Recent work has shown that judgment biases during facial processing serve as a correlate of dysmorphic concern. The current study aimed to replicate and extend this work by further testing larger model-congruent interrelations between judgment biases and intrusions, negative affect, maladaptive behaviors and social media investment.
Methods
Females reporting high (n = 30) vs. low (n = 33) dysmorphic concern categorized photographs of their own and other faces in terms of either attractiveness or gaze direction using the Biased Judgment-Dimension Task. To assess intrusions and maladaptive behaviors, participants completed the Task-Irrelevant Intrusions Task and a 3-day diary, respectively.
Results
Results revealed that, regardless of judgment required, individuals low in dysmorphic concern were faster to make judgments about attractive vs. unattractive photographs. We termed this novel and potentially protective bias “Efficient Processing of Attractive Self” (EPAS). Furthermore, EPAS patterns predicted lower dysmorphic concern, higher self-esteem, and reduced maladaptive behaviors, including reduced screen time.
Conclusions
These findings are consistent with a potential relevance of EPAS in maintaining dysmorphic concern. Associations with cognitive-behavioral factors highlight potential benefits of modifying these bias patterns during visually oriented interventions.
Hinweise
Fanny Alexandra Dietel and Laura Jacobs jointly first-authored this manuscript.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Introduction
Dysmorphic concern—i.e., preoccupation with a perceived flaw in appearance (Oosthuizen et al., 1998)—is a phenomenon underpinning various body image disorders (e.g., Beilharz et al., 2019; Mancuso et al., 2010). Most prominently, dysmorphic concern marks body dysmorphic disorder (BDD), a prevalent, severe mental condition that is associated with substantial chronicity, suicidality, health burden and functional impairment (Angelakis et al., 2016; Marques et al., 2011; Phillips et al., 2013). While dysmorphic concern and BDD may pertain to any body part, they are most often tied to facial features (Phillips et al., 2005). However, the cognitive factors maintaining such facial appearance concerns remain elusive at present. Thus, the current study aims to investigate the role of cognitive biases during facial processing in dysmorphic concern and related symptomatology.
Cognitive-behavioral models of BDD converge on the idea that aberrant information processing is a core mechanism maintaining the disorder (Fang & Wilhelm, 2015; Veale, 2004). Specifically, these models postulate that ordinary appearance concerns become intrusive through cognitive biases fostered by maladaptive beliefs. Overvalued appearance concerns then prompt intense negative emotions (e.g., anxiety or shame) and subsequent maladaptive behaviors, including ritualistic (e.g., mirror checking), safety (e.g., camouflaging) and avoidant behaviors (e.g., avoiding social contacts). While these behaviors reduce distress short-term, they exacerbate symptoms long-term by preventing disconfirmation of maladaptive beliefs, thereby maintaining appearance-related intrusions within a vicious cycle (see Fang & Wilhelm, 2015). Thus, cognitive-behavioral models assume a functional link between cognitive biases, negative emotions, maladaptive behaviors and, eventually, appearance-related intrusions.
Anzeige
Supporting these postulates, research has identified biased processing of appearance-related information as a central component in dysmorphic concern (Fang & Wilhelm, 2015). Particularly, research suggests that individuals with elevated dysmorphic concern show perceptual and cognitive distortions when processing facial stimuli. Specifically, a series of studies revealed that individuals with BDD, compared to mentally healthy individuals, encode faces in a locally oriented (vs. holistic) fashion (e.g., Feusner et al., 2007), display extensive facial scanning behaviors (e.g., Toh et al., 2017) and selectively attend to self-rated unattractive over attractive facial features (e.g., Greenberg et al., 2014; Kollei et al., 2017). Relatedly, individuals with high (vs. low) facial appearance concern have been demonstrated to exhibit speeded attentional orienting towards unattractive faces (Kou et al., 2016). Last, individuals with BDD, vs. those employed in an aesthetic profession and mentally healthy controls, show an absence of self-serving bias when judging photos of oneself as manipulated across a symmetry continuum, vs. other stimuli (Lambrou et al., 2011).
Taken together, these results indicate that preferential encoding of unattractiveness might affect judgments downstream due to increased salience of attractiveness- compared to attractiveness-irrelevant dimensions (e.g., facial expression). Such overrepresentation of attractiveness could be amplified by maladaptive beliefs about the importance of appearance and low self-esteem (Buhlmann et al., 2009a, b, 2011). These aspects may foster lowered attractiveness ratings of one’s own (vs. other) face, as have been identified in BDD (Buhlmann et al., 2008). Negative attractiveness-oriented judgments may further exacerbate negative emotions and maladaptive behaviors, thus promoting increased appearance-related intrusions long-term. Given these possibilities, it is a concern that biases during facial judgment and their etiological links in dysmorphic concern remain insufficiently understood.
A seminal study revealed facial attractiveness judgment biases that differentiate between high and low facial concern (Dondzilo, et al., 2021). In this study, participants underwent a photo shoot and subsequently selected photographs of themselves, vs. another person, that they perceived as the most and least attractive. In a subsequent novel Biased Judgment-Dimension Task (BJDT), participants were then requested to judge these photographs in terms of attractiveness or eye gaze direction. Findings revealed that, regardless of judgment required, individuals high in facial concern were faster to classify unattractive (vs. attractive) photographs of their own face. This potential vulnerability bias was termed the “Efficient Processing of Unattractive Self” (EPUS). Further, results suggested that individuals low in facial dysmorphic concern were faster at making attractive (vs. unattractive) judgments of one’s own and others’ faces, demonstrating a potential protective bias termed the “General Attractiveness Recognition Advantage” (GARA). Additional analyses revealed that variation in EPUS and GARA each meaningfully contributed to variance in facial appearance concern. Critically, a reduced GARA bias was associated with more appearance-related intrusions in high dysmorphic concern (Dondzilo, et al., 2021). This is consistent with a potential etiological link between judgment biases and appearance-related intrusions. However, given the novelty of these findings, replication is warranted, and this forms the first aim of the present study.
As a second aim, the present study seeks to extend prior findings by investigating previously untested interrelations between judgment biases and model-congruent cognitive-behavioral factors. Particularly, this study aims to delineate associations of judgment biases on maladaptive behaviors linked to the assessment of appearance-related visual material. With the burgeoning of social media (e.g., Pew Research Center, 2021), research has evidenced that higher exposure to appearance-related stimuli on these platforms is associated with increased body dissatisfaction (Ryding & Kuss, 2020; Tiggemann & Zaccardo, 2015), more negative emotions (Bennett et al., 2020; Jarman et al., 2021; Ridolfi et al., 2011) and more pronounced appearance comparison, with the latter mediating the relationship between social media usage and depression (Lup et al., 2015). Beyond these detrimental effects of passive consumption, active social media investment (i.e., photo editing) has been linked to body dissatisfaction, lowered mood and disordered eating (e.g., Lonergan et al., 2019, 2020; McLean et al., 2015; Tiggemann et al., 2020). Importantly, harmful effects of social media usage appear more pronounced in individuals exhibiting higher body dissatisfaction and elevated appearance comparison (Fardouly & Vartanian, 2016), a behavioral pattern that marks dysmorphic concern (Anson et al., 2015; Phillips et al., 2006). Overall, these findings suggest that greater exposure to—and evaluation of—appearance-related visual stimuli may exacerbate appearance concerns. Notably, judgment biases may influence this link by promoting salient processing of attractiveness-related stimuli. Specifically, it appears plausible that repeatedly judging appearance-related stimuli on social media (e.g., selfies) in terms of attractiveness may foster negative emotions and maladaptive behaviors, such as extensive photo editing. However, it remains unknown whether judgment biases indeed predict such maladaptive behaviors in elevated dysmorphic concern.
Anzeige
In sum, the first aim of the current study was to examine the replicability of judgment biases during facial processing in low vs. high dysmorphic concern (i.e., within a subclinical analogue sample). Consistent with prior findings and cognitive-behavioral models, the second aim was to investigate the predictive value of judgment biases in (1) negative affect, (2) maladaptive behaviors (i.e., with a focus on social media investment), and (3) appearance-related intrusions. Importantly, social media behaviors were measured as dynamic constructs, using a 3-day daily diary approach.
We predicted that the previously identified judgment bias patterns (i.e., GARA and EPUS) would replicate. Further, we predicted that judgment biases would be associated with dysmorphic symptom severity, maladaptive behaviors (including appearance-related social media behavior), negative affect and self-esteem within meaningful directions. That is, GARA would be associated with lower dysmorphic symptom severity, less maladaptive behaviors, lower negative affect and higher self-esteem, while EPUS would be associated with the reverse patterns.
Methods
Study Design
This was a two-group experimental online study conducted at the University of Muenster between October 2020 and June 2021. The study design was approved by the local ethics committee.
An a priori power analysis using G*Power (Faul et al., 2007) and assuming the smallest three-way interaction effect for bias patterns (Dondzilo et al., 2021; parameters: 1-β = 0.95, α = 0.05, f2 = 0.29, rbetween = 0.5) yielded a required total sample size of n = 40 (i.e., n = 20 per group) to observe these effects.
Participants
As in Dondzilo et al. (2021), participants were undergraduate female students. They were recruited via announcements posted via the Internet, flyers and participant pools. Inclusion criteria were: (1) aged over 18 years; (2) low appearance concern (sum score of ≤ 1 on the Dysmorphic Concern Questionnaire [DCQ; Oosthuizen et al., 1998]) or high appearance concern (DCQ sum score of ≥ 4); (3) main focus of concern within body area above the chest (assessed through the categorical item 2 [“If you find parts of your body unattractive, which parts are they?”] on the Body Symptoms Inventory [“Fragebogen körperdysmorpher Symptome”, FKS; Buhlmann et al., 2009a, b]); (4) no acute suicidality, i.e., a total score of < 3 on the Depressive Symptom Inventory-Suicidality Subscale (DSISS; von Glischinski et al., 2016); (4) no current diagnosis of any mental disorder; (5) no current psychotherapeutic or psychiatric treatment. Participants not meeting inclusion criteria were redirected to a website containing contact information for mental health services.
Of the 366 individuals who were screened, 33 individuals were recruited and designated to the low appearance concern group and a further 30 individuals were recruited and designated to the high appearance concern group. Thus, the final sample consisted of N = 63 individuals.
The four-item DSISS (Joiner et al., 2002; von Glischinski et al., 2016) was employed to screen for suicidality within the past 2 weeks as per the eligibility criteria (von Glischinski et al., 2016).
Dysmorphic Concern Questionnaire
The seven-item DCQ (Oosthuizen et al., 1998) was administered to assess dysmorphic concern severity, basing group allocation on the upper and lower percentile determined in an applicable norming study (Schieber et al., 2018). Individuals falling below a raw score of ≤ 1 (PR = 29) were assigned to the low dysmorphic concern group, and individuals exceeding a raw score of ≥ 4 (PR = 71) were assigned to the high dysmorphic concern group.
Body Dysmorphic Symptoms Inventory
The 18-item FKS (Buhlmann et al., 2009a, b) was employed to assess body dysmorphic symptom severity. To identify primary facial appearance concerns, we used item 2 to isolate main foci of concern. Internal consistency in the current sample was high (Cronbach's α = .94).
Depression and Anxiety Stress Scales-21
The 21-item Depression and Anxiety Stress Scales-21 (DASS-21; Lovibond & Lovibond, 1995; Nilges & Essau, 2015) was employed to assess severity of depression, anxiety and stress in the past week. Internal consistency in the current sample was high (α = 0.92).
Appearance Behavior Checklist
The fourteen-item Appearance Behavior Checklist (ABC; Summers & Cougle, 2018) was used to assess appearance-related maladaptive behaviors. The scale was translated into German using the back-translation method. Moreover, the original response format was modified to an absolute response scale to allow for quantification of behavioral frequency, including the following options: "Not at all", "Yes, less than 5 times", "Yes, 5–10 times", "Yes, 10–15 times", "Yes, more often than 15 times", "Not possible". At baseline, participants were requested to judge behavioral frequency during the last 7 days. During the 3-day diary period, the referential period was the respective day of assessment. Internal consistency in the current sample was high for the baseline (α = .91) and daily diary versions (α = .89 to α = .91).
State Negative Affect
Consistent with precedence (Dietel et al., 2018), state negative affect was assessed by means of a four-item scale requesting participants to rate current fear, shame, joy and depression on a 0–100% visual analog scale. To yield a total score, scores across the four items were averaged, with the joy item being reverse scored. Given the multi-dimensionality of the negative affect construct, internal consistency was conceivably lower than expected (α = 0.64 to α = 0.66).
Social Media Investment Scale
Appearance-related investment in social media was assessed using four items, based on the Self-Photo Manipulation Scale and Self-Photo Investment Scale (McLean et al., 2015), covering (frequency of): (1) sharing images on social media, (2) selecting images for social media sharing, (3) editing attractiveness-related aspects in images for social media sharing, and (4) screen time. Response options were identical to the ABC, except for ‘screen time’ using a seven-point frequentist response scale ranging from ‘less than 30 min’ to ‘more than 5 h’. Internal consistency in the current sample was acceptable at baseline (α = .72), but low for the daily diary versions (α = .49 to α = .57).
State Self-esteem Scale
We used the 15-item State Self-Esteem Scale (SSES; Heatherton & Polivy, 1991) to assess self-esteem. Internal consistency in the current sample was high (α = .92).
Stimulus Development Procedure: Photograph Generation and Selection
Following Dondzilo et al. (2021), the BJDT required a total of 16 photographs, comprising eight photographs depicting the participant’s face and eight photographs depicting the face of a gender-matched confederate. Further, per identity category (self vs. other), the task required an equal number of attractive (N = 4) and unattractive (N = 4) photographs.
To generate stimuli, participants underwent a standardized, online photo shoot (Dondzilo et al., 2021). Photo shoots were executed in a well-lit room, with participants sitting upright on a chair and facing their webcam. During the photo shoot, participants were asked to adopt the following four poses while gazing to the right, with four pictures being taken per pose: (1) passport photograph; (2) challenging verbal task; (3) blowing out cheeks; and (4) smiling. The resulting subset of 16 photographs was flipped vertically to generate the equivalent 16 mirrored “gaze left” images, resulting in a total of 32 photographs.
In a subsequent image selection task (see Fig. 1), participants were required to select the most attractive and least attractive photographs out of their photo shoot (N = 32) and a content-matched confederate stimulus set (N = 32). A 4 × 2 grid containing a selection of eight photographs was presented in each trial. Above the grid, an instruction cue indicated whether selections should be based on attractiveness or eye gaze direction (i.e., select the most/least attractive photograph vs. select the photograph where gaze is directed to the right/left). Judgment instructions were presented at an equal probability to prevent induction of a specific processing style prior to the main task.
Fig. 1
Trial procedure of the photo selection task and the BJDT (exemplary trial procedures with placeholder images). ITI Inter-trial interval
×
Overall, the selection procedure yielded a total of 16 individualized stimuli per participant, comprising eight “self” and eight “other” images, i.e., four pictures perceived as attractive and four pictures perceived as unattractive per identity category. Stimuli thus could be represented along the two within-subject factors Identity (Self vs. Other) and Attractiveness Category (Attractive vs. Unattractive Set).
Experimental Tasks
Biased Judgment-Dimension Task (BJDT)
A German version of the BJDT (Dondzilo, et al., 2021) was employed to assess judgment biases during facial processing (see Fig. 1 for the experimental trial procedure of the BJDT). Participants first received eight practice trials employing silhouetted placeholder faces. Test trials started with a central fixation cross (+) presented for 1000 ms that was subsequently replaced by an image (i.e., taken at equal probability from the eight ‘self’ vs. the eight ‘other’ photograph pool). After 1000 ms, the judgment question addressing either attractiveness (i.e., “attractive set?” or “unattractive set?”) or eye gaze direction (i.e., “gaze right?” or “gaze left?”) was presented at equal frequency. Participants were asked to respond as quickly as possible by pressing the up (i.e., indicating “yes”) or down arrow (i.e., indicating “no”) on the keyboard. Upon response, the next trial commenced after a 500 ms inter-trial interval, with judgment latencies being recorded for all trials.
Overall, the BJDT comprised 256 trials presented across four 64-trial blocks that were separated by a self-timed break. Trial presentation was randomized. Per block, each unique picture was displayed four times and at equal frequencies for each image category (self vs. other), attractiveness category (attractive vs. unattractive set), eye gaze direction (gaze right vs. gaze left) and subsequent judgment category (attractiveness-relevant vs. attractiveness-irrelevant judgment).
Task-Irrelevant Intrusions Assessment Task (TIIT)
Appearance-related intrusions were assessed using an intrusion assessment task (Dondzilo et al., 2021). At the beginning of each trial, a string of five characters was presented centrally. Out of the five characters, there was a 50% probability that three were numbers and two were letters, and a 50% probability of the inverse ratio. Participants were requested to determine as fast as possible whether the majority of characters were numbers or letters. Upon button press, the next trial commenced after a 1500 ms inter-trial interval. While participants were instructed to concentrate on this task, they were asked to press the space bar whenever task-irrelevant thoughts entered their minds. When pressing the space bar, participants were asked if the thoughts they were experiencing related to the photo shoot or photo evaluation task, yielding a frequency index of appearance-related intrusions. After each response, participants resumed the initial number/letter task and had to complete at least four more trials before being able to report more task-irrelevant intrusions. The TIIT had a set duration of 10 min per participant.
Procedure
Figure 2 illustrates the study procedure. The assessment session was conducted online via secure video call. Upon informed consent, participants were requested to complete all self-report questionnaires, including the ABC and the social media investment scale, followed by the photo shoot, the first assessment of negative state affect and the TITT. Upon completion of the photo shoot, experimenters prepared and uploaded the individualized stimulus set for the subsequent experimental tasks. Participants then completed the image selection task, the BJDT, the TIIT again, and the second assessment of state negative affect. The entire baseline assessment session lasted approximately 70 min. At the end of the session, participants received instructions for completing the subsequent 3-day diary. Participants received daily text message prompts to complete the online diary surveys at 7 pm for 3 days following the baseline session. Each diary survey comprised the social media investment scale and the ABC. After completing the final diary, participants were debriefed via e-mail, receiving either time-contingent course credit or financial reimbursement (8 Euro per hour).
Fig. 2
Study procedure. ABC Appearance Behavior Checklist, BJDT Biased Judgment-Dimension Task, DASS-21 Depression Anxiety and Stress Scales-21, DCQ Dysmorphic Concerns Questionnaire, DSISS Depression Symptoms Inventory - Suicidality Subscale, FKS Body Dysmorphic Symptoms Inventory, SSES State Self-esteem Scale, TIIT Task-irrelevant Intrusions Task
×
Statistical Analysis
Data analysis was carried out using IBM SPSS Statistics, Version 28, and R (R Core Team, 2018)1. First, judgment decisions on the BJDT were screened for accuracy, and trials recording an incorrect response were discarded. Participants with an accuracy below 75% (i.e., n = 1 in this study) were excluded from further analysis. The remaining participants exhibited correct responses in 93.2% of the trials, demonstrating high accuracy. To eliminate spurious responses, judgment latencies below 200 ms and above 3000 ms were omitted (Dondzilo et al., 2021). Further, diary questionnaires answered later than within 1 day of receiving the link were excluded.
Anzeige
To test for between-group differences in judgment latencies, median judgment latencies were subjected to a four-way mixed model ANOVA including the within-subject factors Judgment Required (Attractiveness vs. Gaze Direction), Identity (Self vs. Other), and Attractiveness Category (Attractive Set vs. Unattractive Set), and the between-subjects factor Dysmorphic Concern Group (High Appearance Concern vs. Low Appearance Concern). Significant interactions were examined using lower-order mixed model ANOVAs.
Next, the attractiveness-linked processing biases GARA (i.e., reflecting speeded categorization of attractive vs. unattractive pictures during judgments on the attractiveness-related dimension) and EPUS (i.e., reflecting speeded categorization of unattractive vs. attractive pictures when judging self vs. other pictures) were computed using the established equations:
GARA Index = (Judgment latency for attractiveness judgments on unattractive images minus judgment latency for attractiveness judgments on attractive images) minus (judgment latency for gaze direction judgments on unattractive images minus judgment latency for gaze direction judgments on attractive images).
EPUS Index = (Judgment latency for judgments on attractive images of self minus judgment latency for judgments on unattractive images of self) minus (judgment latency for judgments on attractive images of other minus judgment latency for judgments on unattractive images of other).
Anzeige
We further identified a novel bias pattern that indexed relative speeding of responding to attractive (vs. unattractive) pictures when judging self (vs. other) pictures. This bias was termed the Efficient Processing of Attractive Self (EPAS) and was calculated as follows:
EPAS Index = (Judgment latency for judgements on unattractive images of self minus judgment latency for judgments on attractive images of self) minus (judgment latency for judgments on unattractive images of other minus judgment latency for judgments on attractive images of other).
To examine whether bias indices contributed unique variance to dependent variables, we conducted regression analyses with dysmorphic symptom severity, repetitive behavior frequency (baseline), state negative affect, self-esteem and social media investment (baseline) as the dependent variables. Further, we computed binomial regression analyses for intrusion frequency as a dependent variable. Last, correlation analyses were conducted to examine associations between bias indices and diary versions of the ABC and social media investment scale.
Tests were two-tailed using a significance value of p < 0.05, with Bonferroni-corrections for multiple testing.
Anzeige
Results
Sample Characteristics
As shown in Table 1, groups did not differ regarding age. However, as expected, groups differed significantly on measures of psychopathology. Of note, regarding social media investment, groups differed significantly in photo choosing and screen time, but not in photo sharing or editing.
Table 1
Demographic and psychometric characteristics at baseline
Low dysmorphic concern (n = 33)
High dymorphic concern (n = 30)
Total sample (n = 63)
Group comparisons
M (SD)
M (SD)
M (SD)
T
p
d
Age
23.09 (3.23)
23.67 (3.20)
23.37 (3.20)
− .710
.481
− .179
FKS
6.33 (3.85)
25.57 (6.52)
15.49 (11.01)
− 14.08
< .001
− 3.636
DASS-21 total
8.94 (8.12)
18.23 (11.15)
13.37 (10.68)
− 3.751
< .001
− .961
SSES total
56.82 (8.68)
41.50 (7.91)
49.52 (11.30)
7.331
< .001
1.841
Intrusion frequency (baseline)
3.88 (4.97)
5.00 (8.05)
4.41 (6.59)
− .657
.514
− .169
ABC (baseline)
26.61 (7.73)
37.37 (9.30)
31.73 (10.03)
− 4.969
< .001
− 1.265
State negative affect (total)
61.79 (35.41)
123.90 (59.45)
91.37 (57.25)
− 4.97
< .001
− 1.825
Social media investment (baseline)
Sharing photos
1.82 (1.13)
1.93 (0.83)
1.87 (0.99)
− .464
.644
− 0.115
Choosing photos
1.48 (0.83)
2.03 (1.19)
1.75 (1.05)
− 2.101
.041
− 0.539
Editing photos
1.30 (0.69)
1.53 (0.97)
1.41 (0.84)
− 1.077
.287
− 0.273
Screentime
2.97 (1.05)
4.07 (1.53)
3.49 (1.40)
− 3.291
.002
− 0.845
Notes: FKS is not printed in Italics
Intrusions (baseline) refer to sum of intrusions reported on the Task-Irrelevant Intrusions Task at baseline. State negative affect denotes sum scores across four items: shame, fear, joy (reverse scored), depression.
ABC Appearance Behavior Checklist, DASS-21 Depression Anxiety and Stress Scales-21, FKS Body Dysmorphic Symptoms Inventory, SSES State Self-esteem Scale.
Average medians (SDs) of judgement latencies (ms) obtained from the Biased Judgment-Dimension Task by dysmorphic concern group, judgment required, attractiveness, and identity
Attractiveness Judgment
Gaze Direction Judgment
Self
Other
Self
Other
Attractive
Unattractive
Attractive
Unattractive
Attractive
Unattractive
Attractive
Unattractive
Low concern
1034.85 (224.75)
1078.08 (171.71)
1033.55 (227.72)
1121.92 (182.42)
1017.21 (188.21)
1076.03 (201.27)
1017.95 (177.13)
1041.20 (210.05)
High concern
1114.00 (256.15)
1041.67 (219.94)
1034.47 (206.31)
1105.52 (264.41)
1051.73 (247.37)
1078.22 (247.02)
1043.40 (229.14)
1047.40 (215.07)
Analyses on judgment latencies revealed main effects of Judgment Required (F(1, 61) = 4.30, p = 0.042, η2p = 0.006) and Attractiveness Category (F(1, 61) = 12.912, p < 0.001, η2p = 0.175). These main effects indicated that, overall, gaze-related responses were faster than attractiveness-related responses (t(62) = 2.111, p = 0.039, Cohen’s d = 0.266, 95% CI [1.28, 46.94]), and that attractive photographs were judged more swiftly than unattractive photographs (t(62) = − 3,55, p < 0.001, d = -0.447, 95% CI [− 49.18, − 13.73]).
Group Differences Relevant to GARA
Present findings did not replicate the three-way interaction of Judgment Required × Attractiveness Category × Dysmorphic Concern Group (F(1, 61) = 1.539, p = 0.220, η2p = 0.25), as previously obtained by Dondzilo et al. (2021). Thus, there was no evidence for the GARA bias.
Group Differences Relevant to EPUS
Obtained findings did, however, show a three-way interaction of Identity × Attractiveness Category × Dysmorphic Concern Group, (F(1, 61) = 4.594, p = 0.036, η2p = 0.07). To further understand this interaction, we computed separate two-way interactions of Attractiveness Category × Dysmorphic Concern Group at each level of the Identity factor. This two-way interaction was evident only when participants judged “self” photographs (F(1,61) = 12.489, p < 0.001, η2p = 0.17), but not “other” photographs (F(1, 61) = 0.711, p = 0.402, η2p = 0.12). This former interaction reflected the fact that individuals low in dysmorphic concern were faster to make judgements about attractive vs. unattractive photos of oneself (t(32) = − 4.013, p < 0.01, d = -0.698, 95% CI [− 76.92, -25.12]), whereas this effect was absent in high dysmorphic concern (t(29) = 1.355, p = 0.186, d = 0.274, 95% CI [− 11.67, 57.52]). Thus, this effect pointed in an opposing direction from the EPUS pattern.
While this pattern of results did not evidence the EPUS, a novel bias pattern marking low dysmorphic concern—i.e., the EPAS—was identified, indicating relative speeding to respond to attractive images of oneself.
Attractiveness-Linked Processing Bias Scores
Despite non-replication of the GARA and EPUS, for exploratory reasons, and following Dondzilo et al. (2021), we computed both indices and the EPAS.
As shown in Fig. 3, we only found significant between-group differences on the EPAS measure (t(61) = 3.534, p < 0.01, d = 0.891, 95% CI [32.11, 115.79]), but not the GARA measure (t(61) = 1.240, p = 0.220, d = 0.689, 95% CI [− 24.89; 106.20]), or EPUS index (t(61) = − 2.143, p = 0.036, d = 0.689, 95% CI [− 107.59, − 3.73]; p-values adjusted for multiple testing, i.e., padj = 0.017).
Fig. 3
Mean judgment bias indices per dysmorphic concern group. EPAS Efficient Processing of Attractive Self, EPUS Efficient Processing of Unattractive Self, GARA General Attractiveness Recognition Advantage
×
Predictive Value and Associations of Judgment Biases with Cognitive-Behavioral Features
To investigate whether the EPAS pattern contributed unique variance to model-congruent dependent variables, i.e., dysmorphic symptom severity, baseline repetitive behavior frequency, baseline social media behavior, and intrusion frequency, several regression analyses were conducted (see Tables 3 and 4). To determine whether the EPAS index was associated with the diary measures, we computed correlation analyses using Spearman’s rho (see Table 5).
Table 3
Results from linear regression analyses
Linear regressions
Predictor: EPAS
Dependent variable
B
95% CI
β
T
p
R2adj
p
FKS
− .031
[− .049;− .012]
− .506
− 3.359
.001*
.149
.003
State negative affect (pre-TIIT)
− .115
[− .212;− .018]
− .371
− 2.371
.021
.080
.031
DASS-21
− .020
[− .039;− .001]
− .335
− 2.098
.040
.045
.095
SSES
.031
[.012;.050]
.500
3.314
.002*
.144
.004
ABC (baseline)
− .028
[− .044;− .011]
− .496
− 3.295
.002*
.150
.003
Screentime (baseline)
− .003
[− .006;− .001]
− .396
− 2.511
.015
.066
.048
*< .008 (p-value adjusted for multiple testing)
ABC Appearance Behavior Checklist, DASS-21 Depression Anxiety and Stress Scales-21, FKS Body Dysmorphic Symptoms Inventory, SSES State Self-Esteem Scale, TIIT Task-Irrelevant Intrusions Task
Table 4
Results from negative binomial regression analysis on intrusion frequency on the TIIT
Dependent variable
Predictor: EPAS
B
SE
95% CI
p
Intrusion frequency
− 0.01
.0017
[− .004;.003]
.704
Table 5
Results from correlational analyses (Spearman’s rho) with daily diary measures
ABC
Day 1
ABC
Day 2
ABC
Day 3
Screentime
Day 1
Screentime
Day 2
Screentime
Day 3
EPAS
− .436**
− .441**
− .304*
− .093
− .259*
− .281*
*p < .05, **p < .01
ABC Appearance Behavior Checklist, EPAS Efficient Processing of Attractive Self
Results revealed that the EPAS served as a significant predictor of baseline dysmorphic concern, self-esteem and maladaptive behaviors, while all other effects remained non-significant (see Tables 3 and 4). Correlations between the EPAS and daily diary measures of screentime were small- to medium-sized and significant, apart from screentime on day 1 (see Table 5).
Discussion
Extending prior findings (Dondzilo et al., 2021), the present study examined judgment biases during facial processing in individuals who differed in dysmorphic concern, and investigated their associations with dysmorphic symptom severity, negative affect, maladaptive behaviors, social media investment and intrusions.
We first sought to determine the prior results by Dondzilo et al. (2021) that the EPUS bias is a vulnerability marker, such that higher scores are associated with higher dysmorphic concerns, and that the GARA bias is a protective marker, such that higher scores are associated with lower dysmorphic concern. Results did not demonstrate either of these two effects. Despite lack of replication, a novel bias was revealed, i.e., the EPAS, reflecting an enhanced readiness to process attractive over unattractive representations of self. The EPAS was found to be more pronounced in low, vs. high, dysmorphic concern.
This finding is in line with previous research indicating that enhanced processing of one’s attractive body features serves as a potential protective mechanism against body image concerns (Jansen et al., 2016; Smeets et al., 2011). The fact that individuals with high dysmorphic concern showed reduced evidence of the EPAS bias corroborates previous results indicating a lack of self-serving evaluative bias during photographic assessment (Lambrou et al., 2011) and extensive attentional engagement with unattractive facial features in BDD (Greenberg et al., 2014; Grocholewski, et al., 2012; Kollei et al., 2017). Given the overlap of these patterns, future research might explore the serial interplay of preceding perceptual and attentional biases with judgment biases (as plausible within a “combined cognitive bias” approach; Everaert et al., 2012) through comprehensive assessment using the BJDT.
Failure to demonstrate both the GARA and EPUS may indicate the fragility of these effects, but could also reflect methodological disparities between the present study and that of Dondzilo et al. (2021). Specifically, group allocation in this study was based on general dysmorphic symptoms with primary facial concerns, while group allocation in Dondzilo et al. (2021) was based on facial appearance concern only. Relatedly, a substantial portion of the high dysmorphic concern group (76.7%) reported additional foci in other bodily areas. Thus, while this group might have qualified more clearly as an analog-BDD sample, heterogenous foci of concern might have contributed to the divergence of results. Further replication in diverse samples, using fine-grained quantification of facial appearance concern, is needed to demonstrate the robustness of judgment biases and their relationship with dysmorphic symptoms.
The novel EPAS bias revealed in this study showed systematic relationships with cognitive-behavioral features of dysmorphic concern (Fang & Wilhelm, 2015). Specifically, an enhanced EPAS bias (i.e., greater readiness to process attractive representations of self) was associated with reduced BDD symptom severity, higher self-esteem, reduced appearance-related behaviors and reduced screen time. These results illustrate the potential etiological role of EPAS regarding cognitive-behavioral factors, e.g. self-esteem, which has been identified as a hallmark in maintaining BDD symptoms (Kuck et al., 2021; Schulte et al., 2021). Further, they suggest that EPAS meaningfully relates to maladaptive behaviors during social media investment. Overall, these findings corroborate cognitive-behavioral model predictions (Fang & Wilhelm, 2015; Veale, 2004), wherein the lack of protective judgment biases during facial processing contributes to an exacerbation of BDD symptoms. Consistently, recent evidence suggests that judgment biases may play a potentially causal role in dysmorphic concern (Dondzilo et al., 2021). Specifically, participants induced to make ‘attractive’ judgments (i.e., only attractive faces were judged in terms of their attractiveness) showed reduced appearance-related intrusion frequency. Future researchers are thus encouraged to explore the causal links between previously identified judgement biases (i.e., EPUS, GARA and EPAS) and associated factors, e.g., self-esteem and social media investment.
None of the bias indices were significantly related to intrusion frequency or state negative affect. Of note, his study did not assess negative affect in response to a standardized stressor. In this respect, the photo shoot session might have been too brief to evoke substantial negative affect, and allowed maladaptive behaviors to be performed for emotion regulation. Future studies might utilize established stressor paradigms (e.g., in-lab mirror exposure) to examine associations between judgment biases and negative affect. A further possibility is that the measures used to assess intrusions and negative affect lacked ecological validity. Future studies might consider using ecological momentary assessment to sample intrusions and negative affect across an extended period.
Overall, the results of this study suggest the intriguing possibility that BDD symptomatology may benefit from strengthening potentially protective biases through interventions. Based on the components involved in EPAS, interventions could reinforce attentional engagement with attractive facial stimuli—e.g., through perceptual retraining or Cognitive Bias Modification (CBM)—to induce faster processing of attractive facial stimuli. Thus, future studies could investigate the clinical utility of interventions targeting this processing style.
There are some limitations associated with this study. First, the web-based assessment setting, while reliable for experimental tasks (e.g., Hilbig, 2016), could have limited generalizability of stimuli and outcome measures, e.g., via differences in photo quality and missing data on diary surveys. The latter aspect could have further contributed to low internal consistencies on these scales, compromising their interpretability. Second, this was an unregistered replication study conducted in a subclinical sample, thus further restricting generalizability. Third, inherent features of dysmorphic concern, such as perceived flaws appearing objectively slight or absent, were not validated through clinician-administered interviews, which might have introduced spurious sample heterogeneity. Taken together, these limitations highlight the important need for future, preregistered replication studies in clinical BDD, including clinician-based diagnostic procedures. Third, this study featured an exclusively female sample. As dysmorphic concerns have been found to be equally prevalent across genders, yet potentially different regarding foci of concern (e.g., Veale et al., 2016), future studies should examine judgment biases for facial stimuli in more diverse samples. Fourth, attractiveness ratings of photographs in this study were categorical, not continuous, which precludes conclusions on rating differences in the low vs. high dysmorphic group, and could have influenced judgment latencies. Last, future studies might implement different stimulus duration times to differentiate controlled and automatic features of judgment biases.
In conclusion, this study provides further evidence that biased judgment patterns can explain variance in dysmorphic concern and associated cognitive-behavioral features. Thus, findings suggest the existence of an additional protective judgment bias, EPAS, characterizing low dysmorphic concern. This bias—reflecting enhanced processing of attractive representations of oneself—appears to be meaningfully linked to cognitive-behavioral factors. Overall, strengthening EPAS could prove useful in alleviating appearance-related maladaptive behaviors (e.g., mirror checking) and underlying features (e.g., low self-esteem). Future studies could now evaluate this possibility by developing interventions designed to target judgment biases, such as CBM, and evaluating whether these interventions yield benefits in reducing dysmorphic concern.
Acknowledgements
We would like to thank Tom Augstein and Ivelin Georgiev for their assistance in data collection. We would further like to thank the anonymous reviewers for their constructive comments on former versions of this manuscript.
Declarations
Informed Consent
Informed consent was obtained from all individual participants included in the study.
Animal Rights
This study was approved by the Ethics Committee (Faculty 7) of the University of Muenster.
Conflict of Interest
Fanny Alexandra Dietel, Laura Jacobs, Rebecca Onken, Ulrike Buhlmann, Colin MacLeod and Laura Dondzilo have declared no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Mit e.Med Neurologie & Psychiatrie erhalten Sie Zugang zu CME-Fortbildungen der Fachgebiete, den Premium-Inhalten der dazugehörigen Fachzeitschriften, inklusive einer gedruckten Zeitschrift Ihrer Wahl.
Mit e.Med Psychiatrie erhalten Sie Zugang zu CME-Fortbildungen des Fachgebietes Psychiatrie, den Premium-Inhalten der psychiatrischen Fachzeitschriften, inklusive einer gedruckten Zeitschrift Ihrer Wahl.
Erwachsene, die Medikamente gegen das Aufmerksamkeitsdefizit-Hyperaktivitätssyndrom einnehmen, laufen offenbar erhöhte Gefahr, an Herzschwäche zu erkranken oder einen Schlaganfall zu erleiden. Es scheint eine Dosis-Wirkungs-Beziehung zu bestehen.
Menschen mit Typ-2-Diabetes sind überdurchschnittlich gefährdet, in den nächsten Jahren auch noch eine Depression zu entwickeln – und umgekehrt. Besonders ausgeprägt ist die Wechselbeziehung laut GKV-Daten bei jüngeren Erwachsenen.
In einer Leseranfrage in der Zeitschrift Journal of the American Academy of Dermatology möchte ein anonymer Dermatologe bzw. eine anonyme Dermatologin wissen, ob er oder sie einen Patienten behandeln muss, der eine rassistische Tätowierung trägt.
Sportliche Betätigung hilft nicht nur bei Depression, sondern auch in Gruppen von Patientinnen und Patienten mit unterschiedlichen psychischen Erkrankungen, wie Insomnie, Panikattacken, Agoraphobie und posttraumatischem Belastungssyndrom. Sie alle profitieren längerfristig.
Update Psychiatrie
Bestellen Sie unseren Fach-Newsletterund bleiben Sie gut informiert.
