What is the difference between fear and disgust

Researchers led by Stella Lourenco, a psychologist at Emory University in Atlanta, GA, decided to dig deeper into trypophobia and asked why it might occur. Specifically, the team wanted to get to grips with the physiological and psychological drivers of this rather odd — and currently unofficial — phobia. Their results are published this week in the journal PeerJ.

Both fear and disgust impart an evolutionary advantage — fear helps us to avoid peckish predators, while disgust steers us away from eating perished plums. It is now established that the physiological responses are different: fear activates the sympathetic nervous system, and disgust triggers the parasympathetic nervous system.

The sympathetic nervous system prepares the body for threat or injury by increasing heart rate and contracting muscles. The parasympathetic nervous system controls general body functions at rest, making muscles relax and heart rate decrease.

The first question to ask is why groups of holes and irregular repeating patterns are frightening to our primal, caveman brains. Some psychologists believe that the high contrast seen in trypophobia-inducing images is similar to patterns found on some dangerous animals, such as snakes.

It has been argued that this similarity could be the driving force behind the negative response. Low-level visual properties can convey a lot of meaningful information. These visual cues allow us to make immediate inferences — whether we see part of a snake in the grass or a whole snake — and react quickly to potential danger.

The study was designed to identify whether a trypophobic reaction is triggered by the sympathetic or parasympathetic nervous system. The team wanted to know whether this odd reaction is based on disgust or fear. Pupillometry — which is an eye-tracking technique that measures pupil size and reactivity — let the scientists glimpse the physiology behind the emotion.

Earlier work had shown that a fear response induces an increase in pupil size while, conversely, disgust causes pupil size to decrease. If, however, trypophobia is a disgust-based response, the pupils would behave differently between the two experimental image types. After analysis, it was clear that both the images of dangerous animals and trypophobic patterns triggered a response.

However, they were not the same: pictures of snakes and spiders caused an increase in pupil size, whereas images of holes caused the pupils to constrict. Our findings, however, suggest that the physiological underpinnings for these reactions are different, even though the general aversion may be rooted in shared visual-spectral properties.

Specifically, van Gaal et al. In the weakly masked condition, a go or no-go prime was presented for ms, followed by a metacontrast masking annulus presented for 17 ms. In the strongly masked condition, the durations of the go or no-go prime and annulus were 17 ms and ms, respectively. Participants are instructed to respond as fast as possible to the annulus response signal but withhold their response when a no-go signal briefly precedes the annulus. However, when a go signal precedes the annulus, they are instructed to respond as quickly as possible [ 12 ].

While in the strongly masked condition, participants would make a go response to the strongly masked no-go trials because the no-go signal could not be perceived consciously therefore the RT for unconscious no-go trials can be directly measured. Furthermore, the RT slowing i. These results confirmed the existence of unconscious inhibitory control. Several studies have investigated the relationship between unconsciously and consciously triggered inhibitory control and have indicated that they are strongly related [ 12 , 15 ].

On one hand, unconscious and conscious inhibitory control can similarly activate prefrontal control networks [ 12 , 16 ]. On the other hand, unconscious and conscious inhibitory controls differ in the strength, duration, and scope of neural activity [ 15 ]. Recent studies on conflict adaptation [ 17 ] and linguistic operation [ 18 ] have demonstrated that unconscious and conscious executive control are related.

For example, Desender et al. They observed conflict adaptation in both unconscious and conscious conditions, but found that the conflict adaptation effect was smaller in the unconscious condition than it was in the conscious condition. Therefore, it appears that while unconscious and conscious inhibitory control differs in the degree of information processing, they might be similar in many aspects.

Therefore, although previous work has investigated the divergent effects of fear and disgust on perception, attention, and memory, little is known about the differential effects of fear and disgust on inhibitory control, not to mention whether these effects are similar for conscious and unconscious inhibition.

As inhibitory control processes are important in daily life e. Due to their high temporal resolution, event-related potentials ERPs are particularly well suited to studying emotion-modulated response inhibition [ 20 ]. In affective science, researchers have pointed out that P2, a positive peaking potential between and ms which is typically located over the centro-parietal and parieto-occipital regions [ 21 ], shows significant amplitude increments when a negative stimulus automatically attracts attention in a wide variety of tasks [ 9 , 22 ].

These results indicated that disgusting distracters were more efficient at attracting attention. The difference waves of P3 amplitude no-go condition minus go condition can be used as an index of inhibiting ability, with larger P3 difference waves representing a stronger ability to suppress the prepotent response [ 24 ]. In sum, the current study aimed to investigate whether fear and disgust exert different effects on inhibitory control, and whether these effects are similar for conscious and unconscious inhibitory control.

ERPs were measured concurrently. Building on previous studies, we hypothesized that fear and disgust has distinct effects on attention and inhibitory control. On one hand, previous studies have shown that disgusting stimuli attracted more attention than fearful ones [ 9 ].

On the other hand, cognitive resources are limited [ 25 ]. Therefore, when limited attention is diverted to aid in the processing of disgusting stimuli, the subsequent inhibitory control might be impaired [ 25 , 26 ].

Specifically, at the behavioural level, given that the RT slowing was used as an index of inhibiting ability, we hypothesized that the RT slowing under disgusting contexts would be smaller than that under fearful contexts. At the neural level, given that P2 acts as an index of attention allocation [ 9 ], we hypothesized that disgusting distracters would elicit larger P2 than fearful ones would; and given that the difference wave of P3 is an index of inhibiting ability [ 12 , 16 ], we hypothesized that this difference would be smaller under disgusting contexts compared to that under fearful contexts.

Furthermore, as some researchers have indicated that conscious and unconscious inhibitory control are similar in many aspects [ 12 , 15 ], we hypothesized that the effects of fear and disgust on conscious and unconscious inhibition might be similar. The ethics committee of Southwest University of China approved this experiment. Written informed consent was obtained from all participants in compliance with the principles of the Declaration of Helsinki. All participants were right-handed with normal or corrected-to-normal vision.

Upon completion of the task, they received 30 RMB for their participation. We selected only women for this experiment because previous studies have shown that women display greater vigilance to emotional stimuli [ 27 ]; therefore, using an all-female sample removed gender as a confounding variable.

The stimuli consisted of three geometric figures and 30 emotional images. The geometric figures consisted of an annulus visual angle of 0. Therefore, they were clearly differentiated from the background images on which they were superimposed [ 9 ]. Thirty emotional pictures were selected to generate different background contexts 10 fearful, 10 disgusting, and 10 neutral.

The size of all the pictures was 7. Therefore, we chose fearful pictures that represented more fearfulness than did disgusting and neutral pictures and disgusting pictures that represented more disgustingness than did fearful and neutral pictures, while keeping their valence and arousal ratings constant.

Furthermore, since the complexity of images would affect early visual ERPs [ 29 ], only simple figure-ground images these images have a relatively clear figure-ground composition, just like the clock in Fig 1 were chosen to remove the image complexity as a confounding variable.

All stimuli were displayed on a Dell computer with a inch monitor Hz refresh rate. E-prime software Psychology Software Tools, Inc. Subjects were placed in an electrically shielded, soundproofed room.

Each trial began with a crosshair presented in the centre of the screen ms , followed by the emotional image centrally displayed for ms. Finally, a blank screen was presented for ms. In the strongly masked condition, the durations of the go or no-go prime and the metacontrast masking annulus were 17 ms and ms, respectively.

We used an annulus as a metacontrast mask because it strongly reduces stimulus visibility [ 30 ]. The stimulus parameters were akin to those reported by van Gaal et al. There were five blocks and each contained trials, for a total of trials trials for each condition. The stimulus used as the no-go signal square or diamond was counterbalanced across subjects.

Before the experiment began, participants completed a practice block of 40 trials with 10 neutral images as the background; this was done to ensure that they understood the task instructions. These neutral images were different from those used in the formal experiment. The prerequisite of our study was the effectiveness of our masking procedure. Therefore, to assess whether participants were truly unaware of the strongly masked prime, an alternative forced-choice discrimination task was added, which included 80 strongly masked trials after the formal task.

They were informed that response time was not important and were asked to respond as accurately as possible. Brain electrical activity was recorded at 64 scalp sites using tin electrodes mounted in an elastic cap Brain Product, Munich, Germany , with references on the left and right mastoids, and a ground electrode on the medial frontal aspect.

The vertical electrooculograms EOGs were recorded supra- and infra-orbitally at the right eye. The horizontal EOG was recorded from the left versus the right orbital rim. Only correct responses between and ms were analysed.

The data were referenced to the average of the left and right mastoids average mastoid reference , and a bandpass filter of 0. Eye movement artefacts such as eye movements and blinking were excluded offline.

Only trials with correct responses were analysed. This ROI was selected on the basis of previous studies [ 21 , 23 , 31 ]. In line with previous research, mean amplitudes of specific ERP deflections were measured for different time intervals.

The time windows of the P2 and P3 components of obtained average waveforms were established based on the grand averaged potentials of each task condition. Consequently, for the strongly masked condition, the interval of the P2 component was — ms, and the interval of the P3 was — ms; for the weakly masked condition, the interval of the P2 was — ms, and the interval of the P3 was — ms Figs 2A and 3A. A The averaged ERP under different emotional contexts; B P2 amplitude under different emotional contexts; C P3 amplitude under different emotional contexts; D The difference waves no-go condition minus go condition of P3 amplitude under different emotional contexts.

Then, the difference waves of P3 no-go condition minus go condition under fearful, disgusting, and neutral emotional contexts were analysed using a one-way ANOVA. The latency shifts appear to occur between the fear and disgust conditions. Also, there appear to be large differences in amplitude and latency of the negative peaks N2 that separate the two positive components. Therefore, the emotional effects on P2 latency, N2 latency and P3 latency were analysed.

Besides, the correlations of P2, N2, P3 amplitude and latency under different emotional contexts were computed separately. These results were presented and reported in the S1 Appendix. Therefore, the prime could not be perceived in the strongly masked condition, confirming the effectiveness of the masking procedure. No other significant difference was observed. No other significant difference was observed Table 1.

These results are partly consistent with our hypothesis and suggest that disgusting stimuli attract and consume more attentional resources than do fearful and neutral stimuli.

This suggests the presence of unconscious inhibitory control in this study Fig 2C. These results are consistent with our hypothesis and demonstrate that disgusting stimuli, compared to fearful and neutral ones, impaired unconscious inhibitory control.

Finally, the correlation analysis between the P2 and the behavioral indices of unconscious inhibitory control RT slowing and the accuracy of no-go trials was also made to test whether unconscious inhibitory control was related directly to the extent of attentional resources available for processing. These results supported our hypothesis that disgusting distracters would consume more attentional resources and impair unconscious inhibitory control. These results suggest that both disgusting and fearful distracters may impair conscious inhibitory control.

These results are consistent with our hypothesis and demonstrate that disgusting stimuli impaired conscious inhibitory control to a greater extent than did fearful stimuli.

Consistent with our hypothesis, results showed that disgusting stimuli elicited a larger P2, and the difference waves of P3 amplitude under disgusting contexts were smaller than were those under fearful contexts for both conscious and unconscious inhibitory control. For instance, it has been shown that snake-fearful respondents give more negative and extreme scores to snake stimuli when rating valence and arousal Miltner et al.

People with high fear of snakes also show increased cognitive interference in the Stroop test when confronted with snake-related sentences Constantine et al. Furthermore, high-fear individuals demonstrate higher skin conductance response SCR when confronted with a live snake McGlynn et al. Unconscious presentation of snakes within watched video stimuli attracts attention in form of eye saccades directed toward the areas where the snakes were presented and this effect is again more pronounced in snake-fearful participants Rosa et al.

Flykt et al. Moreover, the neural response of snake-fearful respondents to snake movies is higher or qualitatively different: according to Lueken et al.

Even their brain morphology differs as the gray matter volume in the left postcentral gyrus is increased when compared to control participants Hilbert et al. In short, high-fear participants change many aspects of their behavior when confronted with the feared stimuli, regardless of whether these are presented as live specimens or just moving or still pictures. However, there are many snake species, differing in size, color, shape, texture, and also the actual dangerousness they present to humans Kasturiratne et al.

These findings raise a further question whether snake-fearful participants distinguish particular snake morphotypes and respond comparably to non-fearful respondents, or evaluate all snakes in general negatively. The stimuli were carefully standardized, reduced to differences between specific snake morphotypes but uniform in other aspects such as the background or posture. Such approach presents a great advantage because it offers well-described and characterized stimuli, free from uncertainties about the effects of other factors such as the body size, environment, background color, or lightness on the rankings given by human respondents.

Still, we found that even with this reduced variability, there was a great distinction between the stimuli types as the respondents clearly distinguished and categorized each stimulus into its respective category. However, the snakes were examined using a rank-ordering method, which is optimal for analyzing differences between the stimuli but reduces differences between the raters — the main focus of the present study.

Here we examine the relationship of fear and disgust ratings using the absolute scale Likert-type scores , focused on differences between respondents with high and low fear of snakes. More specifically, we aimed to test the following predictions corresponding to alternative hypotheses pertaining to the effect of snake fear as measured by the Snake Questionnaire SNAQ :.

This would mean that high snake fear and consequently the SNAQ score is strictly saturated by the fear emotion with no disgust component involved. In this case, anxiety provoked by snakes as measured by the SNAQ would in fact result from increased disgust propensity and sensitivity which would corroborate the findings of Klieger and Siejak who argued that some SNAQ items are ambiguous and may tap into disgust. Such results would, in accordance with the study on spider fear Vernon and Berenbaum, , suggest that high fear of snakes is composed of negative evaluation in general i.

Additionally, we examined the effect of high disgust propensity [as measured by the Disgust Scale-Revised DS-R ] following the same pattern. All of the above-mentioned predictions would also mean that snake-fearful subjects do not treat various snake morphotypes as distinct categories.

Should the contrary be the case, the intact ability to categorize the snakes would be predicted by the following possible outcomes:. The sets contained snakes standardized for size and placed on a blank white background. In the present study, both F and D snakes were mixed into one picture set and presented to the respondents. A total of respondents women, 51 men, aged 18—65; mean age The instructions were to first score each stimulus randomly presented on a seven-point scale according to elicited fear.

Half of the respondents received the task in a counter-balanced order, i. The Likert-type scale, which helps to acquire absolute scores for each stimulus, is a very sensitive method when considering differences among respondents.

In contrast, the rank-ordering method, in which the respondents sort all of the stimuli in an ascending or descending order according to a specific dimension e. By choosing the upper quartile, we could balance between an individual fear level significant enough to discover its potential effect and a statistically sufficient number of subjects within the high-fear category.

Table 1. Descriptive statistics of the study sample. In order to quantify and test congruence in species ranking provided by different respondents, we adopted a two-way, consistency, average-measures intra-class correlation ICC; McGraw and Wong, ; Hallgren, computed in R irr package.

Prior to the analyses, the raw order-ranks were transformed as follows: each value minus 1 was divided by the number of evaluated species minus 1 and square-root arcsin transformed to achieve a normal distribution. A principal component analysis PCA was performed to visualize the multivariate structure of the data sets. Friedman test and Mann—Whitney U -test were used as a non-parametric alternative for variables deviating from normality raw sores. Effect sizes for the Mann—Whitney U -tests were computed as normal approximation z to r Pallant, ; Field, Pairwise comparisons of the means were done using the post hoc Nemenyi multiple comparison test.

Contribution of the explanatory variables constrains to the scorings and rankings of the snakes was examined using the redundancy analysis RDA as implemented in the R package vegan Oksanen et al. RDA is a multivariate direct gradient method.

It extracts and summarizes the variation in a set of response variables subjective evaluation of fear and disgust evoked by snakes that can be explained by a set of explanatory variables.

Statistical significance of the gradients was confirmed by permutation tests. Principal component analyses of the fear scores generated 80 axes, 12 of which were of an eigenvalue higher than 1. The most variability was explained by the first two axes: PC1 explained The second axis clearly separated the stimuli into the two groups of fear-evoking and disgust-evoking snakes.

Very similar results were found when analyzing the disgust scores: 80 PC axes, the eigenvalues of 13 of which were higher than 1; PC explained Again, the PC2 axes separated the stimuli into the two groups. Figure 1.

Results of PC analyses of the A fear scores and B disgust scores of snake stimuli. The colored triangles refer to the pictures of fear-eliciting snakes red and disgust-eliciting snakes green. In both cases, PC2 axis contributed to the separation of the snakes into their respective categories. Results revealed considerable congruence among the respondents in fear scores. These results indicate that there was a high degree of agreement within the group of respondents and suggest that the snake stimuli were rated similarly in terms of evoked fear.

In the case of disgust scores, biologists scored the majority of fear-evoking snakes 30 out of 40 as less disgusting than did the non-biologists, and also three disgust-evoking snakes ware rated as less disgusting.

Additionally, the respondents who first evaluated the stimuli according to fear scored 26 of the disgusting snakes as less fear-evoking and three of the fear-evoking snakes as more fear-evoking.

For more detailed statistics including effect sizes computed as normal approximation z to r Pallant, ; Field, , see Supplementary Material 1. In the case of disgust scores, no snake was significant. We have also performed the same analysis for women only, but this approach yielded comparable results for more details, please see Supplementary Material 2. It is possible that the effect of gender was not significant because the gender ratio in our sample was very unbalanced 51 men, women.

Because of that and to control for the effect of SNAQ, which was the strongest predictor, see also the RDA analyses below , we randomly selected 51 women from the sample with the corresponding SNAQ scores, pooled them together, and re-analyzed the data. However, a univariate analysis of the disgust scores revealed that the effect of gender was significant in neither case snake , and these results were confirmed by Mann—Whitney U -tests Bonferroni corrected.

This suggests that no strong effect of certain species contributes to the results, but rather that it is constructed by a combination of a number of small effects.

This could be, however, also an artifact of the statistical method. A redundancy analysis confirmed the results of the regressions.

We utilized the automatic model-building feature based on both Akaike criterion but with permutation tests and on permutation P -values. In the case of fear scores, both methods agreed on the inclusion of the following variables into the reduced model: SNAQ scores, age, education, and order of the task. The reduced model has generated four constrained axes that explained Therefore, we have also tried to recalculate the analysis using scores on three individual DS-R subscales known as core, animal reminder, and contamination-based disgust instead of DS-R total scores.

Interestingly, this model better explained the full variability than the one using DS-R total scores For more details, see Table 2 and Figures 2A,B. Figure 2. Redundancy analysis RDA of the respondents characteristics determining their ratings [scores A,B and rank-orderings C,D ] of fear and disgust elicited by snake stimuli. However, in the case of Likert-type scoring A,B , the effect was much higher than in the case of the rank-ordering, because the latter method uses relative ranks and minimizes variability among the respondents.

It is thus more suitable for analyses of variability among the stimuli. Next, we analyzed the effect of dimension i. For each respondent within the high-fear and low-fear category, we computed mean fear scores fear and mean disgust scores disgust separately for disgust-evoking snakes D and fear-evoking snakes F; these variables are further referred to as fear-D, fear-F, disgust-D, and disgust-F.

For a graphical summary, see Figure 3A. Figure 3. Comparison of snake ratings of high-fear and low-fear respondents. A post hoc Nemenyi test revealed that within the low-fear subjects, out of six comparisons, only one was not significant disgust-F vs fear-D, the capital letter marks stimulus category. In the case of the high-fear subjects, disgust-F vs fear-D and disgust-D vs disgust-F were not significant.

Principal component analyses of the fear and disgust ranks were also very similar to each other: each analysis generated 79 axes, none of which had the eigenvalue higher than 1. The first and second axes explained A corresponding regression analysis of the effect of age, gender, SNAQ and DS-R scores, education, and order of the task on the disgust ranks revealed no significant effect of any of these factors.

A redundancy analysis of the fear and disgust rankings did not confirm the regressions. Only the SNAQ scores significantly explained the rankings, but the effect was very small: the constrained axes explained 1.

In the case of disgust dimension, the results were similar but opposite. Whenever the respondent misplaced a snake outside of its place category , we counted this as a miscount. The total number of miscounts was collected for each respondent and further analyzed. In the subsequent glm analysis, we examined the effect of dimension, SNAQ score, and their interaction to the number of miscounts quasipoisson model.

And although such results suggested that the fear-eliciting snakes do not elicit any disgust and the disgust-eliciting snakes do not elicit any fear, one could not be entirely sure as the evaluation was done using a relative scale. In this study, we confirmed that this was true for low-fear subjects by asking the participants to score the same set of mixed F-D snake stimuli on an absolute scale.

The results showed that the fear-eliciting snakes received significantly much higher scores of fear and lower scores of disgust than the disgust-eliciting snakes, for which the opposite was true.

Moreover, the fear scores of the D-snakes and disgust scores of the F-snakes did not significantly differ from each other — both were very low and indicated that the F-snakes elicited no disgust and D-snakes elicited no fear. In comparison, the same scores given by the high-fear respondents also did not differ significantly from each other, but both were significantly higher than those of the low-fear respondents.

In other words, high-fear respondents find snakes that usually i. These results suggest that both fear and disgust propensity are involved in high snake fear and possibly phobia. Moreover, disgust elicited by the F and D snakes did not significantly differ, which points out that the high-fear respondents do not distinguish between the two categories of snake types when considering their emotional effect.

However, they still distinguish the F snakes as significantly more fear-eliciting than the D-snakes. Analysis of the number of miscounts from the rank-ordering data also confirmed that the high-fear respondents partially lose the ability to distinguish between the two snake categories as they misplaced the snakes from one category into the other category significantly more often than the low-fear subjects.

Klieger and Siejak argue that the SNAQ is not a good measurement of snake fear because it is strongly biased by false positives. These authors found out that respondents undergoing a behavioral approach test BAT with a live snake often facially expressed disgust, and performed a study examining the relationship between SNAQ scores and disgust determined using various measurements. They showed that many respondents with high SNAQ scores actually approached the live snake during the BAT with no avoidance and their results suggested that it was either because the SNAQ was affected by disgust of snakes or because fear and disgust might be inseparably connected in this case.

In our study, we asked the respondents to rate both fear and disgust of snakes and compared these data with their SNAQ and DS-R scores. The DS-R scores seem to be a good measurement of disgust propensity because it affected only the scores of disgust of snakes. Moreover, our results show that when scoring and rank-ordering the snake pictures according to perceived disgust, the high-fear high-SNAQ respondents are not able to distinguish between the particular snake morphotypes.

This may be due to the fact that high-fear respondents feel strong disgust not only from the disgust-eliciting snakes, but also from the fear-eliciting ones vipers , otherwise rated as not-disgusting at all by low-fear respondents. In comparison, low-SNAQ respondents only rated the viper as fear-eliciting. Another explanation of this phenomenon is that the high-fear respondents cannot distinguish the emotions and only evaluate the snakes according to negative valence Barrett, ; Barrett and Wager, It may be possible that simply seeing the snake stimuli made the high-fear respondents feel miserable cf.

Each method has its advantages and disadvantages, and should be thus used in purposely designed experiments. The absolute Likert-type scale is sensitive to the differences between respondents and is better to be used in experiments in which differences between two groups of respondents, e. However, when a study uses a block of similar stimuli that is treated and measured as a single condition e. Failing to do so may lead to high noise, skewed results, or even a wrong interpretation.

However, these effects were not strong enough to survive a different type of analysis: the RDA only revealed the effect of SNAQ in both cases, and it was very small 1.