Introduction

Individuals with Autism Spectrum Disorder (ASD) are characterized by difficulties in social communication skills and dysregulation of emotions throughout their life span1,2. According to Happe and Frith3, these deficits are related to unique social cognition processing, often described as a poor understanding of Theory of Mind (ToM)—the ability to attribute mental states, such as beliefs, intentions, and emotions—to explain human behavior4. ToM is a multi-dimensional concept encompassing processes based on cognitive or affective assessments. Cognitive ToM is the ability to make inferences about other people's thoughts, intentions, and beliefs5,6. Affective ToM is the process that infers the emotions and feelings of the other7,8. These two processes, comprising ToM, involve overlapping yet dissociable brain networks9,10. Most studies suggest that both cognitive and affective ToM are impaired in ASD11,12,13. Yet other evidence shows that ASD affects only cognitive ToM, leaving affective ToM intact in ASD14,15. Still others claim that ASD individuals normally process some aspects of cognitive ToM, such as moral violations16.

In line with this, the difficulties in cognitive and affective ToM that exist in ASD might impact not only the ability to refer to others' mental states but also influence their self-awareness and perspective-taking abilities regarding their unique characteristics. For example, Johnson et al.,17 found that older children and adolescents with ASD presented limited self-awareness of their expression of autistic and empathic traits, self-reporting themselves as having fewer autistic features and more empathic traits than their parents reported. In contrast, individuals with high-functioning ASD (HF-ASD), namely individuals who have ASD in the absence of intellectual disability18, develop more awareness of their distinctive traits and challenges during young adulthood while consolidating their self-identity19. This psychological process is perhaps enhanced by their closer engagement with typical peers during college or daily work. Yet, at the same time, this process might harm these young adults' self-confidence, creating greater vulnerability to social isolation experiences. Indeed, according to interviews, students with ASD feel inferior in their academic abilities, social behavior, and mental health compared to their typical peers20,21. In addition, they are apprehensive about disclosing their diagnosis to their environment, fearing that it would set them apart from typical peers22. In line with these reports, research evidence presents ineffective attempts of students with ASD to socialize with peers, including being intimidated by bullying and mockery, due to their behavioral and social unique characteristics23, making successful social integration challenging to achieve24,25.

One possible consequence of experiencing unsuccessful social interaction is developing shame26. Shame is an intense negative emotional response triggered by self-behavior inconsistent with social or moral norms27. Its appearance requires both introspecting on different aspects of the self and a complex appraisal of how others evaluate one's behavior28,29. The limited ability of individuals with ASD to examine their emotional world may also affect how they perceive and recognize shame30,31,32. For example, Heerey et al.33 found that children with ASD are impaired, relative to typically developing children, in recognizing shame while watching social situations. Furthermore, according to Gaziel-Guttman et al.34, individuals with ASD rated imaginary social shame situations with lower shame scores than typical peers.

However, while these findings point to the reduced ability of individuals with ASD to identify social shame situations compared to typical peers, other studies show high shame proneness among individuals with ASD. This inclination is more apparent during young adulthood while trying to socialize with typical peers. For example, Wu et al.35 showed that young adults with high ASD traits feared being ridiculed and ashamed in social situations. In addition, Cage and Howes36 discovered that students with ASD described their social experiences at university as accompanied by long-lasting feelings of shame. Furthermore, Green et al.37 claimed that women with ASD are more likely to feel shame due to their social or behavioral difficulties manifested in their inability to manage close relationships. Consistent with these findings, young women with ASD presented internalization of negative statements regarding their diagnosis, resulting in experiencing feelings of guilt and shame38.

Although seemingly at odds, the conflicting findings may be reconciled by distinguishing between perception of shame and experience of shame; While individuals with ASD perceive shame-evoking situations less accurately than their typical peers, on the one hand, they feel shameful, sometimes even more than non-ASD individuals, when the inappropriateness of their social behavior is apparent to them, on the other hand. This distinction is partly analogous to the one made in ToM in which cognitive ToM is impaired in ASD while affective ToM is not14,15.

One possible way of coping with negative emotions such as shame is using emotional regulation strategies. Emotion regulation is a mental skill defined by extrinsic and intrinsic processes responsible for monitoring, evaluating, and modifying predominantly adverse emotional reactions to accomplish one's goals39. According to prior literature, individuals with ASD tend to feel more negative emotions, such as fear, anger, and depression, and have difficulties regulating them 2,40,41. These problems may be due to the limited ability to label and assess emotions, essential for successful emotion regulation32,42,43. One emotion regulation strategy is reappraisal, which involves reinterpreting an emotion-inducing situation by altering its meaning in less threatening ways44,45 and, by that, improving emotional and physiological outcomes46,47,48. Thus, the adoption of reappraisal strategies may aid individuals with ASD in producing a balanced interpretation of shameful situations that they encounter.

The prevalence of reappraisal in ASD is currently unclear. Samson et al.2 found that adolescents with ASD reported lower levels of reappraisal self-efficacy than typical development participants while reading ambiguous social scenarios designed to elicit negative emotional reactions. This finding was explained by their lower ability to change perspective-taking processes, which is essential for cognitive reappraisal. Yet, Cai et al.49 found that young adults with ASD that reported using reappraisal expressed lower levels of emotional regulation difficulties such as depression symptoms, anxiety, and better well-being, similar to reports in the general population47. Most studies on reappraisal in ASD were based on correlative designs, using questionnaires to assess emotion-regulation abilities. In the present research, we employed an experimental paradigm, focusing exclusively on the skills of individuals with ASD to perform a shame reappraisal strategy.

Cognitive reappraisal also depends on narrative capacities, which may differ in individuals with ASD compared to typically developing individuals. While children and adolescents with ASD confirm lower narrating abilities due to language and communication difficulties50,51, adults with HF-ASD present similar narrative abilities to their typical peers52. Therefore, qualitatively analyzing young adults with ASD narratives is appropriate when exploring their social and emotional experiences22,53,54. Here, in addition to the quantitative analysis of shame and reappraisal in ASD, we examined verbatim narratives of the participants of both groups as they reappraised shame-arousing pictures. This qualitative analysis is based on a methodology that conceptualizes and explores human experience represented in oral form55,56 and was used in analyzing narratives of students with Asperger's syndrome57.

Based on previous findings33,34, we hypothesized that young adults with HF-ASD would rate shame-arousing pictures with lower shame levels than typical adults due to their limited cognitive ToM. Regarding the role of reappraisal in shame processing, we did not propose a specific hypothesis due to the scarcity of studies exploring reappraisal in ASD, their conflicting results, and their correlative nature. Finally, we compared the reappraisal contents supplied in the two groups in response to shame-arousing pictures to examine whether HF-ASD adults generate thematic contents in their reappraisal that mirror those produced by typical young adults.

Results

There were no data exclusions. To examine the reappraisal task's influence on shame ratings, a 2 × 2 repeated-measures ANOVA was conducted, with Group (ASD/typical participants) as the between-subjects variable and Time (before and after reappraisal) as the within subject variable. The dependent variable was the mean shame ratings. Figure 1 presents the means and standard errors of shame ratings before and after reappraisal.

Figure 1
figure 1

Means and standard errors of shame ratings before and after reappraisal of shame-arousing pictures in typical and ASD groups.

Full size image

Consistent with our hypothesis regarding the differences between the shame ratings of the two groups, we found a significant main effect of Group: typical adults rated the shame-arousing pictures with significantly higher shame scores than the ASD group (F(1,69) = 12.50, p < 0.001, η2 = 0.15). In addition, the analysis revealed a significant main effect of Time (F(1,69) = 703.57, p < 0.001, η2 = 0.91), indicating that shame ratings were lower after the reappraisal than before it. No significant interaction between Group and Time was found (F(1,69) = 1.89 ,p > 0.05, η2 = 0.03).

The narratives were audiotaped and transcribed verbatim. Then, two coders, blind to the research goals, listened to each narrative several times while transcribing it. Then they read each participant's narrative transcripts, trying to identify repeated main themes. These central themes were later converted into meaningful units with readable content that was compared and examined for their central topic. Next, each unit was changed into a category with a label and a matching code. Five semantic types were identified: Category A consisted of content units that expressed positive social interactions between characters (e.g., "The boss in the picture points out on me because he wants to flatter me"). Category B referred to narratives representing undirected shameful interactions that occur to the other characters in the picture but not the main character (e.g., "This guy isn't joking at me but at somebody standing next to me"). Category C consisted of meaning units describing social interactions between all characters regarding neutral issues (e.g., "The girl in the room wants to show me her polished nail instead of pointing her finger at something wrong about me"). Category D includes phrases showing unrealistic social interactions when all characters know that they are involved in a fictional scenario (e.g., "I know that they are laughing at me, but we all know that this is part of a play or a show"). Finally, Category E consisted of themes focusing on the inferiority of the peers' characters instead of the supremacy of the main character involved in the social interactions. (e.g., "The colleagues at the office are talking behind my back because they know that I'm more talented than them.”) We calculated category frequencies for each participant to examine differences in narrative types between ASD and typical groups. Figure 2 presents the means and standard errors of the frequency of the categorized narratives.

Figure 2
figure 2

Means and standard errors of the mean frequency per narrative type. The ASD group generated more unrealistic and neutral narratives and less positive narratives than the typical group (**< .01, *< .05).

Full size image

A mixed design ANOVA with repeated measures was performed with Group (ASD/typical participants) as the between subject variable and Category (A-E) as the within-subject variable. The dependent variable was the mean categories' frequency. The main effect of Category was statistically significant, F(4,66) = 52.53, p < 0.001, η2 = 0.76, with positive social interaction narratives as the most frequent type of reappraisal and the inferiority narrative the lowest frequency. Bonferroni-adjusted paired t-tests demonstrated a significant frequency difference between most narrative categories except the undirected/neutral and unrealistic/inferior categories.

In addition, a significant interaction was found between Group and Category (F(4,66) = 6.43, p < 0.01, η2 = 0.28): The ASD group reported less positive interaction narratives than the typical group (t(69) = 4.36, p < 0.01). On the other hand, the young adults with ASD reported significantly more neutral (t(69) = -2.80, p < 0.01) and unrealistic (t(69) = -2.05, p < 0.05) social interaction narratives than the typical group.

Discussion

The present study explored shame feelings elicited while watching social shame-evoking pictures among young adults with HF-ASD and typical adults. In addition, we examined the effect of pictures' reappraisal on these feelings. Lastly, we investigated the type of narratives produced by individuals with ASD in their reappraisal process. Consistent with our hypotheses, we found that participants with ASD judged social shame situations with significantly lower shame ratings than the typical group. Also, reappraising social shame pictures with less negative interpretations reduced shame feelings in both groups to the same degree (~ 43%). Finally, compared to typical participants, younger adults with HF-ASD reappraised these scenarios using less positive, more neutral, and unrealistic narratives than the control group.

The current findings present a complex picture regarding the shame experiences of young adults with ASD. Supporting prior research22,34,58, we found that the HF-ASD group perceived social shame situations as evoking less shame. We suggest that this stems from their diminished cognitive assessment of emotional situations, resulting in a lower ability to identify shame than their typical peers. This reduced capacity is displayed in experimental settings and daily social situations28. Considering prior findings showing the high proneness of individuals with ASD to experience shame, especially during young adulthood35,36,37, we suggest that our results validate the discrimination between intense shame experiences that accompany this group due to their uniqueness on the one hand and lower ability to identify shame in relevant social situations than their typical peers, on the other hand.

Our study supports the existing research, in social cognition in general and shame in particular, that individuals with ASD suffer from impairments in cognitive ToM, especially in complex social contexts and feelings, such as shame59,60,61. A recent exception is a study by Osler et al.16, who found that individuals with ASD activate, as their neurotypical peers, implicit harm representations in moral violation acts, even if it was harmless62. These findings suggest that some aspects of moral cognition, namely moral violation, which elicits emotions such as shame, may be preserved in ASD. However, the phenomenon reported by Osler and colleagues, termed dyadic completion63, focuses on just a part of a moral act. The cognitive evaluation of the complete moral situation may still be diminished in ASD. Also, the findings reported by Osler and colleagues show that ASD and non-ASD individuals make the same incorrect conclusions, namely the existence of harm in a harmless scenario. Concluding about correct shame processes in ASD based on similar faulty moral judgments should be substantiated.

While the HF-ASD participants exhibited lower shame ratings of shame scenarios than the typical participants, both groups exhibited reappraisal abilities of social situations that affected their shame feelings similarly. The reappraisal performed by the participants with ASD shows their capacity to implement flexible thinking regarding the mental states of others, which is essential to creating new narratives. This finding is important because it demonstrates that although HF-ASD individuals may not use reappraisal, they can be trained to apply it. Furthermore, the reappraisal resulted in lower shame feelings, showing that reappraising these scenarios with new narratives benefited individuals with ASD who frequently experience intense shame feelings of shame due to their diagnosis and its accompanying behavioral characteristics36,38. Prior findings also illustrate the contribution of reappraisal and narrative therapy to the emotional regulation of individuals with ASD49,64. Thus, although some studies report less use of reappraisal strategies in individuals with ASD32, they may be able to employ reappraisal when motivated. Moreover, the visual stimuli in the present study may have assisted the participants with ASD in processing the social information and reinterpreting it65,66.

Although both groups used positive social interaction narratives to interpret the shame-arousing pictures, the group with HF-ASD used significantly fewer positive narratives of positive social interactions than their typical peers. The lower ability to produce positive descriptions of social interactions might indicate that young adults with ASD lack internal narrative schemes of positive social interactions, perhaps due to their reported proneness to negative social experiences in young adulthood23,38. Moreover, the differences between the groups are pronounced even more by the higher use of unrealistic and neutral narratives by the participants with ASD, implying imaginative impairments and limited emotional expression66.

Indeed, according to Craig and Baron-Cohen67, children with ASD produce shorter stories with fewer elements than controls. Considering this evidence, the ability to create stories rich with descriptions and details may remain diminished even during young adulthood. This lack of detail provision may be expressed by the ASD group's preference for fictional and unrealistic narratives, characterized by limited characters' descriptions. In addition, the higher rates of neutral stories among the ASD group replicate early evidence regarding this population's difficulty in describing emotions in a straightforward way68. Instead, they interpret positive and negative emotions as more neutral than the typical participants69,70, which may affect their produced narratives.

Integrating the quantitative findings of shame judgments before and after reappraisal and the qualitative self-narratives of both groups allows us to draw some conclusions regarding the social reality of young adults with ASD. First, the restricted shame self-perception among individuals with ASD might lead to a lack of self-improvement motivation, which usually follows ordinary shame experiences and is needed to succeed in social functioning28,71. Moreover, these differences between the two groups are accentuated even more by the higher frequency of unrealistic and neutral narratives used by the participants with ASD, implying unusual imaginative processes and limited emotional expressions. In addition, their lower use of positive social reports may reflect diminished positive social experiences. These narratives reveal a complicated social worldview of young adults with ASD. Moreover, this social reality, reflected through different reappraisal strategies, might eventually lead to a lack of positive social dynamics essential for successfully integrating into society.

Limitations and future directions

There are several limitations to this study. First, we conducted the present research in laboratory conditions, where shame induction was manipulated artificially. This procedure may have reduced the ecological validity of our findings. In addition, participants were asked in the current study to supply their narrative only in the reappraisal stage. Future research should collect these narratives before the reappraisal and qualitatively examine the influence of reappraisal on the thematic contents of the descriptions supplied by the participants with ASD both before and after reappraisal.

Additionally, although reappraisal manipulation has been extensively used in the emotion-regulation literature72, other measures, aside from emotion ratings, such as narratives or electrophysiological observations, are recommended to control or minimize demand effects. Finally, the long-term effects of reappraisal in ASD have not been examined. We suggest broadening future studies into a more extended timeframe to investigate the impact of emotion regulation, and specifically social shame experiences, on adults with ASD.

Conclusions

The present study focused on the ability of young adults with HF-ASD to decrease their feelings of shame by using one emotion regulation tool, namely reappraisal. Our findings reiterate previous findings that HF-ASD is characterized by an impaired capacity to interpret shame-inducing situations, expressed in lower shame ratings. Indeed, HF-ASD participants could implement reappraisal to decrease their feelings of shame, yet they used more neutral or negative reinterpretations than positive ones. Future studies in various relevant disciplines, such as education, psychology, psychiatry, and public health, should examine the use of reappraisal by ASD individuals and its subsequences. In particular, the narratives produced by this ER strategy should be probed more to leverage the reappraisal technique more effectively. Future research should also focus on the effect of reappraisal on other emotions (e.g., anger, depression, etc.) that young adults with ASD encounter in their daily lives. From a clinical and practical perspective, developing efficient emotion-regulation interventions will yield positive interpretations of social situations. This approach will potentially strengthen the social capacities of adults (and children) with ASD, allowing them to navigate more efficiently in the social and vocational milieu and help enrich their lives.

Method

Participants

G*power73 software was used to determine a-priori the sample size of the study. For primary analyses of a mixed-design ANOVA with repeated measures (2X2) and an effect size of 0.30 (small-medium), α error = 0.05, and power = 0.80, the total sample size required was 68 participants. In the current research, the participants were recruited via social media postings during 2020. All participants were Caucasian. The study group consisted of 5 women and 28 men diagnosed with ASD (M age = 23.54), and the control group included 6 women and 32 men (M age = 24.28). All participants with ASD met ADOS74 criteria before participating in the research. In addition, participants of both groups were assessed using the autism spectrum quotient questionnaire75. The ASD group presented significantly higher AQ scores compared with the typical group (t(70) = -13.36, p < 0.001). Moreover, all participants with ASD scored above 26 (M = 30.69, SD = 3.63), which is the score that is clinically acceptable as a cutoff for HF-ASD76,77.

Thirteen participants with ASD reported using medications (SSRI, n = 8, Ritalin, n = 5). In this group, ADHD was reported by eight participants, anxiety disorders by five, and depression by four. One typical participant was diagnosed with ADHD. Based on self-reports, individuals were excluded if they had diagnoses of epilepsy, chromosomal or psychotic disorders.

All participants were screened with a nonverbal intelligence test78, presenting no significant differences in the scores between the two groups (t(70 = 1.17, p > 0.05). They were also screened on the vocabulary subset from the Wechsler Adult Intelligence Scale WAIS-III79. No significant differences between their scores were discovered (t(70) = 1.83, p > 0.05). Participants were paid for their participation and signed an informed consent form at the beginning of the session. The study was approved by the institutional ethics committee.

Stimuli

A total of 60 shame-arousing pictures were selected from a pool depicting social rejection scenarios. They were used in previous studies and shown to arouse moderate levels of shame and negativity80,81 It included scenes of people being pointed and laughed at, people using hostile hand gestures toward others, and people with facial expressions of contempt. We validated the stimuli more and presented to fifteen participants (5 men, 10 women, M = 28.40 years, SD = 3.5, range 23–33) the 60 shame-arousing pictures. They were instructed to view the picture from a personal perspective and imagine themselves as the main character being humiliated in the picture. Then, they were asked to rate the shame-arousing levels of each picture (1 = very proud, 5 = neutral, 9 = very ashamed) and rate its pleasantness (1 = very pleasant, 5 = neutral, 9 = very unpleasant). The ten pictures with the lowest shame-arousing and unpleasantness scores were excluded from the study. The remaining 50 pictures had a shame mean of 6.91 (SD = 0.48, range 5.87–8.40) and an unpleasantness mean of 7.12 (SD = 0.56, range 6.13–8.47). These pictures were divided into two lists of 25 scenarios. Five were used for practice, and the rest were assigned randomly for either the baseline, before-reappraisal shame-rating, or the reappraisal condition. No significant differences were found in shame (t(48) = 0.899 p > 0.05) and unpleasantness ratings (t(48) = 0.288, p > 0.05) between the two lists.

Twenty-two neutral pictures depicting animate and inanimate objects were selected from the International Affective Picture System database (IAPS) pictures database82. These pictures were chosen to minimize the crossover effects of shame pictures from those presented in the first block to those shown in the second block. To examine the neutral level of these pictures, the same 15 participants were now instructed to rate these pictures' shame and unpleasantness levels. Two neutral pictures with the highest rates of shame and unpleasantness were excluded from the study. The 20 remaining neutral pictures had a mean shame rating of M = 4.90 (SD = 0.15) and a mean unpleasantness rating of M = 4.92 (SD = 0.20). The shame-arousing pictures of the two lists were rated significantly higher on the shame scale relative to the neutral pictures (F(2,68) = 138.72, p < 0.01, η2 = 0.803). Similarly, the unpleasantness ratings of the shame-arousing pictures were significantly higher than the neutral pictures (F(2,69) = 166.619, p < 0.01, η2 = 0.828).

Procedure

The session consisted of three parts. In the first part, participants were instructed to imagine themselves as the primary character in the social shame scenarios to be presented. Then, they viewed 20 social-shame-arousing pictures from the first list. After watching each picture, they rated their feelings regarding it on a proud to shame rating scale ("1–very proud", through "5–neutral", to "9–very ashamed").

Each trial consisted of the following sequence: a 1000 ms display of a fixation cross and a 4500 ms display of a shame picture, followed by the proud to shame scale (Fig. 3). Before running the experimental trials, participants viewed and rated five shame arousing pictures for practice.

Figure 3
figure 3

Depiction of the experiment design. The colors are presented to clarify the different stages of the experiment. In the experiment itself, they were exhibited in black and white.

Full size image

In the second part, the participants viewed and rated the 20 neutral pictures on the same shame to proud scale. Finally, in the third part, participants viewed the 20 pictures of social shame from the second list. They were instructed to imagine themselves as the main character in the scenario and then interpret what happened in the photo less negatively (e.g., "This girl is not laughing at me, but rather is enjoying a joke I told her").

In addition, they were asked to state aloud the new interpretation they came up with for each picture. Before participating in the third trial condition, the experimenter demonstrated to each participant how to reappraise shame-arousing pictures less negatively. Then the participants practiced reappraising five shame arousing pictures less negatively. While they watched the practice pictures, the experimenter asked two opening questions: 1. "Imagine yourself in the center of the picture - now, what do you see"? 2. "Can you describe this picture less negatively?". The answers to the second question were analyzed later. Each shame-arousing picture was presented at this reappraisal phase until the participant finished stating the new interpretation. Finally, participants rated their emotions toward the picture on the proud to shame scale. In each phase, pictures were randomly presented. Stimulus presentation and responses were controlled and recorded with SuperLab software 5.083 using a 15-inch PC laptop computer. The results were analyzed using IBM SPSS software (v. 23). The reported narratives were recorded on audiotape and saved as audio files by the experimenter.

Ethics approval

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Bar-Ilan University School of Education institutional ethics committee (#195–2020).

Consent

Informed consent was obtained from all individual participants included in the study.