

A. C.

Blöte

A. W.

Heyne

D. A.

Westenberg

P. M.

(2014). Developmental pathways of social avoidance across adolescence: The role of social anxiety and negative cognition. Journal of Anxiety Disorders, 28, 787–794. doi:10.1016/j.janxdis.2014.09.008

27.

Mineka

Zinbarg

(2006). A contemporary learning theory perspective on the etiology of anxiety disorders: It’s not what you thought it was. American Psychologist, 61, 10–26. doi:10.1037/0003-066X.61.1.10

28.

Pearce

J. M.

Hall

(1980). A model for Pavlovian learning: Variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychological Review, 87, 532–552. doi:10.1037/0033-295X.87.6.532

29.

Pejic

Hermann

Vaitl

Stark

(2013). Social anxiety modulates amygdala activation during social conditioning. Social Cognitive and Affective Neuroscience, 8, 267–276. doi:10.1093/scan/nsr095

30.

Quené

Van Den Bergh

(2004). On multi-level modeling of data from repeated measures designs: A tutorial. Speech Communication, 43, 103–121. doi:10.1016/j.specom.2004.02.004

31.

R Core Team. (2019). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from https://www.R-project.org/

32.

Rescorla

R. A.

Wagner

A. R.

(1972). A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and nonreinforcement. In Black

A. H.

Prokasy

W. F.

(Eds.), Classical conditioning II: Current research and theory (pp. 64–99). New York: Appleton-Century-Crofts. doi:10.1101/gr.110528.110

33.

Rodebaugh

T. L.

Heimberg

R. G.

Brown

P. J.

Fernandez

K. C.

Blanco

Schneier

F. R.

Liebowitz

M. R.

(2011). More reasons to be straightforward: Findings and norms for two scales relevant to social anxiety. Journal of Anxiety Disorders, 25, 623–630. doi:10.1016/j.janxdis.2011.02.002

34.

Royston

Altman

D. G.

Sauerbrei

(2006). Dichotomizing continuous predictors in multiple regression: A bad idea. Statistics in Medicine, 25, 127–141. doi:10.1002/sim.2331

35.

Schlenker

B. R.

Leary

M. R.

(1982). Social anxiety and self-presentation: A conceptualization model. Psychological Bulletin, 92, 641–669. doi:10.1037/0033-2909.92.3.641

36.

Schneider

Weiss

Kessler

Müller-Gärtner

H. W.

Posse

Salloum

J. B.

… Birbaumer

(1999). Subcortical correlates of differential classical conditioning of aversive emotional reactions in social phobia. Biological Psychiatry, 45, 863–871. doi:10.1016/S0006-3223(98)00269-8

37.

Slavich

G. M.

O’Donovan

Epel

E. S.

Kemeny

M. E.

(2010). Black sheep get the blues: A psychobiological model of social rejection and depression. Neuroscience and Biobehavioral Reviews, 35, 39–45. doi:10.1016/j.neubiorev.2010.01.003

38.

Spence

S. H.

Rapee

R. M.

(2016). The etiology of social anxiety disorder: An evidence-based model. Behaviour Research and Therapy, 86, 50–67. doi:10.1016/j.brat.2016.06.007

39.

Stein

M. B.

Walker

J. R.

Forde

D. R.

(1994). Setting diagnostic thresholds for social phobia: Considerations from a community survey of social anxiety. American Journal of Psychiatry, 151, 408–412. doi:10.1176/ajp.151.3.408

40.

Thorwart

Livesey

E. J.

(2016). Three ways that non-associative knowledge may affect associative learning processes. Frontiers in Psychology, 7, 1–15. doi:10.3389/fpsyg.2016.02024

41.

Tinoco-González

Fullana

M. A.

Torrents-Rodas

Bonillo

Vervliet

Pailhez

… Torrubia

(2015). Conditioned subjective responses to socially relevant stimuli in social anxiety disorder and subclinical social anxiety. Clinical Psychology and Psychotherapy, 22, 221–231. doi:10.1002/cpp.1883

42.

Veit

Flor

Erb

Hermann

Lotze

Grodd

Birbaumer

(2002). Brain circuits involved in emotional learning in antisocial behavior and social phobia in humans. Neuroscience Letters, 328, 233–236. doi:10.1016/S0304-3940(02)00519-0

43.

Vervliet

Craske

M. G.

Hermans

(2013). Fear extinction and relapse: State of the art. Annual Review of Clinical Psychology, 9, 215–248. doi:10.1146/annurev-clinpsy-050212-185542

44.

Wiggert

Wilhelm

F. H.

Boger

Georgii

Klimesch

Blechert

(2017). Social Pavlovian conditioning: Short- and long-term effects and the role of anxiety and depressive symptoms. Social Cognitive and Affective Neuroscience, 12, 329–339. doi:10.1093/scan/nsw128

45.

Wiggert

Wilhelm

F. H.

Reichenberger

Blechert

(2015). Exposure to social-evaluative video clips: Neural, facial-muscular, and experiential responses and the role of social anxiety. Biological Psychology, 110, 59–67. doi:10.1016/j.biopsycho.2015.07.008

Social anxiety compared to depression better accounts for enhanced acquisition of self-reported anxiety toward faces paired with negative evaluation in a conditioning task

Abstract

Keywords

Introduction

Method

Participants and procedure

Materials

Depression questionnaire

Social anxiety questionnaire

Conditioning task

Results

Subjective ratings of anxiety as the dependent variable

Subjective ratings of probability of hostile reaction as the dependent variable

Reaction time for hostile reaction (RThostile) as the dependent variable

Post hoc power analysis

Discussion

Footnotes

Author’s note

Acknowledgments

Data accessibility statement

Declaration of Conflicting Interests

Funding

ORCID iD

References

Reaction time for hostile reaction (RT_hostile) as the dependent variable