The concept of “Anxiety sensitivity” in social anxiety disorder presentations, symptomatology, and treatment: A theoretical perspective

Abstract Social Anxiety Disorder (SAD) is considered as prevalent, chronic, and debilitating, with complex symptomatology. Various theories and psychopathological models have been purposed to explain SAD. This article will try to focus and emphasis on the concept of Anxiety Sensitivity (AS) and its role in perceiving anxiety and its treatment. The AS refers to the fear of “somatic sensations” related to anxiety and seems to play an important role in the onset and maintenance of anxiety disorders, including SAD. It seems that AS can play a role in various ways, such as fear of showing anxiety related somatic symptoms, fear of cognitive dyscontrol, and fear of losing control of mental abilities. Considering the SAD etiological models, they generally emphasize on the role of cognitive processing and underlying vulnerabilities in the development of this disorder. However, these models have not considered the role of AS directly, but the precise investigation of etiological models showed that the core elements of these models have implicitly regarded this construct. Altogether, incorporating Anxiety Sensitivity into the treatment model could further enhance the understanding of the symptomatology, as well as its dynamic and treatment modality.


Introduction
Social anxiety disorder (SAD) is a common, chronic, and debilitating disorder with lifetime prevalence and persistence rate of 4% and 60%, respectively (Stein, Lim, & Roest et al., 2017). The SAD is characterized by excessive fear of being humiliated, fear of social interactions and performance situations, avoidance of these situations, and physiological arousal (Aderka, McLean, & Huppert et al., 2013;Association AP, 2013). Socially anxious individuals are afraid that their behavior or anxiety-related symptoms are negatively being evaluated by the others and try to avoid anxiety provocative social situations, as much as possible. Avoiding social situations itself intensifies the symptoms and interferes with fear extinction. There are also some underlying factors for SAD including high physiological arousal, biased information processing, being alerted and fear of being scrutinized by others, low self-evaluation, and Anxiety sensitivity (AS) (Allan, Cooper, & Oglesby et al., 2018;Clark & McManus, 2002;Liebowitz, Gorman, & Fyer et al., 1985).
AS defined as the fear of anxiety-related sensations which are believed to have physical, cognitive, and social consequences (Deacon & Abramowitz, 2006;Taylor, 2014). For individuals with high AS, anxiety symptoms themselves are distressing and maintain the anxiety. Though AS was initially discussed with regard to panic disorder, transdiagnostic approaches have accumulated that it is an important transdiagnostic factor in etiology, assessment, and treatment of multiple emotional disorders including SAD (Panayiotou, Karekla, & Panayiotou, 2014), obsessivecompulsive disorder (OCD) (Timpano, Raines, & Shaw et al., 2016), depression (Capron, Allan, & Ialongo et al., 2015), post-traumatic stress disorder (PTSD) (Wald & Taylor, 2008), generalized anxiety disorder (GAD) (Schmidt, Mitchell, & Richey, 2008). Some studies, employing factor analyses, have also revealed the hierarchical and multifaceted nature of AS, consisting of a higher order factor, the general AS, and lower order factors, including fear of somatic sensations, fear of cognitive dyscontrol and fear of socially observable anxiety symptoms (Asmundson, Weeks, & Carleton et al., 2011;Wheaton, Deacon, & McGrath et al., 2012). Fear of somatic sensations is associated with worries about consequences of arousal sensations. Fear of cognitive dyscontrol represent worries about consequences of psychological symptoms. Finally, fear of socially observable anxiety symptoms results from thoughts that these symptoms might lead to social evaluation or rejection. This multi-dimensional conceptualization of AS shows that the relationships between AS and different anxiety-related psychopathologies are attributable to specific AS dimensions (Schmidt et al., 2008). Research on AS dimensions has demonstrated a different relationship between AS and various anxiety disorders. The AS fear of somatic sensations dimension, for example, is strongly associated with panic disorder (McNally, 2002), whereas the AS fear of cognitive dyscontrol dimension is moderately associated with depression and GAD (Rector, Szacun-Shimizu, & Leybman, 2007). Additionally, the AS social concerns dimension is most strongly related to the fear of negative evaluation and SAD (Allan et al., 2018).
The review of the literature indicates that there are limited studies on the relationship between AS and SAD. Boswell, Farchione, Sauer-Zavala, Murray, Fortune, and Barlow found that socially anxious individuals suffer from higher AS even more than those with panic disorder (Boswell, Farchione, & Sauer-Zavala et al., 2013). Alkozei, Cooper, and Creswell reported high levels of AS in children with SAD (Alkozei, Cooper, & Creswell, 2014). These studies suggested an association between high AS and SAD. Because socially anxious individuals often get involved with their anxiety symptoms, AS may be a strong candidate as a maintaining factor in SAD. For example, a person with a high level of AS may be afraid of a heartbeat because he/she is worried about the likelihood of a heart attack, or another one, who is worried about negative social evaluations, may be afraid of sweating in public places. In fact, they experience intense fear when exposed to a social situation that elicits anxiety and find their own anxiety symptoms to be particularly aversive (Reiss, 1991).
In this regard and according to the multi-dimensional and transdiagnostic applicability of AS, investigation of its potential associations with SAD would be informative. The present study is aimed to identify the relationship between AS and SAD, from perspectives of the onset, maintenance, and treatment of this disorder. We begin with describing three cognitive models, focusing on SAD underlying processes and then discuss SAD underlying factors and symptoms in relation with AS. Finally, we take a brief overview on various evidence-based treatments for AS and SAD.

Social anxiety etiological models
The search for a better understanding of underlying mechanisms that develop and maintain SAD has gained considerable attention over the last decades and several etiological models have been presented. These models mainly emphasize on the role of cognitive processing during a provocative event as well as underlying vulnerabilities. We are going to discuss three cognitive behavioral models which have contributed to the understanding of SAD.

Clark and Wells: self-focused attention
Clark and Wells' cognitive model specifically focuses on maintaining factors involved in SAD (Clark & Wells, 1995). They state that individuals' initial experiences form controversial assumptions about themselves and the social world. These assumptions make individuals perceive the social world as a threat and consider themselves as social objects, which in turn generate social anxiety symptoms. Clark and Wells consider two main factors in maintaining the SAD: highly self-focused attention (biased information processing) and safety behaviors (behavioral symptom).
The model mainly emphasizes on the attentional bias. Socially anxious individuals excessively shift their attention from the social cues toward the "inner symptoms" (self-focused attention), which can lead to increased awareness of feared anxiety responses. High SAD individuals use this internal information (such as anxious feelings) to inference about the situation and others' behavior (Clark & McManus, 2002). McEwan and Devins argue that socially anxious individuals are concerned about their observable symptoms because they experience elevated somatic anxiety symptoms (McEwan & Devins, 1983). In fact, fear of anxiety symptoms consequences (AS) can provoke more anxiety. These individuals are more likely to selectively pay attention to threatening anxiety symptoms and interpret anxiety-related sensations as potential dangers. For example, they perceive blushing as a sign of social embarrassment or confusion as a sign of mental instability. This shift in attention toward inside leads to experience more anxiety-related symptoms and maintain SAD in a vicious cycle (Clark & Wells, 1995). Another SAD maintaining factor is related to safety behaviors. High SAD individuals may avoid or, if necessary, use safety behaviors to reduce anxiety in social situations. These behavioral strategies negatively reinforce anxiety symptoms and negative beliefs which contradictorily exaggerate and maintain them (Clark & Wells, 1995).

Rapee and Heimberg: mental representation of self
Rapee and Heimberg (Rapee & Heimberg, 1997) suggested a similar model, in which social anxiety is initially provoked by the perception of others' evaluations. When socially anxious individuals encounter a social situation, they define themselves in the same ways as are seen by the others. They form this mental representation of themselves through information from past memories and internal and external cues. Unlike Clark and Wells's model, Rapee and Heimberg argued that, in addition to examining the internal symptoms of anxiety, people with the SAD seek out the external environment to find cues for their poor performance, and link this inaccurate information with the self-perception in a manner that is seen by others (Rapee & Heimberg, 1997). This self-perception then is compared to a perceived standard of audience expectations. The discrepancy between the two determines the probability of anticipating negative social outcomes, which in turn results in cognitive, physical, and behavioral symptoms of anxiety. Again, this model considers two maintain factors for SAD: biased information processing and cognitive, physical, and behavioral symptoms.
It has been shown that people with SAD have difficulties in shifting their attention from socially threatening subjects (Amir, Elias, & Klumpp et al., 2003;Heinrichs & Hofmann, 2001) and interpret and judge obscure or neutral signs in a threatening way (Musa & Lepine, 2000). They also avoid paying attention to positive environmental stimuli and form a weaker memory for that (Clark & McManus, 2002;Hunt, Keogh, & French, 2006;Musa & Lepine, 2000). In fact, they seek confirmatory information for negative consequences and selectively focus on the information, which results in less attention to important information around them (Amir et al., 2003;Hirsch & Clark, 2004).
Lucock and Salkovskis argued that, in comparison with the control group, people with SAD overestimate the probability of negative social consequences and underestimate the probability of positive social consequences (Lucock & Salkovskis, 1988). These people believe that in social situations, they will behave in a way that will increase the probability of adverse social consequences, which in turn influence the individual's representation of her/himself and maintain the pathological cycle of the SAD. On the other hand, socially anxious individuals, like individuals with high AS, overestimate their anxiety symptoms and perceive them as a threat. For example, people with SAD are ashamed and anxious about having observable feelings like embarrassment. They believe that other people are aware of their anxiety (e.g., "if I blush, others will immediately find out") and this awareness results in negative evaluations (Carpenter, Curtiss, & Hofmann, 2017). They reintegrate these symptoms into the mental representation of the self, resulting in maintenance of the social anxiety cycle. In fact, it is maladaptive beliefs, such as that anxiety symptoms are dangerous and have negative consequences (AS), in contribution with biased ways of processing information that leads individuals to maintain SAD.

Hofmann and Barlow: general and specific vulnerabilities
Based on the triple vulnerability theory, Hofmann and Barlow have proposed a model to explain SAD (Barlow, 2013). In their view, the development of SAD requires a biological and psychological vulnerability for the anxiety perception. They assumed dispositional factors such as shyness or behavioral inhibition as antecedents of social anxiety that can evoke false or correct alarm when faced with social or performance situations. This model assumes that SAD is more likely to occur without alarm or via false alarms associated with social-evaluative situations. Their findings indicated that individuals with SAD believe that they do not have any internal control over the situation because they had been experiencing false alarms repeatedly (Barlow, 2013;Hofmann, 2007). They proposed cognitive processes such as attentional bias, perception bias, implicit memory, and information retrieval, as effective components in maintaining this disorder (Barlow, 2013).
The main difference between this model and previous ones is consideration of the underlying vulnerabilities, such as behavioral inhibition, as a vulnerability for emotional disorders. Behavioral inhibition is an inherent trait characterized by relatively stable patterns and maladaptive behavioral and emotional responses to new and unfamiliar individuals, stimuli, and situations (Muris, van Brakel, & Arntz et al., 2011). This vulnerability shows a general tendency to react cautiously and with restraining to novel stimuli. This trait is relatively constant during childhood and is related to personality traits in adulthood (Kagan, Snidman, & Arcus, 1998). Some studies indicated the relationship between behavioral inhibition and SAD (Clauss & Blackford, 2012;White, Degnan, & Henderson et al., 2017). Over the time, behavioral inhibition evolves into an extreme fear of social interactional and/or performance situations, which results in maintaining SAD. Moreover, a positive association between behavioral inhibition and AS has been observed (Viana & Gratz, 2012;Viana, Kiel, & Alfano et al., 2017); thus, it is worthwhile to investigate the exact relationship between these vulnerabilities and SAD. We explore this issue more specifically in the next section.
In conclusion, all of these models illustrate some predisposing and perpetuating factors for SAD. They especially emphasize cognitive processes such as self-focused attention, perception bias, and overestimation of negative consequences which lead to social anxiety symptoms. They also demonstrated that, socially anxious individuals, when face social situations, show cognitive, physiological (interoceptive), and behavioral responses. These responses become a source of danger themselves, contributing to a vicious cycle of anxiety symptoms.

Anxiety sensitivity and social anxiety disorder
Some studies have investigated the relationship between AS and SAD and indicate that AS can predict social anxiety symptomatology (Allan et al., 2017;Allan et al., 2018;Barlow, 2013). On the other hand, theoretical models implicitly consider AS in psychopathology of SAD. These models generally emphasize on three factors in the development and maintenance of SAD: biased information processing, behavioral inhibition, and SAD symptoms. Given that there is a close conceptual relationship between AS and these factors, AS can be considered as an important vulnerability in SAD psychopathology. In this section, we are going to take a closer look over the SAD cognitive processes and symptoms in relation with AS.

AS and information processing
As indicated in the SAD cognitive models, the bias in the attention to and interpretation of the information plays a prominent role in the etiology and maintenance of SAD. Bias in information processing is characterized by a particular way individuals process information in a given cognitive domain (e.g., attention and interpretation) (Hirsch & Clark, 2004). This bias in information processing causes a vicious cycle of negative thoughts, behavioral avoidance, and increased anxiety sensations. It can be deduced from the available evidence and theories that anxiety-related attentional and interpretative biases reflect a cognitive vulnerability factor, AS, toward negative responses to anxiety symptoms. Research in the field of cognitive processing suggests an association between attentional and interpretative biases with AS and indicate that these biases are a potential explanatory mechanism for the link between AS and anxiety symptoms (Keogh, Dillon, & Georgiou et al., 2001;Viana & Gratz, 2012).
Individuals with high AS show vigilance for anxiety-related symptoms as well as catastrophic interpretations of the meaning of these symptoms (Peterson & Reiss, 1992), which leads to experience more anxiety symptoms. On the other hand, when individuals high in AS experience anxiety interpret anxiety symptom as a threat and underestimate their ability to manage possible negative consequences or social situation. These cognitive responses provide exaggerated information for individuals' information processing system and exacerbate information processing biases, through a vicious cycle. In this regard, it can be inferred that AS is related to SAD through bias in processing information. However, more studies are needed to clarify the nature of the relationship between them.

AS and behavioral inhibition
As we explained in Hofmann and Barlow's model, behavioral inhibition is an important risk factor in developing the SAD; thus clarifying the nature of the relationship between AS and behavioral inhibition is important. In some studies AS has considered as a potential cognitive factor relating to behavioral inhibition (Viana & Gratz, 2012;Viana et al., 2017). These studies suggested that behavioral inhibition is a more fundamental temperamental trait than AS (Pickett, Lodis, & Parkhill et al., 2012). Papachristou, Theodorou, Neophytou, Panayiotou also illustrated a model in which behavioral inhibition predicts SAD through AS (Papachristou, Theodorou, & Neophytou et al., 2018). They stated that behaviorally inhibited adolescents fear their anxiety sensations which in turn, could predict social anxiety levels. However, their results have indicated that AS does not completely explain the association between behavioral inhibition and SAD, and other mechanisms may account for the remaining effect.
AS is conceptually similar to behavioral inhibition (Hagopian & Ollendick, 1996). Both constructs emphasize on the increased sensitivity, either to the anxiety symptoms (AS) or to the stimuli (behavioral inhibition) and indicate that heightened sensitivity leads to the learned tendency to avoid contact with unpleasant experiences which paradoxically results in experiencing more anxiety and inhibition. Given that, the evidence for the relationship between these two constructs is inconsistent, therefore future studies should examine this connection more accurately.

AS and social anxiety symptoms
Individuals with SAD experience intense distress and anxiety when exposed to fearful situations or anticipating these situations (Association, 2013). Most of the previous studies have propounded a relationship only between AS social concern and SAD, while their relationship seems to be more complicated. Socially anxious individuals experience different kinds of physiological, cognitive, and behavioral symptoms which could be related to different aspects of AS. In this section, we are going to explore SAD symptoms from the perspective of different AS dimensions.

AS and physiological symptoms of SAD
The SAD physiological symptoms, (such as sweating and blushing that are experienced in social situations, are associated with cognitive, emotional, and behavioral responses. Socially anxious individuals worry about looking anxious and being humiliated, so they are alert towards their physiological changes (Boswell et al., 2013) and explore the external environment for cues of other's negative evaluation of their observable symptoms. They perceive their physiological arousal as an observable sign of anxiety and interpret it as a potential source of embarrassment, which is associated with the AS (Anderson & Hope, 2009). They experience high physiological arousal while attending social situations (Chalmers, Quintana, & Abbott et al., 2014) and interpret it as a sign of threat which predisposes individuals for more social anxiety symptoms (Gerlach, Wilhelm, & Roth, 2003;Noël, Lewis, & Francis et al., 2013). It seems that perceived physical sensations are associated with the extent of estimating the negative aspects of one's appearance (seemed anxious or weird), fear of negative evaluation, performance fear, and trait anxiety (Mansell & Clark, 1999;Stevens, Gerlach, & Cludius et al., 2011). Taken together, it could be said that if physiological arousal is considered as a dangerous sign (high AS somatic concerns), it can play a prominent role in the maintenance of social anxiety symptoms.

AS and cognitive symptoms of SAD
Another feature of SAD is the concern about the social consequences of anxiety. Socially anxious individuals catastrophize social consequences and become sensitive to their anxiety symptoms. In other word, they are afraid of these consequences because they worry about their observable symptoms. From this perspective, AS can contribute to anxiety symptoms through fear of negative evaluation of observable symptoms (Carpenter et al., 2017). Studies have suggested that the AS social concerns subscale can significantly predict social anxiety (Allan et al., 2018;Olthuis, Watt, & Stewart, 2014;Wheaton et al., 2012) and AS was the best predictor of anxiety related to performing in public and anxiety related to interacting with others (Norton, Cox, & Hewitt et al., 1997).
It should be noted that the relationship between AS and SAD is more complicated than it seems. Grant, Beck, and Davila demonstrated that AS social concerns cannot significantly predict the symptoms of social anxiety in a non-clinical group (Grant, Beck, & Davila, 2007). This finding indicates that, AS may be a consequence, not a predictor of the SAD. According to Reiss's theory of Expectancy (Reiss, 1991), people with high AS believe that anxiety resulted in negative consequences, so they afraid of becoming anxious. However, this characteristic feature paradoxically increases their anxiety symptoms and AS. Future studies should evaluate the current finding by examining the way in which AS is connected with the social concerns in SAD.

AS and behavioral symptoms of SAD
Social avoidance is one of the most important maintaining factors in SAD. Studies have illustrated different kind of strategies, such as behavioral avoidance, EA, and safety behaviors to minimize anxious feelings or to reduce one's involvement in a social situation such as social events, public speaking, and meetings (Mesri, Niles, & Pittig et al., 2017;Piccirillo, 2016). These strategies limit the chances of developing social skill and prevent habituation from occurring. Recent studies suggest a link between AS and these strategies (Lebowitz, Shic, & Campbell et al., 2015;Mahoney, Segal, & Coolidge, 2015). Evidence indicates a relationship between AS cognitive concerns subscale and the fear and avoidance of performance and social situations (Olthuis et al., 2014;Wilson & Hayward, 2006;Zvolensky & Forsyth, 2002). People with high cognitive concerns in AS worry about losing their control of mental abilities or behave strangely. They ultimately get embarrassed and tend to avoid situations which provoke these sensations-a behavior that intensifies the vicious cycle of anxiety and maintains SAD (Barlow, 2013). Findings also show a specific relationship between AS and EA. Panayiotou, Karekla, and Panayiotou surveyed the direct and indirect predictors of social anxiety and indicated that AS could predict social anxiety through EA (Panayiotou et al., 2014). Individuals with high anxiety sensitivity tend to avoid unpleasant experiences that may provoke their anxious symptoms, a strategy that precedes and exacerbate a cycle of avoidance and anxiety.
Taken together, it can be concluded that different dimensions of AS can also explain various aspects of SAD. From the AS conceptualization, it seems that fear of experiencing anxiety-related symptoms and their negative consequences result in maintaining SAD.

Anxiety sensitivity and treatment of social anxiety disorder
Several evidence-based cognitive behavioral therapies (CBT) have been proposed for SAD (Gould, Buckminster, & Pollack et al., 1997;Hofmann & Otto, 2017;Mayo-Wilson, Dias, & Mavranezouli et al., 2014). Although many of these treatments do not directly target AS, it has been shown that they have a positive effect on the various aspects of this construct. Aderka, McLean, Huppert, Davidson, and Foa demonstrated that CBT can significantly affect fear, avoidance, and physiological symptoms in patients with SAD (Aderka et al., 2013). CBT based on Hoffman's model, that do not explicitly target AS, could also reduce AS and social anxiety symptoms in students with social anxiety (Roushani, Nejad, & Arshadi et al., 2017).
Other treatments address AS as one component of their interventions. For example, Unified Protocol (UP) which has been proposed for the treatment of emotional disorders, directly targets AS through one of its modules, Interoceptive Exposure (Barlow, 2013;Newby, McKinnon, & Kuyken et al., 2015). Interoceptive exposure is aimed to increase the tolerance of anxiety symptoms and reduce the distress by target sensitivity to the somatic sensation as an emotional state (Keough & Schmidt, 2012;Sabourin, Watt, & Krigolson et al., 2016). These results suggest that interoceptive exposure exercises could address social anxiety symptoms such as fears of blushing, sweating, or trembling, which are consistent with the definition of AS (Collimore & Asmundson, 2014;Dixon, Kemp, & Farrell et al., 2015;Plotkin, 2002). Boswell et al. showed that interoceptive exposure as a treatment strategy in UP can reduce the severity of social anxiety symptoms through decreases in AS (Boswell et al., 2013).
It also has been found that the treatments developed to target AS directly, like Anxiety Sensitivity Amelioration Training (ASAT) or Cognitive anxiety sensitivity treatment (CAST), can reduce the severity of the symptoms of axis "I" disorders through AS (Norr, Allan, & Macatee et al., 2014;Schmidt, Capron, & Raines et al., 2014;Schmidt, Eggleston, & Woolaway-Bickel et al., 2007). These treatments use some strategies like psychoeducation, interoceptive exposure, and cognitive restructuring in order to reach treatment outcomes. They indicate AS as a critical risk factor for anxiety psychopathology and assume that changes in AS result in symptoms improvement.
It should be noted that since interoceptive exposure can elicit anxiety responses in individuals with SAD (Collimore & Asmundson, 2014), high AS may become an intrusive factor during the treatment. According to Wald's observation, the high sensitivity to the interoceptive anxiety-related symptoms is interfered with imaginal and in vivo exposure for obsessive-compulsive disorders, social anxiety, and general anxiety (Wald, 2008). AS could contribute to a reluctance to participate in exposure exercises and can prevent patients from achieving full therapeutic gains. In this regard, therapists should identify and address issues like this as a critical part of the treatment in order to facilitate the effective delivery of the treatment.
Altogether, these studies suggest different ways to target AS and indicate that reduction in AS could be considered as a potential underlying mechanism of change in treating anxiety disorders like SAD. However, there is limited evidence to specify the mediational role of AS in the connection between specific intervention and changes in social anxiety symptoms. It would be beneficial for future studies to determine which strategy or combination of strategies target which dimensions of the AS.

Conclusion
Social anxiety is a complicated and debilitating disorder, which perpetuate by some cognitive processes. In order to better conceptualize this disorder, it is vital to identify its underlying factors. According to empirical and theoretical evidence, we proposed that AS can be considered as a potential risk factor in the development and maintenance of SAD. AS can interact with some constructs like information processing or behavioral inhibition to develop and exaggerate anxiety in social situations. AS in combination with cognitive, somatic, and behavioral symptoms of SAD could also lead to the cascading cycle of SAD.
Given the relationship between AS and some of the SAD features and underlying factors outlined in the current theoretical overview, further research is warranted to integrate AS with existing cognitive behavioral SAD models. Longitudinal studies could address the moderational and mediational role of AS, where an early evaluation of temperamental characteristics and a later assessment of AS can improve our understanding of the psychopathology of SAD.
In addition, the information provided in this study can be used in clinical practices. Integrated strategies for reducing AS with SAD treatment protocols may facilitate the treatment process and help patients to achieve successful outcomes from interventions. Moreover, due to the difficulty in changing temperamental factors such as behavioral inhibition (Degnan & Fox, 2007), it is worth targeting more amenable factors like AS in interventions for SAD. A reduction in AS can provide a necessary context for increasing the therapeutic responses. Finally, clinicians should consider both predisposing and disturbing roles of AS in interviewing and evaluating SAD and develop brief and accessible AS screening tools.