Facial affect recognition in body dysmorphic disorder versus obsessive-compulsive disorder: An eye-tracking study

Journal of Anxiety Disorders 35 (2015) 49–59

Contents lists available at ScienceDirect

Journal of Anxiety Disorders

Facial affect recognition in body dysmorphic disorder versus obsessive-compulsive disorder: An eye-tracking study Wei Lin Toh a,b,c,∗ , David J. Castle a,d , Susan L. Rossell a,b,c,d a

Departments of Psychological Sciences and Psychiatry, University of Melbourne, Parkville, VIC 3010, Australia Cognitive Neuropsychiatry, Monash Alfred Psychiatry Research Centre, Level 4, 607 St. Kilda Road, Melbourne, VIC 3004, Australia Brain and Psychological Sciences Research Centre, Swinburne University, PO Box 218, Hawthorn, VIC 3122, Australia d Department of Psychiatry, St. Vincent’s Mental Health, PO Box 2900, Fitzroy, VIC 3065, Australia b c

a r t i c l e

i n f o

Article history: Received 26 November 2014 Received in revised form 12 August 2015 Accepted 18 August 2015 Available online 24 August 2015 Keywords: Body dysmorphic disorder Obsessive-compulsive disorder Eye-tracking Visual scanpaths Facial affect recognition

a b s t r a c t Background: Body dysmorphic disorder (BDD) is characterised by repetitive behaviours and/or mental acts occurring in response to preoccupations with perceived defects or flaws in physical appearance (American Psychiatric Association, 2013). This study aimed to investigate facial affect recognition in BDD using an integrated eye-tracking paradigm. Method: Participants were 21 BDD patients, 19 obsessive-compulsive disorder (OCD) patients and 21 healthy controls (HC), who were age-, sex-, and IQ-matched. Stimuli were from the Pictures of Facial Affect (Ekman & Friesen, 1975), and outcome measures were affect recognition accuracy as well as spatial and temporal scanpath parameters. Results: Relative to OCD and HC groups, BDD patients demonstrated significantly poorer facial affect perception and an angry recognition bias. An atypical scanning strategy encompassing significantly more blinks, fewer fixations of extended mean durations, higher mean saccade amplitudes, and less visual attention devoted to salient facial features was found. Conclusions: Patients with BDD were substantially impaired in the scanning of faces, and unable to extract affect-related information, likely indicating deficits in basic perceptual operations. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction The human face is integral as a primary means of conveying social information (Zebrowitz, 1997). In body dysmorphic disorder (BDD), not only do key elements of social perception underpin significant concerns involving physical appearance, the content of patients’ preoccupations are also often centred on the face (Buhlmann, Etcoff, & Wilhelm, 2008; Phillips, Menard, Fay, & Weisberg, 2005). A study of how people with BDD process faces is therefore especially informative. 1.1. Face processing in BDD Accordingly, BDD is characterised by repetitive behaviours and/or mental acts occurring in response to preoccupations with perceived defects or flaws in physical appearance (American

∗ Corresponding author at: Monash Alfred Psychiatry Research Centre, Level 4, 607 St. Kilda Road, Melbourne, VIC 3004, Australia. Fax: +61 3 9076 6588. E-mail address: [email protected] (W.L. Toh). http://dx.doi.org/10.1016/j.janxdis.2015.08.003 0887-6185/© 2015 Elsevier Ltd. All rights reserved.

Psychiatric Association, 2013). Face processing research in BDD has garnered interest because of its direct significance to clinical features of the disorder. Yet there have been limited studies along three dominant themes: (i) aesthetic sensitivity, (ii) affect recognition, and (iii) selective attention. In an early study, Yaryura-Tobias et al. (2002) reported that when presented with a choice to undertake digital modification based on whether each image was perceived to be distorted (no images were distorted), significantly more BDD (50%) and obsessive-compulsive disorder (OCD; 40%) patients digitally modified their own facial photographs relative to healthy controls (HC; 0%). When asked to judge the attractiveness of their own and others’ faces, persons with BDD, but not OCD or HC, overestimated the good looks of others, and underrated their own physical attractiveness (Buhlmann et al., 2008). Likewise, Reese, McNally, and Wilhelm (2010) found BDD patients were not better at detecting symmetry differences in dot arrays and faces of unfamiliar others relative to OCD and HC groups. Interestingly, BDD participants rated identical facial (but not object) images as significantly more often altered relative to individuals without BDD (Buhlmann, Rupf, Gleiss, Zschenderlein, & Kathmann, 2014). Several studies have

50

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59

examined other aspects of aesthetic perception in BDD (Lambrou, Veale, & Wilson, 2011; Stangier, Adam-Schwebe, Muller, & Wolter, 2008). Buhlmann, McNally, Etcoff, Tuschen-Caffier, and Wilhelm (2004) examined the ability of BDD patients to (i) discriminate facial features, and (ii) identify facial affect. Despite an absence of feature discrimination deficits, BDD patients demonstrated significantly impaired affect recognition, especially misidentifying emotional expressions as angry. Follow-up research uncovered an added contemptuous recognition bias in BDD, which were only significant for self-referent, as opposed to other-referent, scenarios (Buhlmann, Etcoff, & Wilhelm, 2006). It was hypothesised poor insight and ideas of reference underpinned these difficulties; perceptions of anger and rejection bolstered existing beliefs of personal ‘ugliness’ and social undesirability in this disorder. A small series of case studies in BDD also showed evidence of a similar angry recognition bias (Labuschagne, Castle, & Rossell, 2011). Only two known studies to date have attempted to explore attentional biases in BDD with the aid of eye-tracking. Grocholewski, Kliem, and Heinrichs (2012) asked BDD participants to gaze at photographs of themselves as well as unfamiliar others, and found that they, but not participants with social anxiety disorder (SAD) or HC, focused disproportionate visual attention on their perceived facial defect and to the corresponding area of others’ faces. When participants with BDD were asked to view images of themselves and a neutral control face, they predictably had a negative mean bias relative to HC; persons with BDD displayed heightened selective visual attention toward unattractive features of their own face as well as attractive features of another’s face (Greenberg, Reuman, Hartmann, Kasarskis, & Wilhelm, 2014). Collectively, these studies implicate the role of specific attentional biases in precipitating and maintaining symptoms of the disorder. 1.2. Face processing in OCD The majority of face processing research in OCD has centred on affect recognition, specifically implicating the emotion of disgust, but with mixed findings. A number of studies have reported impaired disgust recognition (Corcoran, Woody, & Tolin, 2008; Rector, Daros, Bradbury, & Richter, 2012; Sprengelmeyer et al., 1997), whereas other work has failed to establish a clear disgust deficit in OCD (Buhlmann et al., 2004; Grisham, Henry, Williams, & Bailey, 2010; Jhung et al., 2010; Parker, McNally, Nakayama, & Wilhelm, 2004; Rozin, Taylor, Ross, Bennett, & Hejmadi, 2005). Nevertheless, BDD remains the focus of the current study, with OCD acting as an appropriate clinical control group due to considerable overlaps in terms of clinical features, familial loading, symptomatology, and psychiatric comorbidity between these disorders (Simeon, Hollander, Stein, Cohen, & Aronowitz, 1995; Wilhelm, Otto, Zucker, & Pollack, 1997). This is reflected by the reclassification of BDD under the umbrella of obsessive-compulsive and related disorders (American Psychiatric Association, 2013). For a meta-analytic review of facial emotion recognition in OCD, see Daros, Zakzanis, and Rector (2014).

patients displaying significantly fewer fixations of extended durations, shorter scanpaths and a marked avoidance of salient facial features (Bestelmeyer et al., 2006; Gordon et al., 1992; Loughland, Williams, & Gordon, 2002a, 2002b; Manor et al., 1999; Williams, Loughland, Gordon, & Davidson, 1999). Preliminary investigations into SAD conversely uncovered a hyperscanning strategy, comprising fewer fixations of shorter durations, longer scanpaths and an avoidance of salient features offset by extensive scanning of non-salient features (Horley, Williams, Gonsalvez, & Gordon, 2003, 2004). Typical scanpath variables examined were number of fixations (i.e. frequency of stationary gaze points acquired during scanning), mean fixation durations (i.e. average time length per fixation, usually denoted in ms) and mean saccade amplitudes (i.e. average summed distance travelled by the eye during scanning, typically measured in degrees of visual angle). 1.4. Aims and hypotheses Preliminary studies have indicated significant deficits in facial affect recognition in BDD, especially implicating an angry recognition bias, which may be more pronounced during the scrutiny of one’s own facial image or in self-referent scenarios. Visual scanpath research in BDD is still in its early stages. A combination of these two lines of research would therefore not only enable verification of affect recognition anomalies in BDD, but also aid in possibly identifying underlying mechanisms. Several pertinent research questions exist: (i) Can tentative emotion recognition biases detected in BDD be corroborated? (ii) If so, do these exist alongside aberrant eyetracking strategies? (iii) What is the nature of such eye movement dysfunction? This study endeavoured to answer these questions by examining eye-tracking during a facial affect recognition task. Three hypotheses were postulated: (i) Relative to HC, BDD participants would exhibit poorer affect recognition and a significant angry recognition bias. (ii) Relative to HC, BDD participants would utilise atypical visual scanning strategies, especially in response to negative facial affect (i.e. anger, disgust, fear and sadness). This means their scanpath variables were expected to be significantly different from those found in HC. Basing predictions on phenomenological overlaps between BDD, schizophrenia and SAD however, suggested restricted or extensive scanning was possible. (iii) In the presence of atypical scanning strategies, added scanpath deficits relating to salient facial features (i.e. eyes, nose, mouth) would be expected, likely involving decreased visual attention to one or more of these facial regions. No a priori hypotheses with respect to BDD versus OCD group contrasts were offered for two reasons: (i) the pre-existing literature surrounding OCD is inconsistent at best, and (ii) the OCD group forms the clinical control group, so we would plausibly expect their performance to fall between that of the BDD and HC groups (though we were unable to postulate specific outcomes for given variables). 2. Method 2.1. Participants

1.3. What is eye-tracking? Eye-tracking refers to the monitoring of an individual’s eye movements with the use of specialised equipment during scanning of visual stimuli. The theoretical framework underlying eye-tracking is beyond the scope of the current paper (see Toh, Rossell, & Castle, 2011). Instead, our purpose is to provide a brief overview of findings within comparable disorders to facilitate later discussion as well as define key eye-tracking parameters. A review of existing scanpath literature converged on generalised scanning deficits of a restricted nature in schizophrenia, with affected

Twenty-one BDD patients and 19 OCD patients were recruited from a specialised outpatient psychiatric service and community sources. Twenty-one HC participants were recruited via a voluntary healthy participant database, based on a null personal and immediate family history of diagnosed psychiatric disorders. Axis I diagnoses were verified with the Body Dysmorphic Disorder-Diagnostic Module (BDD-DM) for BDD (Phillips, 2005) and Mini International Neuropsychiatric Interview-English Version 5.0.0 (MINI500) for OCD and other major Axis I disorders (Sheehan et al., 1998). Based on symptom severity, all BDD and OCD patients

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59

received primary clinical diagnosis of their respective disorders. Comorbid Axis I conditions in these groups, with the exception of the psychotic disorders, were permitted given their high frequency. Five BDD patients were diagnosed with OCD, but no OCD patients had current BDD. Notably, 33.5% and 10.5% of BDD and OCD participants respectively had comorbid SAD, whereas 42.9% and 31.6% of BDD and OCD participants respectively had recurrent MDD over their lifetime. Primary disorder severity was assessed using the Yale-Brown Obsessive-Compulsive Scale Modified for Body Dysmorphic Disorder (BDD-YBOCS; Phillips, Hollander, Rasmussen, Aronowitz, DeCaria, & Goodman, 1997) and Yale-Brown Obsessive-Compulsive Scale (YBOCS; Goodman et al., 1989). Anxiety and mood ratings were collected with the Social Interaction Anxiety Scale (SIAS; Mattick & Clarke, 1998), Brief Fear of Negative EvaluationRevised (BFNE-II; Carleton, McCreary, Norton, & Asmundson, 2006), and Self-rating Depression Scale (SDS; Zung, 1965). Most clinical participants were undergoing psychiatric treatment. Prescribed medications included selective serotonin reuptake inhibitors (SSRIs; BDD: n = 11; OCD: n = 10), serotonin-norepinephrine reuptake inhibitors (SNRIs; BDD: n = 6; OCD: n = 3), and tricyclic antidepressants (TCAs; BDD: n = 1; OCD: n = 1), with some receiving atypical antipsychotic augmentation (BDD: n = 6; OCD: n = 4). All participants met the following inclusion criteria: (i) aged 18–65 years, (ii) spoke English as preferred language, (iii) no known history of neurological disorders or serious ocular conditions, (iv) normal visual acuity (at least 20/100) and colour vision, and (v) estimated IQ above 70 based on the Wechsler Test of Adult Reading (WTAR; Wechsler, 2001). The study received ethics approval from the Human Research Ethics Committee at the Alfred Hospital, Melbourne, Australia, and conformed to the Declaration of Helsinki (World Medical Association, 1995). Participants also provided written, informed consent. 2.2. Materials Visual scanpaths were recorded using the SR Research Eyelink II system, comprising an eye-tracking apparatus linked to interconnected host and display computers (2004). Stimuli were presented on a 17 CRT monitor. Recordings were performed in pupil-only mode to ensure optimal sampling rate of 500 Hz and ultra-low noise <0.01◦ RMS, with monocular monitoring of the dominant eye. Stimuli were from the Pictures of Facial Affect by Ekman and Friesen (1975). Forty-two black-and-white photographs of six models (three males, three females) were selected based on the highest interrater agreement (76–100%) for expression valence in the norms. Every model displayed six universal emotions (i.e. anger, disgust, fear, happiness, sadness, surprise), with a neutral expression serving as the control condition. Each photograph was 512 × 768 pixels, and presented for 8 s under free viewing conditions, followed by a 2 s interval for participant response. Each photograph was displayed twice with the order of presentation pseudo-randomised, yielding four blocks of 21 trials. 2.3. Procedure Acceptable visual acuity and colour vision were established using the Snellen chart and Ishihara colour test. Participants were seated in line with the screen at a viewing distance of 30 cm. A calibration/validation process was performed by visual tracking of a 3 × 3 target display. For each trial, stimulus presentation was initiated only when gaze fixation converged on a central dot area (±.75◦ ) for at least 1 s. Task instructions were to look at the fixation target until the stimulus appeared, whereupon participants were asked to examine the photograph, bearing in mind the task objective was to decide what facial expression was being displayed.

51

After the photograph had been removed, they were asked to identify the exhibited emotion by entering an appropriate response on the keyboard. Based on a forced-choice paradigm, seven keys had been labelled ‘angry’, ‘disgusted’, ‘happy’, ‘neutral’, ‘sad’, ‘scared’, and ‘surprised’. A series of practice trials was administered.

2.4. Statistical analysis Scanpath data was managed using the SR Research Data Viewer software (2004). Primary performance measures were face processing (i.e. recognition accuracy) and eye-tracking (i.e. number of blinks, number of fixations, mean fixation durations, mean saccade amplitudes) characteristics. Based on the creation of three rectangular interest areas encompassing the salient facial features (i.e. eyes, nose, mouth), two secondary spatial-temporal parameters namely, feature fixation index (FFI)1 and feature duration index (FDI)2 , were derived (range −1.00 to +1.00; negative values indicating proportionately more visual attention to non-salient feature areas, zero indicating an even split between salient and non-salient feature areas, and positive values indicating proportionately more visual attention to salient feature areas). Statistical analyses were performed using the software package PASW® , v.18. Group-wise comparisons of face processing and eye-tracking characteristics were performed using a series of mixed between-within subjects ANOVAs, with primary performance measures as the dependent variable, group as the between-subjects factor, and facial affect as the within-subjects factor. Interest area analysis was conducted with spatial-temporal parameters as the dependent measures, group as the between-subjects factor, interest area as the first within-subjects factor, and facial affect as the second-within subjects factor. Significant main effects were managed with Tukey’s HSD test and/or within-subjects contrasts. Follow-up simple effects analysis using one-way ANOVAs was conducted for significant interactions. To limit Type I error, Bonferroni adjusted values were applied (p = .025 for recognition biases; p = .007 for scanpath characteristics; p = .017 for interest areas). In line with Cohen (1988), effect sizes were based on partial eta-squared. Pearson’s correlation analysis was used to explore relationships between face processing and eye-tracking parameters as well as participant clinical information (see Table 4). To limit Type I error here, only correlations significant at the .01 level were considered. The use of ANCOVAs to statistically control for extraneous variables, such as anxiety and depression, was deemed inappropriate (see Miller & Chapman, 2001 for statistical justification).

3. Results 3.1. Demographic and clinical characteristics Participant groups were well-matched on demographic variables involving age, sex and IQ, though BDD participants received fewer years of education (see Table 1). Clinical groups experienced similar levels of comorbid Axis I pathology, primary disorder severity as well as anxiety and depression. Though OCD participants had significantly longer illness durations relative to BDD participants, similar proportions of these groups were receiving psychiatric medication.

1 2

Numberof fixations to salient facial features−Number of fixations to non-salient facial features Total number of fixations Fixation durations to salient facial features−Fixation durations to non-salient facial features Total fixations durations

52

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59

Table 1 Participant demographic and clinical characteristics. Mean (standard deviation) BDD (n = 21)

OCD (n = 19)

HC (n = 21)

Demographic Age (years) Sex (% males) IQ Education (years)

34.3 (11.9) 23.8 105 (11.3) 13.6 (1.8)

37.0 (10.4) 26.3 106 (7.1) 14.3 (1.9)

35.7 (10.6) 38.1 110 (6.8) 15.1 (1.8)

Clinical Number of Axis I diagnoses (MINI500) Primary disorder severity (BDD-YBOCS/YBOCS) Social anxiety (SIAS) Fear of negative evaluation (BFNE-II) Depression (SDS) Illness duration (years) Medication (%)

2.8 (1.6) 23.8 (8.9) 36.8 (19.7) 31.3 (13.1) 47.6 (10.5) 10.1 (7.0) 95.2

1.7 (1.1) 20.3 (10.2) 32.6 (15.2) 29.0 (12.2) 44.1 (8.8) 17.5 (10.8) 73.7

2.5 (4.3)f 20.7 (9.3) 14.0 (9.2) 32.0 (6.5)

Statisticsa

p

Effect sizec

F(2, 60) = 0.31 2 (2, N = 61) = 1.17 F(2, 37.5) = 2.51b F(2, 60) = 3.72

.738 .558 .095 .03

.01 .14d .07 .11

U = 131.0, z = −1.90 F(2, 32.4) = 63.15b F(2, 34.5) = 8.23b F(2, 60) = 13.64 F(2, 36.7) = 21.88b F(1, 37)=6.35 2 (2,n = 40) = 2.14

.057 <.001 .001 <.001 <.001 .016 .143

.24e .59 .18 .32 .39 .15 .30g

Note: MINI500 = mini international neuropsychiatric interview-English Version 5.0.0; BDD-YBOCS = Yale-Brown obsessive-compulsive scale modified for body dysmorphic disorder (0–48); YBOCS = Yale-Brown obsessive-compulsive scale (0–48); SIAS = social interaction anxiety scale (0–80); BFNE-II = brief fear of negative evaluation-revised (0–48); SDS = Zung self-rating depression scale (0–60). Clinical instruments rated on 5-point Likert scales ranging from 0 = minimal symptoms to 4 = extreme symptoms, except SDS rated on a 4-point Likert scale ranging from 1 = none to 4 = always. a Statistics refer to one-way ANOVAs unless otherwise stated; chi-square tests for independence (sex, medication) and Mann-Whitney U test (number of Axis I diagnoses). b Welch F ratio. c Effect size refers to eta-squared unless otherwise stated. d Cramer’s V (sex). e r (number of Axis I diagnoses). f BDD-YBOCS administered based on specially elicited appearance-related concern over past week; score of 0 assigned in the absence of any concerns. g phi (medication).

3.2. Facial affect recognition: hypothesis 1 Descriptive statistics and ANOVAs are presented in Tables 2 and 3, respectively. For facial affect recognition, there was a significant main effect for group; the BDD group showed reduced accuracies relative to the OCD and HC groups. There was also a significant main effect for facial affect; reduced accuracies for anger, fear and sadness relative to neutral were found. To investigate whether angry recognition biases existed, one-way ANOVAs with planned comparisons were performed. There was a significant angry recognition bias (F2,58 = 8.74, p = .004); BDD participants misinterpreted more faces as angry relative to OCD and HC participants.

there were significant main effects for group and facial affect; greater saccade amplitudes for anger, fear, sadness and surprise, but smaller saccade amplitudes for disgust relative to neutral were found. There was a significant group by facial affect interaction. Simple effects analysis indicated significant group differences in saccade amplitudes across all facial affect (Fanger;2,60 = 8.46, p = .001; Fdisgust;2,60 = 7.66, p = .001; Ffear;2,60 = 5.37, p = .007; Fhappiness;2,60 = 10.3, p < .001; Fneutral;2,60 = 8.34, p = .001; Fsadness;2,60 = 6.06, p = .004; Fsurprise;2,60 = 10.0, p < .001). The BDD and OCD groups displayed greater saccade amplitudes to angry, happy, neutral, sad and surprised expressions relative to the HC group (see Table 2). The BDD group also displayed greater mean saccade amplitudes to disgusted and fearful expressions relative to the HC group.

3.3. Scanpath characteristics: hypothesis 2 3.4. Interest areas: hypothesis 3 For number of blinks, there was a significant main effect for group (see Table 3); the BDD group displayed more blinks relative to the OCD and HC groups. For number of fixations, there were significant main effects for group and facial affect (see Table 2); more fixations for anger, fear and sadness, but less fixations for happiness relative to neutral were found. There was a significant group by facial affect interaction. Simple effects analysis indicated significant group differences in fixations across all facial affect (Fanger;2,60 = 6.53, p = .003; Fdisgust;2,60 = 5.94, p = .004; Ffear;2,60 = 11.6, p<.001; Fhappiness;2,60 = 6.52, p = .003; Fneutral;2,60 = 7.91, p = .001; Fsadness;2,60 = 7.74, p = .001; Fsurprise;2,60 = 6.39, p = .003). The BDD group displayed fewer fixations to angry, disgusted, happy, neutral and surprised expressions relative to the HC group (see Table 2). The BDD group also displayed fewer fixations to fearful and sad expressions relative to the OCD and HC groups. For mean fixation durations, there was a significant main effect for group; the BDD group displayed greater fixation durations relative to the HC group. There was also a significant main effect for facial affect; reduced fixation durations for anger, fear and sadness, but increased fixation durations for happiness relative to neutral were found. For mean saccade amplitudes,

For FFI, there was a significant main effect for group (see Table 3); the BDD group showed reduced visual attention to salient facial features relative to the OCD and HC groups. Significant main effects for interest area and facial affect were also observed; increased visual attention to the eyes relative to the nose and mouth as well as for disgust, fear and surprise relative to neutral were found. There was a significant group by interest area interaction. Simple effects analysis indicated significant group differences in mean percentage fixation across all interest areas (Feyes;2,60 = 7.99, p = .001; Fnose;2,60 = 7.36, p = .001; Fmouth;2,60 = 7.69, p = .001). The BDD group displayed lower percentage fixations to the eyes, nose and mouth relative to the OCD and HC groups (see Table 2). For FDI, there was a significant main effect for group; the BDD group showed reduced visual attention to salient facial features relative to the OCD and HC groups. Significant main effects for interest area and facial affect were also observed; increased visual attention to the eyes relative to the nose and mouth as well as for fear and surprise relative to neutral were found. There was a significant group by interest area interaction. Simple effects analysis indicated significant group differences in mean percentage fixation durations across all interest areas (Feyes;2,60 = 6.81, p = .002;

Table 2 Group descriptive statistics by facial affect. Anger

Eye-tracking Number of blinks Number of fixations Mean fixation durations (ms) Mean saccade amplitudes (◦ ) Interest areas FFI Eyes (%) Nose (%) Mouth (%) FDI Eyes (%) Nose (%) Mouth (%)

Fear

Happiness

Neutral

Sadness

Surprise

OCD

HC

BDD

OCD

HC

BDD

OCD

HC

BDD

OCD

HC

BDD

OCD

HC

BDD

OCD

HC

BDD

OCD

HC

.75 (.17) 1339 (184)

.74 (.16) 1325 (239)

.85 (.10) 1255 (172)

.82 (.21) 1346 (164)

.82 (.15) 1383 (210)

.88 (.13) 1366 (198)

.65 (.21) 1368 (202)

.67 (.20) 1381 (196)

.73 (.24) 1294 (180)

.84 (.16) 1416 (220)

.85 (.14) 1445 (196)

.92 (.12) 1354 (214)

.80 (.19 1325 (178)

.85 (.15) 1356 (195)

.94 (.07) 1282 (184)

.67 (.23) 1393 (221)

.70 (.19) 1380 (169)

.79 (.16) 1348 (214)

.83 (.15) 1300 (182)

.79 (.19) 1371 (222)

.86 (.16) 1277 (208)

2.00 (1.37) 19.95 (3.96) 398.40 (149.3) 4.43 (.99)

1.11 (1.06) 22.32 (3.68) 337.70 (69.50) 4.23 (.94)

1.18 (.88) 23.97 (3.18) 317.20 (63.30) 3.35 (.77)

2.20 (1.61) 19.46 (3.82) 396.30 (130.60) 4.16 (.93)

1.13 (1.08) 21.54 (4.10) 360.10 (95.70) 3.79 (.89)

1.17 (.96) 23.38 (3.11) 329.60 (67.50) 3.13 (.78)

2.17 (1.64) 19.97 (4.25) 396.40 (169.10) 4.59 (.97)

1.10 (1.11) 22.85 (3.42) 325.30 (61.50) 4.33 (1.06)

1.00 (.86) 25.35 (3.07) 293.10 (52) 3.66 (.81)

2.04 (1.47) 18.62 (3.33) 430.80) (128.60) 4.41 (1.02)

1.19 (1.27 20.84 (4.29) 426 (245.20) 3.97 (1.07)

1.17 (.83) 22.95 (4.02) 336.80 (78.10) 3.10 (.74)

1.85 (1.30) 19.06 (3.66) 420.10 (149.80) 4.39 (1.00)

1.05 (1.00) 21.89 (4.00) 390 (201.70) 4.05 (.98)

1.20 (.89) 23.76 (3.90) 326.60 (70.90) 3.26 (.78)

2.15 (1.44) 19.63 (3.88) 393.90 (138.40) 4.38 (1.00)

1.09 (1.00) 22.71 (3.92) 333.20 (80.70) 4.15 (.86)

1.23 (1.01) 24.06 (3.39) 312.60 (59.50) 3.45 (.81)

2.33 (1.64) 19.73 (3.67) 402.30 (168.7) 4.83 (1.04)

1.21 (1.19) 21.84 (4.38) 364.80 (118.3) 4.34 (1.02)

1.12 (.83) 23.91 (3.30) 314.80 (60.9) 3.52 (.78)

.55 (.14) .47 (.14) .18 (.14) .13 (.07) .57 (.17) .48 (.17) .18 (.15) .13 (.08)

.68 (.17) .61 (.15) .11 (.06) .12 (.08) .69 (.17) .63 (.17) .10 (.06) .12 (.09)

.69 (.14) .62 (.12) .13 (.07) .10 (.06) .71 (.15) .65 (.13) .11 (.08) .09 (.06)

.57 (.15) .47 (.16) .21 (.12) .10 (.07) .59 (.17) .48 (.17) .22 (.15) .10 (.06)

.70 (.16) .59 (.17) .16 (.08) .10 (.07) .73 (.15) .61 (.18) .16 (.1) .10 (.08)

.73 (.13) .59 (.15) .19 (.09) .08 (.05) .75 (.13) .62 (.16) .18 (.1) .09 (.06)

.63 (.13) .53 (.15) .16 (.12) .12 (.08) .64 (.17) .54 (.16) .16 (.13) .13 (.09)

.75 (.13) .65 (.12) .11 (.05) .12 (.09) .76 (.14) .67 (.13) .10 (.06) .12 (.09)

.78 (.11) .67 (.12) .12 (.07) .10 (.06) .79 (.12) .70 (.13) .10 (.07) .10 (.07)

.56 (.16) .48 (.15) .16 (.1) .14 (.07) .59 (.18) .49 (.17) .15 (.12) .15 (.09)

.67 (.16) .58 (.16) .13 (.07) .12 (.08) .70 (.17) .60 (.18) .13 (.08) .12 (.09)

.71 (.15) .63 (.13) .13 (.07) .09 (.06) .73 (.15) .66 (.14) .11 (.06) .10 (.06)

.55 (.18) .52 (.16) .17 (.12) .08 (.06) .57 (.19) .53 (.17) .17 (.14) .08 (.06)

.66 (.19) .64 (.15) .12 (.06) .08 (.06) .70 (.20) .67 (.16) .10 (.07) .08 (.06)

.69 (.13) .65 (.12) .13 (.08) .06 (.04) .72 (.13) .68 (.12) .11 (.08) .06 (.05)

.59 (.15) .55 (.16) .17 (.13) .08 (.07) .61 (.17) .56 (.18) .17 (.14) .08 (.07)

.69 (.18) .65 (.14) .12 (.06) .08 (.06) .71 (.19) .68 (.15) 0.11 (.08) .07 (.06)

.70 (.14) .67 (.12) .13 (.07) .05 (.03) .72 (.14) .70 (.13) 0.11 (.08) .05 (.04)

.64 (.15) .56 (.15) .14 (.12) .12 (.07) .68 (.16) .57 (.17) 0.14 (.14) .12 (.07)

.75 (.13) .68 (.14) .09 (.05) .10 (.07) .79 (.14) .71 (.15) 0.08 (.06) .10 (.07)

.79 (.10) .70 (.11) .11 (.06) .08 (.05) .81 (.11) .73 (.12) 0.10 (.06) .08 (.05)

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59

Face processing Recognition Accuracy Response Latency (ms)

Disgust

BDD

Note: Recognition accuracies (range 0–1) calculated as mean proportion of correct responses by facial affect. Other variables expressed as the mean of total number of trials by facial affect. FFI = Feature fixation index; FDI = feature duration index.

53

54

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59

Table 3 Mixed between-within subjects ANOVAs for performance measures. F Face processing Recognition accuracy

Partial 2

p

Tukey’s HSD

Group Facial affect Group × facial affect

5.96 13.42 .32

.004 <.001 .984

.17 .60 .04

BDD,OCD < HC

Group Facial affect Group × facial affect Group Facial affect

4.86 2.14 1.41 8.1 8.05

.011 .064 .172 .001 <.001

.14 .2 .14 .22 .48

BDD > OCD,HC

Mean fixation durations (ms)

Group × facial affect Group Facial affect

1.83 3.26 4.16

.042 .045 .002

.17 .1 .32

BDD > HC

Mean saccade amplitudes (◦ )

Group × facial affect Group Facial affect

1.49 8.52 27.35

.139 .001 <.001

.14 .23 .76

BDD,OCD > HC

Group × facial affect

1.81

.046

.17

7.99 246.5 9.8 2.87 .25 16.02 1.02 6.77 229 8.48 2.96 .19 12.96 1.07

.001 <.001 <.001 .026 .995 <.001 .451 .002 <.001 <.001 .023 .999 <.001 .394

.22 .90 .53 .09 .03 .8 0.2 .19 .89 .49 .09 .02 .77 .21

Eye-tracking Number of blinks

Number of fixations

Interest areas FFI

FDI

Group Interest area Facial affect Group × interest area Group × facial affect Interest area × facial affect Group × interest area × facial affect Group Interest area Facial affect Group × interest area Group × facial affect Interest area × facial affect Group × interest area × facial affect

Contrasts

Ang,Fea,Sad < Neua

BDD < HC Ang,Fea,Sad > Neua Hap < Neua

Ang,Fea,Sad < Neua Hap > Neua

Ang,Fea,Sad,Sur > Neua Dis < Neua

BDD < OCD,HC Nos,Mou < Eyeb Dis,Fea,Sur > Neua

BDD < OCD,HC Nos,Mou < Eyeb Fea,Sur > Neua

Note: FFI = Feature fixation index; FDI = feature duration index; Ang = anger; Dis = disgust; Fea = fear; Hap = happiness; Neu = neutral; Sad = sadness; Sur = surprise; Eye = eyes; Nos = nose; Mou = mouth. Post-hoc comparisons and contrasts only shown for significant findings (p < .05). Follow-up simple effects analysis performed for significant interactions. a Simple contrasts performed with neutral as reference category. b Simple contrasts performed with eyes as reference category.

Fnose;2,60 = 6.18, p = .004; Fmouth;2,60 = 6.47, p = .003). The BDD group displayed smaller percentage fixation durations to the eyes, nose and mouth relative to the OCD and HC groups (see Table 2). 3.5. Qualitative scanpath analysis Examples of scanning styles and fixation heat maps are illustrated in Fig. 1 and 2 (obtained by random sampling of three individuals from each group based on facial affect). In Fig. 1, BDD participants exhibited relatively fewer fixations of greater durations, with higher saccade amplitudes. Their scanpath strategy can be described as disorganised, with a mixture of increased staring (for disgust, fear, sad) and extended scanning (for anger, happiness, neutral, surprise) as well as avoidance of salient facial features. In contrast, HC participants demonstrated the expected inverted triangular pattern of scanning, with the majority of fixations directed at salient features involving the eyes, nose and mouth. This scanning style was also replicated in OCD participants, albeit in a slightly less ordered manner. For instance, added extensive scanpaths could be observed for faces displaying negative affect (i.e. anger, disgust, fear, sadness). A fixation heat map creates a facial landscape to facilitate identification of the most informative parts of a display based on composite trial recordings. A typical map is expected to comprise a majority of fixations centred on the eyes and mouth. Extra fixations between the eyes, along the length of the nose, and lower portion of the forehead are respectively likely in response to angry,

disgusted and sad faces. In Fig. 2, though requisite fixations to the eyes and mouth were present for BDD participants, added fixations often encompassed non-salient facial regions, including the top of the head (for anger, fear, happiness), cheeks (for disgust, surprise) and chin (for fear, sadness, surprise). The overall pattern of visual attention can once again be described as disorganised. This was in contrast to fixation heat maps of HC participants, for which the expected patterns of visual attention were clearly demonstrated. The OCD participants also displayed similar visual configurations, though these were slightly more diffuse, and extended beyond salient features at times (for disgust, fear, happiness, surprise). 3.6. Correlations and covariates To explore whether affect recognition deficits were related to patterns of visual scanning, correlation analyses were done. Accuracy was negatively correlated with saccade amplitudes (R61 = −.35, P = .006), whereas latency was positively correlated with blinks (R61 = .34, P = .008), but negatively correlated with FFI (R61 = −.34, P = .007) and FDI (R61 = −.36, P = .004). Illness severity (i.e. BDD-YBOCS/YBOCS) was negatively correlated with accuracy and fixations, but positively correlated with saccade amplitudes (see Table 4). Individuals with severe BDD or OCD were poor in recognising facial affect and exhibited fewer fixations and greater scanpath lengths. In fact, illness severity explained 12.3% of the variance in participants’ accuracies. Anxiety (i.e. BFNE-II) was negatively correlated with fixations, whereas depression (i.e. SDS) was

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59

55

Fig. 1. Scanning styles by group and facial affect (n = 3). Group from left to right: BDD, OCD, and HC. Fixations marked in blue, and saccades marked in yellow. Sizes of circles in proportion to fixation durations, and directions of saccades indicated by relevant arrow heads. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

56

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59

Fig. 2. Fixation heat maps by group and facial affect (n = 3). Group from left to right: BDD, OCD, and HC. Routinely fixated areas marked in red, and infrequently fixated areas marked in green. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59 Table 4 Correlation analysis between performance measures and clinical variables. BDD-YBOCS/YBOCS

SIAS

BFNE-II

SDS

Face processing Recognition accuracy

−.35**

−.32*

−.30*

−.39**

Eye-tracking Number of blinks Number of fixations Mean fixation durations (ms) Mean saccade amplitudes (◦ )

.01 −.44** .30* .47**

−.12 −.32* .23 .25*

−.07 −.41** .31* .29*

.11 −.38** .25 .41**

Interest areas FFI FDI

−.29* −.27*

−.2 −.19

−.15 −.11

−.31* −.28*

Note: FFI = Feature fixation index; FDI = feature duration index; BDD-YBOCS = YaleBrown obsessive-compulsive scale modified for body dysmorphic disorder; YBOCS = Yale-Brown obsessive-compulsive scale; SIAS = social interaction anxiety scale; BFNE-II = brief fear of negative evaluation-revised; SDS = Zung self-rating depression scale. * Significant at .05 level (two-tailed). ** Significant at .01 level (two-tailed).

negatively correlated with accuracy and fixations, but positively correlated with saccade amplitudes. 4. Discussion This study employed a facial affect task with integrated eyetracking paradigm to investigate affect recognition and scanpath deficits in BDD relative to OCD and HC. In agreement with hypothesis 1, the BDD group demonstrated diminished accuracies and an angry recognition bias relative to the OCD and HC groups. The BDD participants were not only significantly less accurate in deciphering facial expressions, but also tended to misinterpret more faces as angry. This finding is in line with past research (Buhlmann et al., 2006; Buhlmann et al., 2004; Labuschagne et al., 2011), where preconceived notions of projected anger are believed to reinforce beliefs regarding one’s appearance-related flaws and consequent social rejection. As predicted in hypothesis 2, BDD participants displayed an atypical scanning strategy encompassing more blinks, fewer fixations of extended duration and higher saccade amplitudes relative to HC participants, whereas OCD participants exhibited an intermediate pattern. Specifically, the BDD group showed significantly more blinks relative to the OCD and HC groups. The BDD group also had significantly fewer fixations than the HC group across all facial affect, but also significantly fewer fixations than the OCD group for the emotions of fear and sadness. Fixation durations were significantly prolonged for the BDD group relative to the HC group, and clinical groups showed significantly greater scanpath lengths relative to the HC group for the emotions of anger, happiness, neutral, sadness and surprise, with the BDD group further showing significantly greater scanpath lengths relative to the HC group for the emotions of disgust and fear. This atypical scanning strategy in the BDD group was notably inclusive of (but not solely restricted to) negative affect. In contrast, the OCD group displayed only a slightly anomalous visual scanning strategy (i.e. significantly longer saccade amplitudes) relative to the HC group. In agreement with hypothesis 3, interest area analysis revealed reduced FFI and FDI in the BDD group relative to the OCD and HC groups. The BDD participants exhibited significantly fewer fixations of shorter durations to the eyes, nose and mouth. Correlation analysis demonstrated a significant negative association between accuracy and saccade amplitudes across participant groups. This suggests uncharacteristically extended scanpaths may be related to poor face processing. Illness severity was also associated with specific recognition and scanpath characteristics. Individuals with

57

severe BDD/OCD tended to exhibit lower accuracies, fewer fixations and greater saccade amplitudes. The scanpath strategy adopted by the current BDD sample did not fit wholly within restricted and extensive patterns respectively described in schizophrenia and SAD. The data in fact suggests BDD participants displayed a combination of enhanced staring (i.e. reduced fixations of extended durations) and extensive scanning (i.e. more blinks and longer scanpaths), which can be interpreted in two ways. First, some persons with BDD had a propensity for excessive staring, whereas others employed extensive scanning. This means that some BDD participants could have fixated excessively on their perceived flaw, but other BDD participants could have avoided their perceived flaw altogether, looking elsewhere except at said facial region. Second, disproportionate staring coupled with widespread scanning co-existed within the same individuals. Indeed a closer inspection of the disorganised scanpath strategy adopted by certain BDD participants (see Fig. 1 and 2) indicated both possibilities were equally likely. Such atypical scanpath characteristics point to deficits in basic ocular perception. Faulty perceptual operations impede the uptake and processing of visual information, leading to the formation of erroneous conclusions, evidenced by poor affect recognition. Interest area analysis showed added impairments in attentional processes, with less visual attention assigned to salient facial features in BDD. Similar findings have been documented in preliminary eye-tracking research in BDD, albeit based on an attractiveness paradigm (Greenberg et al., 2014). A deeper understanding of face processing disruptions would inform etiological foundations of BDD, and carry therapeutic implications. A direct clinical application could aim to enhance affect recognition in BDD. Specialised training of affect recognition (TAR; Frommann, Streit, & Wölwer, 2003; Wölwer et al., 2005), extraocular muscle (EOM) proprioception retraining (McCabe et al., 2007), and micro-expression training tool (METT; Russell, Green, Simpson, & Coltheart, 2008) developed for schizophrenia with promising results, may also prove useful in BDD. These existing clinical applications could be tailored to suit the needs of a BDD cohort, for example, by focusing on appearance-related concerns and associated elements of social perception. In particular, perceptual retraining with mirrors (as a component of cognitive-behavioural therapy) during which persons with BDD are taught to look at the ‘big picture’ rather than focusing on single, disliked aspects of their physical appearance, has been modestly effective (Phillips & Rogers, 2011; Wilhelm, Phillips, Fama, Greenberg, & Steketee, 2011). 4.1. Strengths and limitations This study possessed numerous strengths. It added to scant literature on face processing in BDD, and was one of few studies examining visual scanpaths in this population. The eye-tracking technique served as a unique psychophysiological measure, and the Eyelink II system, with its associated advantages of non-invasiveness, robustness, compatibility with a range of participants, and comprehensive data yield, was an added asset. It was noted most clinical participants had comorbid Axis I disorders, with few experiencing co-occurring BDD/OCD (n = 5). This limited the generalisability of findings, and raised the question of whether observed deficits could be attributed to other disorders, for instance, given elevated rates of SAD (33.3% BDD; 10.5% OCD) in these samples. A modest literature implying scanpath deficits in SAD meant extra caution needed to be exercised during interpretation. However, the BDD and OCD groups were well-matched on major demographic and clinical variables, making it more likely observed deficits were a product of the clinical disorders under study, rather than comorbid psychiatric conditions. This assertion was further supported by significant correlations between BDD symptomatology and performance measures.

58

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59

Yet the study was subject to a number of methodological limitations. The first of these relates to its consideration of interventions, and the second was in terms of the stimuli and clinical measures employed. More specifically, the majority of clinical participants were taking (multiple) psychiatric medications. Given the range of medications and doses involved, there was no way to statistically account for this. Yet research has concurred modern drugs exert minimal impact on eye movements (Gaebel, Ulrich, & Frick, 1987; Streit, Wölwer, & Gaebel, 1997; Wölwer, Streit, Polzer, & Gaebel, 19966), though some studies have reported otherwise (de Wilde et al., 2007; Williams, Loughland, Green, Harris, & Gordon, 2003). Furthermore, no information on current psychological therapies was collected, though once again, it would have been difficult to statistically manage this. It is noted that the Pictures of Facial Affect were in black-and-white (Ekman & Friesen, 1975). These stimuli were chosen for their high interrater agreement, excellent definition and placement, and broad range of facial affect. Coloured updates to reflect the times however, would not be amiss. Moreover, Buhlmann et al. (2006) raised a potential limitation in the use of neutral facial emotions as a control condition. This could be problematic, as neutral expressions are open to interpretation, and are at times, misinterpreted to reflect negative emotions especially by clinical participants. Yet till the use of a more appropriate control emotion can be put forward, neutral control faces will continue to serve as a potential drawback to research in the area. The BDD-YBOCS (based on a specially elicited appearance-related concern) was administered to the HC group; this instrument has yet to be validated in non-clinical populations (though Ilhan, Demirbas, & Dogan, 2006 did effectively use this tool to obtain a measure of craving in a group of Turkish men with alcohol dependence). 4.2. Directions for future research Replication studies are essential to uncover whether observed recognition and eye-tracking difficulties are robust effects demonstrable in other samples. Employing ‘pure’ BDD and OCD groups as well as those with comorbid BDD/OCD, and involving drug naive patients would also serve as fruitful replication efforts. Expanded research protocols could engage added facial emotions, such as contempt or shame. In fact, a contemptuous recognition bias has been tentatively identified in BDD (Buhlmann et al., 2006). To enhance ecological validity, digital displays involving dynamic facial interactions and social scenes could be utilised. In-depth investigations could also focus on establishing how and when these attentional biases and perceptual anomalies operate. For instance, the impaired mechanisms involved in configural processing and/or gestalt visual perception have been implicated, and deserve further study. Conflict of interest None. Funding Funding for this project was partially provided by Monash University Research Initiatives Scheme awarded to Susan Rossell and Monash University Strategic Grant awarded to Susan Rossell and Jerome Maller. Equipment was financed by Neuroscience Victoria Grant awarded to Susan Rossell. Role of sponsor The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of Monash University or Neuroscience Victoria.

Data access and responsibility Susan Rossell and Wei Lin Toh had full access to all data in the study and take responsibility for the integrity of data and accuracy of data analysis. Other submission This project was completed in partial fulfilment of a research thesis toward an M.Psych/PhD (clin. psychol.) degree awarded to Wei Lin Toh by University of Melbourne, Australia. References Eyelink II. [Apparatus and software]. SR Research Ontario. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing. Bestelmeyer, P. E. G., Tatler, B. W., Phillips, L. H., Fraser, G., Benson, P. J., & St. Clair, D. (2006). Global visual scanning abnormalities in schizophrenia and bipolar disorder. Schizophrenia Research, 87, 212–222. Buhlmann, U., Etcoff, N. L., & Wilhelm, S. (2006). Emotion recognition bias for contempt and anger in body dysmorphic disorder. Journal of Psychiatric Research, 40, 105–111. Buhlmann, U., Etcoff, N. L., & Wilhelm, S. (2008). Facial attractiveness ratings and perfectionism in body dysmorphic disorder and obsessive-compulsive disorder. Journal of Anxiety Disorders, 22, 540–547. Buhlmann, U., McNally, R. J., Etcoff, N. L., Tuschen-Caffier, B., & Wilhelm, S. (2004). Emotion recognition deficits in body dysmorphic disorder. Journal of Psychiatric Research, 38, 201–206. Buhlmann, U., Rupf, L., Gleiss, M. J. L., Zschenderlein, K., & Kathmann, N. (2014). Seeing changes that aren’t there: facial and object discrimination in body dysmorphic disorder. Comprehensive Psychiatry, 55, 468–474. Carleton, R. N., McCreary, D. R., Norton, P. J., & Asmundson, G. J. G. (2006). Brief fear of negative evaluation scale–revised. Depression and Anxiety, 23, 297–303. Cohen, J. (1988). Statistical power analysis for the behavioural sciences (2nd ed.). New Jersey: Lawrence Erlbaum Associates. Corcoran, K. M., Woody, S. R., & Tolin, D. F. (2008). Recognition of facial expressions in obsessive-compulsive disorder. Journal of Anxiety Disorders, 22, 56–66. Daros, A. R., Zakzanis, K. K., & Rector, N. A. (2014). A quantitative analysis of facial emotion recognition in obsessive-compulsive disorder. Psychiatry Research, 215, 514–521. de Wilde, O., Dingemans, P., Linszen, D. H., Bour, L. J., & Boerée, T. (2007). Visual scan paths in young patients with schizophrenia, healthy siblings and controls. Schizophrenia Research, 89, 362–363. Ekman, P., & Friesen, W. (1975). Unmasking the face: a guide to recognising emotions from facial cues. New Jersey: Prentice-Hall. Frommann, N., Streit, M., & Wölwer, W. (2003). Remediation of facial affect recognition impairments in patients with schizophrenia: a new training program. Psychiatry Research, 117, 281–284. Gaebel, W., Ulrich, G., & Frick, K. (1987). Visuomotor performance of schizophrenic patients and normal controls in a picture viewing task. Biological Psychiatry, 22, 1227–1237. Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleischmann, R. L., Hill, C. L., Heninger, G. R., et al. (1989). The Yale-Brown obsessive-compulsive scale: development, use, and reliability. Archives of General Psychiatry, 46, 1006–1011. Gordon, E., Coyle, S., Anderson, J., Healey, P., Cordaro, J., Latimer, C., et al. (1992). Eye movement response to a facial stimulus in schizophrenia. Biological Psychiatry, 31, 626–629. Greenberg, J. L., Reuman, L., Hartmann, A. S., Kasarskis, S., & Wilhelm, S. (2014). Visual hot spots: an eye tracking study of attention bias in body dysmorphic disorder. Journal of Psychiatric Research, 57, 125–132. Grisham, J. R., Henry, J. D., Williams, A. D., & Bailey, P. E. (2010). Socioemotional deficits associated with obsessive-compulsive symptomatology. Psychiatry Research, 175, 256–259. Grocholewski, A., Kliem, S., & Heinrichs, N. (2012). Selective attention to imagined facial ugliness is specific to body dysmorphic disorder. Body Image, 9, 261–269. Horley, K., Williams, L. M., Gonsalvez, C., & Gordon, E. (2003). Social phobics do not see eye to eye: A visual scanpath study of emotional expression processing. Journal of Anxiety Disorders, 17, 33–44. Horley, K., Williams, L. M., Gonsalvez, C., & Gordon, E. (2004). Face to face: visual scanpath evidence for abnormal processing of facial expressions in social phobia. Psychiatry Research, 127, 43–53. Ilhan, I. O., Demirbas, H., & Dogan, Y. B. (2006). Validation study of the Turkish version of the Yale-Brown Obsessive Compulsive Scale for heavy drinking in a group of male patients. Drug Alcohol Review, 25, 357–360. Jhung, K., Namkoong, K., Kang, J. I., Ha, R. Y., An, S. K., Kim, C., et al. (2010). Perception bias of disgust in ambiguous facial expressions in obsessive-compulsive disorder. Psychiatry Research, 178, 126–131. Labuschagne, I., Castle, D. J., & Rossell, S. L. (2011). What the cognitive deficits in body dysmorphic disorder tell us about the underlying neurobiology: an investigation of three cases. International Journal of Cognitive Therapy, 4, 21–33.

W.L. Toh et al. / Journal of Anxiety Disorders 35 (2015) 49–59 Lambrou, C., Veale, D., & Wilson, G. (2011). The role of aesthetic sensitivity in body dysmorphic disorder. Journal of Abnormal Psychology, 120, 443–453. Loughland, C. M., Williams, L. M., & Gordon, E. (2002a). Schizophrenia and affective disorder show different visual scanning behaviour for faces: a trait versus state-based distinction? Biological Psychiatry, 52, 338–348. Loughland, C. M., Williams, L. M., & Gordon, E. (2002b). Visual scanpaths to positive and negative facial emotions in an outpatient schizophrenia sample. Schizophrenia Research, 55, 159–170. Manor, B. R., Gordon, E., Williams, L. M., Rennie, C. J., Bahramali, H., Latimer, C. R., et al. (1999). Eye movements reflect impaired face processing in patients with schizophrenia. Biological Psychiatry, 46, 963–969. Mattick, R. P., & Clarke, J. C. (1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behaviour Research and Therapy, 36, 455–470. McCabe, K., Loughland, C. M., Cohen, M., Hunter, M., Lewin, T. J., & Carr, V. J. (2007). Remediation of facial affect decoding and visual scanpath deficits in schizophrenia. Schizophrenia Bulletin, 33, 531–532. Miller, G. A., & Chapman, J. P. (2001). Misunderstanding analysis of covariance. Journal of Abnormal Psychology, 110, 40–48. Parker, H. A., McNally, R. J., Nakayama, K., & Wilhelm, S. (2004). No disgust recognition deficit in obsessive-compulsive disorder. Journal of Behaviour Therapy & Experimental Psychiatry, 35, 183–192. Phillips, K. A. (2005). The broken mirror: understanding and treating body dysmorphic disorder. New York: Oxford University Press. Phillips, K. A., Hollander, E., Rasmussen, S. A., Aronowitz, B. R., DeCaria, C., & Goodman, W. K. (1997). A severity rating scale for body dysmorphic disorder: development, reliability, and validity of a modified version of the Yale-Brown obsessive compulsive scale. Psychopharmacology Bulletin, 33, 17–22. Phillips, K. A., Menard, W., Fay, C., & Weisberg, R. B. (2005). Demographic characteristics, phenomenology, comorbidity, and family history in 200 individuals with body dysmorphic disorder. Psychosomatics, 46, 317–325. Phillips, K. A., & Rogers, J. (2011). Cognitive-behavioural therapy for youths with body dysmorphic disorder: current status and future directions. Child and Adolescent Psychiatric Clinics of North America, 20, 287. Rector, N. A., Daros, A. R., Bradbury, C. L., & Richter, M. A. (2012). Disgust recognition in obsessive-compulsive disorder: diagnostic comparisons and posttreatment effects. Canadian Journal of Psychiatry, 57, 177–183. Reese, H. E., McNally, R. J., & Wilhelm, S. (2010). Facial asymmetry detection in patients with body dysmorphic disorder. Behaviour Research and Therapy, 48, 936–940. Rozin, P., Taylor, C., Ross, L., Bennett, G., & Hejmadi, A. (2005). General and specific abilities to recognise negative emotions, especially disgust, as portrayed in the face and the body. Cognition & Emotion, 19, 397–412. Russell, T. A., Green, M. J., Simpson, I., & Coltheart, M. (2008). Remediation of facial emotion perception in schizophrenia: concomitant changes in visual attention. Schizophrenia Research, 103, 248–256. Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, E., et al. (1998). The Mini international neuropsychiatric interview (MINI): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. Journal of Clinical Psychiatry, 59, 22–33.

59

Simeon, D., Hollander, E., Stein, D. J., Cohen, L., & Aronowitz, B. (1995). Body dysmorphic disorder in the DSM-IV field trial for obsessive-compulsive disorder. American Journal of Psychiatry, 152, 1207–1209. Sprengelmeyer, R., Young, A. W., Pundt, I., Sprengelmeyer, A., Calder, A. J., Berrios, G., et al. (1997). Disgust implicated in obsessive-compulsive disorder. Proceedings of the Royal Society of London B: Biological Sciences, 264, 1767–1773. Stangier, U., Adam-Schwebe, S., Muller, T., & Wolter, M. (2008). Discrimination of facial appearance stimuli in body dysmorphic disorder. Journal of Abnormal Psychology, 117, 435–443. Streit, M., Wölwer, W., & Gaebel, W. (1997). Facial affect recognition and visual scanning behaviour in the course of schizophrenia. Schizophrenia Research, 24, 311–317. Toh, W. L., Rossell, S. L., & Castle, D. J. (2011). Current visual scanpath research: a review of investigations into the psychotic, anxiety, and mood disorders. Comprehensive Psychiatry, 52, 567–579. Wechsler, D. (2001). Wechsler Test of Adult Reading (WTAR). San Antonio: Psychological Corporation. Wilhelm, S., Otto, M. W., Zucker, B. G., & Pollack, M. H. (1997). Prevalence of body dysmorphic disorder in patients with anxiety disorders. Journal of Anxiety Disorders, 11, 499–502. Wilhelm, S., Phillips, K. A., Fama, J. M., Greenberg, J. L., & Steketee, G. (2011). Modular cognitive-behavioral therapy for body dysmorphic disorder. Behavior Therapy, 42, 624–633. Williams, L. M., Loughland, C. M., Gordon, E., & Davidson, D. (1999). Visual scanpaths in schizophrenia: is there a deficit in face recognition? Schizophrenia Research, 40, 189–199. Williams, L. M., Loughland, C. M., Green, M. J., Harris, A. W., & Gordon, E. (2003). Emotion perception in schizophrenia: an eye movement study comparing the effectiveness of risperidone versus haloperidol. Psychiatry Research, 120, 13–27. Wölwer, W., Frommann, N., Halfmann, S., Piaszek, A., Streit, M., & Gaebel, W. (2005). Remediation of impairments in facial affect recognition in schizophrenia: efficacy and specificity of a new training program. Schizophrenia Research, 80, 295–303. Wölwer, W., Streit, M., Polzer, U., & Gaebel, W. (1996). Facial affect recognition in the course of schizophrenia. European Archives of Psychiatry and Clinical Neuroscienc, 246, 165–170. World Medical Association, 1995. Declaration of Helsinki: Ethical principles for medical research involving human subjects. Yaryura-Tobias, J. A., Neziroglu, F., Chang, R., Lee, S., Pinto, A., & Donohue, L. (2002). Computerised perceptual analysis of patients with body dysmorphic disorder: a pilot study. CNS Spectrums, 7, 444–446. Zebrowitz, L. A. (1997). Reading faces: window to the soul? Colorado: Westview Press. Zung, W. (1965). A self-rating depression scale. Archives of General Psychiatry, 12, 63–70.