Looking at the self in front of others: Neural correlates of attentional bias in social anxiety

Journal of Psychiatric Research 75 (2016) 31e40

Contents lists available at ScienceDirect

Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires

Looking at the self in front of others: Neural correlates of attentional bias in social anxiety Soo-Hee Choi a, Jung-Eun Shin b, Jeonghun Ku c, Jae-Jin Kim b, d, * a

Department of Psychiatry, Seoul National University College of Medicine and Institute of Human Behavioral Medicine, SNU-MRC, Seoul, Republic of Korea Brain Korea 21 PLUS Project for Medical Science, Yonsei University, Seoul, Republic of Korea c Department of Biomedical Engineering, Keimyung University, Daegu, Republic of Korea d Department of Psychiatry and Institute of Behavioral Science in Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 June 2015 Received in revised form 4 January 2016 Accepted 4 January 2016

In social anxiety disorder (SAD), anxiety reactions are triggered by attentional bias to social threats that automatically appear in social situations. The present study aimed to investigate the neural basis and underlying resting-state pathology of attentional bias toward internal and external social threats as a core element of SAD. Twenty-two patients with SAD and 20 control subjects scanned functional magnetic resonance imaging during resting-state and while performing the visual search task. During the task, participants were exposed to internal threat (hearing participants’ own pulse-sounds) and external threat (crowds in facial matrices). Patients showed activations in the lateral orbitofrontal cortex, rostral anterior cingulate cortex and insula in response to internal threat and activations in the posterior cingulate cortex and middle temporal gyrus in response to external threat. In patients, neural activity related to combined internal and external threats in the posterior cingulate cortex was inversely correlated with the functional connectivity strengths with the default mode network during restingstate. These findings suggest that attentional bias may stem from limbic and paralimbic pathology, and the interactive process of internally- and externally-focused attentional bias in SAD is associated with the self-referential function of resting-state. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Attentional bias Social anxiety disorder Lateral orbitofrontal cortex Posterior cingulate cortex

1. Introduction Social anxiety disorder (SAD) is characterized by fear and avoidance of social situations (Stein and Stein, 2008). The roles of emotional hyper-reactivity and emotional dysregulation in response to threat stimuli have been highlighted in SAD (Etkin and Wager, 2007; Freitas-Ferrari et al., 2010). Social anxiety-provoking situations have shown to be related to increased activity in limbic and paralimbic regions, including the amygdala, insula and anterior cingulate cortex (ACC) (Freitas-Ferrari et al., 2010), and reduced cognitive reappraisal-related neural activation in the dorsolateral prefrontal cortex (DLPFC) and dorsal ACC in SAD (Goldin et al., 2009). However, neural correlates of cognitive distortion, another key feature of SAD (Clark and McManus, 2002; Rapee and Heimberg, 1997), have not been studied.

* Corresponding author. Department of Psychiatry, Yonsei University Gangnam Severance Hospital, 211 Eonju-ro, Gangnam-gu, Seoul 135-720, Korea. E-mail address: [email protected] (J.-J. Kim). http://dx.doi.org/10.1016/j.jpsychires.2016.01.001 0022-3956/© 2016 Elsevier Ltd. All rights reserved.

Attentional bias to social threats, as an initiator of cognitive distortion, is central to social fear (Clark and Wells, 1995; Heeren et al., 2015b; Rapee and Heimberg, 1997). The role of attentional bias to threat has been implicated in the development, maintenance, and remediation of anxiety pathology (Dudeney et al., 2015). Furthermore, a growing body of research has accumulated on the clinical capability of attentional bias modification in SAD (Heeren et al., 2015c). Patients with SAD often pay attention to how they come across to audiences through self-monitoring of inner threat cues, such as interoceptive information, self-imagery and negative thoughts, and if sufficient attentional resources are allocated to this self-monitoring, social performance is impaired (Schultz and Heimberg, 2008). One of the critical attentional processes in social situations is an “internal focus” on these inner cues, and the other is an “external focus” on environmental threats, such as negative evaluations of others (Schultz and Heimberg, 2008). Contrary to the relatively consistent findings for the excessive internal focus in SAD (Hackmann et al., 2000; Hirsch et al., 2003; Ingram, 1990; Mansell et al., 2003; McEwan and Devins, 1983), mixed results including attention towards or away from external

32

S.-H. Choi et al. / Journal of Psychiatric Research 75 (2016) 31e40

threats have been reported for the external focus (Chen et al., 2002; Mansell et al., 2003; Mogg and Bradley, 2002; Mogg et al., 2004; Pineles and Mineka, 2005; Spector et al., 2003). Internal and external focuses have been explored in relative isolation, providing little opportunity for the examination of the potential interactive relationship. Besides, the corresponding neuronal mechanisms underlying these attentional biases in SAD remain unknown. Neurocognitive model of selective attention to threat suggests that the amygdala has a role in strengthening a threat-detection and the lateral prefrontal cortex and rostral ACC involve in attentional control to focus target stimuli rather than threat-related distractor (Bishop, 2008). Thus, it is expected that these regions would contribute to attentional biases in SAD. When it comes to the global brain dynamics, frontoparietal control network and dorsal attention network have shown a contrary action to the default mode network (DMN) (Hellyer et al., 2014). The DMN is a group of regions including the midline structures that are more active at rest than during attention-based tasks (Raichle et al., 2001). Since this intrinsically organized network is responsible for a processing of introspective thoughts during resting-state (Mason et al., 2007), the DMN would have a pivotal role in the first step of allocation of attentional process when facing social threats. A simultaneous task- and resting-state study showed a dynamic change in the relationship between frontoparietal control and default networks during social working memory as social load increased from resting-state (Xin and Lei, 2015). Resting-state connectivity between the DMN and task-positive network also predicted behavioral performance in working memory (Sala Llonch et al., 2012). Therefore, resting-state connectivity between the DMN and regions related to attentional bias may reflect a pathological precondition of SAD. In order to investigate the neural basis of attentional bias toward internal and external social threats in SAD, the present study was designed to combine activation and resting-state studies using functional magnetic resonance imaging (fMRI). As facial crowds of emotion are linked to fears of socially anxious individuals, a face-inthe-crowd-effect task (Gilboa-Schechtman et al., 2005; Pinkham et al., 2010) for detecting a target face among distracter faces was used as a social anxiety-provoking paradigm during fMRI. During the task, participants’ own pulse-sounds and the crowdedness of the faces were used as internal and external focuses, respectively. We hypothesized that the provocation of attentional bias would be related to activations in attention modulation-related and limbic regions, including the lateral prefrontal and cingulate cortices, amygdala, and insula. In addition, these activations would have specific associations with intrinsic DMN activity. We also predicted that neural activations during attentional bias would contribute to clinical manifestation, including cognitive/emotional distress and functional disability, as well as social anxiety symptom. 2. Materials and methods 2.1. Participants and measurements Participants (22 patients with SAD and 20 healthy controls) were recruited from the community through an advertisement on the internet message board of the local universities and the internet board for part-time job which is popular for undergraduate students. They were screened using self-report questionnaires, including the Social Interaction Anxiety Scale (SIAS) and Social Phobia Scale (SPS) (Mattick and Clarke, 1998), Brief version of the Fear of Negative Evaluation scale (B-FNE) (Leary, 1983), and Beck Depression Inventory (BDI) (Beck et al., 1961). SIAS is a 20-item questionnaire measuring the level of anxiety in interpersonal interactions with cutoffs of 34 or more indicative of SAD. SPS consists

of 20 items that measure performance anxiety with cutoffs of 24 indicative of SAD. FNE was developed to measure apprehension about negative evaluation (Watson and Friend, 1969), and we used cutoffs of 48 on a 12-item of its brief version to exclude social anxiety trait in healthy controls. Twenty-one-item BDI was employed to exclude depressive disorder in any participant, and a cutoff score of 21 was proposed for Korean population (Hahn et al., 1986). The inclusion scores were SIAS
34 and/or SPS
24 in patients; SIAS <34, SPS <24, and B-FNE <48 in controls; and BDI <21 and years of education
12 in both groups. The exclusion criteria for both groups included any past or present history of medical or neurological or psychiatric illness (other than SAD). Patients were diagnosed using DSM-IV-TR criteria (First et al., 1997), and controls were given a clinical assessment to confirm absence of psychiatric illnesses by a clinical interview with a psychiatrist (Choi SH). Psychopathological symptoms were assessed using the Liebowitz Social Anxiety Scale (LSAS; a 24-item measure on fear and avoidance of a range of social interaction and performance situations) (Liebowitz, 1987) and Hamilton Anxiety Scale (HAS, a 14-item scale for psychic and physical anxiety) (Hamilton, 1959). Functional disability and intelligence were measured using the Global Assessment of Functioning (American Psychiatric Association, 2000) and Raven's Progressive Matrices (Raven et al., 1988), respectively. Socioeconomic status of participants and their parents were collected using a 5-level scale (Class I ¼ highest level, Class V ¼ lowest level). All participants were medication-naïve and righthanded. As presented in Table 1, demographic variables including age, sex, education, ratio of undergraduate students, socioeconomic status, and intelligence level were not statistically different between the two groups. Patients showed higher scores on HAS and BDI as well as social anxiety scores than controls. This study was approved by the institutional review board of Severance Hospital, and written informed consent was obtained from all participants. 2.2. The behavioral task during fMRI Before scanning, all participants completed the Subjective Units of Disturbance Scale (Wolpe, 1969) to assess subjective distress, and their baseline pulse rates were measured. At first, resting-state fMRI was scanned for 5 min, in which participants were instructed to take a rest and allow thoughts to come and go freely with their eyes fixed on a cross-hair on the screen. Then, task fMRI was scanned during the face-in-the-crowd-effect task (Fig. 1). Participants' responses were to press a left or right button, depending on whether they detected the target or not during a 2.5-s presentation of facial crowds. A 0.5-s visual feedback followed after each trial. The target was a contemptuous face over the entire period of the task because it was previously shown to be a salient emotional cue in SAD (Goldin et al., 2009; Stein et al., 2002). Given that patients with SAD have enhanced vigilance and strong neural responses to angry faces relative to happy faces (Goldin et al., 2009; Mogg et al., 2004; Phan et al., 2006), the angry- and happy-facial crowds were used as the experimental and control conditions of distracters, respectively. The angry-versus happy-crowd contrast was used for the single-subject first-level analyses to obtain estimated parameter weights for the next step. This contrast is needed because task fMRI only provides information about changes in activation over time. The angry-versus happy-crowd contrast used to obtain linear contrasts distinguished the brain's response to aversive stimuli (angry faces) and its response to non-aversive stimuli (happy faces) while controlling for other physical features and cognitive processes (e.g., Phan et al., 2006). Although this contrast can yield neural responses that originate from the difference in attentional capture by angry crowds as well as in vigilance to angry faces (Pinkham et al., 2010), both would be suitable for the present study

S.-H. Choi et al. / Journal of Psychiatric Research 75 (2016) 31e40

33

Table 1 Demographic and descriptive characteristics of participants. Characteristics

Social anxiety disorder (N ¼ 22)

Control (N ¼ 20)

Male, n (%) Undergraduate students, n (%) Socioeconomic status - self, I/II/III/IV/V Socioeconomic status - parents, I/II/III/IV/V

13 (65.0) 19 (86.4) 0/3/6/2/10 2/4/12/3/1

12 (54.5) 12 (60) 0/2/4/3/10 1/5/6/3/5

Mean (s.d.)

Mean (s.d.)

Age, in years Educational level, in years Intelligence,a 0-60 Onset age, in years Global Assessment of Functioning, 0-100 Liebowitz Social Anxiety Scale, 0-144 Social Interaction Anxiety Scale, 0-80 Social Phobia Scale, 0-80 Brief Fear of Negative Evaluation Scale, 12-60 Hamilton Anxiety Scale, 0-56 Beck Depression Inventory, 0-63 Subjective Units of Disturbance Scale, 0-10

24.1 (2.8) 15.4 (1.2) 57.1 (2.8) 16.1 (4.2) 70.5 (11.5) 71.0 (20.0) 43.6 (10.8) 35.0 (14.6) 45.5 (7.0) 14.8 (9.3) 12.0 (7.3) 2.9 (1.2)

24.1 (1.8) 14.8 (1.6) 57.0 (2.2) e e 17.4 (10.9) 10.9 (4.4) 3.5 (3.7) 25.0 (7.0) 0.9 (1.7) 2.9 (3.2) 0.7 (0.7)

c2or t 0.004 0.081 0.702 5.027

0.117 1.390 0.238 e e 10.735* 13.066* 9.793* 9.525* 6.827* 5.345* 6.033*

*p < 0.001. a Intelligence was measured using the Raven's Progressive Matrices.

Fig. 1. An example of the face-in-the-crowd-effect task. Participants were instructed to report whether there was a target (contemptuous face) or not during the 2.5-s of presentation of 4- or 8-crowd facial matrices. During the task, participants heard their own pulse-sounds or control-sounds through the headphones in each block. Promptly after the stimuli presentation, feedback was displayed as “O” or “X” for 0.5 s.

as attentional bias to external threat includes detection of threat and difficulty ignoring threat (Schultz and Heimberg, 2008). The pictures were from the Korean Facial Expressions of Emotion (Park et al., 2011), which contains sets of standardized seven basic emotions of eight-male and seven-female actors and valence values obtained in a reliability study of 105 undergraduate students. Pictures with contempt, angry and happy emotions of four males and four females were selected to make visual stimuli with 8- or 4person crowd in order of the mean valence value. The task was

projected through a MR-compatible head-mounted display (VisuaStim XGA, Resonance Inc.). The visual angles on the desktop computer (14.3  18.2 ) were carried over to the head-mounted display. The stimuli were presented as a block design in two runs. Each run included twenty 18-s task blocks in a randomized order and ten 18-s resting blocks. Each task block consisted of two target-absent and four target-present 3-s trials. The task included 4 different conditions according to the internal (pulse- or control-sounds) and

34

S.-H. Choi et al. / Journal of Psychiatric Research 75 (2016) 31e40

external (8- or 4-person crowd) focuses. Each condition was repeated ten times. To examine the external focus, there were two levels of crowding: an 8-person crowd with two targets or 4-person crowd with one target. Considering that a greater number of facial crowds can induce more intense anxiety in socially anxious people (GilboaSchechtman et al., 2005; Santos et al., 2010), the 8-person crowd was used as the experimental condition and the 4-person crowd as the control condition. For the internal focus study, participants were forced to hear their own pulse-sounds while performing the face-in-the-crowd-effect task. They were instructed to focus on the variations of their pulse-sounds in response to whether their answers were correct or incorrect. Before scanning, participants were given a chance to become aware of the variability of their heartbeat according to their performance through practice blocks. We used this paradigm because prior experiments detected attentional bias toward internal sources of potential threats as represented by pulse-rate information in SAD (Mansell et al., 2003; Pineles and Mineka, 2005). The pulse signal was collected at 50 Hz using BIOPAC Systems during the entire session in the scanner. A controlsound was a predetermined rate with subtle variations in the beep sounds, which were distinguishable from and irrelevant to participants’ pulse-sounds. Participants also familiarized themselves with the control-sound, as well as their pulse-sounds, during the practice blocks on the desktop prior to the scanning. To boost performance anxiety, participants were initially informed that their performance and pulse rates were being observed by researchers. 2.3. fMRI acquisition and preprocessing A whole-body 1.5T MRI system (Signa Eclipse; GE Medical System) was used to obtain blood oxygen level-dependent signals using an echo-planar imaging (64  64  30 matrix, TR/TE ¼ 2000/ 22 ms, FOV ¼ 240 mm, and FA ¼ 90 ) for both resting-state and task fMRI. High-resolution T1-weighted images were obtained using a fast spoiled gradient echo sequence (256  256  116 matrix, TR/ TE ¼ 8.5/1.8 ms, FOV ¼ 240 mm, and FA ¼ 12 ). Image processing was performed using Statistical Parametric Mapping (SPM) 8 (http://www.fil.ion.ucl.ac.uk/spm/software/ spm8/) for task fMRI and Analysis of Functional NeuroImages (Cox, 1996) for resting-state fMRI. The first 30-s data were discarded to eliminate any signal decay. All remaining data were corrected with regard to slice timing and were coregistered to the first remaining time sample. Images were realigned to correct for motion artifacts using a 6-parameter rigid body affine transformation. The functional images were coregistered to the T1weighted image for each subject. The T1-weighted images were normalized to the MNI T1 template, and the resulting transformation matrices were applied to the coregistered functional images. These images were smoothed using a Gaussian kernel of 8mm full width at half-maximum. For task fMRI, high-pass filtering was performed to remove the low frequency drifts which might be made by physiological and physical noise, and autoregressive model was also used for serial correlations in first-level design matrix. Single-subject first-level analyses generated images of parameter estimates for each of internal (pulse- or control-sounds) and external (8- or 4-person crowd) conditions using contrast images between the angryversus happy-crowd. These estimated parameters were used for the subsequent second-level statistical analysis. Resting-state fMRI data were filtered for low frequency fluctuations and artificial signals (Choi et al., 2012). To obtain the averaged signal time course from the DMN, we used a seed-based approach using the posterior cingulate cortex (PCC) and medial prefrontal cortex (MPFC), which have been commonly used as a seed in defining the DMN (Tomasi

and Volkow, 2011). To avoid seed selection bias, we followed a previous work (Hahn et al., 2012) that used both regions as a seed. The DMN mask was generated by merging the two DMN components. The masks were created from a seed (3  3  3 voxels) within the PCC (centered at [0, 52, 30]) and MPFC (centered at [0, 50, 22]) (Hahn et al., 2012). Voxel-wise correlations between the time-series for each seed and the whole-brain were converted to zvalues, and group-level significance was determined using onesample t-tests with P-value < 0.05 after family-wise error correction in each group. The final DMN mask was an intersection of whole-brain maps created from the two masks. We obtained the DMN mask separately in each group because we predicted that the two groups would show significant differences in DMN activity. The DMN seed covered a volume of 4396 voxels in patients and 3267 voxels in controls. Then the representative time series for each participant was extracted. Voxel-wise correlations were conducted between the time series for DMN seed and the whole brain, yielding an individual DMN map. Functional connectivity strengths were generated by converting the correlation coefficients to z values with Fisher's r-to-z transformation.

2.4. Statistical analyses The activation maps of the task fMRI were analyzed using the flexible factorial model in each group to confirm the main and interactive effect of the external and internal focus, as SPM tutorial recommended (Glascher and Gitelman, 2008). The estimated parameter weights (b) were entered into the imaging statistics. Scores on the HAS and BDI were used as covariates to eliminate an influence of group difference in these variables. Uncorrected p < 0.001 with a 10-voxel (80 mm3) extent threshold was considered significant. We also performed a small volume correction (family-wise error corrected p < 0.05) for the lateral prefrontal cortex, rostral ACC, PCC, amygdala, and insula, given our hypothesis described above. We used 10-mm radius spheres for each small volume, and coordinates were taken from previous studies (Bishop et al., 2004; Gentili et al., 2009; Straube et al., 2004). A two-sample t-test was also performed to test the betweengroup difference in the DMN map with a significance threshold of p < 0.05 after false discovery rate correction. The resting-state functional connectivity strengths were calculated as transformed z values of correlations of the resting-state signal time course between the averaged DMN seed and the clusters identified by the flexible factorial model. A between-group comparison was performed using multivariate analysis of variance, which was followed by two-sample t-tests at a threshold of p < 0.05. Regional activities (% signal changes) for the internal focus (pulse-versus control-sounds) and external focus (8- versus 4person crowd) and the regional activity differences during the interactive process of the internal and external focuses were calculated on the clusters identified by the flexible factorial model using MarsBaR v0.43 (http://marsbar/sourceforge.net/). Partial correlations of regional activities with clinical scores and restingstate functional connectivity strengths were calculated using the HAS and BDI scores as covariates. Additionally, adjusted p-values for multiple correlations using the Benjamini-Hochberg procedure were considered to hold the false discovery rate at a q-value threshold of 0.05. T-tests, Chi-square tests and repeated measures analyses of variance (ANOVA) for analyzing clinical and behavioral data were also performed using the HAS and BDI scores as covariates at a threshold of p < 0.05.

S.-H. Choi et al. / Journal of Psychiatric Research 75 (2016) 31e40

3. Results 3.1. Behavioral findings A 2 (group)  2 (internal focus: pulse-versus controlsounds)  2 (external focus: 8- versus 4-person crowd) repeated measures ANOVA on accuracy showed a significant main effect of the external factor (F1,38 ¼ 7.39, p ¼ 0.010, hr2 ¼ 0.163) and interaction effects for internal  external (F1,38 ¼ 4.89, p ¼ 0.033, hr2 ¼ 0.114) and group  external (F1,38 ¼ 5.53, p ¼ 0.024, hr2 ¼ 0.127) factors. It was significantly higher in the 4-person (90.3 ± 8.2) than 8-person (82.4 ± 11.1) crowd (t ¼ 7.04, p < 0.001, d ¼ 1.168). The accuracy difference between the pulsesounds and control-sounds was greater in the 8-person (4.0 ± 8.6) than 4-person (0.5 ± 5.9) crowd (t ¼ 2.03, p ¼ 0.049, d ¼ 0.320). However, the group comparison of the accuracy difference between the 8-person and 4-person crowd (6.1 ± 7.9 in patients and 9.9 ± 6.1 in controls) did not reach significance (t ¼ 1.73, p ¼ 0.091, d ¼ 0.821). Response time revealed a main effect of the external factor (F1,38 ¼ 62.42, p < 0.001, hr2 ¼ 0.622), but no main effect of group and no significant interaction effect. It was faster in the 4-person (1.7 ± 0.2) than 8-person (1.8 ± 0.1) crowd (t ¼ 16.16, p < 0.001, d ¼ 2.320). As shown in Table 2, a within-group analysis on accuracy showed a significant internal  external interaction effect only in patients. Response time revealed no significant interaction effect in both groups. Significant main effects for the external factor on accuracy and response time were found only in the control group. Pulse rates were greater during the task than baseline in patients, but were comparable in controls. 3.2. Neural activities during the task and resting-state Due to technical problems, task fMRI data for 1 patient and 1 control and resting-state fMRI data of 1 patient were excluded from imaging analyses. Results for the within-group 2 (internal focus: pulse-versus control-sounds)  2 (external focus: 8- versus 4person crowd) flexible factorial model analysis is presented in Table 3. During the pulse-sounds, patients showed increased activation mainly in the limbic and paralimbic regions, including the lateral orbitofrontal cortex (OFC), rostral ACC and insula when compared to the control-sounds, whereas controls showed increased activation in the temporo-parietal junction (TPJ) and inferior frontal gyrus. A comparison of the 4-person crowd and 8person crowd conditions revealed that only patients showed increased activation in several regions, such as the PCC, middle temporal gyrus, DLPFC, and inferior occipital gyrus. Significant interaction effects were observed in the PCC and inferior occipital gyrus in patients and the TPJ and frontopolar cortex in controls. The two-sample t-test of the DMN itself revealed no significant group differences. In the group comparison of the regional restingstate functional connectivity strengths with the DMN revealed the following: reduced in patients in the orbitofrontal cortex (F ¼ 4.43, p ¼ 0.042, hr2 ¼ 0.107; t ¼ 1.79, p ¼ 0.082), posterior insula (F ¼ 4.38, p ¼ 0.043, hr2 ¼ 0.106; t ¼ 2.11, p ¼ 0.041), DLPFC (F ¼ 30.07, p ¼ 0.001, hr2 ¼ 0.46; t ¼ 3.252, p ¼ 0.002), and inferior occipital gyrus (F ¼ 4.28, p ¼ 0.046, hr2 ¼ 0.109; t ¼ 1.71, p ¼ 0.096) than controls. 3.3. Correlations between imaging and clinical data In patients, insula activation during the internal focus was positively correlated with SIAS scores (r ¼ 0.486, p ¼ 0.035, q > 0.05) (Fig. 2) and subjective distress scores (r ¼ 0.634, p ¼ 0.004, q < 0.05), and negatively correlated with functioning level

35

(r ¼ 0.559, p ¼ 0.013, q < 0.05). In controls, TPJ activation during the internal focus was positively correlated with fear of negative evaluation scores (r ¼ 0.557, p ¼ 0.020, q > 0.05) and negatively correlated with subjective distress scores (r ¼ 0.718, p ¼ 0.001, q < 0.05). During the external focus, significant correlations were observed only in patients; between PCC activation and SIAS score (r ¼ 0.606, p ¼ 0.006, q > 0.05), between PCC activation and fear of negative evaluation score (r ¼ 0.479 p ¼ 0.038, q > 0.05), between PCC activation and functioning level (r ¼ 0.461, p ¼ 0.047, q > 0.05), and between middle temporal gyrus activation and LSAS score (r ¼ 0.468, p ¼ 0.043, q > 0.05). With regard to the relationship between activations during the task and functional connectivity strengths with the DMN, significant correlations were observed in OFC activation during the internal focus only in patients (r ¼ 0.530, p ¼ 0.024, q > 0.05). The signal differences during the interactive process in the PCC showed a significant inverse correlation with functional connectivity strengths with the DMN in the corresponding regions in patients (r ¼ 0.739, p < 0.001, q < 0.05) and positive correlation in controls (r ¼ 0.539, p < 0.026, q < 0.05) (Fig. 3). Other than the correlations listed above, there were no additional associations in either group. 4. Discussion In this study, we provide evidence for limbic and paralimbic abnormalities in association with attentional bias in SAD. To investigate a neural basis of distorted attention, we generated a task with internal and external threats. Increased pulse rates during the task compared to a baseline illustrated the induction of physiological anxiety in patients, who showed better performance when paid attention to their own pulse-sounds than to the controlsounds especially in the 8-person crowd. As socially anxious individuals have better performance in search capabilities and in emotional recognition during conditions of enhanced social fear due to compensatory strategies (Amir et al., 2005; Eysenck et al., 2007; Juth et al., 2005; Richards et al., 2011; Yang et al., 2013), our results imply elevated anxiety combined with the interactive processes of internal- and external-focused attention towards social threats. Additionally, meaningful results regarding neural responses during interactions of internal monitoring of the self and monitoring of external threat cues support the hypothesis that selffocused attention is modulated by external threats (Rapee and Heimberg, 1997). Although some prior studies reported consistent attentional bias for internal stimuli regardless of external social probe (Kanai et al., 2012; Mansell et al., 2003), there were methodological differences with the present study. We recruited patients with SAD, rather than high socially anxious individuals. Also, participants were involved in active searching of a face among the facial crowd, rather than passive viewing of a face for the duration of several seconds. When patients were urged to pay attention to their own heartbeats, they exhibited notable responses in the lateral OFC. This finding is in line with those prior studies, which found that lateral OFC hyperactivity may be responsible for anxiety-laden cognitions, while medial OFC hypoactivity is associated with a failure of anxiety inhibition in phobic patients (Guyer et al., 2008; Hahn et al., 2011; Milad and Rauch, 2007). Our task with real-time feedback may lead to negative self-representations associated with heightened physiological anxiety rather than fear extinction during the processing of internally-driven threat in SAD. The lateral OFC is engaged in multimodal sensory/evaluative processing during selfreferential applications of interoceptive senses (Monk et al., 2008), which can result in affective dysregulation. Positive association of lateral OFC activation during the internal focus with

36

S.-H. Choi et al. / Journal of Psychiatric Research 75 (2016) 31e40

Table 2 Behavioral and physiological results for the face-in-the-crowd-effect task.

functional connectivity strength with the DMN in patients may imply that aberrant activation of the corresponding region when facing internal social threat would originate from the close connection with the DMN in SAD. Given that structural abnormalities of the lateral OFC in patients with SAD and adults with behaviorally inhibited temperament probably originate early in life (Schwartz et al., 2010; Talati et al., 2013), altered responsiveness in this region might begin early before repetitive exposures to social threats. Our findings of ACC and insula activations in patients during exposure to social threats are consistent with those of previous studies (Goldin et al., 2009; Quadflieg et al., 2008; Straube et al., 2005). Considering the role of the rostral ACC in the affective processing of negative information in SAD (Amir et al., 2005), this

region would be responsible for distorted responses to perceived negative self-evaluations in socially anxious situations. The aberrant rostral ACC activation may be the reason for patients having more negative interpretations of emotional information during self-focused attention than are controls (Schultz and Heimberg, 2008). The insula is also implicated in the processing of threat signals in SAD (Straube et al., 2004), especially in allocating attentional resources (Straube et al., 2006). Our findings of associations of insula activation during the internal focus with the level of social anxiety, subjective anticipatory distress, and functional disability may be related to evaluative functions of the insula monitoring the emotion-relevant physiological state of an individual (Critchley et al., 2004; Damasio et al., 2000; Paulus and Stein, 2006).

S.-H. Choi et al. / Journal of Psychiatric Research 75 (2016) 31e40

37

Table 3 Brain regions showing significant main effects of the internal and external focuses and an interaction effect of these during the face-in-the-crowd-effect task.a Brain region, Brodmann area

Social anxiety disorder (N ¼ 21) Coordinates (x, y, z)

Main effect of the internal focus (pulse-versus control-sounds) B Lateral orbitofrontal cortex, 47/11 30, 14, 22 36, 4, 20 Rostral anterior cingulate cortex, 32 0, 44, 10 L Insula, anterior, 13 34, 12, 6 L Insula, posterior, 13 36, 12, 14 R Temporo-parietal junction, 39/40 L Inferior frontal gyrus, 44/45 L Precentral gyrus, 4 Cerebellum 20, 36, 20 Main effect of the external focus (8- versus 4-person crowd) L Posterior cingulate cortex, 30/31 18, 56, 18 14, 22, 46 R Middle temporal gyrus, 22 58, 42, 4 R Dorsolateral prefrontal cortex, 9/45 20, 60, 30 54, 38, 6 L Inferior occipital gyrus, 17/18 42, 74, 18 24, 94, 2 10, 98, 2 R Lateral orbitofrontal cortex, 47/11 42, 36, 18 Dorsal anterior cingulate cortex, 32 10, 30, 32 L Cerebellum 30, 74, 12 0, 76, 26 Interaction effect of the internal £ external focus L Inferior occipital gyrus, 17/18 26, 74, 2 0, 76, 22 R Posterior cingulate cortex, 30 22, 52, 18 L Parahippocampal gyrus, 36 28, 38, 18 R Temporo-parietal junction, 39/40 R Frontopolar cortex, 10 R Postcentral gyrus, 3

Control (N ¼ 20)

Voxel

Zmax

73 67 73 74 31

3.784 3.775* 3.825* 3.629* 3.519*

19

3.406

27 50 13 15 17 97 90 22 17 71 27 27

3.526 3.797 3.433 3.490 3.319 3.919 3.769 3.510 3.683 3.758 3.485 3.349

22 29 19 12

3.531 3.638 3.324 3.337

Coordinates (x, y, z)

Voxel

Zmax

40, 46, 30 32, 34, 12 40, 12, 54 14, 80, 32 0, 62, 34

63 24 56 114 125

3.915 3.367 3.587 3.558 3.504

15 30 47

3.515 3.682 3.932

54, 62, 28 2, 62, 2 52, 14, 42

*p < 0.05 after family-wise error correction for small volumes. Abbreviations: B, Bilateral; R, Right; L, Left. a The 2  2 (internal focus  external focus) flexible factorial model analysis with scores on the Hamilton Anxiety Scale and Beck Depression Inventory as covariates in each group, with a threshold of uncorrected p < 0.001 and more than 10 voxels.

Otherwise, as attentional bias for internal vibratory stimuli was revealed in early somatosensory event-related potentials in individuals with social anxiety (Kanai et al., 2012), insular activation might reflect interoceptive perception of palpitation during the internal focus. Meanwhile, as one component of the salience network, which plays a critical role during resting-state in switching the DMN to the task-related network when it is needed (Sridharan et al., 2008), the insula has showed decreased functional connectivity during resting-state in SAD (Liao et al., 2010). In the present study, patients also revealed reduced functional connectivity strength with the DMN in this region, suggesting that declined functional connection of the insula with the DMN may precipitate overreaction to threat. Taken together, the insula may be a neural correlate of instantaneous anxiety reactions related to automatic emotional processing of internal attentional bias (Lorberbaum et al., 2004). Unlike controls, patients revealed increased activation in cortical regions when the crowd was larger during the task. Patients showed neuronal activation in the PCC and middle temporal gyrus during the processing of external threats, and we detected positive associations between these activations and the levels of social anxiety. Increased activation of the occipital and temporal cortices in response to external threats in patients is consistent with the findings of prior studies (Doehrmann et al., 2013; Kleinhans et al., 2010; Straube et al., 2004, 2005). Distinctive neuronal responses to external threats in patients can be determining bases of social anxiety, despite conflicting results for the external focus conditions in previous behavioral studies (Schultz

and Heimberg, 2008). Patients with SAD showed a distinctive pattern of significant interaction effect of the internal and external focuses. Previously, hyper-activation of the visual cortex was demonstrated in SAD patients when viewing the recorded video of self-performing cognitive task in front of others (Pujol et al., 2013). In line with this, occipital activations related to the interactive internal and external focuses in this study may also represent the neural pathophysiology of SAD patients while looking at their own self-images from the observers’ perspectives in front of audiences. Meanwhile, signal difference in the PCC during the interactive process of the internal and external focuses was inversely correlated with the functional connectivity strength with the DMN in patients. It suggests that the weaker functional connection between the PCC and the DMN was associated with larger discrepancy of neural activity in response to internally- and externally-high threats versus internally- and externally-low threats. In consistence with a previous study (Pannekoek et al., 2013), we found no significant group differences in the DMN, suggesting that the DMN function per se is not impaired in SAD. However, the reduced functional connection between the PCC and the DMN would explain the vulnerability to social threats. In the perceptual processing and attentional aspect, the PCC has been implicated as one of neurobiological correlates of SAD (Gentili et al., 2009; Liao et al., 2010; Schneier et al., 2011). As a pivotal hub of the DMN, the PCC is also involved in pathological anxiety related to self-referential functions (Hahn et al., 2011). Additionally, an impairment of the PCC was reported to involve in self-focused attention in SAD, because of its role in self-state

38

S.-H. Choi et al. / Journal of Psychiatric Research 75 (2016) 31e40

Fig. 2. Correlations between regional signal changes of the insular cortex during the internal focus condition of the face-in-the-crowd-effect task (contrast of pulse-versus control-sounds) and Social Interaction Anxiety Scale scores. Abbreviations: SAD, patients with social anxiety disorder; CON, control subjects; R: right; L, left. *p < 0.05.

perception and attribution (Gentili et al., 2009). Taken together, the attenuated resting-state connection of the PCC with the DMN may contribute to the aberrant response of this region associated with interactive process of attentional bias in SAD. Interestingly, controls exhibited regional activation related to the theory of mind and mirror neurons, such as the TPJ, inferior
zes et al., 2003; Schultefrontal gyrus and precentral gyrus (Gre Rüther et al., 2007), during the processing of internal threats. Moreover, TPJ activation during internal-focused attention showed a positive correlation with the fear of negative evaluation and an inverse correlation with subjective distress in controls. When healthy individuals view their own experiences from a third person perspective, they would also experience greater social anxiety. However, the TPJ activation provides them with an understanding of objective perspectives of others, thus preventing excessive social anxiety. In addition, the present result of TPJ activation during the interactive process implicates that the TPJ has a role in processing of internal and external social threats in healthy population. There are several limitations to our study. First, we neither directly measured the extent of internally and externally focused attention in social anxiety, nor simultaneously investigated the

balance between these attentional biases in our participants. Second, we informed participants that they would be observed during the task by researchers to boost performance anxiety. However, this impact could not be assessed since a comparison of the level of state anxiety before and after the instruction was not performed in the present study. Third, the symptom severity of the patient group should be noted. They were medication-naïve and had no comorbid mental illnesses. Less severe form of performance-only subtype was also included in this study. Thus our sample seemed to be skewed toward less severely ill cases. However, there exists a cultural issue: whereas prevalence of social anxiety disorder in the Western society is high (13.3%), it is only 0.5e0.6% in Korea and Taiwan (Sadock et al., 2014). As Rapee and Spence described, this difference may originated from the cultural/ethnic distinction on the existence, expression, or reporting of social fears (Rapee and Spence, 2004). In fact, the mean LSAS score of the patients was 71.0. This was similar to the mean LSAS score of 71.1 found in the meta-analytic study of attention bias modification for social anxiety (Heeren et al., 2015). Additionally, this was only slightly lower than those obtained in Western studies; c.f., Pujol et al. 80.7, Sladky et al., 75.6, and Talati et al., 81.4, but higher than those obtained in Eastern studies; c.f., Ding et al., 52.6 and Liu et al., 53.9. On the contrary, the mean LSAS score of the controls, 17.4, noted an opposite tendency; c.f., Pujol et al. 11.8, Sladky et al., 5.3, and Talati et al., 8.1; c.f., Ding et al., 19.2 and Liu et al., 20.0 (Ding et al., 2011; Liu et al., 2015; Pujol et al., 2013; Sladky et al., 2012; Talati et al., 2013). In addition, although patients included both generalized and performance-only subtypes, behavioral and neural differences between the subtypes could not be analyzed because of the limited sample size. Finally, relatively liberal, uncorrected significance levels were used for the analyses. In summary, patients with SAD experience elevated anxiety responses due to the interactive processes of internal- and external-focused attention towards social threats, demonstrate deviant activations in the lateral OFC, rostral ACC and insula during internal monitoring of the self and PCC activation during monitoring of external threat cues, and may have resting-state pathology of functional connection of the PCC with the DMN that is related to peculiar neural activities during the interactive process of internaland external-attentional biases. These results suggest that patients are susceptible to internal and external social threats because they recruit aberrant neural activities stemming from the hyper-reactive

Fig. 3. (A) The posterior cingulate cortex showing significant interaction effect of the internal and external focus conditions of the face-in-the-crowd-effect task. (B) The bar graphs indicate signal changes ± standard error across four conditions in the corresponding regions. (C) Correlation coefficients between signal differences during the interactive process of the internal and external focuses and functional connectivity strengths with the default mode network (DMN) in the corresponding regions. Abbreviations: SAD, patients with social anxiety disorder; CON, control subjects; R: right; L, left. *p < 0.05, **p < 0.01.

S.-H. Choi et al. / Journal of Psychiatric Research 75 (2016) 31e40

attentional system to monitor themselves in social situations. Increasing knowledge of the attentional bias could help to elucidate the psychopathology of SAD and suggest avenues for the cognitive and pharmacological treatment. Self-focused attention with vigilance to external threats should be emphasized in clinical applications for the treatment of SAD. Relatively small effect size of attentional bias modification on SAD symptoms (Heeren et al., 2015b) could be enhanced through an intervention on internal attentional bias together with external attentional bias. The present results would also complement the recent effort of applying neuromodulation tools on attentional bias modification to improve therapeutic effect. Researchers have been using the DLPFC as a target region to improve top-down attentional control (Clarke et al., 2014; Heeren et al., 2015a). However, reduction of aberrant activations of the OFC, rACC, insula, PCC, and visual cortex and/or enhancement of TPJ activation in the presence of socially threatening stimuli would serve better as targets in future work. Further, neurobiological evidence of the distorted attention can be used as a marker for the course or treatment response of this disorder. Contributors Soo-Hee Choi and Jae-Jin Kim designed research. Soo-Hee Choi and Jung-Eun Shin performed research. Soo-Hee Choi, Jung-Eun Shin and Jeonghun Ku analyzed data. Soo-Hee Choi and Jae-Jin Kim wrote the paper. Conflict of interest The authors report no conflicts of interest. Role of the funding source This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (NRF-2013R1A2A2A03068342) and Aspiring Researcher Program through Seoul National University (SNU) in 2014. Acknowledgments The authors would like to thank Dr. Kang Joon Yoon from St. Peter's Hospital, and Ph.D. candidate Yongseok Shin and M.S. candidate Ho-Jong Lee from the Department of Biomedical Engineering, Hanyang University for their technical support. References American Psychiatric Association, 2000. Diagnostic and Statistical Manual of Mental Disorders. Text Revision (DSM-IV-TR), fourth ed. American Psychiatric Publishing, Inc., Arlington, VA. Amir, N., Klumpp, H., Elias, J., Bedwell, J.S., Yanasak, N., Miller, L.S., 2005. Increased activation of the anterior cingulate cortex during processing of disgust faces in individuals with social phobia. Biol. Psychiatry 57, 975e981. Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., Erbaugh, J., 1961. An inventory for measuring depression. Arch. Gen. Psychiatry 4, 561. Bishop, S., Duncan, J., Brett, M., Lawrence, A.D., 2004. Prefrontal cortical function and anxiety: controlling attention to threat-related stimuli. Nat. Neurosci. 7, 184e188. Bishop, S.J., 2008. Neural mechanisms underlying selective attention to threat. Ann. N. Y. Acad. Sci. 1129, 141e152. Chen, Y.P., Ehlers, A., Clark, D.M., Mansell, W., 2002. Patients with generalized social phobia direct their attention away from faces. Behav. Res. Ther. 40, 677e687. Choi, S.H., Lee, H., Chung, T.S., Park, K.M., Jung, Y.C., Kim, S., et al., 2012. Neural network functional connectivity during and after an episode of delirium. Am. J. Psychiatry 169, 498e507. Clark, D.M., McManus, F., 2002. Information processing in social phobia. Biol. Psychiatry 51, 92e100. Clark, D.M., Wells, A., 1995. A cognitive model of social phobia. In: Heimberg, R.G., Liebowitz, M., Hope, D.A., Schneier, F.R. (Eds.), Social Phobia: Diagnosis, Assessment and Treatment. Guilford Press, NewYork, pp. 69e93.

39

Clarke, P.J., Browning, M., Hammond, G., Notebaert, L., MacLeod, C., 2014. The causal role of the dorsolateral prefrontal cortex in the modification of attentional bias: evidence from transcranial direct current stimulation. Biol. Psychiatry 76, 946e952. Cox, R.W., 1996. AFNI: software for analysis and visualization of functional magnetic resonance neuroimages. Comp. Biomed. Res. 29, 162e173. Critchley, H.D., Wiens, S., Rotshtein, P., Ohman, A., Dolan, R.J., 2004. Neural systems supporting interoceptive awareness. Nat. Neurosci. 7, 189e195. Damasio, A.R., Grabowski, T.J., Bechara, A., Damasio, H., Ponto, L.L., Parvizi, J., et al., 2000. Subcortical and cortical brain activity during the feeling of self-generated emotions. Nat. Neurosci. 3, 1049e1056. Ding, J., Chen, H., Qiu, C., Liao, W., Warwick, J., Duan, X., et al., 2011. Disrupted functional connectivity in social anxiety disorder: a resting-state fMRI study. Magn. Reson. Imaging 29, 701e711. Doehrmann, O., Ghosh, S., Polli, F., Reynolds, G., Horn, F., Keshavan, A., et al., 2013. Predicting treatment response in social anxiety disorder from functional magnetic resonance imaging. JAMA Psychiatry 70, 87e97. Dudeney, J., Sharpe, L., Hunt, C., 2015. Attentional bias towards threatening stimuli in children with anxiety: a meta-analysis. Clin. Psychol. Rev. 40, 66e75. Etkin, A., Wager, T.D., 2007. Functional neuroimaging of anxiety: a meta-analysis of emotional processing in PTSD, social anxiety disorder, and specific phobia. Am. J. Psychiatry 164, 1476e1488. Eysenck, M.W., Derakshan, N., Santos, R., Calvo, M.G., 2007. Anxiety and cognitive performance: attentional control theory. Emotion 7, 336e353. First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 1997. Structured Clinical Interview for DSM-iv Axis I Disorders. American Psychiatry Press, Washington, DC. Freitas-Ferrari, M.C., Hallak, J.E., Trzesniak, C., Filho, A.S., Machado-de-Sousa, J.P., Chagas, M.H., et al., 2010. Neuroimaging in social anxiety disorder: a systematic review of the literature. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 34, 565e580. Gentili, C., Ricciardi, E., Gobbini, M., Santarelli, M., Haxby, J., Pietrini, P., et al., 2009. Beyond amygdala: default mode network activity differs between patients with social phobia and healthy controls. Brain Res. Bull. 79, 409e413. Gilboa-Schechtman, E., Presburger, G., Marom, S., Hermesh, H., 2005. The effects of social anxiety and depression on the evaluation of facial crowds. Behav. Res. Ther. 43, 467e474. Glascher, J., Gitelman, D., 2008. Contrast Weights in Flexible Factorial Design with Multiple Groups of Subjects. Unpublished tutorial. http://www.jiscmail.ac.uk/ cgi-bin/webadmin?A2¼ind0803&L¼SPM&P¼R16629. Goldin, P.R., Manber, T., Hakimi, S., Canli, T., Gross, J., 2009. Neural bases of social anxiety disorder: emotional reactivity and cognitive regulation during social and physical threat. Arch. Gen. Psychiatry 66, 170e180.
zes, J., Armony, J.L., Rowe, J., Passingham, R.E., 2003. Activations related to Gre “mirror” and “canonical” neurones in the human brain: an fMRI study. Neuroimage 18, 928e937. Guyer, A.E., Lau, J.Y., McClure-Tone, E.B., Parrish, J., Shiffrin, N.D., Reynolds, R.C., et al., 2008. Amygdala and ventrolateral prefrontal cortex function during anticipated peer evaluation in pediatric social anxiety. Arch. Gen. Psychiatry 65, 1303e1312. Hackmann, A., Clark, D.M., McManus, F., 2000. Recurrent images and early memories in social phobia. Behav. Res. Ther. 38, 601e610. Hahn, A., Stein, P., Windischberger, C., Weissenbacher, A., Spindelegger, C., Moser, E., et al., 2011. Reduced resting-state functional connectivity between amygdala and orbitofrontal cortex in social anxiety disorder. Neuroimage 56, 881e889. € flich, A., Losak, J., et al., Hahn, A., Wadsak, W., Windischberger, C., Baldinger, P., Ho 2012. Differential modulation of the default mode network via serotonin-1A receptors. Proc. Natl. Acad. Sci. U. S. A. 109, 2619e2624. Hahn, H.M., Yum, T.H., Shin, Y.W., Kim, K.H., Yoon, D.J., Chung, K.J., 1986. A standardization study of Beck depression inventory in Korea. J. Korean Neuropsychiat. Assoc. 25, 487e500. Hamilton, M., 1959. The assessment of anxiety states by rating. Br. J. Med. Psychol. 32, 50e55. Heeren, A., Baeken, C., Vanderhasselt, M.A., Philippot, P., de Raedt, R., 2015a. Impact of anodal and cathodal transcranial direct current stimulation over the left dorsolateral prefrontal cortex during attention bias modification: an eyetracking study. PLoS One 10 e0124182. Heeren, A., Mogoas¸e, C., McNally, R.J., Schmitz, A., Philippot, P., 2015b. Does attention bias modification improve attentional control? A double-blind randomized experiment with individuals with social anxiety disorder. J. Anxiety Disord. 29, 35e42. Heeren, A., Mogoas¸e, C., Philippot, P., McNally, R.J., 2015c. Attention bias modification for social anxiety: a systematic review and meta-analysis. Clin. Psychol. Rev. 40, 76e90. Hellyer, P.J., Shanahan, M., Scott, G., Wise, R.J., Sharp, D.J., Leech, R., 2014. The control of global brain dynamics: opposing actions of frontoparietal control and default mode networks on attention. J. Neurosci. 34, 451e461. Hirsch, C.R., Clark, D.M., Mathews, A., Williams, R., 2003. Self-images play a causal role in social phobia. Behav. Res. Ther. 41, 909e921. Ingram, R.E., 1990. Self-focused attention in clinical disorders: review and a conceptual model. Psychol. Bull. 107, 156e176. Juth, P., Lundqvist, D., Karlsson, A., Ohman, A., 2005. Looking for foes and friends: perceptual and emotional factors when finding a face in the crowd. Emotion 5, 379e395. Kanai, Y., Nittono, H., Kubo, K., Sasaki-Aoki, S., Iwanaga, M., 2012. Early

40

S.-H. Choi et al. / Journal of Psychiatric Research 75 (2016) 31e40

somatosensory event-related potentials reveal attentional bias for internal stimuli in social anxiety. Biol. Psychol. 89, 591e597. Kleinhans, N., Richards, T., Weaver, K., Johnson, L.C., Greenson, J., Dawson, G., et al., 2010. Association between amygdala response to emotional faces and social anxiety in autism spectrum disorders. Neuropsychologia 48, 3665e3670. Leary, M.R., 1983. A brief version of the fear of negative evaluation scale. Pers. Soc. Psychol. Bull. 9, 371e375. Liao, W., Chen, H., Feng, Y., Mantini, D., Gentili, C., Pan, Z., et al., 2010. Selective aberrant functional connectivity of resting state networks in social anxiety disorder. Neuroimage 52, 1549e1558. Liebowitz, M.R., 1987. Social phobia. Mod. Probl. Pharmacopsychiat. 22, 141e173. Liu, F., Zhu, C., Wang, Y., Guo, W., Li, M., Wang, W., et al., 2015. Disrupted cortical hubs in functional brain networks in social anxiety disorder. Clin. Neurophysiol. 126, 1711e1716. Lorberbaum, J.P., Kose, S., Johnson, M.R., Arana, G.W., Sullivan, L.K., Hamner, M.B., et al., 2004. Neural correlates of speech anticipatory anxiety in generalized social phobia. Neuroreport 15, 2701e2705. Mansell, W., Clark, D.M., Ehlers, A., 2003. Internal versus external attention in social anxiety: an investigation using a novel paradigm. Behav. Res. Ther. 41, 555e572. Mason, M., Norton, M., Van Horn, J., Wegner, D., Grafton, S., Macrae, C.N., 2007. Wandering minds: the default network and stimulus-independent thought. Science 315, 393e395. Mattick, R.P., Clarke, J.C., 1998. Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behav. Res. Ther. 36, 455e470. McEwan, K.L., Devins, G.M., 1983. Is increased arousal in social anxiety noticed by others? J. Abnorm Psychol. 92, 417e421. Milad, M.R., Rauch, S.L., 2007. The role of the orbitofrontal cortex in anxiety disorders. Ann. N. Y. Acad. Sci. 1121, 546e561. Mogg, K., Bradley, B.P., 2002. Selective orienting of attention to masked threat faces in social anxiety. Behav. Res. Ther. 40, 1403e1414. Mogg, K., Philippot, P., Bradley, B.P., 2004. Selective attention to angry faces in clinical social phobia. J. Abnorm Psychol. 113, 160e165. Monk, C.S., Telzer, E.H., Mogg, K., Bradley, B.P., Mai, X., Louro, H.M., et al., 2008. Amygdala and ventrolateral prefrontal cortex activation to masked angry faces in children and adolescents with generalized anxiety disorder. Arch. Gen. Psychiatry 65, 568e576. Pannekoek, J.N., Veer, I.M., van Tol, M.J., van der Werff, S.J., Demenescu, L.R., Aleman, A., et al., 2013. Resting-state functional connectivity abnormalities in limbic and salience networks in social anxiety disorder without comorbidity. Eur. Neuropsychopharmacol. 23, 186e195. Park, J.Y., Oh, J.M., Kim, S.Y., Lee, M., Lee, C., Kim, B.R., et al., 2011. Korean Facial Expressions of Emotion (KOFEE). Section of Affect & Neuroscience, Institute of Behavioral Science in Medicine, Yonsei University College of Medicine, Seoul. Paulus, M.P., Stein, M.B., 2006. An insular view of anxiety. Biol. Psychiatry 60, 383e387. Phan, K.L., Fitzgerald, D.A., Nathan, P.J., Tancer, M.E., 2006. Association between amygdala hyperactivity to harsh faces and severity of social anxiety in generalized social phobia. Biol. Psychiatry 59, 424e429. Pineles, S.L., Mineka, S., 2005. Attentional biases to internal and external sources of potential threat in social anxiety. J. Abnorm Psychol. 114, 314e318. Pinkham, A.E., Griffin, M., Baron, R., Sasson, N.J., Gur, R.C., 2010. The face in the crowd effect: anger superiority when using real faces and multiple identities. Emotion 10, 141e146. nez, M., Ortiz, H., Soriano Mas, C., Lo  pez Sola
, M., Farre , M., et al., 2013. Pujol, J., Gime Neural response to the observable self in social anxiety disorder. Psychol. Med. 43, 721e731. Quadflieg, S., Mohr, A., Mentzel, H.J., Miltner, W.H., Straube, T., 2008. Modulation of the neural network involved in the processing of anger prosody: the role of task-relevance and social phobia. Biol. Psychol. 78, 129e137. Raichle, M.E., MacLeod, A.M., Snyder, A.Z., Powers, W.J., Gusnard, D.A., Shulman, G.L., 2001. A default mode of brain function. Proc. Natl. Acad. Sci. U. S. A. 98, 676e682. Rapee, R.M., Heimberg, R.G., 1997. A cognitive-behavioral model of anxiety in social

phobia. Behav. Res. Ther. 35, 741e756. Rapee, R.M., Spence, S.H., 2004. The etiology of social phobia: empirical evidence and an initial model. Clin. Psychol. Rev. 24, 737e767. Raven, J.C., Court, J.H., Raven, J., 1988. Manual for Raven's Progressive Matrices and Vocabulary Scales. Lewis, London. Richards, H.J., Hadwin, J.A., Benson, V., Wenger, M.J., Donnelly, N., 2011. The influence of anxiety on processing capacity for threat detection. Psychon. Bull. Rev. 18, 883e889. Sadock, B., Sadock, V., Ruiz, P., 2014. Kaplan and Sadock's Synopsis of Psychiatry: Behavioral Sciences/Clinical Psychiatry. ~ a-Go  mez, C., Arenaza-Urquijo, E., Vidal-Pin ~ eiro, D., Bargallo , N., Sala-Llonch, R., Pen , C., et al., 2012. Brain connectivity during resting state and subsequent Junque working memory task predicts behavioural performance. Cortex 48, 1187e1196. Santos, A., Silva, C., Rosset, D., Deruelle, C., 2010. Just another face in the crowd: evidence for decreased detection of angry faces in children with Williams syndrome. Neuropsychologia 48, 1071e1078. Schneier, F.R., Pomplun, M., Sy, M., Hirsch, J., 2011. Neural response to eye contact and paroxetine treatment in generalized social anxiety disorder. Psychiatry Res. 194, 271e278. Schulte-Rüther, M., Markowitsch, H.J., Fink, G.R., Piefke, M., 2007. Mirror neuron and theory of mind mechanisms involved in face-to-face interactions: a functional magnetic resonance imaging approach to empathy. J. Cogn. Neurosci. 19, 1354e1372. Schultz, L.T., Heimberg, R.G., 2008. Attentional focus in social anxiety disorder: potential for interactive processes. Clin. Psychol. Rev. 28, 1206e1221. Schwartz, C., Kunwar, P., Greve, D., Moran, L., Viner, J., Covino, J., et al., 2010. Structural differences in adult orbital and ventromedial prefrontal cortex predicted by infant temperament at 4 months of age. Arch. Gen. Psychiatry 67, 78e84. € flich, A., Atanelov, J., Kraus, C., Baldinger, P., Moser, E., et al., 2012. Sladky, R., Ho Increased neural habituation in the amygdala and orbitofrontal cortex in social anxiety disorder revealed by FMRI. PLoS One 7 e50050. Spector, I.P., Pecknold, J.C., Libman, E., 2003. Selective attentional bias related to the noticeability aspect of anxiety symptoms in generalized social phobia. J. Anxiety Disord. 17, 517e531. Sridharan, D., Levitin, D.J., Menon, V., 2008. A critical role for the right fronto-insular cortex in switching between central-executive and default-mode networks. Proc. Natl. Acad. Sci. U. S. A. 105, 12569e12574. Stein, M.B., Goldin, P.R., Sareen, J., Zorrilla, L.T., Brown, G., 2002. Increased amygdala activation to angry and contemptuous faces in generalized social phobia. Arch. Gen. Psychiatry 59, 1027e1034. Stein, M.B., Stein, D.J., 2008. Social anxiety disorder. Lancet 371, 1115e1125. Straube, T., Kolassa, I., Glauer, M., Mentzel, H., Miltner, W.H., 2004. Effect of task conditions on brain responses to threatening faces in social phobics: an eventrelated functional magnetic resonance imaging study. Biol. Psychiatry 56, 921e930. Straube, T., Mentzel, H.J., Miltner, W.H., 2006. Neural mechanisms of automatic and direct processing of phobogenic stimuli in specific phobia. Biol. Psychiatry 59, 162e170. Straube, T., Mentzel, H., Miltner, W.H., 2005. Common and distinct brain activation to threat and safety signals in social phobia. Neuropsychobiol 52, 163e168. Talati, A., Pantazatos, S., Schneier, F., Weissman, M., Hirsch, J., 2013. Gray matter abnormalities in social anxiety disorder: primary, replication, and specificity studies. Biol. Psychiatry 73, 75e84. Tomasi, D., Volkow, N.D., 2011. Functional connectivity hubs in the human brain. Neuroimage 57, 908e917. Watson, D., Friend, R., 1969. Measurement of social-evaluative anxiety. J. Consult. Clin. Psychol. 33, 448e457. Wolpe, J., 1969. The Practice of Behavior Therapy. Pergamon Press, New York, NY. Xin, F., Lei, X., 2015. Competition between frontoparietal control and default networks supports social working memory and empathy. Soc. Cogn. Affect Neurosci. 10, 1144e1152. Yang, J.W., Yoon, K.L., Chong, S.C., Oh, K.J., 2013. Accurate but pathological: social anxiety and ensemble coding of emotion. Cogn. Ther. Res. 37, 572e578.