Social anxiety disorder: Amygdala activation and connectivity

(1)

Social anxiety disorder:

Amygdala activation and connectivity

Bachelor Degree Project in Cognitive Neuroscience First Cycle 22.5 credits

Spring term 2021

Student: Amanda Fällmark Supervisor: Oskar MacGregor Examiner: Joel Gerafi

Degree p roject

(2)

2

Abstract

Social anxiety disorder (SAD) interferes with everyday life. It can, for instance, hinder careers, relationships, and leisure time. It is a common anxiety disorder that was neglected for decades. SAD individuals crave and fear social interactions simultaneously, leading to isolation in our highly social world. Therefore, research surrounding these kinds of disorders is essential. This systematic review has focused on the neural aspects and differences between SAD and healthy controls surrounding amygdala activation and connectivity. Functional magnetic resonance imaging (fMRI) studies conducted using social and emotional tasks were included. Findings include increased amygdala activation to fearful faces and words and a positive correlation between amygdala activation and symptom severity. Further, deficits in emotion regulation and a finding of gradual habituation have been found in SAD compared to healthy controls. Some limitations to this research are the small sample sizes used in the included articles and the use of both SAD and individuals with generalized SAD. The study is essential to assess future questions and directions regarding diagnosis, treatment, and understanding of SAD.

Keywords: amygdala, fMRI, neural connectivity, social anxiety disorder, SAD

(3)

3

Social Anxiety Disorder: Amygdala Activation and Connectivity

Every time I walked down the street, I believed people were watching me from cars or buildings, waiting for me to trip over and make a fool of myself. Every time I entered a room full of people, I believed they were all staring at me, judging me, thinking, "What is she doing here?" Every time I spoke with someone, I believed they could see straight through my eyes and into my soul and they knew I was a quivering mess inside. (Cleland, 2003, para. 9)

This systematic review revolves around social anxiety disorder (SAD), also called social phobia, which is a debilitating disorder that can lead to feelings of evaluation, isolation, and lack of relationships which will be further detailed and discussed. The review starts by describing SAD, what it is, diagnosis, threat perception in the minds of SAD individuals, etc.

Furthermore, a section on amygdala and functional magnetic resonance imaging (fMRI) will describe research and findings regarding amygdala activation and connectivity that have already been found in SAD. This thesis will then describe the method used during the literature research as well as the present findings and results and end with a discussion and conclusion regarding SAD and amygdala activation.

Social Anxiety Disorder

SAD is characterized by feelings of fear or anxiety during social situations, the fear of being negatively evaluated by others. Even though SAD individuals crave social company, the fear of being scrutinized may lead to avoidance behaviors such as not going to a social event or doing so but avoiding eye contact and conversations (American Psychiatric Association, 2013;

Stein & Gorman, 2001; Stein & Stein, 2008). People with SAD typically have high self-criticism and low self-esteem (Stein & Stein, 2008). They can be shy when meeting new people, quiet in groups, avoid public speaking, and withdrawn in unfamiliar settings (Stein & Gorman, 2001;

Stein & Stein, 2008). Furthermore, they can experience physiological symptoms such as blushing, heart racing, sweating, and trembling (Stein & Stein, 2008). It is a common, disabling disorder with loss of opportunities, leading to isolation and loneliness (Stein &

Gorman, 2001).

To confirm a diagnosis, high SAD symptom severity should be present, to the extent that it interferes significantly with regular routines such as occupational functioning (work or school), relationships, or other vital areas in life (American Psychiatric Association, 2013;

Morrison & Heimberg, 2013). The anxiety is persistent and usually lasts six months or longer, and the 12-month prevalence estimate of SAD in the United States is about 7 %. About 30 % of SAD individuals experience remission of symptoms within one year, approximately 50 % experience it within a few years. For nearly 60 % of SAD individuals without specific treatment, remission can take several years. Moreover, SAD is often comorbid with major depressive disorder, other anxiety disorders, bipolar disorder, body dysmorphic disorder, and substance use (American Psychiatric Association, 2013). SAD is also associated with greater rates of

(4)

4

school dropouts, as well as decreased well-being and quality of life (American Psychiatric Association, 2013; Stein & Gorman, 2001).

Despite the above, only about half of SAD individuals in Western societies seek treatment after having symptoms for 15-20 years already (American Psychiatric Association, 2013; Stein & Stein, 2008). Stein and Stein (2008) state that the underlying condition can lead to a fear of talking to authorities and embarrassment about one’s symptoms. In other words, the individual's fear of being scrutinized acts like a restraint against seeking help: worries about a care provider not taking one seriously can be sufficient to avoid asking for help. Further, the researchers claim that it is difficult to get a SAD individual to open up about their problems, making it more difficult to diagnose. To get them talking, asking if they have thought about having social anxiety, can be the first step to open a dialogue (Stein & Stein, 2008). In other words, SAD is a debilitating disorder that can lead to personal, societal, and economical issues and impacts a large segment of the general population.

Researchers only began scrutinizing SAD during the last few decades, as it did not historically hold the same interest for psychopathology researchers as other anxiety disorders (Rapee & Heimberg, 1997; Stein & Stein, 2008). Rapee and Heimberg (1997) argue that one reason for this was that the American Psychiatric Association did not characterize it as its own defined disorder until the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM) in 1980. Regardless, recent research into SAD has shown that the median age at onset in the United States is 13 years, and 75% of SAD individuals are between the ages 8-15 at onset. It can emerge from a childhood history of shyness or as a consequence following stressful events such as bullying. Further, SAD is heritable and more common in females than males. Females also report higher numbers of social fears and comorbid disorders, while males are more likely to fear dating and to use alcohol or drugs to relieve their symptoms (American Psychiatric Association, 2013).

Rapee and Heimberg (1997) developed a model for SAD where they argue that SAD individuals go through different steps when encountering a social situation. First, they form a mental representation of their appearance (looks and behavior) from others’ perspectives.

This mental representation is built on different inputs such as long-term memory of, for instance, experiences of similar situations, internal cues such as proprioception or physical symptoms, and external cues like others’ reactions. Second, they indicate that SAD individuals formulate a performance norm prediction for what the others expect in the given situation.

Third, with this prediction, they will then evaluate their appearance with a focus on and exaggeration of the negative aspects. The discrepancy of this determines the perceived likelihood of the socially anxious individual being negatively evaluated. Finally, this mental scenery leads to increased anxiety with further behavioral, cognitive, and physiological consequences, which changes the mental representation of oneself and starts the circle again.

(5)

5

Rapee and Heimberg (1997) further argue that SAD individuals will scan their environment looking for social threats (negative evaluations), finding perceived threats quite rapidly and then having a difficult time directing their attention away from the perceived danger. Additionally, SAD individuals’ avoidance behavior and deficits in social performance and communication are not signs of poor social skills, but depend on the degree of anxiety felt in the current situation (Morrison & Heimberg, 2013; Rapee & Heimberg, 1997). The perceived lack of social skills comes to be known as safety behaviors, and SAD individuals use them to avoid social threats. Safety behaviors differ between individuals and situations, such as avoiding eye contact, rehearsing sentences, speaking quickly, or looking for exits. This keeps the individual in an anxious state which may end in a self-fulfilling prophecy of negative evaluation from others and lead to other anxious situations. Safety behaviors thus help conceal the SAD individuals’ anxiety (Morrison & Heimberg, 2013; Wells et al., 1995). Wells et al.

(1995) showed that modified or decreased safety behaviors were correlated with less anxiety in SAD individuals. The researchers therefore argued that patients should be encouraged to disengage from their safety behaviors to challenge the negative beliefs of the social situation and threats to invalidate their beliefs. More research on SAD is needed to manage the disorder, help more of those affected by it, and understand its underlying causes. It is a disorder that seems to hinder people from living their lives to the fullest and making contacts in society.

Amygdala Activation in SAD

The amygdala is a brain structure located in the temporal lobe. It is involved in expressing aggression, fear, and anxiety (Davis & Shi, 2000). It is described as consisting of several distinct groups of cells (the basal, lateral, and accessory basal nuclei) called the basolateral amygdala. Additionally, the so-called “amygdaloid complex” has gradually come to be seen as a part of the amygdala too. This complex refers to structures surrounding the basolateral amygdala, such as the central, medial, and cortical nuclei (Davis & Shi, 2000). The amygdala nuclei can be divided into subnuclei which can then be further distinguished. For example, the lateral nucleus can be divided into dorsal, ventromedial, and medial subnuclei (LeDoux, 2007).

The different nuclei receive inputs from other brain regions, allowing a lot of information to be processed in the amygdala. For instance, the central nucleus is believed to be necessary to express innate emotional and associated physiological responses, while the basal nucleus is involved in connecting with the striatum and controlling actions (LeDoux, 2007). Additionally, the amygdala is involved in detecting threats and responding accordingly;

projections from the amygdala to the brainstem are involved in expressions of fear responses.

Projections from the amygdala to the cortex is linked to the experience of cognitive aspects of emotion processing (LeDoux, 2002). Given all these connections, the amygdala has been implicated in reward learning, motivation, emotional states, and the regulation of different

(6)

6

cognitive functions, for instance perception and attention. Further, it has a role in unconscious memory and implicit information processing, and has also been involved in different psychiatric conditions, such as anxiety disorders, schizophrenia, and depression (LeDoux, 2007).

To measure brain activity, functional magnetic resonance imaging (fMRI) can be used.

The technique is currently the best option for insights into brain function. It measures the hemodynamic changes in the brain caused by neural activation. This means that changes in oxygenation are measured in neuronal populations that are hypothesized to be involved in the function of interest. It is used to research neural mechanisms of different cognitive capacities, such as functional localization, physiological properties, neuropsychology etc. The advantages of using fMRI compared to other techniques are its noninvasive nature, high availability, high spatiotemporal resolution, and capacity to show the entire network of brain areas involved in particular tasks (Logothetis, 2008).

Using the fMRI technique and looking into the amygdala have shown increased amygdala activation in SAD compared to healthy controls while looking at neutral faces. The finding concerned faces that did not differ in subjective ratings between the groups. Showing that even if SAD individuals perceive the faces as equally threatening or non-threatening to healthy controls, they still have an affective, neural response to them (Birbaumer et al., 1998).

Phan et al. (2005) found that, compared to happy faces, gSAD¹individuals showed greater amygdala activation to harsh expressions (angry, fearful, disgusted) compared to healthy controls. Amygdala activation was found to be positively correlated with social anxiety symptom severity (evaluated by the Liebowitz Social Anxiety Scale; LSAS) but not with trait or state anxiety, or depression symptoms. Further, harsh compared to happy faces led to increased activation in the left parahippocampal gyrus and the right dorsal anterior cingulate cortex. Studies using fMRI have shown differences in brain activation in SAD individuals compared to healthy controls. This has been shown in different situations, such as imagining social situations (Nakao et al., 2011) and categorizing emotional (and neutral) faces vs. ovals (Cooney et al., 2006). Nakao et al. (2011) found evidence for dysfunction in a broad neuronal network, including the cerebellum, limbic system, and parieto-posterior cortex, in SAD individuals when imagining social situations. Cooney et al. (2006) found that, to the contrast of neutral faces vs. ovals, SAD individuals activated the right amygdala while controls activated the left amygdala. Further, Kreifelts et al. (2019) showed pictures of distinct objects and faces

1 Generalized social anxiety disorder is a subtype of social anxiety disorder characterized by fear and avoidance of a large variety of social situations. It is thought of as the most disabling and severe form of SAD (Stein & Stein, 2008). gSAD is not present in the current Diagnostic and Statistical Manual of Mental Disorders (DSM-5) but several studies used in the present research are old enough to have used the older version (4th edition) and some refers to gSAD as it is defined as a subtype of SAD in that edition, thus it is included in this research.

(7)

7

and various sounds (both human and nonhuman) to investigate socially anxious people’s brains. The researchers found amygdala sensitivity for threatening responses to neutral faces and voices in socially anxious individuals.

This systematic review aimed to see how the amygdala is activated and connected to other brain areas in individuals with SAD. Studies comparing individuals with SAD to healthy controls, by looking at their neural activation in fMRI during a social or emotional task, were of interest. Since previous research showed that amygdala is implicated in SAD individuals’

perception of and response to threat, the present hypothesis is that activation of the amygdala will be greater in individuals with SAD than in healthy controls when encountering more loaded emotional or social stimuli. Further, other areas that connect and work with the amygdala and areas that seem necessary or essential for any unusual amygdala activation, such as deficits or enhancements in connectivity, will also be discussed.

Methods Search Strategy

Regarding the literature search, the databases Web of Science and Scopus were used to find English sources. The following search string was used on the 23d of February 2021:

(“social anxiety disorder” OR “social phobia”) AND (fMRI OR “functional magnetic resonance imaging”) AND amygdala NOT treat*. The term “NOT treat*” was added to exclude studies concerning the treatment of the disorder. The time range was set to 2011-2021 to narrow the research to more recent advances in the field. All sources were exported to a Google Sheets spreadsheet for further processing. The total number of references found was 254, Web of Science gave 146 sources, and Scopus 108 sources. Duplicates were removed by comparing DOI numbers in the spreadsheet. Reviews and meta-analyses were saved to be used to explore for other relevant research on SAD and its neural correlates. By scanning the references used in the four reviews and meta-analyses that fit the present study, seven additional articles were found that could be of interest to this research. However, after screening the abstracts of those articles, all were excluded. After duplicates were removed and articles taken from reviews were counted, 199 articles were included.

Inclusion and Exclusion Criteria

To be included in this research, articles had to fit the PICO of studying (Population) individuals with SAD or gSAD²and (Intervention) a task involving social or emotional stimuli, such as viewing angry or fearful faces or imagining being in a social situation. The population had to be compared (Comparison) to healthy controls (HCs; individuals with no DSM-IV or DSM-V disorder) undergoing the same task. Finally, the studies had to have (Outcome) fMRI results

2 Additionally, the participants had to be primarily non-medicated, and SAD participants had to be checked for diagnosis and comorbid disorders. If comorbid disorders just involved a few participants it was not sufficient for exclusion, if the primary diagnosis was SAD.

(8)

8

related to the amygdala, and other brain areas involved in, affecting, or affected by amygdala activation.

During screening, articles were excluded for not being relevant to the study, focusing on the wrong disorder, having no healthy controls, containing mixed patient groups, including no task (only resting-state fMRI), being treatment-oriented, looking at neurotransmitters, not looking at the amygdala, etc. By reading the titles and abstracts, 168 articles could be excluded, leaving 31 for eligibility assessment. These articles were all downloaded and read in full, after which 17 more articles were excluded for not meeting the proper PICO criteria. Exclusion were for instance, for not having individuals with SAD or gSAD and healthy controls as participants, not having an emotional or social task, or not using fMRI during the tasks to see brain activation changes, particularly in the amygdala. From the 14 remaining articles, PICO criteria were extracted into the spreadsheet to give an overview of the similarities and differences of the articles. Throughout the included articles, there were different tasks, such as visual stimuli of faces, words or scenes, and auditory stimuli of words, which shows SAD individuals in different situations that may or may not give different neural correlations. This can provide an overall view of how the brain of an individual with SAD processes different social threats and how the amygdala and other brain regions react and interact.

Data Extraction

Data extraction from the accepted articles focused on the PICO defined above.

Extracted participant information included the number of participants from each group (SAD and HC), mean age and gender distribution of each group, etc. Information regarding the task included details of experimental condition and control condition, etc. For the outcome, data extraction primarily focused on results concerning the amygdala. However, it also comprised other areas related to the amygdala’s activation in SAD.

(9)

9

Full-text articles assessed for eligibility

(n = 31 )

Full-text articles excluded, with reasons

(n = 17 )

Studies included in qualitative synthesis

(n = 14 ) Figure 1

PRISMA flow diagram used to document the literature search process

Note: Standard flow diagram used to document the literature search process. From Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): e1000097.

doi:10.1371/journal.pmed1000097

Records identified through database searching

(n = 254 )

Additional records identified through other sources

(n = 7 )

Records after duplicates removed (n = 199 )

Included

Records excluded (n = 168 ) Records screened

(n = 199 )

EligibilityScreeningIdentification

(10)

10

Results

The studies included in this review have used social and emotional tasks to study the neural correlates of SAD compared to HCs. The tasks used differed between using pictures of faces or scenes as well as using words. Faces were sometimes used as distractors in a task or as the task itself, by answering what gender a displayed face had or by matching expressions on faces. Scenes were also used, by passively watching and answering if they were pleasant or unpleasant. Finally, words were occasionally used, either by audio or by reading them. All the different studies used SAD-related stimuli as well as neutral. An overview of the articles included can be found in table 1.

Faces

One way to elicit social anxiety is by presenting emotional faces to participants. This have, for instance, been done by using it as a distraction during a task, asking participants to identify the gender of a face, or matching facial expressions from different faces. The tasks will guide the presentation of the studies.

Minkova et al. (2017) investigated how simultaneous processing of cognition and face perception affected amygdala and orbitofrontal cortex (OFC) connectivity. Two displays portraying numbers and asterisks were presented separately with pictures of faces (angry, disgusted, or scrambled) in between. The participants’ task was to decide which display showed the highest quantity of numbers, the first or the second. There were three different conditions;

congruent, incongruent, and passive viewing. In the congruent condition, the quantity of numbers was the same as the value. In the incongruent condition, the quantity of numbers was not the same as the value. The participants were shown just asterisks in the passive viewing condition which required no response.

SAD, compared to HCs, showed hyperactivity in the bilateral amygdala during passive viewing, especially in the right amygdala. The congruent condition activated the left amygdala more in SAD than in HCs, but the incongruent condition showed no between-group difference.

Further, increased connectivity between OFC to left amygdala was found in SAD and increased connectivity between OFC to the right amygdala in HCs in the congruent condition. The researchers also found negative connectivity between OFC to the right amygdala in HCs during passive viewing, not found in SAD. Because of this, the researchers suggested an emotion regulation deficit in SAD individuals since their amygdala was not down-regulated during the passive viewing condition as HCs’ amygdala were (Minkova et al., 2017).

Faces and Gender

Four studies (Blair et al., 2011; Demenescu et al., 2013; Frick et al., 2013; Kraus et al., 2018) asked their participants to identify the gender of faces. Kraus et al. (2018) also used aversive and neutral, non-social scenes as contrasts. The participants in this study were asked to answer if the scenes were pleasant or unpleasant. Kraus et al. (2018) and Frick et al. (2013)

(11)

11

Table 1

Overview of Articles as Presented in Text (SAPS-M = Social Anxiety Picture Set Muenster, LSAS = Liebowitz Social Anxiety Scale, SPS = Social Phobia Scale)

Studies Patients Controls Task Outcome/Main findings

Faces Minkova et al., 2017

15 SAD 8 females Mean age: 26.6 (8.6)

15 HC (matched) 7 females

Mean age: 25.4 (3.4)

Faces (disgusted, angry, or scrambled) shown between two displays of numbers and asterisks.

Task: Had the first or second display the greater quantity of digits?

SAD showed significant bilateral amygdala hyperactivity during passive viewing, especially in the right amygdala, compared to HCs.

Increased connections from OFC to the left amygdala in SAD, OFC to the right amygdala in HC in the congruent condition.

Blair et al., 2011

39 gSAD 25 adults:

15 females, Mean age: 32.2 (9.14)

14 adolescents:

7 females, Mean age: 13.3 (3.42)

39 HC (matched) 23 adults:

10 females,

Mean age: 29.7 (8.3) 16 adolescents:

7 females,

Mean age: 14.9 (2.03)

Emotional faces (fearful, angry, and neutral) in static grayscale images.

10 actors, seven expressions: 50 % fear, 100 % fear, 150 % fear, 50 % anger, 100 % anger, 150

% anger, 25 % happy (as neutral faces).

Task: Gender of face?

Increased activity in amygdala and ACC to fearful expressions in SAD compared to HCs.

SAD showed greater activation of the rostral ACC to angry expressions than HCs.

Adults showed a positive correlation between LSAS and BOLD response in ACC to both fearful and angry faces.

Demenescu et al., 2013

16 HC (matched) 11 females Mean age: 33.56 (9.62)

Emotional faces (angry, fearful, sad, happy, and neutral).

Anxiety patients showed a positive correlation between symptom severity and left amygdala connectivity to right rostral ACC and left dorsal mPFC for fearful vs. neutral faces compared to angry or happy faces and HCs.

Frick et al., 2013

Mean age: 28.0 (8.2)

Faces, fearful and neutral.

Greater connectivity between the fusiform gyrus and the right amygdala in SAD compared to HCs for fearful faces.

Positive correlations between LSAS-score and amygdala activity to fearful over neutral faces.

(12)

12

Kraus et al., 2018

Mean age: 28.0 (8.2)

(1) Faces, neutral and fearful.

Task: Identify gender.

(2) Scenes, neutral and aversive, mainly non- social.

Task: Was the scene unpleasant?

Fearful faces vs. aversive (non-social) scenes elicited greater amygdala activation in SAD compared to HCs. This correlated positively with LSAS-score.

Fonzo et al., 2015

15 HC (not matched on education) 9 females Mean age: 30.00 (10.21)

Three emotional faces (happy, angry, or fearful) in a triangle, one reference image (top).

Task: Which matches the reference face?

Control condition: Same task with shapes.

General emotion processing (all faces vs. shapes) activated the amygdala and insula in all participants.

Positive correlation between right amygdala activation and trait anxiety for fearful vs. happy faces.

Prater et al., 2013

20 gSAD 11 females Mean age: 25.95 (5.39)

Three emotional faces (happy, angry, or fearful) in a triangle, one reference image (top).

Decreased connectivity between amygdala to rostral ACC, and amygdala to dlPFC in gSAD when viewing fearful minus happy faces, compared to HCs

Sladky et al., 2012

Mean age: 25.4 (3.4)

Three emotional faces (anger, disgust, fear, happiness, sadness, surprise, calmness) in a triangle, one reference image (top).

A linear decrease in bilateral amygdala, OFC, and pulvinar activation was shown throughout the experiment in SAD individuals compared to HCs (showing gradual habituation).

Sladky et al., 2015

Mean age: 25.4 (3.4)

Three emotional faces (anger, disgust, fear, happiness, sadness, surprise, calmness) in a triangle, one reference image (top).

For emotional faces:

From OFC to amygdala, negative connectivity was found in HC but positive connectivity in SAD individuals.

Positive connectivity from the amygdala to DLPFC was lower in SAD compared to HCs.

(13)

13

Words Simon et al., 2017

Words, 20 neutral nouns (five letters) spoken in angry or neutral prosody, presented at 2 different sound intensity levels (67dB-normal condition, 84dB-loud condition).

Task: Gender of voice?

Loud and angry prosody showed increased activity in the amygdala in SAD than HCs and loud and neutral prosody.

SAD showed greater activation in the left insula to loud stimuli and greater activation in OFC when hearing angry prosody than HC.

Laeger et al., 2014

Words, 96 German nouns (24 negative, 24 SAD-related, 24 positive, 24 neutral).

Task: Read the words attentively.

Patients showed stronger amygdala activation in response to SAD-related vs. negative words than HCs, activation of the right amygdala was associated with the patients' SPS scores.

Scenes Heitmann et al., 2016

Scenes, 50 disorder-related (SAPS-M) and 50 neutral.

Task: Image viewing with non-counted attention task.

Increased activation of right corticomedial amygdala, left insula, left thalamus, left globus pallidus, left dorsal mPFC, bilateral precuneus, and left PCC/precuneus in SAD for the contrast disorder-related vs. neutral scenes compared to HCs.

Heitmann et al., 2017

Scenes, 50 disorder-related (SAPS-M) and 50 neutral as distractors during the task.

Task: Bars comparison - same or different orientations?

Symptom severity correlated positively with amygdala activity, and amygdala-prefrontal coupling and negatively with dlPFC, dorsal ACC, posterior midcingulate cortex, and STS activity for disorder-related vs. neutral scenes in SAD.

Feldker et al., 2017

Mean age: 26.7 (5.94)

Scenes, 50 disorder-related (SAPS-M) and 50 neutral.

Task: Image viewing with non-counted attention task.

To the contrast disorder-related vs. neutral, all patient groups showed increased activity in the central amygdala and right lateral amygdala, compared to HCs.

Right lateral amygdala activation correlated positively with scene-induced anxiety ratings.

No differential effects were found between disorders (SAD, PD, DP, PTSD).

(14)

14

used only fearful and neutral facial expressions while the other two studies used more. Blair et al. (2011) used fearful, angry, and neutral expressions. They had altered the faces digitally to present seven different versions; 50 %, 100 %, and 150 % fear, 50 %, 100 %, and 150 % anger, and 25 % happy as the neutral expression. The study was conducted using 25 % happy as the neutral face to avoid the neutral face appearing slightly cold or threatening. Demenescu et al.

(2013) used angry, fearful, sad, happy, and neutral expressions. Their study additionally looked at panic disorder (PD) and had one group containing both SAD and PD individuals.

Kraus et al. (2018) concluded that increased amygdala reactivity in SAD is associated with social stimuli and not with non-social but still aversive stimuli. This is because they found that viewing fearful faces vs. aversive scenes elicited greater amygdala activation in SAD participants compared to HCs. Blair et al. (2011) found greater amygdala and anterior cingulate cortex (ACC) activation to fearful faces in SAD compared to HCs. Greater connectivity between the right amygdala and the fusiform gyrus in SAD compared to HCs was found for fearful faces (Frick et al., 2013). Additionally, positive correlations were found for symptom severity in the amygdala for fearful faces vs. aversive scenes (Kraus et al., 2018), fearful vs. neutral faces (Frick et al., 2013), and in ACC in adults for fearful and angry faces (Blair et al., 2011). Further positive correlations were found between symptom severity and the connectivity between the left amygdala and right rostral ACC and left dorsal medial prefrontal cortex (mPFC) for fearful vs.

neutral faces compared to angry or happy faces and HCs (Demenescu et al., 2013).

Face Matching

Four studies (Fonzo et al., 2015; Prater et al., 2013; Sladky et al., 2012; Sladky et al., 2015) used a face matching paradigm where participants were presented with three faces in a triangular shape. The top middle face was the reference face and the other two were the alternatives. All three faces had different identities and showed expressions. The task was to match expressions from one of the bottom faces to the reference face. The control condition was the same task but with shapes instead of faces. The studies differed in how many emotions they were using. Sladky et al. (2012, 2015) used seven different emotions; anger, disgust, fear, happiness, sadness, and calmness. The other two studies (Fonzo et al., 2015; Prater et al., 2013) used only three emotions; anger, fear, and happiness. Other differences between these studies were other patient disorder groups (panic disorder and generalized anxiety disorder; Fonzo et al., 2015), testing participants in fMRI during both task and rest (Prater et al., 2013), and different research questions. Sladky et al. (2012) researched habituation of the emotion processing circuitry in SAD. Prater et al. (2013) investigated the amygdala-frontal connectivity in gSAD. Fonzo et al. (2015) studied common and distinct abnormalities in emotional processing in the different disorders. Sladky et al. (2015) hypothesized SAD dysfunctions in amygdala-OFC connectivity.

(15)

15

Sladky et al. (2015) showed that both SAD and HCs showed significant connectivity from the amygdala to OFC for emotional faces, but it was greater in HCs. Reversed (OFC to the amygdala), the researchers found negative connectivity for HCs but positive connectivity in SAD for emotional faces. Further, lower positive connectivity between amygdala to dlPFC was found in SAD compared to HCs to emotional faces. Fonzo et al. (2015) found that general emotion processing activated the amygdala and insula in all participants and a positive correlation between trait anxiety and right amygdala activation for fearful vs. happy faces was detected. No differences between different patient groups were found.

Regarding habituation, Sladky et al. (2012) found that SAD individuals’ activation of the amygdala, OFC, and pulvinar decreased in a linear motion over the five blocks of the face- matching task. According to the researchers, this effect was not found in HCs or the control condition, indicating gradual habituation in SAD. Prater et al. (2013) found greater right amygdala activation in gSAD individuals for the fearful vs. happy contrast. The medial frontal wall in the bilateral amygdala showed less connectivity to rostral ACC during fearful minus happy faces in patients compared to HCs. The same was found in the lateral prefrontal wall to the bilateral dorsolateral PFC (dlPFC). Right amygdala connectivity to right dlPFC also correlated negatively with the LSAS fear subscale.

Words

Simon et al. (2017) argued that faces are not the only way to detect threatening social situations. They brought up that voices are very powerful in conveying others’ emotional states.

The study used angry and neutral prosodic stimuli to investigate how the sound intensity affects brain responses in SAD individuals. The participants listened to two women and two men saying neutral words in either angry or neutral prosody in two different sound intensities (normal: max 67dB; loud: max 84dB). Another study that revolved around words was conducted by Laeger et al. (2014) and had the participants read words instead of listening. This study presented a total of 96 negative, positive, neutral, and SAD-related words to the participants with the task to read attentively. All words were matched on the frequency of use and length.

Simon et al. (2017) found greater amygdala activation to angry prosody, especially to loud voices, in SAD compared to HCs and loud, neutral prosody. The researchers also found increased insula activity to loud voices and increased OFC activation to angry prosody in SAD compared to HCs. Laeger et al. (2014) also found greater amygdala activation in SAD compared to HCs but in their case to reading SAD-related vs. negative words. The study found positive correlations between Social Phobia Scale-score and right amygdala for the contrast SAD- related vs. negative words. The coupling between the right amygdala and left middle frontal gyrus was stronger in HCs than in patients. SPS-score correlated negatively with the connectivity between the right amygdala and the orbital part of the medial frontal gyrus. Social

(16)

16

Interaction Anxiety Scale (SIAS)-score correlated negatively with the coupling between the right amygdala and the middle and superior frontal gyrus. Further, Laeger et al. (2014) also found that high social anxiety correlated negatively with connectivity between left amygdala and ventromedial PFC (vmPFC) and OFC. They concluded that altered amygdala activation in SAD is associated with disorder-related but not generally negative words. Simon et al. (2017) indicated that sound intensity should be considered as a relevant feature that influences neural responses in SAD compared to HCs in threatening situations, such as being spoken to in angry prosody.

Scenes

Three studies (Heitmann et al., 2016; Heitmann et al., 2017; Feldker et al., 2017) used another type of social and emotional stimulation consists of using pictures of different scenes.

These studies presented 100 scenes, 50 disorder-related and 50 neutral, to their participants.

All three studies used the same picture set, the Social Anxiety Picture Set Muenster, for the disorder-related scenes. Two of the studies (Feldker et al., 2017; Heitmann et al., 2016) asked their participants to watch the scenes passively. These studies used five blurred pictures to keep the participants’ attention. The participants were thus asked to push a button whenever they saw a blurred picture. The third study (Heitmann et al., 2017) used a task that involved the orientation of bars. The participants were presented with the scenes as distractors and two circles containing either horizontal or vertical bars. The circles were placed one above and one below the picture of the scene. The participants evaluated if the bars had the same orientation or not.

All three studies showed increased amygdala activation in SAD individuals compared to HCs for disorder-related vs. neutral scenes though two studies specified different regions.

Heitmann et al. (2016) showed, among other brain regions, increases in the right corticomedial amygdala. Feldker et al. (2017) showed increased activity in the bilateral central amygdala and right lateral amygdala. Further, all studies showed a positive correlation between the amygdala and symptom severity. Heitmann et al. (2016) found it between LSAS-, SPS-, and SIAS-score and right corticomedial amygdala. Feldker et al. (2017) found it between scene-induced anxiety and right lateral amygdala. Lastly, Heitmann et al. (2017) found positive correlations between symptom severity and amygdala and amygdala-prefrontal connectivity for disorder-related vs.

neutral scenes.

Other findings in these studies were as follows. Negative correlations were found between symptom severity and activity in dlPFC, dorsal ACC, posterior midcingulate cortex, and STS for disorder-related vs. neutral scenes in SAD (Heitmann et al., 2017). Feldker et al.

(2017) found no significant differences between the different patient groups they studied (panic disorder, SAD, dental phobia, post-traumatic stress disorder). Heitmann et al. (2016) found that, besides the amygdala, SAD individuals showed increased activation in the left

(17)

17

insula, left thalamus, left globus pallidus, left dorsal mPFC, bilateral precuneus, and left PCC/precuneus for disorder-related scenes compared to HCs. These regions also were positively correlated with LSAS-score, and all but PCC/precuneus correlated positively with SPS- and SIAS-score. Heitmann et al. (2016) also found hyperconnectivity for PCC/precuneus with insula and precuneus and globus pallidus with mPFC and ACC in SAD individuals.

(18)

18

Discussion

This thesis aimed to determine how amygdala activation and other areas that relate to or connect to the amygdala are changed in SAD individuals. This was investigated by reviewing fMRI studies on SAD individuals and HCs during an emotional or social task. Even though the studies used different tasks, they found amygdala differences between SAD individuals and HCs, with some even specifying the amygdala subregions involved. Some showed an increased amygdala in SAD compared to HCs to for instance fearful expressions (Blair et al., 2011) and to the contrast disorder-related vs neutral scenes (Feldker et al., 2017) and to disorder-related vs negative words (Laeger et al., 2014). Others showed changed connectivity between the amygdala and other brain regions in for instance increased OFC to left amygdala connectivity in SAD to the congruent condition (Minkova et al., 2017) and decreased amygdala to ACC and dlPFC in SAD compared to HCs for fearful vs happy faces (Prater et al., 2013). Positive correlations between symptom severity and amygdala activation were also found using LSAS, SPS, and SIAS measurements (cf. Frick et al., 2013; Heitmann et al., 2017). In several studies, areas that showed differences in connection with the amygdala in SAD compared to HCs were ACC and frontal areas of the brain such as the OFC, mPFC, and dlPFC. These changes were discovered in fearful faces vs. neutral or happy faces or aversive scenes, disorder-related vs.

neutral or negative scenes, and angry prosody. This revolved around decreased or increased activation and connectivity between the amygdala and these areas and correlations with symptom severity according to different measures such as the LSAS, SPS, and SIAS (cf.

Demenescu et al., 2013; Heitmann et al., 2017). These brain areas (ACC, mPFC, OFC, dlPFC) are involved in important functions, including the social brain, decision-making, reasoning, and interaction which can be discussed as changed or inhibited in SAD individuals. This, and their connection to the amygdala, may explain the behavioral, perceptive, and cognitive differences in SAD compared to HCs.

The present research has shown that small differences can change the neural outcome.

We have, for instance, seen less coupling between the medial frontal wall in the bilateral amygdala to rostral ACC in SAD compared to HCs to fearful vs. happy faces (Prater et al., 2013).

Further, we have seen positive correlations with symptom severity and the connectivity between the left amygdala and right rostral ACC in SAD compared to HCs to fearful vs. neutral faces (Demenescu et al., 2013). Further, Kraus et al. (2018) found that fearful faces elicited increased amygdala activation compared to HCs and compared to aversive, non-social scenes.

Several studies (cf. Feldker et al., 2017; Heitmann et al., 2016) showed that disorder-related vs. neutral scenes elicited increased amygdala activation in SAD compared to HCs. Regarding words, Simon et al. (2017) showed that loud and angry prosody activated the amygdala more in SAD than in HCs and Laeger et al. (2014) showed that SAD individuals elicit greater

(19)

19

amygdala activation compared to HCs when reading disorder-related vs. negative words. This shows that use of words and prosody can influence how one perceives a situation. The results have showed that disorder-related stimuli increase amygdala activation and hinder emotion regulation in SAD individuals while neutral or non-social stimuli do not.

Limitations

Limitations in this research are the small sample sizes throughout the included research articles, the largest sample size is 39 SAD and 39 HCs (Blair et al., 2011), while the smallest is just 14 SAD and 12 HCs (Frick et al., 2013; Kraus et al., 2018). This is a poor representation of the population, and therefore the results should be carefully considered.

Further, “NOT treat*” was used in the search string to exclude articles concerning treatment, but this might also have excluded articles with the proper focus and only conclusions regarding treatment. Another limitation is the differentiation, or in today’s DSM (DSM-V), the lack of differentiation between SAD and gSAD that some researchers use, and others don’t mention.

This causes differences and limitations as to what should be counted as SAD in studies like the present. Comorbidities is another limitation to consider, as mentioned in the intro it is common to have a comorbid diagnosis such as depression, schizophrenia, other anxiety disorders, substance use etc. when having SAD. It is a limitation in part because it can affect the results but also, as you can see in three of the studies (Demenescu et al., 2013; Feldker et al., 2017; Fonzo et al., 2015), the results did not differ throughout different anxiety disorders which makes them hard to distinguish from a neurological perspective.

Society and Ethics

SAD is a common disorder, hindering several citizens from participating in society. The disorder is a challenge for the self and society in work, economy, and other resources.

Therefore, research regarding disorders like SAD is essential to help the individuals who suffer from it, integrate them into society and strengthen their bonds to the community through social relationships, education, and careers. Further, the ethical aspects of this thesis – as a systematic review - are upheld by not needing consent or confidentiality concerning any participants (since there were none) and being clear about references and sources used throughout the work. Another point is the objectivity that results from systematically reviewing several original studies, thereby bringing the field forward and providing examples of research questions for future research.

Future Research

Similar to this review, researchers should look at different situational and emotional tasks to find additional neural activations and connections that differ between SAD and HCs.

This to get a complete and diverse picture of the neural correlates and changes of the disorder.

In the present study, tasks revolved around faces, scenes, and words but not actual interaction or evaluative tasks that should be of interest in future research. Another way could be to

(20)

20

individualize the tasks, not all SAD individuals are anxious in the same situations, and that may influence the results in studies that only have one task. Further, it was difficult to draw conclusions regarding specific amygdala subregions involved in the disorder since only a few studies specify which parts of the amygdala were implicated. This is something for future researchers to think about to evolve the research surrounding the disorder and focus on how the different amygdala subregions are connected to other areas of the brain. Future studies should also expand the research on habituation in SAD to see if that is a solution to alter the brain regions affected in SAD to normalize its activity. As stated above, studies conducted using a greater representation of the population is needed since all studies involved in this systematic review had small sample sizes.

Conclusion

The present research found that the amygdala plays a role in SAD individuals' decreased emotion regulation, because it is more activated and has changed connectivity compared to HCs throughout different social and emotional situations. To specify, amygdala activation was increased for fearful faces, disorder-related scenes and words, and angry prosody. Connectivity between amygdala-rostral ACC, amygdala-dlPFC was decreased and left amygdala-OFC increased in SAD compared to HCs. Further important findings were that symptom severity correlated positively with amygdala activation and some amygdala connectivity. This research can also conclude that more research is needed in subregions of the amygdala and in more SAD-specific and individualized situations to get a better grasp of the disorder, as well as specify neurological differences between SAD and other anxiety disorders.

The findings show that SAD individuals have deficits in emotion regulation. It is important to use these studies for treatment to help SAD individuals socialize and be a part of our high contact- and interaction-focused society. Cleland (2003, cited at the beginning of this thesis) had a long 20 plus years of fighting SAD with low self-esteem, panic attacks, alcohol addiction, isolation, suicidal thoughts, and more. To end this thesis, here are Cleland’s (2003) words on her recovery:

With a strong determination to beat the fear and anxiety I had lived with for so long, and with support from my family and friends, I overcame my social anxiety [...] The main motivation behind my recovery was my determination to never again have to visit that lonely, miserable and fearful place in which I had spent so many years trapped. I wanted to live a more enjoyable life, free from fear, to develop a sense of self-worth, and to learn to take control and understand my emotions. (para. 14)

(21)

21

References

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders. 5th ed.

Birbaumer, N., Grodd, W., Diedrich, O., Klose, U., Erb, M., Lotze, M., Schneider, F., Weiss, U.,

& Flor, H. (1998). fMRI reveals amygdala activation to human faces in social phobics.

NeuroReport, 9(6), 1223-1226. Retrieved from:

https://journals.lww.com/neuroreport/Abstract/1998/04200/fMRI_reveals_amygd ala_activation_to_human_faces_in.48.aspx

Blair, K. S., Geraci, M., Korelitz, K., Otero, M., Towbin, K., Ernst, M., Leibenluft, E., Blair, R.

J. R., & Pine, D. S. (2011). The pathology of social phobia is independent of developmental changes in face processing. American Journal of Psychiatry, 168(11), 1202-1209. https://doi.org/10.1176/appi.ajp.2011.10121740

Cleland, S. (2003). Living with social anxiety disorder. ADAVIC.

https://www.adavic.org.au/PG-personal-stories-living-with-social-anxiety- disorder.aspx

Cooney, R. E., Atlas, L. Y., Joormann, J., Eugène, F., & Gotlib, I. H. (2006). Amygdala activation in the processing of neutral faces in social anxiety disorder: Is neutral really neutral? Psychiatry Research: Neuroimaging, 148(1), 55-59.

https://doi.org/10.1016/j.pscychresns.2006.05.003

Davis, M., & Shi, C. (2000). The amygdala. Current Biology, 10(4), PR131.

https://doi.org/10.1016/S0960-9822(00)00345-6

Demenescu, L. R., Kortekaas, R., Cremers, H. R., Renken, R. J., van Tol, M. J., van der Wee, N. J. A., Veltman, D. J., den Boer, J. A., Roelofs, K., & Aleman, A. (2013). Amygdala activation and its functional connectivity during perception of emotional faces in social phobia and panic disorder. Journal of psychiatric research, 47(8), 1024-1031.

https://doi.org/10.1016/j.jpsychires.2013.03.020

Feldker, K., Heitmann, C. Y., Neumeister, P., Tupak, S. V., Schrammen, E., Moeck, R., Zwitserlood, P., Bruchmann, M., & Straube, T. (2017). Transdiagnostic brain responses to disorder-related threat across four psychiatric disorders. Psychological medicine, 47(4), 730. https://doi.org/10.1017/S0033291716002634

Fonzo, G. A., Ramsawh, H. J., Flagan, T. M., Sullivan, S. G., Letamendi, A., Simmons, A. N., Paulus, M. P., & Stein, M. B. (2015). Common and disorder-specific neural responses to emotional faces in generalised anxiety, social anxiety and panic disorders. The

British Journal of Psychiatry, 206(3), 206-215.

https://doi.org/10.1192/bjp.bp.114.149880

(22)

22

Frick, A., Howner, K., Fischer, H., Kristiansson, M., & Furmark, T. (2013). Altered fusiform connectivity during processing of fearful faces in social anxiety disorder. Translational psychiatry, 3, e312. https://doi.org/10.1038/tp.2013.85

Heitmann, C. Y., Feldker, K., Neumeister, P., Brinkmann, L., Schrammen, E., Zwitserlood, P.,

& Straube, T. (2017). Brain activation to task-irrelevant disorder-related threat in social anxiety disorder: the impact of symptom severity. NeuroImage: Clinical, 14, 323-333.

https://doi.org/10.1016/j.nicl.2017.01.020

Heitmann, C. Y., Feldker, K., Neumeister, P., Zepp, B. M., Peterburs, J., Zwitserlood, P., &

Straube, T. (2016). Abnormal brain activation and connectivity to standardized disorder‐related visual scenes in social anxiety disorder. Human Brain Mapping, 37(4), 1559-1572. https://doi.org/10.1002/hbm.23120

Kraus, J., Frick, A., Fischer, H., Howner, K., Fredrikson, M., & Furmark, T. (2018). Amygdala reactivity and connectivity during social and non-social aversive stimulation in social anxiety disorder. Psychiatry Research: Neuroimaging, 280, 56–61.

Kreifelts, B., Eckstein, K. N., Ethofer, T., Wiegand, A., Wächter, S., Brück, C., Erb, M., Lotze, M., & Wildgruber, D. (2019). Tuned to voices and faces: Cerebral responses linked to

social anxiety. Neuroimage, 197, 450-456.

https://doi.org/10.1016/j.neuroimage.2019.05.018

Laeger, I., Dobel, C., Radenz, B., Kugel, H., Keuper, K., Eden, A., Arolt, V., Zwitserlood, P., Dannlowski, U., & Zwanzger, P. (2014). Of ‘disgrace’ and ‘pain’–corticolimbic interaction patterns for disorder-relevant and emotional words in social phobia. PLoS One, 9(11). https://doi.org/10.1371/journal.pone.0109949

LeDoux, J. (2002). The emotional brain, fear, and the amygdala. Cellular and Molecular Neurobiology, 23, 727-738. https://doi.org/10.1023/A:1025048802629

LeDoux, J. (2007). The amygdala. Current Biology, 17(20), R868. Retrieved from:

https://www.cell.com/current-biology/pdf/S0960-9822(07)01779-4.pdf

Logothetis, N. K. (2008). What we can do and what we cannot do with fMRI. Nature, 453, 869- 878. https://doi.org/10.1038/nature06976

Minkova, L., Sladky, R., Kranz, G. S., Woletz, M., Geissberger, N., Kraus, C., Lanzenberger, R.,

& Windischberger, C. (2017). Task-dependent modulation of amygdala connectivity in social anxiety disorder. Psychiatry Research Neuroimaging, 26, 39-46.

Moher, D., Liberati, A., Tetzlaff, J., & Altman, DG., The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement.

PLoS Med, 6(7): e1000097. https://doi.org/10.1371/journal.pmed1000097

(23)

23

Morrison, A. S., & Heimberg, R. G. (2013). Social anxiety and social anxiety disorder. Annual Review of Clinical Psychology, 9, 249-274. https://doi.org/10.1146/annurev-clinpsy- 050212-185631

Nakao, T., Sanematsu, H., Yoshiura, T., Togao, O., Murayama, K., Tomita, M., Masuda, Y., &

Kanba, S. (2011). fMRI of patients with social anxiety disorder during a social situation

task. Neuroscience Research, 69(1), 67-72.

https://doi.org/10.1016/j.neures.2010.09.008

Phan, K. L., Fitzgerald, D. A., Nathan, P. J., & Tancer, M. E. (2005). Association between amygdala hyperactivity to harsh faces and severity of social anxiety in generalized social

phobia. Biological Psychiatry, 59(5), 424-429.

https://doi.org/10.1016/j.biopsych.2005.08.012

Prater, K. E., Hosanagar, A., Klumpp, H., Angstadt, M., & Luan Phan, K. (2013). Aberrant amygdala–frontal cortex connectivity during perception of fearful faces and at rest in generalized social anxiety disorder. Depression and anxiety, 30(3), 234-241.

https://doi.org/10.1002/da.22014

Rapee, R. M., & Heimberg, R. G. (1997). A cognitive-behavioral model of anxiety in social phobia. Behaviour Research and Therapy, 35(8), 741-756.

https://doi.org/10.1016/S0005-7967(97)00022-3

Simon, D., Becker, M., Mothes-Lasch, M., Miltner, W. H., & Straube, T. (2017). Loud and angry: sound intensity modulates amygdala activation to angry voices in social anxiety disorder. Social cognitive and affective neuroscience, 12(3), 409-416.

https://doi.org/10.1093/scan/nsw131

Sladky, R., Höflich, A., Atanelov, J., Kraus, C., Baldinger, P., Moser, E., Lanzenberger, R., &

Windischberger, C. (2012). Increased neural habituation in the amygdala and orbitofrontal cortex in social anxiety disorder revealed by FMRI. PloS one, 7(11).

https://doi.org/10.1371/journal.pone.0050050

Sladky, R., Höflich, A., Küblböck, M., Kraus, C., Baldinger, P., Moser, E., Lanzenberger, R., &

Windischberger, C. (2015). Disrupted effective connectivity between the amygdala and orbitofrontal cortex in social anxiety disorder during emotion discrimination revealed by dynamic causal modeling for fMRI. Cerebral Cortex, 25(4), 895-903.

https://doi.org/10.1093/cercor/bht279

Stein, M. B., & Gorman, J. M. (2001). Unmasking social anxiety disorder. Journal of Psychiatry & Neuroscience, 26(3), 185-189. Retrieved from:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1408304/pdf/jpn00091-0019.pdf Stein, M. B., & Stein, D. J. (2008). Social anxiety disorder. The Lancet, 371(9618), 1115-1125.

https://doi.org/10.1016/S0140-6736(08)60488-2

(24)

24

Wells, A., Clark, D. M., Salkovski, P., Ludgate, J., Hackmann, A., & Gelder, M. (1995). Social phobia: The role of in-situation safety behaviors in maintaining anxiety and negative beliefs. Behavior Therapy, 26(1), 153-161. https://doi.org/10.1016/S0005-

7894(05)80088-7