Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Feb 14;18(3):604–613. doi: 10.1002/aur.70003

The Impact of Affective Theory of Mind on Autistic Individual's Perception of Remorse

Qi Sheng Tan 1,✉, Nathan Weber 1, Alliyza Lim 1, Robyn L Young 1
PMCID: PMC11928921  PMID: 39953750

ABSTRACT

Previous research indicates that difficulties with expressing remorse may contribute to the longer sentences autistic individuals receive within the criminal justice system. These differences in remorse expression may stem from their reduced ability to perceive emotions in others. This study investigated the association between an individual's level of autistic traits and their remorse perception ability. We also examined the influence of Affective Theory of Mind (ToM), the ability to understand others' emotional experiences, in remorse perception. We thought that the more autistic traits with which a person presented, the poorer their ability to perceive remorse would be, with overall ToM and affective ToM serving as mediating factors. Forty‐five autistic and 47 non‐autistic individuals assessed a series of facial expressions to determine the perceived level of remorse in each face. Results revealed that neither the combination of autistic traits nor any individual trait was significantly correlated with remorse perception ability. Additionally, autistic traits did not indirectly impact the perception of remorse through either overall ToM or affective ToM. These findings imply that individuals with high autistic traits, regardless of their perspective‐taking abilities, exhibit similar capacities for perceiving remorseful expressions to those with less autistic traits—at least when the stimuli are static. Future research should investigate the differences between autistic and non‐autistic individuals in perceiving remorse through different modalities of emotional expression, including behavioral and verbal cues.

Keywords: autism, expressed emotion, facial emotion perception, facial expression, theory of mind

Summary.

  • Both autistic and non‐autistic participants were shown photos of people's faces and asked to rate how remorseful they thought the person in the photo was.

  • There was no significant relationship between the number of autistic traits people have and their ability to perceive remorse.

  • Additionally, the ability to understand others' emotional experience (also known as affective Theory of Mind), did not influence this relationship.

  • These results suggest that individuals may rely on cues other than, or in addition to, facial expressions, such as verbal and behavioral cues, to perceive remorse.

1. Introduction

Research has shown that autistic individuals receive longer sentences in the criminal justice system, which is suspected to be because being perceived as less remorseful for the harm caused (Foster and Young 2022). While it is unknown why autistic people may show less remorse than their non‐autistic counterparts, or at least be perceived that way, it is commonly thought to be linked to their inability to perceive how their actions have affected others (Fitzpatrick et al. 2018). However, we argue that one's inability to express remorse, or mimic such expression, may instead be related to one's inability to perceive remorse in others, thereby affecting one's ability to learn how to express it (Ekman and Oster 1979; Keltner et al. 2019).

2. Remorse

Remorse is an emotion, which is experienced and expressed by an offender when they recognize their wrongdoing and wish to undo the harm caused (Brinke et al. 2012; Fisher and Exline 2006; MacLin et al. 2009; Stokkom 2002). Remorse is a multifaced emotion. First, remorse is unequivocally classified as a negative emotion, given that it causes distress for the individual who experiences it, which is similar to other negative emotions such as sadness, shame, and guilt (Stokkom 2002). Second, remorse is described as a feeling of regret and compunction, specifically during the acknowledgement of one's responsibility for an offense (Fisher and Exline 2006; MacLin et al. 2009). This characteristic is similar to the realization that one's behavior has violated social norms or moral standards, which is experienced in self‐conscious emotions, such as embarrassment and shame (Heerey et al. 2003). The last feature of remorse is its unique communication purpose within a social interaction (Davis and Gold 2011). Remorse is often expressed within an apology to restore the relationship with the victim, the person who was physically or psychologically hurt by the offense (Davis and Gold 2011; George et al. 2023). Therefore, remorse is a social emotion, and its expression is one of the fundamental factors that affect the perceived sincerity of an apology, because it conveys that the offender has an active understanding of the wrongfulness of their behavior (Davis and Gold 2011). In other words, the person who expresses remorse indicates that they feel ‘negative’ regarding themself and their behavior (i.e., negative emotion), understand that they have caused the behavior (i.e., self‐conscious emotion), and have the obligation to express a sincere apology (i.e., the social function of remorse). These three features provide a conceptual framework for remorse. To the best of our knowledge, no previous study has directly investigated the ability of autistic individuals to perceive remorse. However, studies have shown that, compared to non‐autistic individuals, individuals with autism have greater difficulty recognizing negative complex emotions (although their performance is better than for emotions in positive valance; Palmer et al. 2023), social emotions (Williams and Happé 2010), and self‐conscious emotions (Heerey et al. 2003). Considering that these are core features of remorse, we hypothesize that autistic individuals may also struggle to perceive the expression of remorse.

3. Autism and Theory of Mind

In the Diagnostic and Statistical Manual of Mental Disorders—5th Edition—Text Revision (DSM‐5‐TR; American Psychiatric Association [APA] 2022), impairments in non‐verbal communicative behaviors are identified as a diagnostic feature of autism. These impairments often manifest as difficulties understanding facial expressions and interpreting social cues, which may also be associated with difficulties with Theory of Mind (ToM). ToM, known as perspective‐taking, refers to one's ability to make inferences about other's mental states (Pisani et al. 2021). Specifically, it can be separated into two mental aspects, which are cognitive ToM, the understanding of other's beliefs and intentions, and affective ToM, the ability to attribute emotional states of others (Canty et al. 2021; Pisani et al. 2021; Westby and Robinson 2014). While the literature suggests that individuals with autism may exhibit poorer ToM skills compared to non‐autistic individuals (Baron‐Cohen et al. 1997; Dyck et al. 2001), some studies have found no such association, as only a subgroup of autistic individuals experiences difficulties in ToM (Brewer et al. 2017). It has been argued that this variation could be attributed to the heterogeneous nature of autism (Mao et al. 2023). However, other studies propose that these inconsistencies arise due to the diverse designs of ToM assessment tools and the lack of standardized measurements that effectively differentiate between the different aspects of ToM (Pisani et al. 2021).

Pisani et al. (2021) found that only affective ToM, not cognitive ToM, was associated with alexithymia (the inability to recognize and describe one's own emotions). This study indicates that one's inferences regarding their own beliefs and emotions may depend on different mental processes. Caillaud et al. (2019) showed higher activation in several brain regions relevant to affective ToM, during the processing of others' self‐conscious emotions compared to basic emotions. This result suggests that affective ToM is involved in the process of recognizing self‐conscious emotions, aligning with the argument that self‐conscious emotions are experienced when acknowledging the discrepancy between oneself and the social standard (Heerey et al. 2003), and ToM is needed to recognize other's experiences or acknowledgement on this discrepancy. Therefore, it is proposed that one's ability to perceive remorse is likely to be influenced by one's level of affective ToM. However, ToM is included to better understand the influence of the entire construct on remorse perception, as ToM is multifaceted (Westby and Robinson 2014) and includes not only cognitive and affective dimension.

In this study, we investigated the impact of autism on one's ability to perceive remorse using a trait‐based approach. This approach provides a more comprehensive understanding of how each trait is associated with the perception of remorse and presents autistic traits in a continuous form, considering that clinical diagnosis may fail to capture the nuance of various levels and dimensions of autism. Furthermore, we examined the mediation effect of both ToM and affective ToM on this relationship. We hypothesized that greater autistic traits would be associated with poorer remorse perception ability and that this relationship would be mediated by both overall ToM and affective ToM.

4. Method

4.1. Participants

A total of 101 participants (51 autistic individuals and 50 non‐autistic individuals) were recruited through the online survey platform, Prolific. Participation in the study was voluntary, and all collected data were de‐identified to ensure confidentiality. The survey required approximately 30 min to complete, and participants received compensation of AU$8.00 upon completion. The data from participants who did not complete the study in its entirety, had a verbal IQ score below 85 on the Spot‐The‐Word (STW) task, or provided inconsistent responses regarding their autism diagnostic status were excluded from the analysis. As a result, the final sample included 45 autistic and 47 non‐autistic individuals, totaling 92 participants. Participants comprised 45 male, 43 female, and four non‐binary individuals, with ages ranging from 19 to 72 years (M = 39.28, SD = 12.14).

4.2. Materials

4.2.1. Objective Remorse—Facial Photos

Twenty‐five facial photos depicting remorse were developed by Funk et al. (2017) using a data‐driven computational approach, based on neutral facial photos from Walker et al. (2018). The photos comprised five facial identities, with each identity displaying five different degrees of remorse, including high remorse, low remorse, neutral, low happiness, and high happiness (see Supporting Informations S1). Happy faces were employed as distractor stimuli, while the study focused on the analysis of neutral and remorseful faces.

The 25 remorse photos were divided into five sets, with each set containing a unique combination of five photos, ensuring the absence of repetition in terms of facial identity and degree of remorse. This approach was designed to prevent participants from experiencing visual fatigue, as repeated exposure to the same face with similar facial expressions could lead to tiredness and potentially skew the results. Participants encountered only one set of these photos twice during the study. Neutral faces were coded as 0% remorse, low remorse faces as 50% remorse, and high remorse faces as 100% remorse, resulting in a three‐point scale ranging from 0% to 100%. A higher percentage indicates a greater intensity of remorseful expression.

4.2.2. Remorse Perception Scale

To assess participants' ability to perceive remorse, a slider scale was presented after participants viewed each individual photo. The question was: Using the provided scale, please indicate how remorseful you perceive the individual in the picture to be for the proposed crime (0 = Not at all remorseful, 100 = Extremely remorseful).

4.2.3. Autism Spectrum Quotient (AQ)

The AQ is a self‐report questionnaire designed to assess participants' level of autistic traits (Baron‐Cohen et al. 2001). The questionnaire comprises 50 questions, with 10 questions measuring each of five different subscales of traits: social skills, attention switching, attention to detail, communication, and imagination. Responses to each question are scored either 0 (non‐autistic) or 1 (autistic), resulting in a possible range of 0–50. Higher scores indicate greater presence of autistic traits. Studies have demonstrated that the AQ possesses good predictive validity (AUC = 0.91) for screening for autism (Booth et al. 2013), test–retest reliability (r = 0.70), and internal consistency within each trait (Cronbach's alpha ranging from 0.63 to 0.77; Baron‐Baron‐Cohen et al. 2001). Additionally, the AQ showed high convergent validity by displaying a strong correlation with the Social Responsiveness Scale (SRS, r = 0.64, p < 0.01; Armstrong and Iarocci 2013).

4.2.4. Frith‐Happé Animations Task (AT)

The AT is an advanced assessment tool for ToM designed to assess individuals' ToM ability through recognizing the mental states and emotions of others by watching the interaction between two animated triangles (Abell et al. 2000). The videos are classified into three categories, which are: no interaction, physical interaction, and mental interaction. The no interaction videos involve two triangles moving randomly on the screen without any interaction, serving as control trails. In contrast, the physical interaction videos depict one triangle responding to the physical action of the other triangle, while the mental interaction videos feature one triangle responding to the mental state of another triangle.

Each category consists of four videos, and two practice videos are included, for a total of 14 videos in the task. Participants were required to classify each video into one of the three categories, with 1 point given for each correct classification, resulting in a score range of 0–4 for each category. If the participants correctly identified the mental state videos, they were presented with two additional multiple‐choice questions about the expressed feelings in the videos, with score for this feeling categorization ranging from 0 to 8. For all categories, higher scores indicate better ToM.

Studies have found that the AT has strong sensitivity in differentiating between autistic and non‐autistic participants in both the mental interaction subtest (U = 59.5, p = 0.009), feelings identification subtest (U = 28.5, p < 0.001; White et al. 2011), as well as the combination of these subtests (Hedges's g > 0.56; Wilson 2021). However, both the mental interaction and feelings identification subtests only showed weak to moderate correlation with the Adult Theory of Mind (A‐ToM), with r = 0.25 (p < 0.01) and r = 0.33 (p < 0.01), respectively (Brewer et al. 2017).

Despite the concerns about the validity of the AT and the fact that its original design did not explicitly target the mental aspects of ToM, the measurements for mental interaction and recognition of feelings within the task are conceptually aligned with the concept of cognitive ToM and affective ToM, respectively. Correctly classifying mental interaction videos indicates that individuals can recognize the thoughts and intentions of others, which suggests cognitive ToM ability. In contrast, the accurate identification of the feelings of the animated triangles suggests that individuals can understand others' emotions, indicating affective ToM. Thus, a combined score that integrates both mental interaction and feelings identification will be utilized to represent overall ToM capacity. Although it has limitations, this tool is currently the most reliable and valid tool available that separates cognitive ToM and affective ToM for adults.

4.2.5. Adult Theory of Mind Quick Social (A‐ToM‐Q)

The A‐ToM‐Q is another assessment tool for ToM, designed as a multiple‐choice version of the A‐ToM (Brewer et al. 2022), which only assesses ToM as a unitary construct. A‐ToM‐Q Social subscale consists of six short clips demonstrating various real‐life social scenarios. Participants are required to select the answer that best describes each scenario, with 2 points given for the most suitable answer and 1 point given for each acceptable answer. The total possible score range is 0–12 with higher scores indicating greater ToM ability. The A‐ToM‐Q has excellent concurrent validity, as evidenced by its significant correlation with the Strange Stories social sub‐scale (r = 0.45, p < 0.01; Brewer et al. 2022). The original A‐ToM demonstrated excellent inter‐rater reliability (Cohen's kappa = 0.77) and test–retest reliability (r = 0.82; Brewer et al. 2017). Considering the robust psychometric properties of A‐ToM‐Q (Brewer et al. 2022), it has been included as an additional ToM assessment tool to enhance the validity of the current study.

4.2.6. Verbal Intelligent Task (Spot‐The‐Word)

Spot‐The‐Word (STW) is a lexical decision task utilized to assess an individual's verbal intelligence (Yuspeh and Vanderploeg 2000). In this task, participants encounter 100 pairs of items, each including a real word and a non‐word. Their goal is to identify the real word from each pair of items. Correct responses in the STW task are scored as 1, while incorrect responses are scored as 0, resulting in a possible range of 0–100. Scoring 48 and above indicates moderate to high verbal intelligence (with an estimated WASI‐IV VCI equivalent of 85 and above; Baddeley and Crawford 2012). Studies have shown that the STW task shows good convergent validity with the American National Adult Reading Test (r = −0.56, p < 0.001; Yuspeh and Vanderploeg 2000), as well as good test–retest reliability (r = 0.78; Baddeley et al. 1993) and internal consistency with a Cronbach's α of 0.85 (Mackinnon and Christensen 2007).

4.2.7. Demographic Questionnaire

All participants were asked to provide information about their gender, age, first language, and any diagnostic history of autism or other developmental or mental health condition through the online questionnaire.

4.3. Procedure

The current study was approved by the appropriate institutional ethics committee. Prolific users who showed interest and met the inclusion criteria (i.e., aged 18 years and above and used English as their first language), were directed to an online Qualtrics questionnaire through a provided link. After obtaining participants' consent, they were given the following instruction:

These photos were taken after the individuals underwent police interrogation. In each case, the individuals caused minor injuries to the victims, who were subsequently sent to the hospital for treatment. These individuals have confessed to committing crimes against the victims. Your task is to rate the level of remorse you perceive in the criminal's expression and to indicate your confidence in your judgment.

Subsequently, participants encountered one randomly selected set of facial identity photos from a pool of five sets. After viewing each photo, participants completed the Remorse Perception Scale for each photo. To ensure the reliability of the scales and to account for both order and carryover effects, participants completed the same set of data twice in random order with a short break in between. Following these two trials, all participants completed the AQ, AT, A‐ToM‐Q, STW and a demographic questionnaire. Finally, participants were informed that the facial stimuli were not actual photos of real criminals, but rather were developed solely for research purposes.

After the data were collected, a unique remorse perception index for each individual was calculated using regression analysis. This index was derived from the relationship between degree of objective remorse (i.e., the intensity of remorseful facial expression in the photos) and the average of two subjective remorse ratings, quantifying participants' ability to distinguish varying degrees of remorseful expressions.

5. Results

5.1. Data Analysis

We adopted an alpha level of 0.05 to determine the significance of the inferential analysis. Additionally, we utilized a bootstrap procedure to ensure the robustness of the mediation analysis conducted using PROCESS macro (Hayes 2023, Version 4.3) for the effect estimates. Specifically, 10,000 bootstrap samples with 95% Confidence Intervals (CI95) were used to assess the significance of the mediation pathways, wherein the direct or indirect effect was considered significant if the CI95 did not include 0.

The remorse perception index was computed based on the relationship between degree of objective remorse and ratings of subjective remorse using regression analysis, with each participant having a b value to represent their remorse perception ability. As least squares regression assumes that predictors are measured without error, the use of a regression coefficient as a predictor in a different regression can be problematic. Specifically, by ignoring the potential for error in the measurement of the index, regression can overestimate certainty in the relationship between the index and the outcome variable. However, to presage our results, there was no significant relationship found between remorse perception index and either AQ or ToMs. Thus, it appears unlikely that our analysis approach distorted the findings in any meaningful way.

5.2. Hypothesis Testing

5.2.1. Correlation Between AQ and Remorse Perception Index

To test the hypothesis that there would be a negative relationship between level of autistic traits and remorse perception ability, a simple correlation was conducted. As indicated in Table 1, there was no significant association between scores on the AQ and remorse perception index, thus failing to support the hypothesis. Considering the mean and standard deviation of the AQ (M = 26.17, SD = 11.05) and remorse perception index (M = 0.09, SD = 0.24) in the overall sample, it is unlikely that the non‐significant correlation between AQ and remorse perception index is due to limited variability. Additionally, the means, standard deviations, and correlations among all variables for each diagnostic group are also presented in Table 1.

TABLE 1.

Means and standard deviations for each group with the correlations for study variables.

Variable ASD Non‐ASD 1 2 3 4
1. AQ 33.78 (9.05) 18.89 (7.26) —
2. RR 0.11 (0.25) 0.07 (0.24) −0.01 —
3. Age 33.11 (8.78) 45.19 (12.04) −0.42*** −0.08 —
4. VIQ 71.02 (11.73) 73.96 (13.14) 0.04 0.07 0.38*** —
5. AT 5.69 (3.20) 5.93 (3.53) 0.06 0.00 −0.01 0.30**
6. A‐ToM‐Q 10.11 (2.38) 11.06 (1.22) −0.31** 0.04 0.10 0.24*
7. AT feel 3.00 (2.25) 3.32 (2.56) 0.02 0.02 0.03 0.34***
Open in a new tab

Note: *p < 0.05; **p < 0.01; ***p < 0.001.

Abbreviations: AQ, autism spectrum quotient; AT, animation task; A‐ToM‐Q, adult theory of mind quick social; AT feel, animation task feeling subscale; RR, remorse perception index; VIQ, verbal intelligence.

5.2.2. Mediation Effect of AT and A‐ToM‐Q

We used the PROCESS macro (Hayes 2023, Version 4.3) within SPSS to execute the analysis for testing the mediation hypotheses. Three separate simple mediation analyses (Model 4) were conducted. In each model, AQ was included as the predictor variable and remorse perception index was the outcome variable, with each model testing a different mediator. The mediators in each of the three models were AT, A‐ToM‐Q, and AT Feelings. Lastly, age and verbal intelligence were entered as covariates in every model.

The results (see Table 2) showed no evidence that AT or A‐ToM‐Q mediated the relationship between AQ and remorse perception index. These findings were contrary to the prediction that there would be a negative indirect effect from AQ to remorse perception index through AT and A‐ToM‐Q. Furthermore, Figures 1 and 2 illustrate the mediated pathways connecting AQ and remorse perception index through AT and A‐ToM‐Q, respectively. These figures show that after the control of AQ, neither AT nor A‐ToM‐Q was related to the remorse perception index. Additionally, only A‐ToM‐Q shows a significant relationship with AQ. Together, these findings indicate a weak connection within the mediation models.

TABLE 2.

Unstandardized indirect effects and 95% bootstrapped confidence intervals of AQ on remorse perception index with AT, A‐ToM‐Q and AT feelings as the mediator in each of the separate models.

Mediator Indirect effect CI95
AT 0.000 [−0.001, 0.001]
A‐ToM‐Q 0.000 [−0.003, 0.002]
AT Feelings 0.000 [−0.001, 0.001]
Open in a new tab

Note: Given the small coefficients observed in these results, some values are presented as zero. However, the negative signs have been retained in the unstandardized coefficients and indirect effects to indicate the direction of the effect.

FIGURE 1.

FIGURE 1

Open in a new tab

Unstandardized coefficients [95% bootstrapped confidence intervals] for the effect of AQ on remorse perception index, with AT the mediator. Both age and verbal intelligence were included as covariates in these models.

FIGURE 2.

FIGURE 2

Open in a new tab

Unstandardized coefficients [95% Bootstrapped Confidence Intervals] for the Effect of AQ on Remorse Perception Index, With A‐ToM‐Q as the Mediators. Both age and verbal intelligence were included as covariates in these models. *p < 0.001.

5.2.3. Mediation Effect of AT Feelings

Table 2 illustrates that there was no significant indirect effect from AQ to remorse perception index through AT Feelings. Consequently, the hypothesis positing that AT Feelings would significantly mediate the relationship between AQ and remorse perception index was not supported. Furthermore, Figure 3 demonstrates the direct pathways between AQ and the remorse perception index, with the mediation of AT Feelings. Specifically, it shows that there is no significant association between AQ and AT Feelings, nor was any significant relationship evidenced between AT Feelings and the remorse perception index (when controlling for AQ).

FIGURE 3.

FIGURE 3

Open in a new tab

Unstandardized coefficients [95% bootstrapped confidence intervals] for the effect of AQ on Remorse Perception Index, With AT Feelings as the Mediator. Both age and verbal intelligence were included as covariates in these models.

5.3. Follow‐Up Analyses

The results presented above do not support the hypothesis that the AT, A‐ToM‐Q, and AT Feelings would mediate the relationship between scores on the AQ and remorse perception index. Given that the AQ can be divided into five subscales that measure distinct aspects of autistic traits (social skills, attention switching, attention to detail, communication, and imagination), it is plausible that these non‐significant findings are influenced by potential compensatory relationships among various autistic traits. Hence, we repeated the analyses by replacing AQ scores with each AQ subscale scores to further understand whether specific domains of autistic traits have a more robust association, either directly or indirectly, with the remorse perception index via the AT, A‐ToM‐Q, and AT Feelings. Scores on each AQ subscale ranged from 0 to 10, with higher scores indicating higher levels of autistic traits. The means and standard deviation for each autistic trait and their correlations with other variables are included in Supporting Information S2.

5.3.1. Correlation Between Each AQ Subscale and Remorse Perception Index

As shown in Table 3, none of the AQ subscales exhibited a significant relationship with the remorse perception index. These findings are aligned with our initial results, which showed no association between the AQ and remorse perception index.

TABLE 3.

Correlation coefficient for AQ subscales and remorse perception index.

AQ subscale r p
Social skills −0.02 0.86
Attention switching 0.03 0.75
Attention to detail −0.06 0.56
Communication −0.06 0.60
Imagination 0.06 0.55
Open in a new tab

5.3.2. Mediation Effect of ToM Measurements

As illustrated in Table 4, none of the AQ subscale showed a significant indirect effect on remorse perception index through the AT, A‐ToM‐Q, or AT Feelings. These results are similar to our initial findings, which showed no evidence that supports AT, A‐ToM‐Q and AT Feelings mediated the effect of AQ on the remorse perception index. However, it is worth noting that all autistic traits (except attention to detail) were significantly negatively associated with the A‐ToM‐Q. However, the A‐ToM‐Q had no significant association with the remorse perception index. Conversely, there was no significant relationship between any AQ subscale and either AT or AT Feelings, while both AT and AT Feelings were not significantly associated with the remorse perception index. The direct effects of each AQ subscale on ToM measurements and each ToM measurement on the remorse perception index are included in Supporting Informations S3 to S5.

TABLE 4.

Unstandardized indirect effects and 95% bootstrapped confidence intervals of each AQ subscale on remorse perception index with AT, A‐ToM‐Q, and AT feelings as the mediators.

AQ subscale Mediator Indirect effect CI95
Social skills AT 0.000 [−0.003, 0.003]
A‐ToM‐Q −0.000 [−0.008, 0.006]
AT Feelings 0.000 [−0.003, 0.003]
Attention switching AT −0.000 [−0.004, 0.003]
A‐ToM‐Q −0.001 [−0.008, 0.006]
AT Feelings −0.000 [−0.003, 0.003]
Attention to detail AT 0.000 [−0.004, 0.003]
A‐ToM‐Q −0.000 [−0.008, 0.005]
AT Feelings 0.000 [−0.004, 0.003]
Communication AT 0.000 [−0.003, 0.004]
A‐ToM‐Q 0.000 [−0.008, 0.009]
AT Feelings 0.000 [−0.003, 0.004]
Imagination AT 0.000 [−0.003, 0.003]
A‐ToM‐Q −0.001 [−0.014, 0.009]
AT Feelings 0.000 [−0.003, 0.003]
Open in a new tab

Note: Given the small coefficients observed in these results, some values are presented as zero. However, the negative signs have been retained in the unstandardized coefficients and indirect effects to indicate the direction of the effect.

5.3.3. Autistic and Non‐autistic Comparison

Further analyses were conducted by applying the same mediation models but using autism diagnosis instead of AQ scores as the predictive variable. In this analysis, individuals with an autism diagnosis were coded as 0.5, while those without an autism diagnosis were coded as −0.5. These values were chosen because they effectively represent a full unit change due to the difference of 1, with the midpoint being 0. Also, this coding maintains the directionality, as a higher value represents a higher level of autistic traits. Aligned with our prior results, there wass no association between autism diagnosis and remorse perception index (see Supporting Information S6).

6. Discussion

The current study explored the impact of autistic traits on remorse perception ability. Furthermore, we examined the mediation effect of ToM and affective ToM in this relationship. Contrary to expectations, the results showed that autistic traits did not significantly correlate with one's ability to perceive remorse. These results contradicted previous studies suggesting that autistic individuals have difficulty with perceiving complex emotions compared to their non‐autistic peers (Heerey et al. 2003; Palmer et al. 2023; Williams and Happé 2010). Furthermore, neither overall ToM nor affective ToM significantly mediated the relationship between autistic traits and remorse perception. In other words, autistic traits had no significant impact on one's remorse perception ability through ToM and affective ToM. Specifically, both ToMs did not show a significant direct effect on remorse perception ability, which is contrary with prior findings that suggest ToM plays a crucial role in emotional recognition (Lee et al. 2014; Löytömäki et al. 2020).

The mean scores on the remorse perception scale demonstrated that both individuals with higher and lower autistic traits rated perceived remorse similarly, regardless of the degree of remorse presented (see Supporting Information S7). These ratings further support the findings from the mediation models, which indicate that individuals with lower autistic traits do not perform better than those with higher autistic traits in perceiving remorse from facial expressions. These findings contradict our initial assumption that difficulty perceiving remorseful facial expressions is one of the possible reasons why autistic individuals have difficulty expressing remorse, as autistic traits did not have an impact on one's remorse perception ability. However, the results should be carefully interpreted, as they apply only to individuals' ability to perceive static facial expressions of remorse.

The use of static facial expression as stimuli may explain the non‐significant relationship between autistic traits and remorse perception ability. As some studies argue, facial expressions are inherently dynamic in real‐life social situations (Leung et al. 2023; Tang et al. 2019), involving subtle and rapid changes in emotional cues, making them more challenging to recognize (Fridenson‐Hayo et al. 2016). Researchers have suggested that autistic individuals tend to apply rule‐based strategies and focus on specific facial features in emotional recognition (Boraston et al. 2008), which may require more time to process and interpret the emotional cues (Leung et al. 2023). In other words, the impact of autistic traits may emerge when there is a time constraint on the processing of an expression or requirement to rapidly process a series of expressions. This compensatory strategy may explain why all participants, regardless of their level of autistic traits, provided similar ratings for each photo. Although this explanation suggests that the results of the current study may not be generalizable to real life scenarios, it does not invalidate or devalue its findings. Rather, our findings indicate that the difficulties autistic individuals may face in perceiving and expressing remorse are not due to a fundamental inability to perceive remorseful facial expressions, suggesting that other underlying factors may be at play.

Furthermore, there is a possibility that expressing remorse relies heavily on cues beyond facial expressions. As Keltner et al. (2019) suggested, emotions can be expressed through different modalities, such as verbal components, vocal tone, and body language. While some emotions, such as happiness and sadness, can be readily recognized through facial expressions, particularly the mouth region (Boraston et al. 2008), other emotions may require the integration of multiple modalities to accurately capture the expression of the emotion (Tang et al. 2019). Therefore, conveying remorse may depend on other modalities or the integration of various modalities, including facial expressions. For instance, from the facial expression of remorse alone, a victim can only infer that the offender has negative feelings, given the similarity of remorse with the expression of sadness (Funk et al. 2017). However, they cannot tell whether the offender acknowledges the offense and intends to repair the relationship (i.e., the offender's perspective), as these aspects of remorse may be more effectively conveyed through other modalities or a combination of facial expressions and other forms of modalities, rather than through the facial expression alone.

Given that remorse facial expressions alone may not effectively convey an offender's mental and emotional states, this may explain why both overall ToM and affective ToM had no direct effect towards remorse perception ability and no mediation effect on the association between autistic traits and remorse perception. In other words, both cognitive and affective ToM were not required in the process of perceiving remorse facial expression, as there is no mental state to attribute from the facial stimuli. It is suspected that if other modalities were used as stimuli, the mediation effect of ToM might appear. Therefore, the current findings do not fully reject our initial argument that ToM, specifically affective ToM, is an important factor influencing autistic individuals' ability to perceive remorse. For the same reason, the current findings do not suggest that difficulty in perceiving remorse is not a factor in explaining the possible challenges autistic individuals face in expressing remorse, as using only facial stimuli may not adequately capture the impact of autistic traits on remorse perception ability.

However, it remains plausible that there is a dissociation between remorse perception and remorse expression. Given that remorse serves as a communicative tool for acknowledging harm and is often used in apologies (Davis and Gold 2011), experiencing remorse, rather than perceiving remorse may have a greater association with its expression. Autistic individuals' difficulties in expressing remorse could be due to their challenges in recognizing the relevant social cues (i.e., ToM difficulty) that influence the experience of remorse. In other words, it is not that autistic individuals having difficulties perceiving remorse that leads to atypical remorse expression. Rather, it is that events of the same severity may trigger a lower experience of remorse in autistic individuals compared to non‐autistic individuals, resulting in a weaker expression of remorse or being perceived as less remorseful. This may explain why autistic individuals do not exhibit difficulties with remorse perception but do struggle with remorse expression.

One major limitation of the current study is the number of photo stimuli used. To avoid participant fatigue from viewing the same face with different levels of remorse, participants encountered only five facial identities, and all of them were of the same race and gender (i.e., Caucasian male). The few stimuli used may limit the generalizability of the findings.

The current study provides a foundation for future investigations into the challenges autistic individuals face in perceiving remorse. To further our understanding on this topic, future studies should consider using dynamic expressions of remorse as stimuli, given that this may reveal the unique challenges autistic individuals encounter during real life interactions due to their use of different emotional recognition techniques. Additionally, future research should replicate this study using different modalities to identify which modalities are crucial for expressing remorse and whether autistic individuals experience difficulties in perceiving remorse through these modalities. Finally, further studies should investigate the suspected relationship between perceiving and expressing remorse. To justify this relationship, future studies could examine whether the expression of remorse improves with training focused on perceiving remorse but not expressing it. This approach may demonstrate whether improvements in remorse perception lead to improvements in remorse expression.

6.1. Conclusion

In summary, this study found no evidence that level of autistic traits had an impact on the ability to perceive static remorseful facial expressions directly or through ToM and affective ToM. Furthermore, both autistic traits and ToM had no significant direct effect on remorse perception ability. Remorse is a complex emotion, and its underlying features may be difficult to perceive through static facial expression alone. Therefore, future research should expand its focus to include behavioral cues, verbal cues, and dynamic facial expressions to gain a deeper understanding of the variations in remorse perception with autistic traits, as well as the role of affective ToM in the process of remorse perception. Finally, validation is needed for the relationship between remorse perception and expression.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1. Supporting Information.

AUR-18-604-s001.docx (882.9KB, docx)

Acknowledgments

The authors express their gratitude to Professor Friederike Funk for providing the face stimuli used in the study. This research was supported by the Hamish Ramsay Fund. Open access publishing facilitated by Flinders University, as part of the Wiley ‐ Flinders University agreement via the Council of Australian University Librarians.

Funding: This work was supported by Hamish Ramsay Fund.

Data Availability Statement

The data that support the findings of this study are openly available in Open Science Framework at https://www.osf.io/7jhp2/?view_only=f63cb4e100f6436ea7556fd8cc937ac4.

References

  1. Abell, F. , Happé F., and Frith U.. 2000. “Do Triangles Play Tricks? Attribution of Mental States to Animated Shapes in Normal and Abnormal Development.” Cognitive Development 15, no. 1: 1–16. 10.1016/S0885-2014(00)00014-9. [DOI] [Google Scholar]
  2. American Psychiatric Association . 2022. Diagnostic and Statistical Manual of Mental Disorders. 5th ed, text rev. American Psychiatric Association. 10.1176/appi.books.9780890425787. [DOI] [Google Scholar]
  3. Armstrong, K. , and Iarocci G.. 2013. “Brief Report: The Autism Spectrum Quotient has Convergent Validity With the Social Responsiveness Scale in a High‐Functioning Sample.” Journal of Autism and Developmental Disorders 43: 2228–2232. 10.1007/s10803-013-1769-z. [DOI] [PubMed] [Google Scholar]
  4. Baddeley, A. D. , and Crawford J. R.. 2012. Spot the Word 2: Administration, Scoring and Technical Manual. Pearson. [Google Scholar]
  5. Baddeley, A. , Emslie H., and Nimmo‐Smith I.. 1993. “The Spot‐The‐Word Test: A Robust Estimate of Verbal Intelligence Based on Lexical Decision.” British Journal of Clinical Psychology 32, no. 1: 55–65. 10.1111/j.2044-8260.1993.tb01027.x. [DOI] [PubMed] [Google Scholar]
  6. Baron‐Cohen, S. , Jolliffe T., Mortimore C., and Robertson M.. 1997. “Another Advanced Test of Theory of Mind: Evidence From Very High Functioning Adults With Autism or Asperger Syndrome.” Journal of Child Psychology and Psychiatry 38, no. 7: 813–822. 10.1111/j.1469-7610.1997.tb01599.x. [DOI] [PubMed] [Google Scholar]
  7. Baron‐Cohen, S. , Wheelwright S., Skinner R., Martin J., and Clubley E.. 2001. “The Autism‐Spectrum Quotient (AQ): Evidence From Asperger Syndrome/High‐Functioning Autism, Males and Females, Scientists and Mathematicians.” Journal of Autism and Developmental Disorders 31: 5–17. 10.1023/a:1005653411471. [DOI] [PubMed] [Google Scholar]
  8. Booth, T. , Murray A. L., McKenzie K., Kuenssberg R., O'Donnell M., and Burnett H.. 2013. “Brief Report: An Evaluation of the AQ‐10 as Brief Screening Instrument for ASD in Adults.” Journal of Autism and Developmental Disorders 43: 2997–3000. 10.1007/s10803-013-1844-5. [DOI] [PubMed] [Google Scholar]
  9. Boraston, Z. L. , Cohen B., Miles L. K., Skuse D. V., and Blackmore S.. 2008. “Brief Report: Perception of Genuine and Posed Smiles by Individuals With Autism.” Journal of Autism and Developmental Disorders 38: 574–580. 10.1007/s10803-007-0421-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Brewer, N. , Young R. L., and Barnett E.. 2017. “Measuring Theory of Mind in Adults With Autism Spectrum Disorder.” Journal of Autism and Developmental Disorders 47, no. 7: 1927–1941. 10.1007/s10803-017-3080-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Brewer, N. , Young R. L., Norris J. E., Maras K., Michael Z., and Barnett E.. 2022. “A Quick Measure of Theory of Mind in Autistic Adults: Decision Accuracy, Latency and Self‐Awareness.” Journal of Autism and Developmental Disorders 52: 2479–2496. 10.1007/s10803-021-05166-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Brinke, L. T. , MacDonald S., Porter S., and O'Connor B.. 2012. “Crocodile Tears: Facial, Verbal and Body Language Behaviours Associated With Genuine and Fabricated Remorse.” Law and Human Behavior 36, no. 1: 51–59. 10.1037/h0093950. [DOI] [PubMed] [Google Scholar]
  13. Caillaud, M. , Bejanin A., Laisney M., et al. 2019. “Influence of Emotional Complexity on the Neural Substrates of Affective Theory of Mind.” Human Brain Mapping 41, no. 1: 139–149. 10.1002/hbm.24794. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Canty, A. L. , Cao Y., Neumann D., and Shum D. H.. 2021. “The Functional Significance of Cognitive Empathy and Theory of Mind in Early and Chronic Schizophrenia.” Psychiatry Research 299: 113852. 10.1016/j.psychres.2021.113852. [DOI] [PubMed] [Google Scholar]
  15. Davis, J. R. , and Gold G. J.. 2011. “An Examination of Emotional Empathy, Attributions of Stability, and the Link Between Perceived Remorse and Forgiveness.” Personality and Individual Differences 50, no. 3: 329–397. 10.1016/j.paid.2010.10.031. [DOI] [Google Scholar]
  16. Dyck, M. J. , Ferguson K., and Shochet I. M.. 2001. “Do Autism Spectrum Disorders Differ From Each Other and From Non‐spectrum Disorders on Emotion Recognition Tests?” European Child & Adolescent Psychiatry 10: 105–116. 10.1007/s007870170033. [DOI] [PubMed] [Google Scholar]
  17. Ekman, P. , and Oster H.. 1979. “Facial Expressions of Emotion.” Annual Review of Psychology 30, no. 1: 527–554. 10.1146/annurev.ps.30.020179.002523. [DOI] [Google Scholar]
  18. Fisher, M. L. , and Exline J. J.. 2006. “Self‐Forgiveness Verses Excusing: The Roles of Remorse, Effort, and Acceptance of Responsibility.” Self and Identity 5, no. 2: 127–146. 10.1080/15298860600586123. [DOI] [Google Scholar]
  19. Fitzpatrick, P. , Frazier J. A., Cochran D., Mitchell T., Coleman C., and Schmidt R. C.. 2018. “Relationship Between Theory of Mind, Emotion Recognition, and Social Synchrony in Adolescents With and Without Autism.” Frontiers in Psychology 9: 1337. 10.3389/fpsyg.2018.01337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Foster, T. R. , and Young R. L.. 2022. “Brief Report: Sentencing Outcomes for Offenders on the Autism Spectrum.” Journal of Autism and Developmental Disorders 52, no. 7: 3314–3320. 10.1007/s10803-021-05212-4. [DOI] [PubMed] [Google Scholar]
  21. Fridenson‐Hayo, S. , Berggren S., Lassalle A., et al. 2016. “Basic and Complex Emotion Recognition in Children With Autism: Cross‐Cultural Findings.” Molecular Autism 7: 52. 10.1186/s13229-016-0113-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Funk, F. , Walker M., and Todorov A.. 2017. “Modelling Perceptions of Criminality and Remorse From Faces Using a Data‐Driven Computational Approach.” Cognition and Emotion 31, no. 7: 1431–1443. 10.1080/02699931.2016.1227305. [DOI] [PubMed] [Google Scholar]
  23. George, M. , Guilfoyle J. R., Steele J. R., and Struthers C. W.. 2023. “A Sorry Excuse for an Apology: Examining People's Mental Representations of an Apologetic Face.” Journal of Nonverbal Behavior 47, no. 1: 57–81. 10.1007/s10919-022-00422-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hayes, A. F. 2023. “PROCESS Macro for SPSS and SAS (Version 4.3) [Computer Software].” https://www.processmacro.org.
  25. Heerey, E. , Keltner D., and Capps L. M.. 2003. “Making Sense of Self‐Conscious Emotion: Linking Theory of Mind and Emotion in Children With Autism.” Emotion 3, no. 4: 394–400. 10.1037/1528-3542.3.4.394. [DOI] [PubMed] [Google Scholar]
  26. Keltner, D. , Sauter D., Tracy J., and Cowen A.. 2019. “Emotional Expression: Advances in Basic Emotion Theory.” Journal of Nonverbal Behavior 43: 133–160. 10.1007/s10919-019-00293-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lee, S. B. , Koo S. J., Song Y. Y., et al. 2014. “Theory of Mind as a Mediator of Reasoning and Facial Emotion Recognition: Findings From 200 Healthy People.” Psychiatry Investigation 11, no. 2: 105–111. 10.4306/pi.2014.11.2.105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Leung, F. Y. N. , Stojanovik V., Micai M., Jiang C., and Liu F.. 2023. “Emotion Recognition in Autism Spectrum Disorder Across Age Groups: A Cross‐Sectional Investigation of Various Visual and Auditory Communicative Domains.” Autism Research 16, no. 4: 783–801. 10.1002/aur.2896. [DOI] [PubMed] [Google Scholar]
  29. Löytömäki, J. , Ohtonen P., Laakso M. L., and Huttunen K.. 2020. “The Role of Linguistic and Cognitive Factors in Emotion Recognition Difficulties in Children With ASD, ADHD or DLD.” International Journal of Language & Communication Disorders 55, no. 2: 231–242. 10.1111/1460-6984.12514. [DOI] [PubMed] [Google Scholar]
  30. Mackinnon, A. , and Christensen H.. 2007. “An Investigation of Measurement Properties of the Spot‐The‐Word Test in a Community Sample.” Psychological Assessment 19, no. 4: 459–468. 10.1037/1040-3590.19.4.459. [DOI] [PubMed] [Google Scholar]
  31. MacLin, M. K. , Downs C., MacLin O. C., and Caspers H. M.. 2009. “The Effect of Defendant Facial Expression on Mock Juror Decision‐Making: The Power of Remorse.” North American Journal of Psychology 11, no. 2: 323–332. [Google Scholar]
  32. Mao, S. Y. , Chiu H. M., Yu Y. T., and Chen K. L.. 2023. “The Associations of Theory of Mind With Both General and Theory‐Of‐Mind‐Related Social Interaction in Children With Autism Spectrum Disorder.” Research in Autism Spectrum Disorders 102: 102107. 10.1016/j.rasd.2023.102107. [DOI] [Google Scholar]
  33. Palmer, S. J. , Fanucci‐Kiss A., Kipervassar E., et al. 2023. “Effect of Emotional Valence on Emotion Recognition in Adolescents With Autism Spectrum Disorder.” Journal of Autism and Developmental Disorders 54, no. 4: 1494–1506. 10.1007/s10803-022-05831-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Pisani, S. , Murphy J., Conway J., Millgate E., Catmur C., and Bird G.. 2021. “The Relationship Between Alexithymia and Theory of Mind: A Systematic Review.” Neuroscience and Biobehavioral Reviews 131: 497–524. 10.1016/j.neubiorev.2021.09.036. [DOI] [PubMed] [Google Scholar]
  35. Stokkom, B. V. 2002. “Moral Emotions in Restorative Justice Conferences: Managing Shame, Designing Empathy.” Theoretical Criminology 6, no. 3: 339–360. 10.1177/136248060200600306. [DOI] [Google Scholar]
  36. Tang, J. S. , Chen N. T., Falkmer M., Bölte S., and Girdler S.. 2019. “Atypical Visual Processing but Comparable Levels of Emotion Recognition in Adults With Autism During the Processing of Social Scenes.” Journal of Autism and Developmental Disorders 49: 4009–4018. 10.1007/s10803-019-04104-y. [DOI] [PubMed] [Google Scholar]
  37. Walker, M. , Schönborn S., Greifeneder R., and Vetter T.. 2018. “The Basel Face Database: A Validated Set of Photographs Reflecting Systematic Differences in Big Two and Big Five Personality Dimensions.” PLoS One 13, no. 3: e0193190. 10.1371/journal.pone.0193190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Westby, C. , and Robinson L.. 2014. “A Developmental Perspective for Promoting Theory of Mind.” Topics in Language Disorders 34, no. 4: 362–382. 10.1097/TLD.0000000000000035. [DOI] [Google Scholar]
  39. White, S. J. , Coniston D., Rogers R., and Frith U.. 2011. “Developing the Frith‐Happé Animations: A Quick and Objective Test of Theory of Mind for Adults With Autism.” Autism Research 4, no. 2: 149–154. 10.1002/aur.174. [DOI] [PubMed] [Google Scholar]
  40. Williams, D. , and Happé F.. 2010. “Recognising ‘Social’ and ‘Non‐Social’ Emotions in Self and Others: A Study of Autism.” Autism 14, no. 4: 285–304. 10.1177/1362361309344849. [DOI] [PubMed] [Google Scholar]
  41. Wilson, A. C. 2021. “Do Animated Triangles Reveal a Marked Difficulty Among Autistic People With Reading Minds?” Autism 25, no. 5: 1175–1186. 10.1177/1362361321989152. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Yuspeh, R. L. , and Vanderploeg R. D.. 2000. “Spot‐The‐Word: A Measure for Estimating Premorbid Intellectual Functioning.” Archives of Clinical Neuropsychology 15, no. 4: 319–326. 10.1016/S0887-6177(99)00020-7. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1. Supporting Information.

AUR-18-604-s001.docx (882.9KB, docx)

Data Availability Statement

The data that support the findings of this study are openly available in Open Science Framework at https://www.osf.io/7jhp2/?view_only=f63cb4e100f6436ea7556fd8cc937ac4.

Articles from Autism Research are provided here courtesy of Wiley

RESOURCES