Imagine you are being gazed at by multiple individuals simultaneously. Is the provoked anxiety a learned social-specific response or related to a pathological disorder known as trypophobia? A previous study revealed that spectral properties of images induced aversive reactions in observers with trypophobia. However, it is not clear whether individual differences such as social anxiety traits are related to the discomfort associated with trypophobic images. To investigate this issue, we conducted two experiments with social anxiety and trypophobia and images of eyes and faces. In Experiment 1, participants completed a social anxiety scale and trypophobia questionnaire before evaluation of the discomfort experienced upon exposure to pictures of eye. The results showed that social anxiety had a significant indirect effect on the discomfort associated with the eye clusters, and that the effect was mediated by trypophobia. Experiment 2 replicated Experiment 1 using images of human face. The results showed that, as in Experiment 1, a significant mediation effect of trypophobia was obtained, although the relationship between social anxiety and the discomfort rating was stronger than in Experiment 1. Our findings suggest that both social anxiety and trypophobia contribute to the induction of discomfort when one is gazed at by many people.
Trypophobia is an irrational fear of holes. Typically, trypophobic images are composed of holes such as the seed head of the lotus flower or a honeycomb. According to the trypophobia website (
Trypophobic images, as currently understood, usually present hole-like patterns. However, in one pioneering study of trypophobia conducted by
Furthermore,
In the present study, we focused on social anxiety. Social anxiety disorder is the fear of a social situation that may involve negative judgment from others. Social anxiety disorder is a relatively common psychiatric disorder with a lifetime prevalence of 10% to 15% (
Considering the above studies, we hypothesized that social anxiety has an influence on the degree of discomfort with trypophobic images composed of clusters of human eyes and faces. The aim of the current study is to investigate three questions. The first question is whether images that contain clusters of human eyes and faces can induce aversion. The second question is whether the TQ shows validity and reliability in other populations, as suggested by the previous study (
The experiment was conducted online using Qualtrics (Provo, UT:
The survey consists of three parts: the social anxiety scale, the Trypophobia scale, and the discomfort rating of eyes. The order of these three scales was randomized across all participants. There was no time limit to complete each scale.
Liebowitz social anxiety scale: We used the English version of the LSAS (
We employed the original English version of the TQ, and the images of a lotus seed head and a honeycomb for all participants from the previous study (
We exported the real eyes stimuli from the database of ATR-Promotions (DB99;
The letter of each label indicates the Pair-condition (single or paired), and the number in each label indicates the Number-of-eyes-condition (1, 4, 16, or 64). The 256 × 256 pixels of the real eye image were created using the 64 × 64 pixels of one eye in the 1- and 16-eye-condition. The 4-eye-condition used the 128 × 128 pixels of the real eye image. The 64-eye condition used the 512 × 512 pixels of the real eye image.
We performed analyses of variance (ANOVAs) to determine whether the number and pair of eyes affected the participants’ DRS. A three-way within-participant ANOVA with language (English and Japanese), pair (single and paired), and number of eyes (4, 16, and 64) as factors was conducted. To discriminate clearly whether the number of eyes or the pair condition affected the DRS, the one- and two-eye conditions were excluded from the analysis to match the total number of eyes between the single and paired conditions.
As shown in
To demonstrate the validity and reliability of the TQ, a confirmatory factor analysis (with promax rotation, Maximum-likelihood method) was individually conducted on the English and Japanese TQ scores. As in a previous study (
Item | Factor1 | Communality |
---|---|---|
Feel anxious, full of dread or fearful | .91 | .83 |
Feel sick or nauseous | .89 | .79 |
Feel like going crazy | .89 | .79 |
Feel like panicking or screaming | .88 | .77 |
Vomit | .87 | .76 |
Have trouble breathing | .86 | .74 |
Have goosebumps | .86 | .74 |
Feel like crying | .83 | .70 |
Have an urge to destroy the holes | .83 | .68 |
Feel itchiness | .82 | .67 |
Chills | .81 | .66 |
Feel freaked out | .80 | .63 |
Feel nervous (e.g., heart pounding, butterflies in stomach, sweating, stomach ache, etc.) | .79 | .63 |
Shiver | .79 | .62 |
Feel aversion, disgust or repulsion | .68 | .47 |
Feel skin crawl | .65 | .42 |
Feel uncomfortable or uneasy | .59 | .35 |
Factor contribution | 11.23 | |
Cumulative contribution | 68.60 | |
0.97 |
Item | Factor1 | Communality |
---|---|---|
Feel anxious, full of dread or fearful | .91 | .82 |
Feel sick or nauseous | .89 | .80 |
Feel nervous (e.g., heart pounding, butterflies in stomach, sweating, stomach ache, etc.) | .89 | .79 |
Feel like panicking or screaming | .88 | .77 |
Chills | .87 | .75 |
Have trouble breathing | .87 | .75 |
Have goosebumps | .86 | .74 |
Feel like going crazy | .86 | .73 |
Feel itchiness | .82 | .68 |
Feel freaked out | .82 | .67 |
Vomit | .81 | .66 |
Feel uncomfortable or uneasy | .78 | .61 |
Shiver | .77 | .59 |
Feel skin crawl | .73 | .53 |
Feel like crying | .72 | .52 |
Feel aversion, disgust or repulsion | .71 | .51 |
Have an urge to destroy the holes | .50 | .25 |
Factor contribution | 11.16 | |
Cumulative contribution | 67.58 | |
0.97 |
We calculated the DRS by summing the scores from each pair condition (single and paired) across all participants (i.e., data from both language groups were analyzed altogether). The correlations among the DRS (single and paired), the TQ score, and the LSAS score are shown in
1 | 2 | 3 | 4 | |
---|---|---|---|---|
1. DRS for single eye | – | |||
2. DRS for paired eyes | .90 |
– | ||
3. LSAS score | .15 |
.19 |
– | |
4. TQ score | .39 |
.44 |
.27 |
– |
To investigate the direct influence of the LSAS score on the DRS, we conducted a mediation analysis. We set the LSAS and TQ scores as predictors of the two conditions (single & paired) used in the correlation analysis. The mediation model is shown in
DRS | Path | df | |||
---|---|---|---|---|---|
Single eye | a | .27 | .0933 | 206 | 4.09 |
b | .38 | .0050 | 205 | 5.63 |
|
c | .15 | .0073 | 206 | 2.13 |
|
c′ | .04 | .0071 | 205 | 0.66 | |
Paired eyes | a | .27 | .0933 | 206 | 4.09 |
b | .42 | .0048 | 205 | 6.40 |
|
c | .19 | .0070 | 206 | 2.80 |
|
c′ | .08 | .0067 | 205 | 1.20 |
In Experiment 1, we examined the relationship between social anxiety and the discomforting symptoms induced by images of eye clusters. In line with a previous study (
The results from the factor analysis also supported the previous study (
We calculated the correlations among the DRS, the LSAS score, and the TQ score, from which we found a positive correlation (
The mediation analysis showed a significant mediation effect when the TQ score was a mediator; however, the path from the LSAS score to the DRS was not significant. These results were contrary to our hypothesis that there is positive correlation between social anxiety and the DRS concerning trypophobic images composed of eye clusters.
One of the reasons the significant path between social anxiety and discomfort was not observed may be attributed to the stimulus property. In the previous studies that investigated the relationship between social anxiety and gaze perception, stimulus eyes were usually embedded in a face (
In this experiment, a new sample of 499 English speakers was recruited from Qualtrics panels (159 females and 340 males; mean age = 34.6 yrs., SD = 9.8 yrs.). The purpose of the present study was not revealed to the participants.
The procedure was identical to Experiment 1, except that the rating of eye clusters was replaced by the rating of face clusters. As facial stimuli, we selected 32 neutral faces (half males, half females; AF01NES ∼AF16NES, AM01NES ∼AM11NES, AM13NES ∼AM15NES, AM17NES, AM18NES) from the Karolinska Directed Emotion Face Set (
In the rating of face clusters, the number of eyes was the same as in Experiment 1 (i.e., 4, 16 and 64;
(A) 2-face condition, (B) 8-face condition, and (C) 32-face condition. Each face image was scaled to an elliptical shape, 57 × 75 pixels. Realistic faces were used as stimuli in the actual experiment.
As shown in
A confirmatory factor analysis (with promax rotation, Maximum-likelihood method) was conducted on the TQ scores. As shown in
Item | Factor 1 | Communality |
---|---|---|
Feel sick or nauseous | .92 | .85 |
Feel like panicking or screaming | .92 | .84 |
Have trouble breathing | .92 | .84 |
Feel nervous (e.g., heart pounding, butterflies in stomach, sweating, stomach ache, etc.) | .91 | .83 |
Feel like going crazy | .91 | .82 |
Feel anxious, full of dread or fearful | .90 | .82 |
Feel freaked out | .90 | .81 |
Chills | .90 | .81 |
Have goosebumps | .90 | .80 |
Shiver | .87 | .76 |
Feel itchiness | .87 | .76 |
Feel like crying | .87 | .75 |
Have an urge to destroy the holes | .86 | .73 |
Vomit | .85 | .73 |
Feel uncomfortable or uneasy | .80 | .65 |
Feel aversion, disgust or repulsion | .77 | .60 |
Feel skin crawl | .77 | .59 |
Factor contribution | 12.99 | |
Cumulative contribution | 0.76 | |
0.98 |
We calculated the DRS by summing the scores of each of the faces for each participant.
1 | 2 | 3 | |
---|---|---|---|
1. DRS for faces | – | ||
2. LSAS score | .36 |
– | |
3. TQ score | .56 |
.32 |
– |
DRS | Path | df | |||
---|---|---|---|---|---|
Faces | a | .32 | .073 | 497 | 7.58 |
b | .42 | .008 | 496 | 10.47 |
|
c | .36 | .015 | 497 | 8.64 |
|
c′ | .23 | .014 | 496 | 5.66 |
In the mediation analysis, we entered the LSAS and TQ scores as predictors of the DRS of faces. The mediation model was set as in Experiment 1, and the results are shown in
We examined whether social anxiety directly linked to aversion to trypophobic images composed of faces in Experiment 2. The results revealed that the cluster of faces induced aversion and that the DRS increased when the number of faces increased; the validity and reliability of TQ was confirmed, as in Experiment 1. These results were in agreement with Experiment 1 and supported previous findings (
In addition, the correlation analysis showed that the correlation coefficient between the LSAS score and the DRS was not higher than that between the TQ score and the DRS. The results suggested that even when the eyes were embedded in a face, social anxiety was not strongly related to discomfort for facial clusters, although the correlation was significant.
More importantly, not only the indirect effect but also the direct effect from the LSAS score to the DRS was significant in Experiment 2 when the TQ score was set as a mediator. This significant direct effect after mediation was not shown in Experiment 1, where we used only the cropped eye images. Based on these results, it is suggested that the perception of a face strengthens the induction of discomfort to gaze in individuals with social anxiety disorder and that trypophobia mediates this relation.
The present study aimed to investigate whether social anxiety is related to discomfort induced by clusters of eyes (Experiment 1) and faces (Experiment 2). The results suggested that both eyes and faces induced discomfort as the number of images increased. Additionally, this effect was strongly related to the social anxiety trait and was mediated by the trypophobia trait. Moreover, the social anxiety trait more strongly predicted discomfort for the clusters of eyes when the eyes appeared in a face.
Although our hypothesis that social anxiety is directly linked to discomfort due to the trypophobic images of human face clusters was supported, we did not expect the significant mediation effect in both experiments. It does not necessarily mean that only the TQ score mediated the path between the LSAS score and the DRS because general anxiety may have an influence on trypophobia. Nevertheless, we cannot completely exclude the possibility that the relationship between the LSAS score and DRS was mediated by the TQ score. For example,
A limitation of the current study lies in the fact that the differences between individuals and cultures were not examined. For example, in Experiment 1, the results of the ANOVA showed that Japanese speakers feel more discomfort than English speakers. However, the question of whether the difference in the DRS between Japanese and English speakers is caused by the aversion to trypophobic images or the sensitivity to gaze has not yet been determined. If the cultural difference in DRS was specific to the trypophobic images of eye clusters, one can argue that the difference came from cultural or ethnic differences in the sensitivity to gaze. This hypothesis predicts that Japanese speakers would show stronger social anxiety than English speakers, because individuals with social anxiety tend to attend to gaze (
In conclusion, the current study revealed that not only clusters of holes but also clusters of other objects such as eyes and faces induce aversive reactions. The validity and reliability of the TQ and the fact that social anxiety directly involved the aversion to trypophobic images of faces were also established. Further studies are required to explore other factors that influence trypophobia. Such studies could contribute to understanding why some individuals develop an aversion to trypophobic images but others do not.
The file consists of three sheets. First sheet shows Liebowitz social anxiety scale scores. Second has Trypophobia questionnaire scores, and Third sheet contains discomfort rating scores about trypophobic eye-images.
The file consists of three sheets. First sheet shows Liebowitz social anxiety scale scores. Second has Trypophobia questionnaire scores, and Third sheet contains discomfort rating scores about trypophobic eye-images.
The authors would like to thank Drs. Keiko Ihaya and Kyoshiro Sasaki for their solid supports for developing the Japanese version of the Trypophobia Questionnaire.
The authors declare there are no competing interests.
The following information was supplied relating to ethical approvals (i.e., approving body and any reference numbers):
The ethical committees of Kyushu University approved the protocol (approval number: 2013-008).
The following information was supplied regarding data availability:
The raw data has been supplied as