EXAMINING THE IMPACT OF AFFECT LABELING ON EXPOSURE THERAPY FOR PUBLIC SPEAKING ANXIETY.

EXAMINING THE IMPACT OF AFFECT LABELING ON EXPOSURE THERAPY FOR PUBLIC

SPEAKING ANXIETY

Chapter 1 Introduction

Behavioral treatments for anxiety disorders, particularly exposure therapy, are highly

effective in reducing anxiety symptoms (Butler, Chapman, Forman, & Beck, 2006; Hofmann &

Smits, 2008; Tolin, 2010). However, many patients experience limited improvement, drop out of

treatment, or relapse (Loerinc, Meuret, Twohig, Rosenfield, & Craske, submitted for

publication). Exposure therapy, considered the gold standard treatment for anxiety disorders, is

thought to mimic extinction training observed in fear conditioning research. Despite significant

advances in understanding fear extinction in laboratory settings, these findings have not been

widely integrated into clinical practices for treating anxiety disorders.

Given the need for improved treatments, this project aims to bridge the gap between

basic science and intervention research by examining methods to enhance the efficacy of

exposure therapy for public speaking anxiety. Specifically, the study compares the effectiveness

of exposure therapy alone versus exposure combined with affect labeling in reducing anxiety

during public speaking.

Public speaking anxiety is among the most prevalent psychological disorders in the

United States, with estimates ranging from 11% to 30% of the population (Pollard & Henderson,

1988; Stein, Walker, & Forde, 1996; Wittchen, Stein, & Kessler, 1999). For some individuals, it

can be profoundly debilitating, particularly in academic or professional settings where public

presentations are essential. Research indicates that public speaking anxiety correlates with lower

income, reduced educational attainment, and higher unemployment rates (Stein et al., 1996).

Traditional treatments for public speaking anxiety typically combine exposure therapy

(e.g., practicing public speaking) with cognitive restructuring, which encourages patients to

reinterpret feared situations neutrally or positively (Heimberg, 2002; Hofmann & Smits, 2008;

Hope, Heimberg, Juster, & Turk, 2000; Rapee & Heimberg, 1997). Recently, affect labeling—

naming one’s emotions in response to a feared stimulus—has emerged as a promising approach

to emotion regulation (Lieberman et al., 2007). Evidence suggests that affect labeling may

enhance the effectiveness of exposure therapy (Kircanski, Lieberman, & Craske, 2012).

Although exposure therapy alone has demonstrated efficacy in treating social anxiety

disorder (Feske & Chambless, 1995), the potential additive benefits of combining it with affect

labeling for public speaking anxiety remain unexplored. This study seeks to address this gap by

evaluating whether affect labeling can augment exposure therapy's effectiveness, ultimately

contributing to improved interventions for individuals with public speaking anxiety.

Fear and Social Anxiety Disorder

The experience of fear is an adaptive mechanism found in many species and helps an

organism respond to threats in the environment. Fear is characterized by autonomic activity,

subjective experience of distress, and behavioral responses. While one of the most common

experiences of fear occurs in the context of threat to the physical self (e.g. possibility of pain or

injury), humans can also experience an anxiety response during social threat (e.g. possibility of

rejection) (Dickerson, Gruenewald, & Kemeny, 2004; Dickerson & Kemeny, 2004). In fact, the

same neural regions activated by physical pain are activated following social rejection

(Eisenberger & Lieberman, 2004). Maintaining social connections with others is one of the

strongest human motivations (Baumeister & Leary, 1995), which suggests that doing so may

have been adaptive for survival in early humans. Rejection by one’s group may have posed a

significant threat to survival in an evolutionary context, and therefore, humans likely evolved

mechanisms to avoid being rejected from social groups. Therefore, the experience of anxiety,

fear, or shame in situations during which rejection is possible is an adaptive response to social

threat.

For many people, anxiety associated with social threat is excessive and interferes with

the ability to interact with others and perform in social contexts. Anxiety of this magnitude is

diagnosed as social anxiety disorder. Excessive anxiety that occurs only during public speaking

or performance situations is categorized as non-generalized social anxiety and is far more

common than the generalized subtype in which anxiety occurs in multiple social contexts

(Pollard & Henderson, 1988). Speaking in front of an audience is the epitome of social threat

because the attention of the audience is focused solely on the speaker, and there is the possibility

that the audience will dislike or disagree with what the speaker presents. Frequently, those

viewing the speech are peers, colleagues or superiors who may be important to one’s social

network, future career, or educational success. The high prevalence of public speaking fears

may be better understood given the potential for negative evaluation inherent in public speaking.

1.1. Indices of Fear Responding

The defensive motivational system (Masterson & Crawford, 2010) is a system of

behaviors activated in response to environmental threat. Defensive responding includes

autonomic activation, subjective discomfort, and avoidance behavior. The amygdala is a key

region in the experience of and response to threat and has been identified as the mediator

between detection of and response to threat (LeDoux, 2000). In addition, the amygdala plays an

important role in fear learning, memory and extinction (Knight, Smith, Cheng, Stein, &

Helmstetter, 2004; Öhman & Mineka, 2001). Patients with social anxiety disorder show greater

amygdala activity in response to social threat (e.g. angry faces) (Stein, Goldin, Sareen, Zorrilla,

& Brown, 2002), as well as non-social negative stimuli (Brühl et al., 2011; Shah, Klumpp,

Angstadt, Nathan, & Phan, 2009).

1.1.1. Autonomic. The experience of threat in the environment activates the sympathetic

nervous system, which prepares the body and mind to respond to threat. Sympathetic activation

is generally thought of as the “fight or flight” response, and includes changes in heart rate (HR)

and sweat gland activity or galvanic skin response (GSR). HR and GSR have been shown to

increase in response to an unconditioned threat stimulus (Deane, 1969) and during public

speaking (Beatty & Behnke, 1991; Myers, 1974); however decreases in HR (fear brachycardia)

also occur in response to threat due to increased attentional focus (Bradley, Codispoti, Cuthbert,

& Lang, 2001). Autonomic activity during laboratory stressors has also been shown to decrease

following treatment for panic disorder (Craske, Lang, Aikins, & Mystkowski, 2005). GSR in

particular is a measure of sympathetic activation as sweat gland activity is stimulated by the

sympathetic nervous system. HR receives input from the sympathetic and parasympathetic

nervous systems, and therefore, while increases in HR indicate sympathetic activity, HR is not a

pure measure of sympathetic nervous system activation (Hugdahl, 1995).

1.1.2. Self report. Fear responding also includes an individual’s subjective response

such as distress and negative cognitions in feared situations. Individuals with social anxiety

report higher levels of anxiety and more negative cognitions during public speaking than non-

anxious controls (Hofmann & DiBartolo, 2000; Levin et al., 1993). In addition, self-reported

anxiety and negative cognitions decrease after completion of treatment (Craske et al., In Press)

and over the course of treatment (Niles, Mesri, Burklund, Lieberman, & Craske, 2013). Given

that subjective experiences of anxiety and distress are one of the primary targets in treatment,

self-report is an essential indicator of fear responding.

1.1.3. Behavior. Although physiological and cognitive experiences during threat help an

individual respond appropriately (e.g. fight or flight), anxiety disorders are characterized by

maladaptive avoidance of feared stimuli and situations including escape, retreat, caution or

thought suppression (Craske, 2003). Avoidance can negatively impact work or academic

performance, social relationships, or other important domains, and immediate decreases in

autonomic arousal and subjective discomfort as a result of avoidance reinforce avoidance over

the long term.

1.2. Importance of Enhancing Exposure Effectiveness

One of the key ingredients in treatment for social anxiety disorder is repeated exposure to

feared and avoided social situations (Feske & Chambless, 1995; Rodebaugh, Holaway, &

Heimberg, 2004). Exposure treatment is thought to be analogous to fear extinction in laboratory

studies in which subjects are conditioned to fear a cue that signals the onset of a threatening

stimulus, then subsequently exposed to the cue in the absence of the threat (Bouton, 1993). With

repeated exposure to the cue, subjects show a reduction in fear over time (Watson, 1970).

However, while exposure therapy is highly effective in the treatment of social anxiety

disorder, many patients do not complete treatment, do not benefit from exposure, or show a

return of fear after completion of behavioral treatment (Loerinc et al., Submitted for

Publication). Therefore, there is a need to enhance the effects of exposure with the ultimate

goal of improving treatment outcomes and preventing relapse.

Laboratory studies of fear extinction in animals and humans have also shown a

resurgence of fear after fear has been extinguished (Craske & Mystkowski, 2006; Ricker &

Bouton, 1996). This can occur simply after a period of time (Bouton, Woods, Moody, Sunsay,

& García-Gutiérrez, 2006), after re-experiencing of the feared stimulus or reinstatement

(Rescorla & Heth, 1975) and exposure to a different context or renewal (Bouton, 1993). These

phenomena as observed in laboratory studies likely mirror relapse in humans and highlight a

need to find ways of enhancing exposure effectiveness to ultimately reduce relapse rates (Craske

& Mystkowski, 2006).

1.3. Using Language to Enhance Exposure Effectiveness

Although studies of extinction in laboratory animals may simulate exposure for anxiety

disorders, humans have the ability to use language to regulate emotional experiences in ways that

other animals do not. Language is generally used alongside exposure in an attempt to help

patients better understand why the situation is feared and to create more realistic interpretations

of the situation. Although behavioral treatments generally include a cognitive component,

whether this linguistic strategy actually improves the effectiveness of extinction has not been

experimentally assessed in controlled laboratory trials.

Lieberman’s disruption theory of language and emotion (Lieberman, 2003, 2011) posits

that labeling one’s emotional state can disrupt the experience of the emotion. Research in

functional neuroimaging, psychophysiology, and behavior provide evidence that affect labeling

is a promising approach for enhancing emotion regulation during exposure. Linguistic processes

activate areas of the prefrontal cortex, such as the right ventrolateral prefrontal cortex

(RVLPFC) and the medial prefrontal cortex (MPFC), which corresponds to decreases in

activation of limbic emotional response regions such as the amygdala. Activation of the

amygdala in turn affects physiological arousal and subjective experiences of emotion (Kim et al.,

2011), and therefore reductions in amygdala activation should correspond with decreases in

autonomic arousal and subjective experience of distress.

Studies of the neural circuitry underlying extinction suggest that the extent to which

prefrontal regions are activated, and the strength of connectivity between prefrontal regions and

the amygdala affect the success of extinction training. Electrical stimulation of the mPFC led to

reduction of conditioned responding after fear conditioning in rats (Milad & Quirk, 2002).

Greater cortical thickness in the mPFC is associated with better extinction effects in humans

(Hartley, Fischl, & Phelps, 2011; Milad et al., 2005), and greater mPFC activity is associated

with better extinction effects of learned US-CS associations (Delgado, Nearing, LeDoux, &

Phelps, 2008; Milad et al., 2005; Phelps, Delgado, Nearing, & LeDoux, 2004). These findings

suggest that down regulation of amygdala activity by mPFC is responsible for successful

extinction. In addition, evidence suggests that patients with social anxiety disorder have weaker

connectivity between the mPFC and the amygdala (Hahn et al., 2011; Kim et al., 2011).

Therefore, treatments that strengthen connectivity between these regions may prove beneficial

in the treatment of social anxiety. The principle of neural plasticity states that repetition of a

process can increase efficiency and efficacy of that process through changes in neuron function,

chemical profile, and structure (Anderson, 2010; Kandel & Schwartz, 1982). Therefore, verbal

processing may enhance connectivity in PFC-Amygdala pathways therefore improving patients’

ability to regulate emotional responses.

1.3.1. Methods of verbalization to activate the RVLPFC. Verbalization such as

cognitive reappraisal is emphasized in treatments for anxiety disorders (Craske, Antony, &

Barlow, 2006; Novalis, Rojcewicz, & Peele, 1993). Reappraisal is a deliberate attempt to reduce

emotional responding by generating neutral or positive (as opposed to negative) interpretations

of a situation. Reappraisal is consistent with cognitive restructuring in cognitive behavioral

therapies for anxiety disorders in which patients are taught to think more flexibly about feared

stimuli. Affect labeling, another method of verbalization, is the process of identifying

and labeling emotional experiences (Lieberman, 2011; Pennebaker, 1997), and has been shown

to reduce affective responding to negative stimuli (Lieberman, Inagaki, Tabibnia, & Crockett,

2011). It can be argued that negative thoughts and emotions must first be identified before they

can be modified, so affect labeling is also the first step to cognitive reappraisal (Arch & Craske,

2008). Affect labeling has been categorized as an incidental emotion regulation strategy because

affective change is a consequence of labeling but not an explicit goal (Burklund, Creswell, Irwin,

& Lieberman, 2014). Although both these methods of emotion regulation have been shown to

reduce affective responding to unpleasant stimuli (Lieberman et al., 2011), only one study to

date has examined whether using affect labeling and reappraisal during exposure results in

greater fear reduction at retest compared to exposure alone.

Kircanski, Lieberman & Craske (2012) compared the effects of exposure to a live spider

with and without linguistic processing in spider fearful subjects. Participants were randomly

assigned to use reappraisal or affect labeling during exposure or to complete exposure without

verbalization. At re-test the group that completed exposure with affect-labeling had lower

autonomic activity while viewing a spider and moved closer to the spider compared to the

reappraisal and exposure alone groups. In addition, those who used the greatest number of

anxiety and fear related words during affect labeling showed the greatest reductions in fear

responding. This study provides evidence that affect labeling rather than reappraisal may be a

more promising approach to enhancing the effectiveness of exposure. Therefore, the current

project compared exposure plus affect labeling to exposure alone to specifically test affect

labeling as an augmentation strategy for exposure therapy.

1.3.2. Behavioral research on affect labeling. Many studies have demonstrated that

expressing emotions through language is beneficial for psychological wellbeing, physical health,

and cognitive performance. Pennebaker and Beall (1986) were some of the first researchers to

demonstrate through a controlled study that writing about stressful experiences is beneficial for

physical health. Subsequently, hundreds of researchers have assessed the effects of expressive

writing about a wide variety of experiences and within many different populations. Cognitive

processing therapy (Resick & Schnicke, 1993), an effective treatment for post-traumatic stress

disorder, involves systematic written exposure about one’s traumatic experience and the evoked

emotions. Studies that have examined the effect of written or verbal processing of trauma

suggest that linguistic processing of traumatic experiences is beneficial for depression and

anxiety (Hemenover, 2003; Lepore, Silver, Wortman, & Wayment, 1996) and reduces stress

responses to trauma related memories (Smyth, Hockemeyer, & Tulloch, 2008). Studies have also

demonstrated that recording worries about an upcoming exam in writing increases exam

performance (Frattaroli, Thomas, & Lyubomirsky, 2011; Ramirez & Beilock, 2011), and that

writing about negative events decreases intrusive thoughts and increases working memory (Klein

& Boals, 2001). Finally, Lieberman and colleagues (2011) compared participants’ distress while

viewing negative IAPS images with and without affect labeling and found lower reported

distress in the labeling condition despite participants’ predictions that affect labeling would

increase distress.

1.3.3. Neuroimaging and physiological research on affect labeling. Lieberman and

colleagues (2007), demonstrated that when participants label the emotional expression of a face

(e.g. “angry”), they show reduced amygdala activation, and increased activation in the RVLPFC

compared to viewing faces without labeling, or labeling the gender of the face. A significant

number of neuroimaging studies have now demonstrated that labeling one’s emotional

experience activates areas of the prefrontal cortex, and reduces activation in the amygdala

(Gorno-Tempini et al., 2001; Hariri, Bookheimer, & Mazziotta, 2000; Hariri, Mattay, Tessitore,

Fera, & Weinberger, 2003; Narumoto et al., 2000). These findings suggest that downregulation

of amygdala activation occurs through projections from prefrontal regions.

Tabibnia and colleagues (2008) examined the effect of repeated exposure to evocative

images with and without negative affective labels. In study 1 in a non-clinical sample, repeated

presentation of emotionally evocative images paired with an affect label resulted in greater

attenuation of GSR upon presentation of the images without a label at 1-week re-test. In study 2,

the effects were replicated in a spider-fearful sample undergoing exposure to images of spiders.

Participants who went through exposure paired with negative labels showed greater attenuation

of GSR at 1-week re-test compared to those who saw no labels or those who saw neutral labels.

1.4. Moderators of Response to Labeling Versus Exposure Alone

Better matching of treatments to individuals can improve therapy outcomes. However,

researchers have yet to identify whether patients who have deficits in a particular skill are more

likely to benefit from treatments that target that skill (deficit model), or whether those who

already gravitate towards a coping approach will benefit most from a treatment that matches that

approach (matching model). Evidence from the expressive writing literature supports both

models with some studies showing that participants high in trait emotional expressivity benefit

most from expressive writing (Austenfeld, Paolo, & Stanton, 2006; Niles, Haltom, Mulvenna,

Lieberman, & Stanton, 2014; Stanton, Kirk, Cameron, & Danoff-Burg, 2000), while others show

that participants with deficits in emotional expression benefit most from expressive writing (Lu

& Stanton, 2009; Páez, Velasco, & González, 1999). The current study will examine whether

the matching model can be used to predict who will benefit most from exposure combined with

affect labeling.

1.5. Summary

Fear of public speaking is common and in some cases can cause significant impairment

in functioning. Exposure is an effective treatment method for public speaking anxiety, and is

often combined with cognitive restructuring or other methods of verbalization, but few studies

have examined whether verbalization can enhance the effectiveness of exposure. Evidence from

neuroscience suggests that in anxiety disorders, there are deficits in connectivity between regions

of the prefrontal cortex and the amygdala, which may explain deficits in regulating anxious

responding in these individuals. Affect labeling reduces self-reported distress while viewing

negative images, and when combined with exposure in spider fearful subjects, produces greater

fear attenuation for physiological measures of arousal compared to exposure alone. Affect

labeling activates pathways between prefrontal regions and the amygdala, and therefore,

combining exposure with affect labeling may strengthen prefrontal-amygdala connections

thereby enhancing long-term fear reduction.

1.6. Specific Aims and Hypotheses

The first aim of the study will be to assess whether affect labeling enhances the

effectiveness of exposure compared to exposure alone. We hypothesize that participants who

use affect labeling during exposure will show greater attenuation of fear of public speaking

compared to those who undergo exposure alone. The second aim of the study will be to assess

whether the number of anxiety or fear related words used during exposure predicts greater

attenuation of fear responding at re-test. We hypothesize that participants who use more anxiety

or fear related words compared to other emotion words will show the greatest fear reduction at

re-test. The third aim of the study will be to assess whether individual differences in incidental

emotion regulation (i.e. the extent to which affect labeling reduces distress) at baseline moderate

response to exposure with affect labeling versus exposure alone. Consistent with the matching

model, we hypothesize that participants who show strengths in incidental emotion regulation at

baseline will show greater fear attenuation in the labeling condition than in the exposure alone

condition.

CHAPTER 2

Method

2.1 Overview of Design, Independent Variables, and Procedure

This study used a 2 (Group) × 3 (Time) mixed design with speech fearful participants.

Groups included an exposure combined with affect labeling condition, and an exposure alone

condition, and Time was three assessment time-points at baseline (Time 1), following exposure

(Time 2) and at one-week follow-up (Time 3). To test Aim 1, the labeling condition was

compared to the exposure alone condition on fear attenuation immediately following exposure

(Time 2) and at one-week follow-up (Time 3). To test Aim 2, the number of anxiety related

emotion words used by participants during exposure was used to predict fear responding

following exposure (Time 2) at one-week follow-up (Time 3). To test Aim 3, incidental emotion

regulation at baseline was examined as a moderator of response to exposure plus affect labeling

versus exposure alone.

For a flowchart of study procedures, see Figure 1. On day one (Time 1), participants first

completed a series of questionnaires followed by a behavioral approach task (BAT-1) during

which autonomic arousal, self-reported affect, cognitions, self-rated performance, and behavioral

avoidance was assessed. Participants next completed an affect labeling task to assess incidental

emotion regulation. Participants were then assigned to the labeling or exposure alone conditions.

Participants were moved to a different room, and completed 10 exposure trials (Exp-1) either

with or without labeling depending on group assignment. Time 2 was three days later, and

participants again underwent 10 exposure trials (Exp-2) with or without labeling depending on

group assignment. After the exposures, participants moved to the room in which the original

BAT was conducted, and underwent a second BAT using the same protocol as Time 1 (BAT-2).

Time 3 was eight days later, and participants again completed the BAT using the same method as

at Time 1 and Time 2 (BAT-3).

2.2. Participants

One hundred two participants (AL = 52; Control = 50) were recruited to participate. Two

participants assigned to the control group were not included in analyses. Due to experimenter

error, one participant received the affect labeling exposure protocol rather than the control

protocol on the second day of exposure. The other participant appeared to be answering

randomly on questionnaires and fell asleep during the experiment. Therefore, the final sample

included in analyses was 100. See Figure 2 for a consort diagram of flow through study

procedures. Eligible participants reported a six or higher on anxiety and a five or higher on

avoidance of public speaking on a zero to eight scale (see Appendix A). Participants were

recruited from the UCLA Psychology Subject Pool and using flyers posted around campus.

Participants were given 1 hour of research credit per day of participation or were paid $10.00 per

day if they were not enrolled in a course that required research credit. Participants therefore

received 3 research credits or $30.00 for completing all three days.

Participants were over 18 years of age, fluent in English, free of heart, neurological, or

respiratory conditions, hearing impairment, physician recommendation to avoid stressful

situations, current treatment for public speaking anxiety, or psychotropic medication

prescription for an emotional problem. These participants were excluded due to potential

interference with psychophysiological measurement and for safety precautions.

2.3. Materials

A summary of study measures is included in Table 1. Questionnaires are included in Appendix

2.3.1 Indices of fear responding.

2.3.1.1. Physiological activity. Physiological activity recording through the Biopac

system was facilitated using an IBM Pentium II and AcqKnowledge software (AcqKnowledge

4.1 for Windows; BIOPAC Systems, inc). Galvanic skin response (GSR) was recorded from

electrodes attached to the distal phalanges of the second and third finger of a participant’s non-

preferred hand. Heart rate (HR) was recorded from electrodes attached below the right

collarbone and bottom left rib. Physiological data from the final one-minute of the two-minute

baseline (baseline), the one-minute anticipation period prior to the speech (anticipation), and the

one-minute recovery period after completion of the speech (recovery) were analyzed.

All physiological data were first visually inspected to ensure proper measurement. For

HR, one participant at Time 1, and two participants at Time 2 were excluded from analysis due

to recording error. For GSR, nine participants at Time 1, 10 participants at Time 2, and 6

participants at Time 3 were excluded from analysis because no variations in GSR signal were

observed. These numbers are consistent with estimates that approximately 10% (or more in

clinical samples) of individuals do not show a reliable GSR response (Braithwaite, Watson,

Jones, & Rowe, 2013). To assess HR, electrocardiogram signals were collected from two

electrodes, one on the participant’s right clavicle, and one below the bottom left front rib. HR

was defined as the number of heart beats per minute. A band pass filter with a low cutoff of

1.00Hz and a high cutoff of 35.00Hz was applied prior to analysis to limit the effect of signal

noise on the data. GSR was measured using two indices: Skin conductance level (SCL) and

non- specific skin conductance response (SCR-NS). Skin conductance level was defined as the

average skin conductance level over each one-minute assessment period and was measured to the

nearest microsiemen (μs). SCR-NS was assessed by calculating the frequency of non-specific

skin conductance responses during each one-minute period. A skin conductance response was

defined by a minimum increase of .02 μs. Data were analyzed using built in analysis tools in

AcqKnowledge software for HR, SCL, and SCR-NS.

2.3.1.2 Personal Report of Public Speaking Anxiety (PRPSA; McCroskey, 1970). See

Appendix A. PRPSA is a 34-item measure that assesses fear of public speaking. Responders

rate their degree of agreement with each statement on a 5-point Likert scale (1=strongly disagree

to 5=strongly agree). The scale has excellent reliability (α=.90) (McCroskey, 1970). Scores of

131 or higher indicate high public speaking anxiety, scores of 98 – 130 indicate moderate

anxiety and scores below 98 indicate low public speaking anxiety. In the current study, α = .97

at Time 1.

2.3.1.3. Subject Units of Distress Scale (SUDS). SUDS is a single item measure used to

assess state anxiety. Participants were shown a 0 to 8 Likert scale with 0 indicating no anxiety

and 8 indicating extreme anxiety. Participants were then asked to report their SUDS ratings at

various points in the study. SUDS ratings were taken directly prior to and following speech

tasks, and when reported, SUDS ratings are the average of the ratings before and after the

speech.

2.3.1.4. Self Statements During Public Speaking (SSPS; Hofmann & DiBartolo, 2000).

See Appendix A. The 10-item SSPS assesses negative and positive cognitions and was

completed immediately after the public speaking task; participants rated the extent to which they

experienced five negative and five positive thoughts during the speaking task. The SSPS shows

good internal consistency (α = .86) and test-retest reliability (r = .80) (Hofmann & DiBartolo,

2000). In the current study α = .78 for positive cognitions and .84 for negative cognitions at

Time 1.

2.3.1.5. Optional Speech. At each BAT, participants had the option of speaking on an

additional speech topic. Participants received $5.00 for completing the additional speech.

2.3.2. Other Measures.

2.3.2.1. Performance Rating Form (PRF; Rapee & Lim, 1992). See Appendix A. This

17-item questionnaire assesses self-reported public speaking performance. The scale includes 12

specific items to represent individual behaviors or reactions necessary to good public speaking

(e.g. kept eye contact with audience, had a clear voice), and 5 global items to assess overall

evaluations of performance (e.g. generally spoke well). Participants rate items on a 5-point

Likert scale from 0 (not at all) to 4 (very much) and higher scores indicate better rated

performance. In the current study, α = .91 at Time 1.

2.3.2.2. Patient Health Questionnaire (PHQ; Spitzer, Kroenke, & Williams, 1999). See

Appendix A. The PHQ is a widely used 9-item questionnaire used to assess symptoms of

depression. The PHQ is a valid measure of depressive symptoms (Löwe, Spitzer, et al., 2004)

and is sensitive to change (Löwe, Kroenke, Herzog, & Gräfe, 2004). Patients report the

frequency of experiencing each symptom on a four-point Likert scale. In the current study, α =

.84 at Time 1.

2.3.2.3. Exposure Credibility Questionnaire. See Appendix A. This four-item measure

includes 0-8 Likert scale ratings assessing participants’ perceived credibility of the exposure

task. This measure has been used in previous exposure studies as a potential predictor of study

drop out (Craske, Street, Jayaraman, & Barlow, 1991). In the current study, α = .83 at Time 1.

2.3.2.4. Word use during exposure. During exposures, participants assigned to the

verbalization group chose from a set of emotion words displayed on the computer screen using

the keyboard. Their responses were recorded, and the number of anxiety related emotions was

identified for each participant.

2.3.2.5 Mini Social Phobia Inventory (SPIN; Connor, Kobak, Churchill, Katzelnick, &

Davidson, 2001). See Appendix A. This three-item measure assesses symptoms of social

anxiety disorder and is based on a longer 17-item version of the scale. Participants respond to

items on a 0 (not at all) to 4 (extremely) Likert scale. Scores of 6 or higher indicate possible

problems with social anxiety disorder. The scale demonstrates 90% accuracy in identifying a

diagnosis of generalized social anxiety disorder. In the current study α = .86 at Time 1.

2.3.2.6. Mini International Neuropsychiatric Interview (MINI; Lecrubier et al., 1997).

The MINI is a brief fully structured diagnostic interview that can be administered by non-

specialized interviewers. Kappa, specificity, and sensitivity for the MINI compared to diagnoses

obtained on the Composite International Diagnostic Interview are good or very good for all but

three diagnoses (generalized anxiety disorder, agoraphobia, and bulimia), and inter-rater and

test- retest reliability are good (Lecrubier et al., 1997). The current project used only the social

anxiety disorder section of the MINI to determine whether participants met diagnostic criteria for

social anxiety disorder.

2.3.2.7. Incidental Emotion Regulation. Incidental emotion regulation was assessed

using the Affect Labeling Task (Lieberman et al., 2011), which assessed how effective

participants are at decreasing distress using affect labeling. Incidental emotion regulation was

calculated by subtracting participants’ average level of distress when labeling negative images

from their average level of distress when viewing negative images without labeling. Scores

ranged from -1.4 to 2.0, and higher incidental emotion regulation scores indicated more

effective emotion regulation. For more details on the Affect Labeling Task, see procedure

below.

2.4. Procedure

Participants who were eligible to participate were scheduled for three appointment times

on days 1, 4 and 9. Appointment times were in the afternoon between 4:00 and 6:00 pm.

Participants were scheduled at the same time on each of the three days. On day 1,

participants completed the Affect Labeling Task, the first BAT (BAT-1), and the first exposure

session (Exp- 1). On day 4, participants completed the second exposure session (Exp-2) and the

second BAT (BAT-2). On day 9, participants completed the third BAT (BAT-3). See Figure 1

for a diagram of study procedures.

2.4.1. Time 1.

2.4.1.1. BAT-1 (30 minutes). On day 1, participants were located in a small room on the

A Level of Franz Hall. When participants arrived, they were consented. After being consented,

participants were interviewed using the Social Anxiety Disorder section of the MINI

Neuropsychiatric Interview by trained research assistants. Following completion of the

interview, electrodes were attached to participants to begin recording by the Biopac

physiological measurement system. Physiological measurement included HR and GSR.

Although physiological activity was recorded throughout the BAT, only measurements taken

during the baseline, anticipation, and recovery periods were analyzed.

Participants then reported on demographic characteristics and completed the PRPSA,

PHQ, and SPIN while acclimating to the environment for approximately 10 minutes. After

completion of the questionnaires, the two-minute baseline period began. Participants were then

trained on how to use the SUDS to report anxiety level and practiced entering their SUDS ratings

using the computer. The research assistant then provided the participant with instructions for

completing the speech BAT. Once the participant was clear about how the task would proceed, a

one-minute anticipation period began during which the participant sat behind a screen. During

the anticipation period, the three confederates sat in chairs facing the area where the participant

would stand during his/her speech. At the end of the one-minute anticipation, the participant

provided a SUDS rating, and stood in front of the audience. The beginning of the speech task

was signaled by a tone played from the computer, and the first speech topic was displayed on the

computer screen. The participant spoke for one minute on the first topic. The participant then

entered a SUDS rating into the computer and, using the computer, indicated whether or not

she/he was willing to speak on an additional topic for $5.00. If the participant was willing to

continue, a second anticipation period occurred, another SUDS rating was taken, and the next

topic was displayed. After one-minute, a final SUDS rating was taken. Regardless of whether or

not the participant gave the additional speech, a one-minute recovery period followed the final

speech. After completion of the speeches, the participant completed the SSPS, and rated his/her

performance on the PRF. The participant was compensated $5.00 if he/she opted to give the

additional speech.

2.4.1.2. Affect Labeling Task (10 minutes). When the participant was finished speaking,

physiological equipment was removed, and the participant was taken to another location in the

building. The participant then completed a computer task that involved affect labeling and

viewing negative images from the international affective picture system (IAPS; Lang, Bradley, &

Cuthbert, 1999). The Affect Labeling Task followed the procedure of study 1 by Lieberman,

Inagi, Tabibnia, and Crockett (2011). Participants viewed neutral and negative images presented

in two blocks of four images. Each block contained two moderately negative, and two extremely

negative images. Prior to each block, participants were prompted by cues that said either

“scene description” (labeling) or “look and let yourself respond naturally” (watching). Pictures

appeared for 5 seconds. For labeling blocks, participants were asked to choose from three labels

that appeared at the bottom of the screen. One label was relevant to the image, and the other two

were not. Two of the labels were negative, and one was neutral. Participants chose a word by

pressing a key on the keyboard that corresponded to the position of the word on the screen. For

watching blocks, participants simply viewed each image for 5 seconds. Following the

presentation of each image, regardless of block condition, participants were asked “How

distressed did you feel while looking at the picture?” and responded on a 9-point Likert scale

with 0 being not distressed, and 8 being very distressed.

2.4.1.3. Exp-1 (30 minutes). The participant was then randomly assigned to complete

exposure alone, or exposure with affect labeling. The participant completed 10 one-minute

repeated exposure speech trials interspersed with 30-second inter-trial intervals between each

exposure. This protocol was chosen based on previous research findings that this duration is

effective in exposure for phobias (Kircanski et al., 2012; Mineka, Mystkowski, Hladek, &

Rodriguez, 1999; Rowe & Craske, 1998). Participants were provided with instructions by a new

experimenter. Once the participant understood the instructions and practiced responding to

questions using the computer, the participant sat behind a folding screen. Three confederates,

including the experimenter, were seated facing the area where the participant gave the speech. A

tone was played from the computer, which signaled to the participant to step in front of the

audience. All participants were prompted by the computer to provide a SUDS rating.

Participants in the affect labeling condition were then prompted by the computer to choose an

emotion and a feared outcome from four options presented on the screen. At the top of the

screen, the phrase, “I feel ” was presented followed by three possible emotions and

“other.” All three emotions were negative, and one was always anxiety related (see Appendix B

for list of emotions). The phrase “The audience will ” was then presented followed

by three possible feared outcomes related to the audience’s response to the participant (see

Appendix B for list of feared outcomes). An “other” option was also available. Participants

used the keyboard to respond and had up to 15 seconds to make each selection. Participants

were presented with ten different sets of emotion labels and feared outcomes. Participants in the

exposure alone condition completed shape matching. Participants were first presented with a

large black shape at the top of the screen, then were asked to match the shape with one of three

options at the bottom of the screen. If the shape at the top did not match any of the three shapes

at the bottom, the participants were asked to choose “other.” Participants then did the same

shape matching exercise a second time with blue instead of black shapes. Then, a speech topic

appeared on the screen. Participants spoke for 1 minute then were prompted to step behind the

screen for a 30 second inter-trial interval. The speech exposures were repeated 10 times. After

completion of the exposure, participants completed the ECQ.

2.4.2. Time 2.

2.4.2.1. Exp-2 (30 Minutes). The second exposure session was completed three days

after Exp-1. Two exposure sessions were used to parallel multiple exposures as done in therapy.

Participants received the same instructions as at the previous exposure session. Procedures were

identical to those for Exp-1 on Day 1.

2.4.2.2. BAT-2 (30 Minutes). The second BAT was completed directly following Exp-2.

The protocol for BAT-2 followed the exact protocol for BAT-1, however the speech

topics were different. The same audience that was present for BAT-1 was present for BAT-2.

2.4.3. Time 3.

2.4.3.1. BAT-3 (30 minutes). The third and final BAT was completed 8 days after BAT-

1. One-week retest assessed long-term fear reduction and inhibitory learning (Craske et al.,

2008). The protocol for BAT-3 followed the exact protocol for BAT-1 and BAT-2, however the

speech topics were different. The same audience that was present for BAT-1 and BAT-2 was

present for BAT-3.

2.4.4. Speech topics. Speech topics are listed in Appendix C. Across all three BATs,

participants spoke on a total of three to six speech topics depending on how many additional

speeches they were willing to complete. The 6 topics for the BATs were divided into three sets

of two and were counterbalanced across the three time points using a Latin square. Across both

exposure sessions, participants spoke on 20 speech topics. The 20 speech topics for the

exposures sessions were the same across all participants and were administered in the same

order.

2.4.5. Exposure instructions. Copies of exposure instructions are presented in Appendix

D. All participants received the same set of instructions to keep research assistants blind to

condition. Participants were informed that they would be giving 10 speeches in front of a small

audience and that they may be asked to label the emotion that they were feeling. They were

asked to choose the emotion from a list of four emotions that best matched how they were

feeling at that moment. They were encouraged to choose one of the emotions listed, but if they

really felt that the words did not match their experience, they were allowed to choose “other.”

They were also instructed that they may be asked to choose how they thought the audience

might respond to them while they were speaking, and were asked to choose an outcome from the

list that best matched their expectation of how the audience would respond. They were again

encouraged to choose one of the options listed, but could also select other. Finally, they were

informed that they may be asked to match shapes, and were asked to choose the shape from a list

of four possible shapes that matched a large target shape at the top of the screen. If none of the

shapes listed matched the target shape, they were instructed to choose other. Finally, they were

informed that they might see the same shaping matching screen with blue instead of black

shapes.

2.5. Data Analysis

Analyses were conducted using Stata 12. Dependent measures assessed for all three study

aims were HR, SCL, SCR-NS, PRPSA, SUDS, SSPS, PRF, and optional speech. For analyses

including HR, HR2 at baseline was used as a covariate in the models to account for the non-

linear relationship between heart rate and activity in the sympathetic and parasympathetic

nervous systems (Cacioppo, Tassinary, & Berntson, 2007). SCL was log transformed in

accordance with recommendations by Venables and Christie (as cited in Cacioppo et al., 2007).

For models including SCR-NS (a count variable) as the dependent measure, Poisson regression

was used to account for non-normality. For models including optional speech (a dichotomous

variable) as the dependent variable, logistic regression was used. For analyses of the recovery

period (following participants’ decision to complete or avoid the optional speech), optional

speech was tested as a covariate and included in the model when significant. All tests were two-

tailed with an α level of .05.

All three study aims were tested using multi-level modeling (MLM) because

measurement was repeated at three time-points for each participant. MLM accounts for within

and between participant variance. In addition, MLM effectively handles missing data by

including all participants in the model regardless of missing data points. Time was modeled at

level 1, and participant level variables (e.g. Group) were modeled at level two. Time was

modeled using two segments. Once segment modeled change from Time 1 to Time 2, and the

second modeled change from Time 2 to Time 3. This approach was chosen based on a pattern of

results typically observed in intervention studies characterized by an initial steep change in

symptoms from pre to post intervention and a leveling off of change through follow-up. For

each dependent variable, random effects of the intercept and slope and their covariance were

first included in the model. Non-significant random effects (tested using likelihood ratio tests)

were removed from the model. For each dependent variable, the model with the fewest random

effect parameters necessary to achieve optimal model fit was chosen.

For Aim 1, predictors in the model were Time, Group and the Time × Group interaction,

and the Time × Group fixed effect was examined for significance. For Aim 2, predictors in the

model were Time, number of anxiety labels chosen, and the Time × number of anxiety labels

interaction, and the Time × number of anxiety labels fixed effect was examined for significance.

Analyses for aim 2 were only conducted within the AL group. For Aim 3, predictors in the

model were Time, Group, incidental emotion regulation, Time × Group, Time × incidental

emotion regulation, Group × incidental emotion regulation, and Time × Group × incidental

emotion regulation, and the Time × Group × incidental emotion regulation fixed effect was

examined for significance. If the three way interaction (test of moderation) was not significant,

it was dropped from the model and the Time × incidental emotion regulation (test of prediction)

fixed effect was examined for significance.

2.6. Effect Size

For study Aim 1, effect sizes reported are Cohen’s d, and were calculated using an

approach described by Feingold (2009) for estimating group differences in randomized clinical

trials with repeated measures. For study Aims 2 and 3, effect sizes reported are Cohen’s f2 and

were estimated using an approach outlined by Selya and colleagues (2012). Cohen’s f2 uses

residual variance from the model to estimate effect size. However, for multi-level models, effect

sizes calculated using residual variance and proportion of variance explained should be

interpreted with caution because the addition of variables to the model can, in some cases,

increase residual variance resulting in negative estimates of explained variance and even of

effect size (Snijders & Bosker, 1994) In addition, this method cannot be used for non-

continuous dependent measures. As a result, effect sizes are not reported for analyses with

SCR- NS (count variable) or optional speech (dichotomous variable) as the outcome.

2.7. Missing Data

There were missing data on a number of demographic, predictor, and dependent

measures. Reasons for missing data included participant non-response, experimenter error,

inadvertent deletion, and computer program failure. The numbers of missing data points for all

study variables are shown in Table 2.

2.8. Power Analyses

Effect sizes for group comparisons in the study by Kircanski and colleagues (2012), in

which verbalization during exposure was compared to exposure alone in spider phobic

participants, ranged from .58 to .99 depending on which outcome measure was assessed. To

achieve power of 0.8, for an effect size of .60, the goal sample size was 72 participants.

CHAPTER 3

Results

3.1. Preliminary Analyses

3.1.1. Sample characteristics. Groups did not significantly differ on any demographic

or clinical characteristics at baseline (ps > .05). Table 3 includes baseline descriptive statistics

and statistical tests for group differences. At baseline, 54% of the sample fell in the “high”

public speaking anxiety range on the PRPSA (scores above 131), 46% of the sample fell in the

moderate range (scores between 97 and 131), and 0% of the sample fell in the “low” public

speaking anxiety range. The current sample had a mean PRPSA score of 133.0, which is

approximately one standard deviation above the mean of 114.6 (SD = 17.2) observed in a college

sample (McCroskey, 1970). On the mini-SPIN, 65% of the sample fell at or above the clinical

cut-off of 6, indicating possible problems with social anxiety, and 36% met criteria for social

anxiety disorder based on the MINI diagnostic interview. On the PHQ, 49% of the sample fell in

the minimal range, 39% in the mild range, 7% in the moderate range, 4% in the moderately

severe range, and 1% in the severe range for depressive symptoms.

3.1.2. Descriptive statistics and correlations among study variables. Table 4 includes

raw means and standard deviations for all dependent measures at all three time points by group.

Table 5 includes proportions of affect labels and feared outcomes chosen at Exp1 and Exp2 (AL

group only), mean SUDS ratings at the beginning and end of Exp1 and Exp2 by group, and

ECQ ratings at Exp1 by group. Table 6 includes correlations among dependent measures, affect

labeling variables collected during exposure, and incidental emotion regulation.

3.1.3. Comparison of completers vs. dropout. Table 7 includes means and standard

deviations for dependent measures and exposure variables for completers and participants who

dropped from the study as well as statistical tests comparing the two groups. Participants who

dropped reported significantly fewer positive self statements during the public speaking task,

rated their performance significantly lower, had significantly higher HR during anticipation of

giving a speech, had marginally significantly higher HR when recovering after giving a speech,

had marginally significantly higher SCR-NS during anticipation of giving a speech, and were

marginally significantly less likely to give the optional speech compared to participants who

completed the study. Dropout did not differ between the AL group (N=11) and the Control

group (N=8) χ2 = .33, p = .568. Given that significant differences were found between

completers and those who dropped from the study, dropout was tested as a covariate in each

model, and when significant, was included in the final model.

3.2. Primary Analyses Aim 1. Does Affect Labeling Enhance Exposure Effectiveness

Compared to Exposure Alone?

Figures are provided for each dependent variable assessed. For each model tested,

interaction effects from Time 1 to Time 2, Time 2 to Time 3, and Time 1 to Time 3 are reported.

For significant and marginally significant interactions, the following simple effects are tested.

(1) Group differences at Time points 2 and 3; (2) Simple slopes (whether slopes differ

from zero) from Time 1 to Time 2, Time 2 to Time 3, and/or Time 1 to Time 3 (depending on

significant interactions).

3.2.1. Heart rate.

3.2.1.1. Anticipation. Results are displayed in Figure 3. For HR during anticipation of

giving a speech, the Time × Group interactions from Time 1 to Time 2 (p = .934), Time 2 to

Time 3 (p = .191), and from Time 1 to Time 3 (p = .152) were not significant.

3.2.1.2. Recovery. Results are displayed in Figure 4. For HR during recovery following

the speech, the Time × Group interactions from Time 1 to Time 2 (p = .654) and from Time 1 to

Time 3 (p = .153) were not significant. The Time × Group interaction from Time 2 to Time 3

was significant (b = -3.79, 95% Confidence Interval (CI) = -7.4 to -.2, p = .041, d = .33) such

that participants in the AL group showed a steeper decrease in HR from Time 2 to Time 3 than

did participants in the Control group. Tests of group differences revealed no significant

differences at Time 2 (p =.419) or at Time 3 (.281). Tests of simple slopes from Time 2 to Time

3 revealed a significant increase in HR in the Control group (change = 2.77, p = .036) and no

significant change in the AL group (p = .432).

3.2.2. Skin conductance level.

3.2.2.1. Anticipation. Results are displayed in Figure 5. For SCL during anticipation of

giving a speech, the Time × Group interactions from Time 1 to Time 2 (p = .690), Time 2 to

Time 3 (p = .947), and Time 1 to Time 3 (p = .743) were not significant.

3.2.2.2. Recovery. Results are displayed in Figure 6. For SCL during recovery following

the speech, the Time × Group interactions from Time 1 to Time 2 (p = .302), Time 2 to Time 3

(p = .861), and Time 1 to Time 3 (p = .227) were not significant.

3.2.3. Non-specific skin conductance response.

3.2.3.1. Anticipation. Results are displayed in Figure 7. For SCR-NS during anticipation

of giving a speech, the Time × Group interactions from Time 1 to Time 2 (p = .875), Time 2 to

Time 3 (p = .778), and Time 1 to Time 3 (p = .878) were not significant.

3.2.3.2. Recovery. Results are displayed in Figure 8. For SCR-NS during recovery

following the speech, the Time × Group interaction from Time 1 to Time 2 was not significant (p

= .587), whereas the interactions from Time 2 to Time 3 (b = -1.14, CI = -2.1 to -.2, p =

.023, d = 1.0) and Time 1 to Time 3 (b = -.90, CI = -1.8 to 0, p = .023, d = .79) were significant

such that participants in the AL group showed a steeper decrease in SCR-NS from Time 2 to

Time 3 and from Time 1 to Time 3 than did participants in the Control group. Tests of group

differences revealed no significant group difference at Time 2 (p =.273) and a marginally

significant difference at Time 3 (difference = .29, p = .100) such that participants in the AL

group had fewer SCR-NS during recovery than did participants in the Control group. Tests of

simple slopes from Time 2 to Time 3 revealed a marginally significant reduction in SCR-NS

during recovery in the AL group (change = -.61, p = .077), but not in the Control group (p

= .146), and no significant changes from Time 1 to Time 3 in either group (ps > .123).

3.2.4. Personal Report of Public Speaking Anxiety. Results are displayed in Figure 9.

For PRPSA, the Time × Group interactions from Time 1 to Time 2 (p = .196), Time 2 to

Time 3 (p = .875), and Time 1 to Time 3 (p = .394) were not significant.

3.2.5. Subjective Units of Distress. Results are displayed in Figure 10. For SUDS, the

Time × Group interactions from Time 1 to Time 2 (p = .385), Time 2 to Time 3 (p = .328),

and Time 1 to Time 3 (p = .901) were not significant.

3.2.6. Self Statements During Public Speaking.

3.2.6.1. Positive. Results are displayed in Figure 11A (higher scores indicate more

positive cognitions). For SSPS Positive, the Time × Group interactions from Time 1 to Time 2

(p = .837), Time 2 to Time 3 (p = .264), and Time 1 to Time 3 (p = .316) were not significant.

3.2.6.2. Negative. Results are displayed in Figure 11B. For SSPS Negative, the Time ×

Group interactions from Time 1 to Time 2 (p = .952), Time 2 to Time 3 (p = .320), and Time 1

to Time 3 (p = .417) were not significant.

3.2.7. Performance Rating Form. Results are displayed in Figure 13 (higher scores

indicate better self-rated performance). For PRF, the Time × Group interactions from Time 1 to

Time 2 (p = .544) and from Time 1 to Time 3 (p = .353) were not significant whereas, the

interaction from Time 2 to Time 3 was marginally significant (b = -3.06, CI = -6.5 to .34, p =

.077, d = .36) such that participants in the Control group showed a steeper increase in

PRF from Time 2 to Time 3 than did participants in the AL group. Tests of group differences

revealed no significant differences at Time 2 or Time 3 (ps > .210). Tests of simple slopes

revealed a significant increase from Time 2 to Time 3 in PRF in the Control group (change =

4.57, p <

.001), but not in the AL group (p = .217).

3.2.8. Optional speech. Results are displayed in Figure 12. For optional speech, the

model did not converge when all three time points were included. Therefore, the model was run

including Time 1 and Time 3 only. The Time × Group interaction from Time 1 to Time 3 was

not significant (p = .669).

3.3. Secondary Analyses Aim 1. Does Affect Labeling Enhance Exposure Effectiveness for

High Labelers Compared to Exposure Alone?

A number of participants assigned to the affect labeling condition chose no or very few

anxiety labels; we evaluated whether participants who more consistently selected anxiety labels

during exposures showed greater fear reduction compared to those assigned to exposure alone.

Participants were categorized into “low” or “high” labelers based on the median frequency of

affect labels chosen, and the high labelers were compared to participants in the control condition.

3.3.1. Heart Rate.

3.3.1.1. Anticipation. Results are displayed in Figure 14. For HR during anticipation of

giving a speech, the Time × Group interactions from Time 1 to Time 2 (p = .554) and Time 1 to

Time 3 (p = .198) were not significant whereas the interaction from Time 2 to Time 3 was

marginally significant (b = -4.96, CI = -10.3 to .4, p = .071, d = .39) such that participants in the

AL group showed a steeper decrease in HR from Time 2 to Time 3 than did participants in the

Control group. Tests of group differences revealed no significant differences at Time 2 or Time

3 (ps > .174). Tests of simple slopes from Time 2 to Time 3 revealed no significant change in

either group (ps > .191).

3.3.1.2. Recovery. Results are displayed in Figure 15. For HR during recovery following

the speech, the Time × Group interaction from Time 1 to Time 2 was not significant (p = .328).

The interaction from Time 2 to Time 3 (b = -6.47, CI = -10.6 to -2.3, p = .002, Cohen’s d = .56)

was significant and the interaction from Time 1 to Time 3 was marginally significant (b = -4.39,

CI = -8.9 to .1, p = .054, Cohen’s d = .38) such that high labelers in the AL group had a steeper

decline in HR during recovery than did participants in the Control group. Tests of group

differences revealed no group differences at Time 2 or Time 3 (ps > .129). Tests of simple

slopes from Time 1 to Time 3 revealed a marginally significant increase in HR in the Control

group (change = 2.63, p = .057) and no significant change in the AL group (p = .333). Tests of

simple slopes from Time 2 to Time 3 revealed a significant increase in HR in the Control group

(change = 2.70, p = .033) and a significant decrease in HR in the AL group (change = -3.77, p =

.027).

3.3.2. Skin conductance level.

3.3.2.1. Anticipation. For SCL during anticipation of giving a speech, the Time × Group

interactions from Time 1 to Time 2 (p = .281), Time 2 to Time 3 (p = .921) and Time 1 to Time

3 (p = .329) were not significant.

3.3.2.2. Recovery. For SCL during recovery following the speech, the Time × Group

interactions from Time 1 to Time 2 (p = .603), Time 2 to Time 3 (p = .676), and Time 1 to

Time 3 (p = .339) were not significant.

3.3.3. Non-specific skin conductance response.

3.3.3.1. Anticipation. Results are displayed in Figure 16. For SCR-NS during

anticipation of giving a speech, the Time × Group interactions from Time 1 to Time 2 (p = .715)

and Time 2 to Time 3 (p = .128) were not significant. The interaction from Time 1 to Time 3

was significant (b = -.62, CI = -1.2 to 0, p = .049, d = .35) such that high labelers in the AL

group had a steeper decline in SCR-NS during anticipation than did participants in the Control

condition. Tests of group differences revealed no group difference at Time 2 (p = .287), and a

significant group difference at Time 3 (difference = -.89, p = .008) such that the AL group had

fewer SCR-NS than did the Control group. Tests of simple slopes revealed a marginally

significant decrease in SCR-NS from Time 1 to Time 3 in the Control group (p = .055), and a

significant decrease in SCR-NS from Time 1 to Time 3 in the AL group (change = -.93, p =

.001).

3.3.3.2. Recovery. Results are displayed in Figure 17. For SCR-NS during recovery

following the speech, the Time × Group interaction from Time 1 to Time 2 was not significant (p

= .385). The interaction from Time 2 to Time 3 (b = -1.87, CI = -3.3 to -.4, p = .013, d = 1.63)

was significant, and the interaction from Time 1 to Time 3 was marginally significant (b = -.14,

CI = -2.8 to 0, p = .055, Cohen’s d = 1.21) such that high labelers in the AL group had a steeper

decline in SCR-NS during recovery than did participants in the Control group. Tests of group

differences revealed no significant group difference at Time 2 (p = .536), and a significant group

difference at Time 3 (difference = .47, p = .008) such that the AL group had fewer SCR-NS

during recovery than did the Control group. Tests of simple slopes from Time 2 to Time 3

revealed a significant decrease in SCR-NS during recovery in the AL group (change = -1.33, p =

.041), but not in the Control group (p = .143), and no significant change from Time 1 to Time 3

in either group (ps > .131).

3.3.4. Personal Report of Public Speaking Anxiety. Results are displayed in Figure 18.

For PRPSA, the Time × Group interactions from Time 2 to Time 3 (p = .957) and Time 1 to

Time 3 (p = .214) were not significant. The interaction from Time 1 to Time 2 was marginally

significant (b = 5.37, CI = -.9 to 11.7, p = .095, d = .35) such that participants in the Control

group had a steeper decline in PRPSA than did high labelers in the AL group. Tests of group

differences revealed no significant group difference at Time 2 (p = .122) or Time 3 (p = .184).

Tests of simple slopes revealed a significant decrease in PRPSA from Time 1 to Time 2 in the

AL group (change = -5.33, p = .041) and in the Control group (change = -10.70, p < .001).

3.3.5. Subjective Units of Distress. Results are displayed in Figure 19. For SUDS, the

Time × Group interactions from Time 2 to Time 3 (p = .692) and Time 1 to Time 3 (p = .209)

were not significant. The interaction from Time 1 to Time 2 was marginally significant (b = .57,

CI = -.1 to 1.2, p = .079, d = .38) such that participants in the Control group showed a steeper

decrease in SUDS from Time 1 to Time 2 than did participants in the AL group. Tests of group

differences revealed a significant group difference at Time 2 (difference = .81, p = .038) such

that the AL group had higher SUDS ratings than did the Control group. Groups did not

significantly differ at Time 3 (p = .104). Tests of simple slopes revealed a significant decrease in

SUDS from Time 1 to Time 2 in the AL group (change = -.70, p = .007) and in the

Control group (change = -1.28, p < .001).

3.3.6. Self Statements During Public Speaking.

3.3.6.1. Positive. For SSPS positive, the Time × Group interactions from Time 1 to Time

2 (p = .908), Time 2 to Time 3 (p = .240), and Time 1 to Time 3 (p = .418) were not significant.

3.3.6.1. Negative. For SSPS negative, the Time × Group interactions from Time 1 to

Time 2 (p = .853), Time 2 to Time 3 (p = .212), and Time 1 to Time 3 (p = .268) were not

significant.

3.3.7. Performance Rating Form. For PRF, the Time × Group interactions from Time 1

to Time 2 (p = .841), Time 2 to Time 3 (p = .143), and Time 1 to Time 3 (p = .177) were not

significant.

3.3.8. Optional Speech. For the optional speech, the Time × Group interaction from

Time 1 to Time 3 was not significant (p = .649).

3.4. Aim 2. Does the Number of Anxiety Related Labels Used During Exposure Predict

Greater Attenuation of Fear Responding at Re-Test?

For each model tested, the significance of the Time × Number of Anxiety Labels

interaction from Time 1 to Time 2, Time 2 to Time 3, and Time 1 to Time 3 are reported. For

significant and marginally significant interactions, tests of whether participants at one standard

deviation below the mean (-1SD) on Number of Anxiety Labels differ from those at one standard

deviation above the mean (+1SD) at Time points 1, 2 and 3.

3.4.1. Heart rate.

3.4.1.1. Anticipation. For HR during anticipation of giving a speech, the Time × number

of anxiety labels interactions from Time 1 to Time 2 (p = .525), Time 2 to Time 3 (p = .387), and

Time 1 to Time 3 (p = 783) were not significant.

3.4.1.2. Recovery. For HR during recovery following the speech, the Time × number of

anxiety labels interactions from Time 1 to Time 2 (p = .695), Time 2 to Time 3 (p = .325), and

Time 1 to Time 3 (p = .517) were not significant.

3.4.2. Skin conductance level.

3.4.2.1. Anticipation. For SCL during anticipation of giving a speech, the Time × number

of anxiety labels interactions from Time 1 to Time 2 (p = .307), Time 2 to Time 3 (p = .878), and

Time 1 to Time 3 (p = .400) were not significant.

3.4.2.2. Recovery. For SCL during recovery following the speech, the Time × number of

anxiety labels interactions from Time 1 to Time 2 (p = .926), Time 2 to Time 3 (p = .969), and

Time 1 to Time 3 (p = .896) were not significant.

3.4.3. Non-specific skin conductance response.

3.4.3.1. Anticipation. Results are displayed in Figure 20. For SCR-NS during

anticipation of giving a speech, the Time × number of anxiety labels interaction from Time 1 to

Time 2 was not significant (p = .402). The interactions from Time 2 to Time 3 (b = -1.42, CI = -

2.8 to 0, p = .045), and Time 1 to Time 3 were significant (b = -1.92, CI = -3.2 to -.6, p = .004)

such that participants who used more affect labels during exposure had a steeper decline in SCR-

NS over time. Tests of the difference between SCR-NS during anticipation for participants at

+1SD and -1SD from the mean revealed no significant difference at Time 1 or Time 2 (ps >

.117), and a significant difference at Time 3 (difference = 1.70, p = .001) such that participants at

+1SD from the mean on use of anxiety labels had fewer SCR-NS during anticipation of giving a

speech than did participants at -1SD from the mean.

3.4.3.2. Recovery. For SCR-NS during recovery following the speech, the Time ×

number of anxiety labels interactions from Time 1 to Time 2 (p = .574), Time 2 to Time 3 (p =

.321), and Time 1 to Time 3 (p = .151) were not significant.

3.4.4. Personal Report of Public Speaking Anxiety. Results are displayed in Figure 21.

For PRPSA, the Time × number of anxiety labels interaction from Time 2 to Time 3 was not

significant (p = .961). The interaction from Time 1 to Time 2 (b = 15.76, CI = 1.6 to 29.9, p =

.029) was significant, and the interaction from Time 1 to Time 3 was marginally significant (b =

15.40, CI = -2.7 to 33.5, p = .095) such that participants who used fewer affect labels during

exposure had a steeper decline in PRPSA over time (interaction f2 = .05). Tests of the difference

between PRPSA for participants at +1SD and -1SD from the mean on number of anxiety labels

revealed no significant difference at Time 1 (p = .368), and significant differences at Time 2

(difference = 11.70, p = .011) and Time 3 (difference = 11.52, p = .032) such that participants at

+1SD from the mean on use of anxiety labels had higher PRPSA scores than did participants at -

1SD from the mean.

3.4.5. Subjective Units of Distress. Results are displayed in Figure 22. For SUDS, the

Time × number of anxiety labels interaction from Time 2 to Time 3 was not significant (p =

.850). The interaction from Time 1 to Time 2 was significant (b = 1.72, CI = .3 to 3.2, p = .019),

and the interaction from Time 1 to Time 3 was marginally significant (b = 1.58, -.1 to 3.2, p =

.058) such that participants who used fewer affect labels during exposure had a steeper decline in

SUDS over time (interaction f2 = .05). Tests of the difference between SUDS for participants at

+1SD and -1SD from the mean on number of anxiety labels revealed a marginally significant

difference at Time 1 (difference = .67, p = .083), and significant differences at Time 2

(difference = 1.52, p < .001) and Time 3 (difference = 1.45, p = .003) such that participants at

+1SD from the mean on use of anxiety labels had higher SUDS scores than did participants at -

1SD from the mean.

3.4.6. Self-Statements During Public Speaking.

3.4.6.1. Positive. For SSPS Positive, the Time × number of anxiety labels interactions

from Time 1 to Time 2 (p = .887), Time 2 to Time 3 (p = .760) and Time 1 to Time 3 (p =

.938) were not significant.

3.4.6.2. Negative. For SSPS Negative, the Time × number of anxiety labels interactions

from Time 1 to Time 2 (p = .615), Time 2 to Time 3 (p = .966), and Time 1 to Time 3 (p =

.984) were not significant.

3.4.7. Performance Rating Form. Results are displayed in Figure 23 (higher scores

indicate better self-rated performance). For PRF, the Time × number of anxiety labels

interactions from Time 2 to Time 3 (p = .144) and Time 1 to Time 3 were not significant (p

.501). The interaction from Time 1 to Time 2 was significant (b = -9.87, CI = -18.0 to -1.8, p =

.017) such that participants who used fewer affect labels during exposure had a steeper increase

in PRF over time (interaction f2 = .15). Tests of the difference between PRF scores for

participants at +1SD and -1SD from the mean on number of anxiety labels revealed a marginally

significant difference at Time 1 (difference = 4.86, p = .054), and significant differences at Time

2 (difference = 9.76, p < .001) and Time 3 (difference = 6.67, p = .024) such that participants at

+1SD from the mean on use of anxiety labels had lower PRF scores than did participants at -1SD

from the mean.

3.4.8. Optional speech. For optional speech, the Time × number of anxiety labels

interaction from Time 1 to Time 3 was not significant (p = .666).

3.5. Aim 3. Does Incidental Emotion Regulation at Baseline Moderate or Predict Response

to Exposure with Affect Labeling Versus Exposure Alone?

For each dependent variable, incidental emotion regulation is first tested as a moderator

of response to exposure plus affect labeling versus exposure alone. The significance of the Time

× Group × incidental emotion regulation interactions from Time 1 to Time 2, Time 2 to Time 3,

and Time 1 to Time 3 are reported. For significant and marginally significant interactions, the

following simple effects are tested: (1) Group mean differences at Time 1, Time 2, and Time 3 at

+1SD and -1SD from the mean of incidental emotion regulation. (2) Group slope differences

from Time 1 to Time 2, Time 2 to Time 3, and/or Time 1 to Time 3 (depending on significant

interactions) at +1SD and -1SD from the mean of incidental emotion regulation.

When the three-way interaction is either marginally significant or not significant,

incidental emotion regulation is tested as a predictor of treatment outcome, and the significance

of the Time × incidental emotion regulation interaction is reported for Time 1 to Time 2, Time 2

to Time 3, and Time 1 to Time 3. For significant and marginally significant interactions, tests of

whether participants at -1SD from the mean on incidental emotion regulation differ from those at

+1SD from the mean at Time points 1, 2 and 3.

3.5.1. Heart rate.

3.5.1.1. Anticipation. Moderation results are displayed in Figure 24. For HR during

anticipation of giving a speech, the Time × Group × incidental emotion regulation interactions

from Time 1 to Time 2 (p = .269) and Time 2 to Time 3 (p = .344) were not significant,

whereas the interaction from Time 1 to Time 3 was significant (b = 7.56, CI = .3 to 14.9, p

= .042)

(interaction f2 = .04). Tests of group mean differences at -1SD and +1SD from the mean

revealed no differences at Time 1 (ps > .745), or Time 2 (ps > .443). At Time 3, for participants

at -1SD from the mean on incidental emotion regulation, participants in the AL group had

significantly lower HR than did participants in the Control group (difference = 6.77, p = .014).

No group difference was found for participants at +1SD from the mean on incidental emotion

regulation at Time 3 (p = .592). Tests of Group slope differences from Time 1 to Time 3

revealed that for participants at -1SD from the mean on incidental emotion regulation, the AL

group had a significantly more negative slope than did the Control group (slope difference =

7.57, p = .019). No group slope difference was found for participants at +1SD from the mean on

incidental emotion regulation (p = .543).

3.5.1.2. Recovery. Moderation results are displayed in Figure 25. For HR during

recovery following the speech, the Time × Group × incidental emotion regulation interaction

from Time 1 to Time 2 was not significant (p = .446), whereas the interaction from Time 2 to

Time 3 was marginally significant (b = 5.09, CI = -.3 to 12.5, p = .063), and the interaction from

Time 1 to Time 3 was significant (b = 8.48, CI = 2.2 to 14.8, p = .008) (interaction f2 = .07).

Tests of group mean differences at -1SD and +1SD from the mean revealed no differences at

Time 1 (ps > .207) or Time 2 (ps > .301). At Time 3, for participants at -1SD from the mean on

incidental emotion regulation, participants in the AL group had marginally significantly lower

HR than participants in the Control group (difference = 4.72, p = .061). No group mean

differences were observed for participants at +1SD from the mean on incidental emotion

regulation at Time 3 (p = .329). Tests of Group slope differences from Time 1 to Time 3

revealed that for participants at -1SD from the mean on incidental emotion regulation, the AL

group had a significantly more negative slope than did the Control group (slope difference =

7.56, p = .007). No group slope difference was observed for participants at +1SD from

the mean on incidental emotion regulation (p = .262).

3.5.2. Skin conductance level.

3.5.2.1. Anticipation. For SCL during anticipation of giving a speech, the Time × Group

× incidental emotion regulation interactions from Time 1 to Time 2 (p = .602), Time 2 to

Time 3 (p = .218), and Time 1 to Time 3 (p = .449) were not significant.

For SCL during anticipation of giving a speech, the Time × incidental emotion regulation

interactions Time 1 to Time 2 (p = .664), Time 2 to Time 3 (p = .153), and from Time 1 to Time

3 (p = .296) were not significant.

3.5.2.2. Recovery. For SCL during recovery following the speech, the Time × Group ×

incidental emotion regulation interactions from Time 1 to Time 2 (p = .272), Time 2 to Time 3 (p

= .218) and Time 1 to Time 3 (p = .578) were not significant.

For SCL during recovery following the speech, the Time × incidental emotion regulation

interactions from Time 1 to Time 2 (p = .565), Time 2 to Time 3 (p = .153), and Time 1 to Time

3 (p = .120) were not significant.

3.5.3. Non-specific skin conductance response.

3.5.3.1. Anticipation. For SCR-NS during anticipation of giving a speech, the Time ×

Group × incidental emotion regulation interactions from Time 1 to Time 2 (p = .906), Time 2 to

Time 3 (p = .369), and Time 1 to Time 3 (p = .267) were not significant.

Predictor results are displayed in Figure 26. For SCR-NS during anticipation of giving a

speech, the Time × incidental emotion regulation interactions from Time 1 to Time 2 (p = .180)

and Time 1 to Time 3 were not significant (p = .125), whereas the Time × incidental emotion

regulation interaction from Time 2 to Time 3 was significant (b = .53, CI = .1 to .9, p = .009)

such that participants with deficits in incidental emotion regulation had more negative slopes

from Time 2 to Time 3 than did participants with strengths in incidental emotion regulation.

Tests of the difference between SCR-NS during anticipation for participants at +1SD and -1SD

from the mean revealed no significant differences at Time 1 or Time 3 (ps > .331), and a

significant difference at Time 2 (difference = .90, p = .032) such that participants at +1SD from

the mean on incidental emotion regulation had lower SCR-NS than did participants at -1SD

from the mean.

3.5.3.2. Recovery. Moderator results are displayed in Figure 27. For SCR-NS during

recovery following the speech, the Time × Group × incidental emotion regulation interaction

from Time 1 to Time 3 was not significant (p = .419), whereas the interaction from Time 2 to

Time 3 was marginally significant (b = -1.74, CI = -3.7 to .2, p = .080), and the interaction from

Time 1 to Time 2 was significant (b = 2.37, CI = .5 to 4.2, p = .012). Tests of group mean

differences at -1SD and +1SD revealed no differences at Time 1 (ps > .432) or Time 3 (ps >

.200). At Time 2, for participants at +1SD from the mean on incidental emotion regulation,

participants in the AL Group had significantly higher SCR-NS than participants in the Control

group (difference = .48, p = .010). At Time 2, no group differences emerged for participants at -

1SD from the mean on incidental emotion regulation (p = .487). Tests of Group slope

differences from Time 1 to Time 2 revealed that for participants at +1SD from the mean on

incidental emotion regulation, the AL group had a significantly more positive slope than did the

Control group (slope difference = .58, p = .027). No group slope difference was observed for

participants at -1SD from the mean on incidental emotion regulation (p = .236). Tests of Group

slope differences from Time 2 to Time 3 revealed that for participant at +1SD from the mean on

incidental emotion regulation, the AL group had a significantly more negative slope than did the

Control group (slope difference = .71, p = .027). No group slope difference was observed for

participants at -1SD from the mean on incidental emotion regulation (p = .670).

3.5.4. Personal Report of Public Speaking Anxiety. For PRPSA, the Time × Group ×

incidental emotion regulation interactions from Time 1 to Time 2 (p = .389), Time 2 to Time 3 (p

= .787), and Time 1 to Time 3 (p = .656) were not significant.

For PRPSA, the Time × incidental emotion regulation interactions from Time 1 to Time

2 (p = .470), Time 2 to Time 3 (p = .270) and Time 1 to Time 3 (p = .165) were not significant.

3.5.5. Subjective Units of Distress. Moderator results are displayed in Figure 28. For

SUDS, the Time × Group × incidental emotion regulation interactions from Time 1 to Time 3 (p

= .838) and Time 2 to Time 3 (p = .138) were not significant, whereas the interaction from Time

1 to Time 2 was marginally significant (b = -.81, CI = -1.7 to .1, p = .085) (interaction f2 = .05).

Tests of group mean differences at -1SD and +1SD from the mean on incidental emotion

regulation revealed no differences at Time 1 (ps > .457), Time 2 (ps > .102) or Time 3 (ps >

.786). Tests of Group slope differences from Time 1 to Time 2 revealed no differences at +1SD

or -1SD from the mean on incidental emotion regulation (ps > .122).

Predictor results are displayed in Figure 29. For SUDS, the Time × incidental emotion

regulation interactions from Time 1 to Time 2 (p = .376) and Time 2 to Time 3 (p = .248) were

not significant. The interaction from Time 1 to Time 3 was marginally significant (b = .46, CI =

0 to .9, p = .061) such that greater deficits in incidental emotion regulation were associated with

greater decreases in SUDS over Time (interaction f2 = .01). Tests of the difference between

SUDS for participants at +1SD and -1SD from the mean on incidental emotion regulation

revealed no significant differences at Time 1, Time 2, or Time 3 (ps > .113).

3.5.6. Self Statements During Public Speaking.

3.5.6.1. Positive. Moderator results for SSPS Positive are displayed in Figure 30 (higher

scores indicate more positive cognitions). For SPSS Positive, the Time × Group × incidental

emotion regulation interactions from Time 2 to Time 3 (p = .634) and Time 1 to Time 3 (p =

.272) were not significant, whereas the interaction from Time 1 to Time 2 was marginally

significant (b = 2.15, CI = -.1 to 4.3, p = .055) (interaction f2 = .06). Tests of group mean

differences at -1SD and +1SD from the mean on incidental emotion regulation revealed no

differences at Time 1 (ps > .440). At Time 2, for participants at -1SD from the mean on

incidental emotion regulation, participants in the AL Group had marginally significantly lower

SSPS Positive scores than participants in the Control group (difference = 1.77, p = .096). No

group differences emerged for participants at +1SD from the mean on incidental emotion

regulation at Time 2 (p = .585). At Time 3, for participants at -1SD from the mean on incidental

emotion regulation, participants in the AL Group had marginally significantly lower SSPS

Positive scores than participants in the Control group (difference = 2.20, p = .071). No group

differences emerged for participants at +1SD from the mean on incidental emotion regulation at

Time 3 (p = .653). Tests of Group slope differences from Time 1 to Time 2 revealed no

differences at +1SD or -1SD from the mean on incidental emotion regulation (ps > .141).

Predictor results for SSPS Positive are displayed in Figure 31 (higher scores indicate

more positive cognitions). For SSPS Positive, the Time × incidental emotion regulation

interaction from Time 2 to Time 3 was not significant (p = .605). The interactions from Time 1

to Time 2 (b = -1.28, CI = -2.3 to -.2, p = .016) and Time 1 to Time 3 (b = -1.56, CI = -2.9 to -.2,

p = .025) were significant such that greater deficits in incidental emotion regulation were

associated with greater increase in SSPS Positive over Time (f2 = .01). Tests of the difference in

SSPS Positive for participants at +1SD and -1SD from the mean on incidental emotion

regulation revealed no significant differences at Time 2, or Time 3 (ps > .746). At Time 1,

participants at +1SD from the mean on incidental emotion regulation had significantly higher

SSPS Positive scores than did participants at -1SD from the mean (difference = 1.70, p = .025).

3.5.6.2. Negative. Moderator results for SSPS Negative are displayed in Figure 32. For

SPSS Negative, the Time × Group × incidental emotion regulation interactions from Time 1 to

Time 2 (p = .389) and from Time 1 to Time 3 (p = .237) were not significant, whereas the

interaction from Time 2 to Time 3 was marginally significant (b = 2.20, CI = -.2 to 4.6, p = .068)

(interaction f2 = .03). Tests of group mean differences at -1SD and +1SD revealed no differences

at Time 1 (ps > .388), Time 2 (ps > .278), or Time 3 (p > .438). Tests of Group slope differences

from Time 2 to Time 3 revealed a marginally significant difference for participants at +1SD from

the mean on incidental emotion regulation such that the AL group had a significantly more

positive slope than did the Control group (slope difference = 1.96, p = .060). No group slope

difference was found for participants at -1SD from the mean on incidental emotion regulation

from Time 2 to Time 3 (p = .446).

Predictor results for SSPS Negative are displayed in Figure 33. For SSPS Negative, the

Time × incidental emotion regulation interaction from Time 1 to Time 2 was not significant (p =

.241). The interactions from Time 2 to Time 3 (b = 1.13, CI = 0 to 2.2, p = .047) and Time 1 to

Time 3 (b = 1.80, CI = .4 to 3.2, p = .013) were significant such that greater deficits in incidental

emotion regulation were associated with greater decreases in SSPS Negative over Time

(interaction f2 = .00). Tests of the difference in SSPS Negative for participants at +1SD and -

1SD from the mean on incidental emotion regulation revealed no significant differences at Time

2, or Time 3 (ps > .255).

At Time 1, participants at +1SD from the mean on incidental emotion

regulation had significantly lower SSPS Negative scores than did participants at -1SD from the

mean (difference = 1.83, p = .041).

3.5.7. Performance Rating Form. For PRF, the Time × Group × incidental emotion

regulation interactions from Time 1 to Time 2 (p = .274), Time 2 to Time 3 (p = .316), and

Time 1 to Time 3 (p = .944) were not significant.

For PRF, the Time × incidental emotion regulation interactions from Time 1 to Time 2 (p

= .470), Time 2 to Time 3 (p = .216), and from Time 1 to Time 3 (p = .165) were not significant.

3.5.8. Optional speech. For the optional speech, the Time × Group × incidental emotion

regulation interaction from Time 1 to Time 3 was not significant (p = .988).

For the optional speech, the Time × incidental emotion regulation interaction from Time

1 to Time 3 was not significant (p = .689).

CHAPTER 4

Discussion

The current study had three primary aims. The first aim was to test whether exposure

combined with affect labeling resulted in greater attenuation of fear responding than did

exposure alone. The second aim was to assess whether the number of anxiety labels used during

exposure in the affect labeling group predicted greater fear attenuation. The final aim was to test

whether incidental emotion regulation capacity at baseline could be used to predict who would

benefit from exposure plus affect labeling compared to exposure alone. If no moderation was

found, incidental emotion regulation was assessed as a predictor of who would respond to

exposure regardless of group assignment. All three aims were tested on the same set of

dependent variables. These included heart rate, skin conductance level, and non-specific skin

conductance response during anticipation of and recovery from a speech, self-reported trait

public speaking anxiety, anxiety level during public speaking, negative and positive self

statements during public speaking, self rated public speaking performance, and avoidance of

giving an additional speech.

4.1. Aim 1: Does Affect Labeling Enhance Exposure Effectiveness Compared to Exposure

Alone?

Consistent with hypotheses, participants in the exposure plus affect-labeling group had a

steeper decline in heart rate and non-specific skin conductance responses during recovery

following the speech than did participants in the exposure alone condition. This finding is

consistent with previous research showing that exposure combined with affect labeling results in

greater reduction in galvanic skin response than exposure alone (Kircanski et al., 2012; Tabibnia

et al., 2008). It is notable that the effect was found only for skin conductance response and heart

rate during recovery following the speech and not anticipation of the speech. Although research

on public speaking tasks generally focuses on anticipation of rather than recovery from speaking,

post-event processing is common in patients with social anxiety and relates both to the severity

of social anxiety symptoms and predicts subsequent avoidance of similar social situations

(Rachman, Grüter-Andrew, & Shafran, 2000). Therefore, the effect of affect labeling on

reducing physiological activation during recovery following the speech may reflect reduction in

negative post event processing and rumination.

Additional analyses were conducted comparing participants in the exposure alone

condition to those in the exposure plus affect labeling condition who were “high” labelers (i.e. in

the top 50th percentile of number of labels chosen). Because there were participants who chose

no or very few affect labels during exposure, limiting the sample to those in the top 50th

percentile included only those participants who chose at least 7 labels (out of 20 possible) during

exposures. When only high labelers were included in the group comparison, consistent with

hypotheses, participants in the affect labeling group had marginally significantly lower heart rate

and significantly fewer non-specific skin conductance responses during anticipation of giving a

speech, and significantly lower heart rate and marginally significantly fewer non-specific skin

conductance responses during recovery following the speech. Although these findings replicate

previous research, causality cannot be inferred given that participants were not randomly

assigned to the high labeling group within the affect labeling condition. It is possible that high

labelers in the affect labeling condition differed in some meaningful way from participants

assigned to the exposure alone condition that explains the significant group differences.

However, it is important to note that the analytic approach controlled for baseline levels of the

dependent variables, so group differences at time points 2 and 3 cannot be explained by

preexisting baseline group differences on the dependent measures. Because many participants

did not choose the affect labels provided, a more effective approach for future studies may be to

allow participants to generate their own affective labels as in Kircanski et al (2012). However,

researchers have yet to assess whether matching participants’ labels to their own affective

experience produces greater benefit than providing predetermined labels from which participants

can choose. Regardless of whether participants generate their own labels or not, future studies

should ensure that all participants in the labeling group engage, at least to some extent, in affect

labeling and are not given the option to completely refrain from labeling.

Contrary to hypotheses, on self-report measures of anxiety, participants in the affect

labeling condition had marginally significantly less improvement in self rated performance

than participants in the exposure alone condition. In addition, high labelers in the affect

labeling condition had marginally significantly less improvement in trait public speaking

anxiety and anxiety while speaking than did participants in the exposure alone condition.

These findings however were limited to the change from Time 1 to Time 2 and were no longer

significant at Time 3, and no group differences remained by the third time point. In a previous

study, the benefit of affect labeling during exposure was found only for galvanic skin response

and not for self-report measures (Kircanski et al., 2012). One possible explanation provided by

Kircanski and colleagues is that people tend to predict that affect labeling will not be an

effective strategy for reducing distress (Lieberman et al., 2011). Another possibility is that by

repeatedly labeling anxiety and feared outcomes during exposure, high labelers in the affect

labeling group were trained to report increased anxiety symptoms on self report measures

administered during subsequent assessments. Consistent with mindfulness based approaches to

the treatment of

anxiety such as Acceptance and Commitment Therapy (ACT), it is possible that participants

who engaged in affect labeling became more willing to report anxiety symptoms at follow-up

assessments, but were less distressed by these symptoms, and therefore less physiologically

reactive. A phrase often used in ACT regarding mindfulness practice is that the goal is not to

feel better, but to get better at feeling (Hayes et al., 1999). Therefore, although not assessed in

the current study, perhaps through labeling, participants became more attuned to and accepting

of their emotions and anxiety symptoms, which led to increased symptom reporting, but

decreased physiological activation. Future studies that include training in affect labeling should

measure acceptance of anxiety in addition to anxiety symptoms.

Inconsistent with previous research (Kircanski et al., 2012), we did not find that affect

labeling reduced avoidance of public speaking. This failure to replicate previous findings may

be attributable to floor effects on our measure of avoidance that limited statistical power to

detect group differences. Participants were given the option to give an additional speech in

exchange for $5. Although approximately 39% of participants avoided the additional speech at

the baseline assessment, following exposure and at one-week follow-up, only 13% of

participants avoided the additional speech. Given the sample size, this dichotomous outcome did

not allow enough variability to detect group differences. Future studies should use a measure of

avoidance that allows for greater variability (e.g. amount of time speaking) or should use a larger

sample size to increase statistical power.

4.2. Aim 2: Does the Number of Anxiety Related Words Used During Exposure Predict

Greater Attenuation of Fear Responding at Re-Test?

Consistent with hypotheses, the more anxiety labels participants chose during exposure,

the fewer non-specific skin conductance responses participants had during anticipation of giving

a speech by the one-week follow-up. These findings are consistent with previous research in

spider fearful participants where it was found that participants who used more anxiety words

during affect labeling had a greater reduction in spontaneous skin conductance response at re-test

than did participants who used fewer anxiety words (Kircanski et al., 2012). Perhaps the use of

more affect labels during exposure lead to greater activation in PFC-Amygdala pathways. If

participants with anxiety show weaker connectivity between areas of the PFC (such as

RVLPFC) and the amygdala, it is possible that the number of repetitions of activation in these

pathways during exposure is positively correlated with the strength of PFC-amygdala

connectivity following completion of exposure. Consistent with the principle of neural

plasticity, which states that repetition of a process can increase efficiency and efficacy of that

process through changes in neuron function, chemical profile, and structure (Anderson, 2010;

Kandel & Schwartz, 1982), greater activation of PFC-amygdala neural pathways as a result of

more frequent labeling may have produced greater neural change and ultimately more effective

down regulation of physiological fear responding.

Contrary to hypotheses, following exposure, participants who used fewer affect labels

during exposure showed greater improvement in trait public speaking anxiety, anxiety while

speaking, and self reported public speaking performance than did participants who used more

affect labels during exposure. Consistent with findings from study Aim 1, self reported

symptoms did not align with physiological measures of anxious responding. Perhaps

participants who chose more anxiety labels during exposure reported more symptoms in order to

maintain consistency across different self-report assessments. Festinger’s theory of cognitive

dissonance (1957) suggests that individuals are motivated to maintain internal consistency, and

discomfort arises when beliefs and actions are inconsistent across different settings. Therefore,

perhaps when participants engaged in more frequent labeling of negative emotions, in order for

their responses to be consistent from one setting to the next, they reported more anxiety

symptoms despite having less physiological arousal. Given a longer lag time between labeling

and the assessment of self-report symptoms, perhaps the importance of consistency on self-report

measures would be diminished. It is also possible that, had participants been followed for a

longer period of time, those who labeled more frequently would show continued improvement on

self-report measures, while those who labeled less frequently would show a return of symptoms.

4.3. Aim 3: Does Incidental Emotion Regulation at Baseline Moderate or Predict Response

to Exposure with Affect Labeling Versus Exposure Alone?

Contrary to hypotheses, we found that participants who had deficits incidental emotion

regulation at baseline benefited more from exposure combined with affect labeling than exposure

alone on measures of physiology including heart rate during speech anticipation and recovery,

and non-specific skin conductance responses during recovery. Based on previous research that

supports the matching hypothesis (Engebretson, Matthews, & Scheier, 1989; Niles et al., 2014;

Stanton et al., 2000), we hypothesized that participants with strengths in incidental emotion

regulation at baseline would show greater benefit from an intervention that included affect

labeling. Previous research however did not measure physiological activation as the outcome,

but instead measured self-reported symptoms. Therefore, the differences in findings may be

attributed to the different method of outcome assessment. In healthy participants, affect labeling

leads to a reduction in self-reported distress while viewing negative images by increasing

activation in areas of the PFC and decreasing activation in the amygdala. It is possible that

participants who showed less benefit from incidental emotion regulation have greater deficits in

PFC-Amygdala connectivity, and as a result benefited more from an intervention that

specifically targeted these connectivity deficits. Neural activation however was not directly

assessed in the current project, and future studies that include affect labeling as a strategy for

enhancing exposure effectiveness would benefit from inclusion of measures of neural activation

before and after the intervention.

Contrary to hypotheses, from Time 1 to Time 2, we found that participants with greater

deficits in incidental emotion regulation did better in exposure alone than exposure plus affect

labeling on self-reported negative and positive cognitions and anxiety during public speaking.

Results should be interpreted with caution given that effects were marginally significant and

were not maintained at one-week follow-up. Consistent with findings from study Aims 1 and 2,

findings on the self-report measures were not consistent with findings on the physiological

measures. Again, one possibility is that by labeling anxiety and feared outcomes during

exposure, participants in the affect labeling group were trained to report increased anxiety

symptoms on self report measures administered during subsequent BATs. The moderation

effects however suggest that only participants with deficits in incidental emotion regulation later

reported more symptoms. It is possible that these participants’ incidental emotion regulation

ability improved, and as a result, they were more likely to report symptoms because they found

this strategy more effective at reducing physiological arousal.

For marginally significant moderator findings, incidental emotion regulation capacity

was also tested as an overall predictor of outcome (regardless of group). Significant prediction

effects were found for negative and positive cognitions during public speaking, and a marginally

significant effect was found for self-reported anxiety during public speaking. Across all three

dependent measures, participants with deficits in incidental emotion regulation at baseline

showed the greatest improvement in anxious responding following exposure. More specifically,

participants with greater deficits in incidental emotion regulation reported significantly more

anxiety symptoms at baseline than participants with less deficit, but following completion of

exposure, those with deficits no longer differed from those without deficits. Because incidental

emotion regulation is a proxy for prefrontal-amygdala connectivity, these findings imply that

exposure, regardless of the inclusion of affect labeling, is particularly effective for participants

with greater deficits in prefrontal-amygdala circuitry.

4.4. Differential Findings for Self-Report and Physiological Measurement

The lack of synchrony between the physiological measurement and self reported anxiety

is an important consideration for research on the treatment of anxiety. At baseline, the increase

in heart rate, skin conductance level, and non-specific skin conductance responses during

anticipation did not correlate with self-report measures of anxiety. The only significant

correlation with a self-report measure was found between non-specific skin conductance

fluctuation increase during anticipation and depressive symptoms. This suggests that physiology

may index a different facet of anxious responding than does self-reported anxiety. It is also

notable that the only significant correlate of increased avoidance of the optional speech at

baseline was the amount of increase in skin conductance response from baseline to recovery.

Self-report measures of anxiety were not significantly associated with behavioral avoidance.

Given the importance of avoidance in the maintenance of anxiety disorders, physiology,

particularly during recovery following a stressor, may be an important indicator of disorder

severity in addition to self-reported symptoms.

4.5. Summary and Implications

In sum, the current research supports the theory that affect labeling can enhance exposure

effectiveness. Patients with anxiety show deficits in prefrontal amygdala connectivity (Hahn et

al., 2011; Kim et al., 2011), and repeated activation of prefrontal regions that project to the

amygdala through exposure and affect labeling can lead to a reduction in physiological activation

in response to an anxiety provoking stimulus. The results of the current study indicate that the

more a fearful individual labels their emotional experience during exposure, the greater the

reduction in galvanic skin response and heart rate (measures of fear arousal) when they next

encounter the feared stimulus. The benefit of affect labeling was not shown for self-reported

anxiety, which may require longer-term follow up than was possible in the current study.

Finally, adding affect labeling to exposure was particularly beneficial (on physiological

outcomes only) for individuals who showed deficits in incidental emotion regulation, which is an

indicator of poor prefrontal-amygdala connectivity. This finding provides further evidence that

targeting prefrontal-amygdala circuitry in anxiety patients using tasks that activate key regions

involved in emotion regulation can improve treatment effectiveness, and that such interventions

will be particularly effective for patients who show the greatest deficits in this circuit.

Table 1. Measures Used at Each Stage of the Study.

BAT-1 Exp-1 Exp-2 BAT-2 BAT-3

MINI X

PRPSA X X X

SSPS X X X

PHQ X X X

SPIN X X X

PRF X X X

SUDS X X X X X

HR, SCR-NS, SCL X X X

Optional Speech X X X

% Anxiety Labels X X

Incidental Emotion Regulation X

ECQ X

Note. MINI=MINI Neuropsychiatric Interview for Social Anxiety; PRPSA=Personal Report of

Public Speaking Anxiety; SSPS=Self-Statements During Public Speaking Questionnaire;

PHQ=Patient Health Questionnaire; SPIN=Social Phobia Inventory; PRF=Performance Rating

Form; SUDS=Subjective Units of Distress Scale; HR=Heart Rate; SCR-NS=Non Specific Skin

Conductance Response; SCL=Skin Conductance Level; % Anxiety Labels=Percent of anxiety

labels chosen; ECQ=Exposure Credibility Questionnaire

Table 2. Number of Missing Data Points and Reason for Missing Data for Study Variables by

Time Point

Time 1 Time 2 Time 3 Reasons Missing

Age 6 -- -- Participant Did Not Complete

Experimenter Error

Gender 0 -- --

Ethnicity 5 -- -- Participant Did Not Complete

Experimenter Error

Write in a Journal 10 -- -- Participant Did Not Complete

Experimenter Error

Student 3 -- -- Participant Did Not Complete

Experimenter Error

English First Language 5 -- -- Participant Did Not Complete

Experimenter Error

Born in United States 7 -- -- Participant Did Not Complete

Experimenter Error

MINI 1 -- -- Inadvertently Deleted

PRPSA 4 2 1 Inadvertently Deleted

SSPS 5 2 1 Inadvertently Deleted

Experimenter Error

PHQ 4 2 1 Inadvertently Deleted

SPIN 4 2 1 Inadvertently Deleted

PRF 5 2 1 Inadvertently Deleted

Experimenter Error

SUDS 1 0 0 Computer Program Failure

HR 1 2 1 Experimenter Error

SCR-NS/SCL 9 10 7 Participant Non Response

Optional Speech 0 0 0

% Anxiety Labels 0 0 --

Incidental Emotion 8 -- -- Participant Did Not Complete

Regulation Experimenter Error

ECQ 2 -- -- Experimenter Error

Note. MINI=MINI Neuropsychiatric Interview for Social Anxiety; PRPSA=Personal Report of

Public Speaking Anxiety; SSPS=Self-Statements During Public Speaking Questionnaire;

PHQ=Patient Health Questionnaire; SPIN=Social Phobia Inventory; PRF=Performance Rating

Form; SUDS=Subjective Units of Distress Scale; HR=Heart Rate; SCR-NS=Non Specific Skin

Conductance Response; SCL=Skin Conductance Level; % Anxiety Labels=Percent of anxiety

labels chosen; ECQ=Exposure Credibility Questionnaire

Table 3. Descriptive Statistics and Tests of Baseline Group Differences for Demographics

Overall (n=100) AL (n=52) Control (n=48) Statistical Test

Age (mean (sd)) 25.3 (9.1) 24.7 (8.9) 26.1 (9.3) t(92) = .7

Female

Ethnicity

80% (80/100) 77% (40/52) 83% (40/48) χ2(1) = .6

Fisher’s p = .83a

Asian 55% (52/95) 57% (28/49) 52% (24/46)

Hispanic 16% (15/95) 14% (7/49) 17% (8/46)

African American 6% (6/95) 6% (3/49) 7% (3/46)

Caucasian 14% (13/95) 10% (5/49) 17% (8/46)

Other 9% (9/95) 12% (6/49) 7% (3/46)

Write in a Journal 17% (15/90) 14% (6/44) 20% (9/46) χ2(1) = .6

Student 92% (89/97) 94% (47/50) 89% (42/47) χ2(1) = .7

English First Language 63% (60/95) 63% (30/48) 64% (30/47) χ2(1) = 0

Born in United States 75% (70/93) 79% (37/47) 72% (33/46) χ2(1) = .6

Social Anxiety Disorder 36% (36/99) 35% (18/51) 38% (18/48) χ2(1) = .8

PRPSA Score 133.0 (14.7) 133.2 (14.7) 132.8 (16.1) t(94) = -.1

PHQ Score 5.6 (4.3) 5.9 (4.4) 5.3 (4.2) t(94) = -.7

SPIN Score 6.4 (3.0) 6.7 (2.8) 6.1 (3.2) t(94) = -1.0

* p < .05

aFisher’s exact test was used to test for significance due to small cell sizes

Note. AL = Affect Labeling Group; Control = Control Group; PRPSA=Personal Report of Public

Speaking Anxiety; PHQ=Patient Health Questionnaire; SPIN=Social Phobia Inventory

Table 4. Descriptive Statistics by Group and Time for BAT Measures

AL Control

Speech (%)

Note. HR=Heart Rate; SCL=Skin Conductance Level; SCR-NS=Non Specific Skin Conductance

Response; PRPSA=Personal Report of Public Speaking Anxiety; SUDS=Subjective Units of

Distress Scale; SSPS=Self-Statements During Public Speaking Questionnaire;

PRF=Performance Rating Form

T1 (n=52) T2 (n=43) T3 (n=41) T1 (n=48) T2 (n=42) T3 (n=40)

Mean (SD)

Baseline 77.1 (12.6) 77.3 (11.7) 77.4 (12.5) 72.4 (10.0) 73.9 (11.0) 76.5 (13.6)

Anticipation 81.9 (13.3) 81.3 (12.2) 80.4 (13.3) 78.3 (12.1) 78.1 (11.9) 79.8 (13.1)

Recovery 75.5 (13.1) 75.5 (10.5) 74.7 (12.7) 71.4 (9.7) 71.0 (10.3) 72.8 (12.0)

SCL

Baseline 2.0 (1.1) 1.9 (1.2) 1.9 (1.2) 2.1 (1.1) 1.8 (1.1) 2.1 (1.7)

Anticipation 3.0 (1.6) 2.6 (1.6) 2.6 (1.5) 3.0 (1.2) 2.6 (1.2) 2.9 (1.8)

Recovery 3.1 (1.4) 3.2 (1.4) 3.3 (1.2) 3.2 (1.2) 3.0 (1.2) 3.3 (1.8)

SCR-NS

Baseline .8 (1.3) .5 (1.2) .8 (1.3) .3 (.8) .2 (.8) .8 (1.4)

Anticipation 2.5 (1.9) 1.9 (2.0) 1.8 (2.4) 2.5 (1.7) 1.9 (2.2) 2.0 (1.8)

Recovery .8 (1.3) .6 (1.3) .3 (.7) .6 (.9) .4 (.7) .6 (.9)

PRPSA 133.2 (14.7) 126.2 (17.6) 122.7 (17.8) 132.8 (16.1) 121.5 (16.1) 117.3 (15.1)

SUDS 4.9 (1.4) 3.7 (1.8) 3.3 (1.8) 4.7 (1.6) 3.5 (1.5) 3.4 (1.6)

SSPS Positive 14.4 (3.9) 15.7 (3.5) 15.9 (4.1) 14.8 (4.0) 16.5 (3.3) 17.5 (3.1)

SSPS

Negative 16.5 (4.3) 14.3 (4.7) 13.4 (4.5) 16.5 (4.1) 14.2 (3.8) 12.5 (3.8)

PRF 25.7 (9.1) 35.6 (10.1) 36.9 (10.1) 24.0 (8.0) 32.6 (8.3) 37.6 (9.9)

Optional 60% (31/52) 88% (38/43) 88% (36/41) 63% (30/48) 83% (35/42) 85% (34/40)

Table 5. Descriptive Statistics by Group and Time for EXP measures

EXP 1 (n=49) EXP 2 (n=43)

Mean Percentage (SD)

Affect Labels

Anxiety Words 36% (26) 37% (26)

Anger Words 7% (11) 9% (11)

Sadness Words 10% (13) 10% (13)

Other 46% (35) 43% (33)

Feared Outcome 73% (34) 75% (32)

AL (n=49) Control (n=46) AL (n=43) Control (n=43)

Mean (SD)

SUDS

Beginning 4.7 (1.8) 4.0 (1.7) 3.6 (1.9) 4.0 (1.7)

End 5.1 (1.9) 4.4 (2.0) 3.8 (1.8) 3.5 (1.7)

ECQ 19.0 (5.5) 19.6 (6.4) -- --

Note. SUDS = Subjective Units of Distress; ECQ = Exposure Credibility Questionnaire

* p < .05, ** p < .01, *** p < .001

Note. PRPSA = Personal Report of Public Speaking Anxiety; SSPSp = Self Statements During Public Speaking – Positive;

SSPSn = Self Statements During Public Speaking – Negative; PHQ = Patient Health Questionnaire; SPIN = Social Phobia

Inventory; PRF = Performance Rating Form; SUDS = Subjective Units of Distress; HR = Heart Rate; SCL = Skin

Conductance Level SCR-NS = Non Specific Skin Conductance Response; b = anticipation; a-b = anticipation minus

baseline; r-b = recovery minus baseline; Opt = optional speech; % anx = percent anxiety labels chosen; % fo = percent

feared outcomes chosen; IER = Incidental Emotion Regulation

Table 6. Correlations Among Key Study Variables

Table 7. Descriptive Statistics Completers vs. Dropout for Relevant Study Variables

Completers (n=81) Dropout (n=19) t statistic

Mean (SD)

Baseline 74.0 (11.5) 78.7 (11.6) t(97) = -1.6

Anticipation 78.9 (1.4) 85.7 (11.2) t(97) = -2.1*

Recovery

SCLa

72.5 (11.5) 78.1 (12.0) t(96) = -1.9t

Baseline .5 (.5) .5 (.7) t(89) = 0

Anticipation 1.0 (.4) 1.1 (.5) t(89) = -.9

Recovery

SCR-NSb

1.1 (.3) 1.2 (.5) t(89) = -1.0

Baseline .5 (1.2) .6 (1.2) z=.1

Anticipation 2.3 (1.7) 3.4 (1.9) z=-1.8t

Recovery .9 (1.2) .6 (1.1) z=-1.4

PRPSA 131.9 (15.1) 137.7 (16.0) t(94) = -1.4

SUDS (BAT) 4.7 (1.5) 5.0 (1.5) t(97) = -.8

SSPS Pos 15.0 (3.9) 12.8 (3.7) t(93) = 2.1*

SSPS Neg 16.4 (4.1) 16.9 (4.7) t(93) = -.5

PRF 25.7 (8.6) 20.9 (7.6) t(93) = 2.2*

Optional Speech (%) 65% (53/81) 42% (8/19) χ2(1) = 3.5t

PHQ 14.4 (4.0) 15.4 (5.4) t(94) = -.8

SPIN 9.2 (3.1) 10.1 (2.6) t(94) = -1.1

SUDS (exposure)

Beginning 4.2 (1.7) 4.9 (1.9) t(93) = -1.5

End 4.7 (2.0) 5.1 (2.0) t(88) = -.7

ECQ 19.1 (5.6) 19.8 (7.2) t(96) = -.5

t p < .10, * p < .05

a SCL values were log transformed

b Mann-Whitney test was used because variable is non-normal (count)

Note. HR=Heart Rate; SCL=Skin Conductance Level; SCR-NS=Non Specific Skin Conductance

Response; PRPSA=Personal Report of Public Speaking Anxiety; SUDS=Subjective Units of

Distress Scale; SSPS=Self-Statements During Public Speaking Questionnaire; PRF=Performance

Rating Form; PHQ = Patient Health Questionnaire; SPIN = Social Phobia Inventory; ECQ =

Exposure Credibility Questionnaire

Figure 1. Study Procedure

Affect

Labeling

Task

BAT-1

Exp-1 BAT-2

Exp-2

BAT-3

Enrollment

via phone

screening

Time 1 Time 2 Time 3

3 days 5 days

30 Min60 Min

75 Min

Figure 2. Participant Flow Chart

* See Table 2 for additional information regarding missing data

Figure 3. Heart Rate Over Time by Group during Anticipation

Figure 4. Heart Rate During Recovery Over Time by Group

Figure 5. Skin Conductance Level During Anticipation Over Time by Group

Figure 6. Skin Conductance Level During Recovery Over Time by Group

Figure 7. Non-Specific Skin Conductance Response During Anticipation Over Time by Group

Figure 8. Non-Specific Skin Conductance Response During Recovery Over Time by Group

Figure 9. Personal Report of Public Speaking Anxiety Over Time by Group

Figure 10. Subjective Units of Distress Over Time by Group

Figure 11. Self Statements During Public Speaking Over Time by Group: (A) Positive Self

Statements; (B) Negative Self Statements

(A)

(B)

Figure 12. Performance Rating Form Over Time by Group

Figure 13. Optional Speech Over Time by Group

Figure 14. Heart Rate During Anticipation over Time by Group for High Labelers Only

Figure 15. Heart Rate During Recovery over Time by Group for High Labelers Only

Figure 16. Non-Specific Skin Conductance Response during Anticipation over Time by Group

for High Labelers Only

Figure 17. Non-Specific Skin Conductance Response During Recovery over Time by Group for

High Labelers Only

Figure 18. Personal Report of Public Speaking Anxiety Over Time by Group for High Labelers

Only

Figure 19. Subjective Units of Distress Over Time by Group for High Labelers Only

Figure 20. Association Between Number of Anxiety Words Chosen and Non-Specific Skin

Conductance Response during Anticipation over Time

Figure 21. Association Between Number of Anxiety Words Chosen and Personal Report of

Public Speaking Anxiety Over Time

Figure 22. Association Between Number of Anxiety Words Chosen and Subjective Units of

Distress Over Time

Figure 23. Association Between Number of Anxiety Words Chosen and Performance Rating

Form Over Time

Figure 24. Moderation of Heart Rate during anticipation over Time by Incidental Emotion

Regulation and Group

Figure 25. Moderation of Heart Rate during recovery over Time by Incidental Emotion

Regulation and Group

Figure 26. Prediction of Non-Specific Skin Conductance Response during anticipation over

Time by Incidental Emotion Regulation

Figure 27. Moderation of Non-Specific Skin Conductance Response During Recovery Over

Time by Incidental Emotion Regulation and Group

Figure 28. Moderation of Subjective Units of Distress Over Time by Incidental Emotion

Regulation and Group

Figure 29. Prediction of Subjective Units of Distress Over Time by Incidental Emotion

Regulation

Figure 30. Moderation of Self Statements During Public Speaking Positive Over Time by

Incidental Emotion Regulation and Group

Figure 31. Prediction of Self Statements During Public Speaking Positive Over Time by

Incidental Emotion Regulation

Figure 32. Moderation of Self Statements During Public Speaking Negative over Time by

Incidental Emotion Regulation and Group

Figure 33. Prediction of Self Statements During Public Speaking Negative Over Time by

Incidental Emotion Regulation

References

Anderson, M. L. (2010). Neural reuse: A fundamental organizational principle of the brain.

Behavioral and Brain Sciences, 33(4), 245.

Arch, J. J., & Craske, M. G. (2008). Acceptance and commitment therapy and cognitive

behavioral therapy for anxiety disorders: Different treatments, similar mechanisms?

Clinical Psychology: Science and Practice, 15(4), 263–279.

Austenfeld, J. L., Paolo, A. M., & Stanton, A. L. (2006). Effects of writing about emotions

versus goals on psychological and physical health among third-year medical students.

Journal of Personality, 74(1), 267–286. doi:10.1111/j.1467-6494.2005.00375.x

Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal

attachments as a fundamental human motivation. Psychological Bulletin, 117, 497–497.

Beatty, M. J., & Behnke, R. R. (1991). Effects of public speaking trait anxiety and intensity of

speaking task on heart rate during performance. Human Communication Research, 18(2),

147–176. doi:10.1111/j.1468-2958.1991.tb00542.x

Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of

Pavlovian learning. Psychological Bulletin, 114(1), 80.

Bouton, M. E., Woods, A. M., Moody, E. W., Sunsay, C., & García-Gutiérrez, A. (2006).

Counteracting the context-dependence of extinction: Relapse and tests of some relapse

prevention methods. In M. G. Craske, D. Hermans, & D. Vansteenwegen (Eds.), Fear

and learning: From basic processes to clinical implications (pp. 175–196). Washington,

DC: American Psychological Association.

Bradley, M. M., Codispoti, M., Cuthbert, B. N., & Lang, P. J. (2001). Emotion and motivation I:

Defensive and appetitive reactions in picture processing. Emotion, 1(3), 276–298.

doi:10.1037/1528-3542.1.3.276

Braithwaite, J. J., Watson, D. G., Jones, R., & Rowe, M. (2013). A guide for analysing

electrodermal activity (EDA) and skin conductance responses (SCRs) for

psychophysiological experiments via the MP36R and AcqKnowledge software (No. 1).

University of Birmingham, UK: Selective Attention & Awareness Laboratory (SAAL),

Behavioral Brain Sciences Centre, School of Psychology.

Brühl, A. B., Rufer, M., Delsignore, A., Kaffenberger, T., Jäncke, L., & Herwig, U. (2011).

Neural correlates of altered general emotion processing in social anxiety disorder. Brain

Research, 1378, 72–83.

Burklund, L. J., Creswell, J. D., Irwin, M. R., & Lieberman, M. D. (2014). The common and

distinct neural bases of affect labeling and reappraisal in healthy adults. Frontiers in

Psychology, 5. doi:10.3389/fpsyg.2014.00221

Butler, A. C., Chapman, J. E., Forman, E. M., & Beck, A. T. (2006). The empirical status of

cognitive-behavioral therapy: A review of meta-analyses. Clinical Psychology Review,

26(1), 17–31. doi:10.1016/j.cpr.2005.07.003

Cacioppo, J. T., Tassinary, L. G., & Berntson, G. (2007). Handbook of Psychophysiology (3rd

ed.). New York, NY: Cambridge University Press.

Connor, K. M., Kobak, K. A., Churchill, L. E., Katzelnick, D., & Davidson, J. R. (2001). Mini-

SPIN: A brief screening assessment for generalized social anxiety disorder. Depression

and Anxiety, 14(2), 137–140.

Craske, M. G. (2003). Origins of phobias and anxiety disorders: Why more women than men?

Toronto, ON: Elsevier.

Craske, M. G., Antony, M. M., & Barlow, D. H. (2006). Mastering your fears and phobias:

Therapist guide. New York, NY: Oxford University Press.

Craske, M. G., Kircanski, K., Zelikowsky, M., Mystkowski, J., Chowdhury, N., & Baker, A.

(2008). Optimizing inhibitory learning during exposure therapy. Behaviour Research and

Therapy, 46(1), 5–27. doi:10.1016/j.brat.2007.10.003

Craske, M. G., Lang, A. J., Aikins, D., & Mystkowski, J. L. (2005). Cognitive behavioral

therapy for nocturnal panic. Behavior Therapy, 36(1), 43–54. doi:10.1016/S0005-

7894(05)80053-X

Craske, M. G., & Mystkowski, J. L. (2006). Exposure therapy and extinction: Clinical studies. In

M. G. Craske, D. Hermans, & D. Vansteenwegen (Eds.), Fear and learning: From basic

processes to clinical implications (pp. 217–233). Washington, DC: American

Psychological Association.

Craske, M. G., Niles, A. N., Burklund, L. J., Wolitzky-Taylor, K., Plumb, J., Saxbe, D., &

Lieberman, M. (In Press). Randomized controlled trial of cognitive behavioral therapy

and acceptance and commitment therapy for social anxiety disorder: Outcomes and

moderators. Journal of Consulting and Clinical Psychology.

Craske, M. G., Street, L. L., Jayaraman, J., & Barlow, D. H. (1991). Attention versus distraction

during in vivo exposure: Snake and spider phobias. Journal of Anxiety Disorders, 5(3),

199–211. doi:10.1016/0887-6185(91)90001-A

Deane, G. E. (1969). Cardiac activity during experimentally induced anxiety. Psychophysiology,

6(1), 17–30. doi:10.1111/j.1469-8986.1969.tb02879.x

Delgado, M. R., Nearing, K. I., LeDoux, J. E., & Phelps, E. A. (2008). Neural circuitry

underlying the regulation of conditioned fear and its relation to extinction. Neuron, 59(5),

829–838.

Dickerson, S. S., Gruenewald, T. L., & Kemeny, M. E. (2004). When the social self is

threatened: Shame, physiology, and health. Journal of Personality, 72(6), 1191–1216.

doi:10.1111/j.1467-6494.2004.00295.x

Dickerson, S. S., & Kemeny, M. E. (2004). Acute stressors and cortisol responses: A theoretical

integration and synthesis of laboratory research. Psychological Bulletin, 130(3), 355–391.

Eisenberger, N. I., & Lieberman, M. D. (2004). Why rejection hurts: a common neural alarm

system for physical and social pain. Trends in Cognitive Sciences, 8(7), 294–300.

Engebretson, T. O., Matthews, K. A., & Scheier, M. F. (1989). Relations between anger

expression and cardiovascular reactivity: Reconciling inconsistent findings through a

matching hypothesis. Journal of Personality and Social Psychology, 57(3), 513–21.

Feingold, A. (2009). Effect sizes for growth-modeling analysis for controlled clinical trials in the

same metric as for classical analysis. Psychological Methods, 14(1), 43–53.

doi:10.1037/a0014699

Feske, U., & Chambless, D. L. (1995). Cognitive behavioral versus exposure only treatment for

social phobia: A meta-analysis. Behavior Therapy, 26(4), 695–720. doi:10.1016/S0005-

7894(05)80040-1

Festinger, L. (1957). A Theory of Cognitive Dissonance. Stanford: Stanford University Press.

Frattaroli, J., Thomas, M., & Lyubomirsky, S. (2011). Opening up in the classroom: Effects of

expressive writing on graduate school entrance exam performance. Emotion, 11(3), 691.

100

Gorno-Tempini, M. L., Pradelli, S., Serafini, M., Pagnoni, G., Baraldi, P., Porro, C., … Nichelli,

P. (2001). Explicit and incidental facial expression processing: An fMRI study.

NeuroImage, 14(2), 465–473. doi:10.1006/nimg.2001.0811

Hahn, A., Stein, P., Windischberger, C., Weissenbacher, A., Spindelegger, C., Moser, E., …

Lanzenberger, R. (2011). Reduced resting-state functional connectivity between

amygdala and orbitofrontal cortex in social anxiety disorder. NeuroImage, 56(3), 881–

889. doi:10.1016/j.neuroimage.2011.02.064

Hariri, A. R., Bookheimer, S. Y., & Mazziotta, J. C. (2000). Modulating emotional responses:

Effects of a neocortical network on the limbic system. Neuroreport, 11(1), 43–48.

Hariri, A. R., Mattay, V. S., Tessitore, A., Fera, F., & Weinberger, D. R. (2003). Neocortical

modulation of the amygdala response to fearful stimuli. Biological Psychiatry, 53(6),

494–501.

Hartley, C. A., Fischl, B., & Phelps, E. A. (2011). Brain Structure correlates of individual

differences in the acquisition and inhibition of conditioned fear. Cerebral Cortex, 21(9),

1954–1962. doi:10.1093/cercor/bhq253

Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (1999). Acceptance and commitment therapy: An

experiential approach to behavior change. New York, NY: The Guilford Press.

Heimberg, R. G. (2002). Cognitive-behavioral therapy for social anxiety disorder: Current status

and future directions. Biological Psychiatry, 51(1), 101–108.

Hemenover, S. H. (2003). The good, the bad, and the healthy: Impacts of emotional disclosure of

trauma on resilient self-concept and psychological distress. Personality and Social

Psychology Bulletin, 29(10), 1236–1244.

101

Hofmann, S. G., & DiBartolo, P. M. (2000). An instrument to assess self-statements during

public speaking: Scale development and preliminary psychometric properties. Behavior

Therapy, 31(3), 499–515.

Hofmann, S. G., & Smits, J. A. J. (2008). Cognitive-behavioral therapy for adult anxiety

disorders: A meta-analysis of randomized placebo-controlled trials. The Journal of

Clinical Psychiatry, 69(4), 621–632.

Hope, D. A., Heimberg, R. G., Juster, H. A., & Turk, C. L. (2000). Managing social anxiety: A

cognitive-behavioral therapy approach client workbook. New York, NY: Oxford

University Press.

Hugdahl, K. (1995). Psychophysiology: The mind-body perspective. Boston, MA: Harvard

University Press.

Kandel, E. R., & Schwartz, J. H. (1982). Molecular biology of learning: Modulation of

transmitter release. Science, 218(4571), 433–443.

Kim, M. J., Loucks, R. A., Palmer, A. L., Brown, A. C., Solomon, K. M., Marchante, A. N., &

Whalen, P. J. (2011). The structural and functional connectivity of the amygdala: From

normal emotion to pathological anxiety. Behavioural Brain Research, 223(2), 403–410.

doi:10.1016/j.bbr.2011.04.025

Kircanski, K., Lieberman, M. D., & Craske, M. G. (2012). Feelings into words: Contributions of

language to exposure therapy. Psychological Science. doi:10.1177/0956797612443830

Klein, K., & Boals, A. (2001). Expressive writing can increase working memory capacity.

Journal of Experimental Psychology: General, 130(3), 520.

102

Knight, D. C., Smith, C. N., Cheng, D. T., Stein, E. A., & Helmstetter, F. J. (2004). Amygdala

and hippocampal activity during acquisition and extinction of human fear conditioning.

Cognitive, Affective, & Behavioral Neuroscience, 4(3), 317–325.

Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1999). International affective picture system

(IAPS): Technical manual and affective ratings. Gainesville, FL: The Center for

Research in Psychophysiology, University of Florida.

Lecrubier, Y., Sheehan, D., Weiller, E., Amorim, P., Bonora, I., Harnett Sheehan, K., … Dunbar,

G. (1997). The Mini International Neuropsychiatric Interview (MINI). A short diagnostic

structured interview: reliability and validity according to the CIDI. European Psychiatry,

12(5), 224–231. doi:10.1016/S0924-9338(97)83296-8

LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience, 23(1), 155–

184.

Lepore, S. J., Silver, R. C., Wortman, C. B., & Wayment, H. A. (1996). Social constraints,

intrusive thoughts, and depressive symptoms among bereaved mothers. Journal of

Personality and Social Psychology, 70(2), 271.

Levin, A. P., Saoud, J. B., Strauman, T., Gorman, J. M., Fyer, A. J., Crawford, R., & Liebowitz,

M. R. (1993). Responses of “generalized” and “discrete” social phobics during public

speaking. Journal of Anxiety Disorders, 7(3), 207–221. doi:10.1016/0887-

6185(93)90003-4

Lieberman, M. D. (2003). Reflexive and reflective judgment processes: A social cognitive

neuroscience approach. In J. P. Forgas, K. D. Williams, & W. von (Eds.), Social

judgments: Implicit and explicit processes (pp. 44–67). New York, NY: Cambridge

University Press.

103

Lieberman, M. D. (2011). Why symbolic processing of affect can disrupt negative affect: Social

cognitive and affective neuroscience investigations. Social Neuroscience: Toward

Understanding the Underpinnings of the Social Mind, 188–209.

Lieberman, M. D., Eisenberger, N. I., Crockett, M. J., Tom, S. M., Pfeifer, J. H., & Way, B. M.

(2007). Putting feelings into words: Affect labeling disrupts amygdala activity in

response to affective stimuli. Psychological Science, 18(5), 421–428.

doi:10.1111/j.1467- 9280.2007.01916.x

Lieberman, M. D., Inagaki, T. K., Tabibnia, G., & Crockett, M. J. (2011). Subjective responses

to emotional stimuli during labeling, reappraisal, and distraction. Emotion, 11(3), 468–

480. doi:10.1037/a0023503

Loerinc, A., Meuret, A., Twohig, M., Rosenfield, D., & Craske, M. G. (Submitted for

Publication). Response rates in CBT for anxiety disorders: Measurement matters.

Löwe, B., Kroenke, K., Herzog, W., & Gräfe, K. (2004). Measuring depression outcome with a

brief self-report instrument: Sensitivity to change of the Patient Health Questionnaire

(PHQ-9). Journal of Affective Disorders, 81(1), 61–66. doi:10.1016/S0165-

0327(03)00198-8

Löwe, B., Spitzer, R. L., Gräfe, K., Kroenke, K., Quenter, A., Zipfel, S., … Herzog, W. (2004).

Comparative validity of three screening questionnaires for DSM-IV depressive disorders

and physicians’ diagnoses. Journal of Affective Disorders, 78(2), 131–140.

Lu, Q., & Stanton, A. L. (2009). How benefits of expressive writing vary as a function of writing

instructions, ethnicity and ambivalence over emotional expression. Psychology & Health,

25(6), 669–684. doi:10.1080/08870440902883196

104

Masterson, F. A., & Crawford, M. (2010). The defense motivation system: A theory of

avoidance behavior. Behavioral and Brain Sciences, 5(04), 661.

doi:10.1017/S0140525X00014114

McCroskey, J. C. (1970). Measures of communication‐bound anxiety. Speech Monographs,

37(4), 269–277. doi:10.1080/03637757009375677

Milad, M. R., Quinn, B. T., Pitman, R. K., Orr, S. P., Fischl, B., & Rauch, S. L. (2005).

Thickness of ventromedial prefrontal cortex in humans is correlated with extinction

memory. Proceedings of the National Academy of Sciences of the United States of

America, 102(30), 10706–10711.

Milad, M. R., & Quirk, G. J. (2002). Neurons in medial prefrontal cortex signal memory for fear

extinction. Nature, 420(6911), 70–74. doi:10.1038/nature01138

Mineka, S., Mystkowski, J. L., Hladek, D., & Rodriguez, B. I. (1999). The effects of changing

contexts on return of fear following exposure therapy for spider fear. Journal of

Consulting and Clinical Psychology, 67(4), 599.

Myers, R. M. (1974). Validation of systematic desensitization of speech anxiety through galvanic

skin response. Speech Monographs, 41(3), 233–235. doi:10.1080/03637757409375841

Narumoto, J., Yamada, H., Iidaka, T., Sadato, N., Fukui, K., Itoh, H., & Yonekura, Y. (2000).

Brain regions involved in verbal or non-verbal aspects of facial emotion recognition.

Neuroreport, 11(11), 2571–2576.

Niles, A. N., Haltom, K. E. B., Mulvenna, C. M., Lieberman, M. D., & Stanton, A. L. (2014).

Randomized controlled trial of expressive writing for psychological and physical health:

The moderating role of emotional expressivity. Anxiety, Stress, and Coping, 27(1), 1–17.

doi:10.1080/10615806.2013.802308

105

Niles, A. N., Mesri, B., Burklund, L. J., Lieberman, M. D., & Craske, M. G. (2013). Attentional

bias and emotional reactivity as predictors and moderators of behavioral treatment for

social phobia. Behaviour Research and Therapy, 51(10), 669–679.

doi:10.1016/j.brat.2013.06.005

Novalis, P. N., M. D., Rojcewicz, S. J., M. D., & Peele, R., M. D. (1993). Clinical manual of

supportive psychotherapy. Washington, DC: American Psychiatric Press.

Öhman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module

of fear and fear learning. Psychological Review, 108(3), 483.

Páez, D., Velasco, C., & González, J. L. (1999). Expressive writing and the role of alexythimia

as a dispositional deficit in self-disclosure and psychological health. Journal of

Personality and Social Psychology, 77(3), 630–641.

Pennebaker, J. W. (1997). Writing about emotional experiences as a therapeutic process.

Psychological Science, 8(3), 162–166.

Pennebaker, J. W., & Beall, S. K. (1986). Confronting a traumatic event: Toward an

understanding of inhibition and disease. Journal of Abnormal Psychology, 95(3), 274–

281.

Phelps, E. A., Delgado, M. R., Nearing, K. I., & LeDoux, J. E. (2004). Extinction learning in

humans: Role of the amygdala and vmPFC. Neuron, 43(6), 897–905.

doi:10.1016/j.neuron.2004.08.042

Pollard, C. A., & Henderson, J. G. (1988). Four types of social phobia in a community sample.

The Journal of Nervous and Mental Disease, 176(7), 440–445.

Rachman, S., Grüter-Andrew, J., & Shafran, R. (2000). Post-event processing in social anxiety.

Behaviour Research and Therapy, 38(6), 611–617. doi:10.1016/S0005-7967(99)00089-3

106

Ramirez, G., & Beilock, S. L. (2011). Writing about testing worries boosts exam performance in

the classroom. Science, 331(6014), 211.

Rapee, R. M., & Heimberg, R. G. (1997). A cognitive-behavioral model of anxiety in social

phobia. Behaviour Research and Therapy, 35(8), 741–756.

Rapee, R. M., & Lim, L. (1992). Discrepancy between self-and observer ratings of performance

in social phobics. Journal of Abnormal Psychology, 101(4), 728.

Rescorla, R. A., & Heth, C. D. (1975). Reinstatement of fear to an extinguished conditioned

stimulus. Journal of Experimental Psychology: Animal Behavior Processes, 1(1), 88.

Resick, P. A., & Schnicke, M. (1993). Cognitive processing therapy for rape victims: A

treatment manual (Vol. 4). Newbury Park, CA: Sage Publications.

Ricker, S., & Bouton, M. (1996). Reacquisition following extinction in appetitive conditioning.

Learning & Behavior, 24(4), 423–436. doi:10.3758/BF03199014

Rodebaugh, T. L., Holaway, R. M., & Heimberg, R. G. (2004). The treatment of social anxiety

disorder. Clinical Psychology Review, 24(7), 883–908. doi:10.1016/j.cpr.2004.07.007

Rowe, M. K., & Craske, M. G. (1998). Effects of varied-stimulus exposure training on fear

reduction and return of fear. Behaviour Research and Therapy, 36(7-8), 719–734.

Selya, A. S., Rose, J. S., Dierker, L. C., Hedeker, D., & Mermelstein, R. J. (2012). A practical

guide to calculating cohen’s f(2), a measure of local effect size, from PROC MIXED.

Frontiers in Psychology, 3, 111. doi:10.3389/fpsyg.2012.00111

Shah, S. G., Klumpp, H., Angstadt, M., Nathan, P. J., & Phan, K. L. (2009). Amygdala and

insula response to emotional images in patients with generalized social anxiety disorder.

Journal of Psychiatry and Neuroscience, 34(4), 296–302.

107

Smyth, J. M., Hockemeyer, J. R., & Tulloch, H. (2008). Expressive writing and post-traumatic

stress disorder: Effects on trauma symptoms, mood states, and cortisol reactivity. British

Journal of Health Psychology, 13(1), 85–93.

Snijders, T. a. B., & Bosker, R. J. (1994). Modeled variance in two-level models. Sociological

Methods & Research, 22(3), 342–363. doi:10.1177/0049124194022003004

Spitzer, R. L., Kroenke, K., & Williams, J. B. W. (1999). Validation and utility of a self-report

version of PRIME-MD. JAMA: The Journal of the American Medical Association,

282(18), 1737–1744.

Stanton, A. L., Kirk, S. B., Cameron, C. L., & Danoff-Burg, S. (2000). Coping through

emotional approach: Scale construction and validation. Journal of Personality and

Social Psychology, 78(6), 1150–69.

Stein, M. B., Goldin, P. R., Sareen, J., Zorrilla, L. T. E., & Brown, G. G. (2002). Increased

amygdala activation to angry and contemptuous faces in generalized social phobia.

Archives of General Psychiatry, 59(11), 1027.

Stein, M. B., Walker, J. R., & Forde, D. R. (1996). Public-speaking fears in a community

sample. Prevalence, impact on functioning, and diagnostic classification. Archives of

General Psychiatry, 53(2), 169–174.

Tabibnia, G., Lieberman, M. D., & Craske, M. G. (2008). The lasting effect of words on

feelings: Words may facilitate exposure effects to threatening images. Emotion, 8(3),

307.

Tolin, D. F. (2010). Is cognitive-behavioral therapy more effective than other therapies? A meta-

analytic review. Clinical Psychology Review, 30(6), 710–720.

doi:10.1016/j.cpr.2010.05.003

108

Watson, J. B. (1970). Behaviorism. New York, NY: W.W. Norton.

Wittchen, H. U., Stein, M. B., & Kessler, R. C. (1999). Social fears and social phobia in a

community sample of adolescents and young adults: Prevalence, risk factors and co-

morbidity. Psychological Medicine, 29(2), 309–323.

Students also viewed