The biosocial correlates and predictors of emotion dysregulation in autistic adults compared to borderline personality disorder and nonclinical controls

doi:10.21203/rs.3.rs-3128484/v1

Download PDF

Research Article

The biosocial correlates and predictors of emotion dysregulation in autistic adults compared to borderline personality disorder and nonclinical controls

https://doi.org/10.21203/rs.3.rs-3128484/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 18 Dec, 2023

Read the published version in Molecular Autism →

You are reading this latest preprint version

Background

Emotion dysregulation (ED) is a core symptom of borderline personality disorder (BPD), whose etiology has been attributed to biosocial factors. In autism spectrum disorder (ASD), although ED is prevalent and is associated with decreased well-being (e.g., self-harm, suicidality), it has been understudied, especially in adults. It is therefore crucial to further understand ED in autistic adults to improve its treatment. Our study investigates ED, its behavioural correlates (e.g., self-harm, suicidality) and biosocial predictors in autistic adults relative to BPD and nonclinical controls (NC).

Methods

724 participants (ASD = 154; BPD = 111; NC = 459) completed 11 self-reported questionnaires assessing ED, ASD and BPD traits, co-occurring disorders, alexithymia, emotional vulnerability and invalidating experiences (e.g., bullying, autistic camouflaging). The occurrence of ED behavioural correlates (i.e., self-harm, history of suicide attempts, and psychiatric hospitalizations) was collected. In addition, between-groups analyses, linear regressions and machine learning models (ML) were used to identify ED predictors in each group.

Results

ED and its behavioural correlates were higher in ASD compared to NC, but milder than in BPD. While gender did not predict ED scores, autistic women had increased risk factors to ED, including sexual abuse and camouflaging. Interestingly, BPD traits, emotional vulnerability, and alexithymia strongly predicted ED scores across the groups. Using ML models, sensory sensitivity and autistic camouflaging were associated with ED in ASD, and ADHD symptoms with ED in BPD.

Limitations:

ASD and BPD diagnoses were self-reported, which did not allow us to check their accuracy. Additionally, we did not explore the transactional and the moderating/mediating relationships between the different variables. Moreover, our research is cross-sectional and cannot draw conclusions regarding the direction and causality of relationships between ED and other clinical dimensions.

Conclusions

ED and its behavioural correlates are heightened in BPD compared to ASD and nonclinical controls. In the ASD group, there were no gender differences in ED, despite the heightened exposure of autistic women to ED risk factors. BPD traits, emotional vulnerability, and alexithymia are core to ED regardless of diagnosis. Although less central, sensory sensitivity and autistic camouflaging seem to be specific predictors of ED in autistic adults.

Emotion dysregulation (ED) is defined as patterns of emotional experience or expression that interfere with goal-directed behaviour [1]. ED is recognized as a core etiological and maintenance mechanism of borderline personality disorder (BPD) [2, 3]. Recently, ED has been found to be a key transdiagnostic mechanism of psychopathology involved in the development and maintenance of several psychiatric disorders, such as depression, eating disorders and complex post-traumatic stress disorder (cPTSD) [4, 5, 6, 7, 8]. Interestingly, in recent years, ED has also become a central area of research in autism spectrum disorder (ASD) [9, 10]. Indeed, findings support that ED is more prevalent in autistic people compared to the general population [9, 10, 11, 12]. Of particular importance, recent findings have reported a high prevalence of self-harm [13, 14, 15] and suicidality in ASD [16, 17, 18], particularly in autistic adults without intellectual disability presenting with high levels of ED [19, 20, 21]. Indeed, similar to findings in BPD and in the general population [14, 22], ED has been associated with self-harm with or without suicidal intent in ASD [19, 21]. Interestingly, akin to BPD [22], self-harm and suicidal behaviour have also been found to be strongly linked in ASD [14, 23]. This suggests that ED is a risk factor for suicidality and self-harm in ASD, and that autistic people may develop capability for suicide through self-harm [14, 23]. Autistic women in particular have been reported to be at greater risk of developing severe ED compared to autistic men [24, 25, 26], which suggests that gender-related factors might be involved in ED in ASD [27].

Few studies have focused on interventions targeting ED in autistic adults. Pharmacological treatments have shown limited effectiveness in this context [28]. Therefore, there is a critical need to develop psychological interventions targeting ED in autistic people given the multitude of downstream effects on adaptive functioning and quality of life [29]. Interventions based on cognitive behavioural therapy (CBT) have shown encouraging outcomes [30]. This is particularly the case for dialectical behaviour therapy (DBT) [31, 32], the most studied treatment targeting ED in BPD [33, 22]. Recently, DBT has proved to be feasible and acceptable in autistic adults without intellectual disability, as well as potentially effective to reduce ED in the presence of self-harm and suicidal behaviour [31, 32].

Linehan’s biosocial theory, which underlies DBT, conceptualizes that biological and environmental factors in childhood are involved in the emergence of ED in BPD [33]. According to this theory, ED emerges from an interaction between: (a) emotional vulnerability that stems from biological factors impacting the functioning of brain areas (e.g., prefrontal regions and amygdala) involved in emotional processing, and (b) an invalidating environment that refers to chronic and inadequate responses of the environment to the emotional needs of the child (i.e., neglect, minimization or punishment, including physical and sexual abuse) [22, 33]. Temperamental impulsivity has been subsequently added to the model as an additional risk factor for BPD [22, 34]. According to Linehan, invalidation early in life maintains and may exacerbate the pre-existing biological vulnerability in the child [33]. It also shapes maladaptive coping responses, such as using self-harm with or without suicidal intent when experiencing emotional pain [22, 35]. Most empirical findings support the relevance of Linehan’s model to conceptualize ED in BPD [e.g., 36,37,38,39,40].

Although DBT has been adapted to and studied in clinical conditions other than BPD [e.g., 41,42,43], few studies have focused on the pertinence of Linehan's biosocial model to conceptualize ED beyond BPD. Nevertheless, there are findings that support the implication of emotional vulnerability [44], including temperamental impulsivity [45], and invalidation (e.g., childhood maltreatment) [46, 47, 48] in the development of ED across psychopathology.

In ASD, studies investigating the factors involved in the emergence of ED have mainly focused on ASD-related particularities (e.g., social cognition peculiarities, sensory sensitivity, cognitive inflexibility) [9, 49] and the role played by co-occurring disorders (e.g., anxiety, depression) [11, 20, 50, 51, 52]. Thus, to our knowledge, there is a lack of comprehensive models which integrate biological and psychosocial factors potentially involved in the emergence of ED in ASD [e.g., 21,53,54,55].

Based on the existing literature, including Mazefsky and White’s model for ED in autistic youth [49], we recently proposed an application of Linehan’s biosocial model to ED in ASD [56] to provide a comprehensive conceptualization of the biosocial factors (i.e., emotional vulnerability and invalidating environment) involved in the emergence of ED in ASD. More specifically, we considered the interplay between ASD traits and both emotional vulnerability and the experience of invalidation. Indeed, in addition to the biological vulnerability similar to that found in BPD (i.e., hypersensitivity, hyperreactivity and slow return to emotional baseline) [57], ToM peculiarities, sensory sensitivity, lack of cognitive flexibility, change-related anxiety and repetitive behaviour have been associated with ED in ASD [49, 11]. Alexithymia, prevalent in ASD, has also been reported to be linked to ED in autistic adults [20], especially in autistic women [27]. Moreover, ASD-related difficulties seem to interfere directly with the ability to self-regulate [10, 11, 49], but also contribute to high levels of anxiety and fatigue making emotion regulation costly for autistic people [11, 49, 58]. In terms of invalidating experiences, autistic children are highly exposed to different early stressful and traumatic experiences (e.g., physical and emotional maltreatment from caregivers and school bullying), because of their atypical functioning that cause misunderstanding and rejection from others [59, 60, 61]. Autistic girls seem to be particularly vulnerable to the experience of adverse events [62, 63]. Among other environmental factors potentially involved in the emergence of ED, lack of parental scaffolding and modeling (i.e., support provided by the parent to help the child regulate their emotions) have been pinpointed as risk factors for dysregulated behaviour in autistic youth [64, 65]. Additionally, recent studies have reported that autistic camouflaging, i.e., behaviour and/or strategies used to appear “less autistic”, is associated with lifetime suicidality, especially in autistic women [66, 67, 68]. Given this, our extension of the biosocial model to ASD includes excessive autistic camouflaging as a form of self-invalidation resulting from internalized invalidation from others [69].

The application of Linehan’s model to ASD requires an empirical test of its validity as well as an assessment of its specificity to ED in ASD, particularly in comparison with BPD. The latter point is crucial given that ED is highly associated with BPD [70] and that individuals with BPD may exhibit ASD-like traits that may lead to misdiagnosis with ASD [71]. For instance, some studies found that individuals with BPD might also have sensory sensitivity and social cognition peculiarities [72, 73]. Despite the overlap between ASD and BPD, studies comparing ED and its etiological factors in ASD and BPD are lacking.

The aim of the current study is to evaluate the relevance and the specificity of factors involved in Linehan’s model applied to ED in ASD [56]. To do so, autistic adults without intellectual disability (ASD group), adults with BPD (BPD group), and adults without any known diagnoses (nonclinical controls group; NC group) completed a battery of self-report scales measuring the model’s components and indicated the occurrences of the ED behavioural correlates (i.e., hospitalizations, self-harm and suicidal behaviour). We did not assess parental scaffolding as relevant measures are observational in the context of child-parent interaction [64, 74]. Specifically, we were interested in the characteristics of ED and its behavioural correlates in each clinical group. Given that ED is a core feature of BPD and that self-harm and suicidal behaviour are strongly associated with BPD [2, 3, 22], we hypothesized that ED and its behavioural correlates would be higher in the BPD group compared to the ASD group, whereas both clinical groups would have heightened ED scores compared to the NC group (H1). In addition, since self-harm has been associated with ED in BPD [22] and ASD [14, 20], we expected to find an association between self-harm and suicidal behaviour in both groups (H2). We also expected ED to be a predictor of the presence of self-harm and/or suicidal behaviour in both clinical groups (H3). Moreover, we hypothesized that ED would be heightened in the BPD group compared to the ASD group, whereas both clinical groups would have higher ED scores compared to the NC group (H4). Additionally, in the autistic group only, we expected to observe gender differences, with autistic women presenting with higher ED than autistic men (H5) [27]. Regarding the predictors of ED, we hypothesized that emotional vulnerability, impulsivity and invalidation – which are direct measures of the biosocial model – would predict ED in both clinical groups, but ASD-related factors would be specific predictors of ED in the ASD group compared to the BPD group (H6). To assess the specific load of the predictors of ED, machine learning (ML) models were used, and we expected emotional vulnerability and invalidation to emerge among the largest contributors of ED in both clinical groups, while ASD-related factors were expected to rank among the largest ED predictors for the ASD group only (H7).

Participants’ recruitment and study sample

This cross-sectional study was conducted online from January 2, 2023, to April 26, 2023. Data were collected anonymously through LimeSurvey using a 45-minute battery of standardized scales. The call for participants was advertised by different means: emails, leaflets distribution, poster display in mental health institutions, posts on social networks (Facebook, LinkedIn and Discord), targeting various communities: mental health professionals (psychologists and psychiatrists) in institutions or in private practice, researchers in the field of adult ASD and/or BPD, associations in the field of adult ASD or BPD, student communities, and people from the general population more broadly. The call for participants was advertised mainly in France.

The inclusion criteria were: being aged ≥ 18 years old and being fluent in French. For the ASD and BPD groups, having received a formal diagnosis by a physician was required. The co-occurrence of ASD and BPD was an exclusion criterion. For the no-diagnosis group (NC), participants were required to have no self-reported psychiatric or neurodevelopmental diagnosis.

The sample size has been estimated using G*Power 3.1.9.7 [75] for the three groups comparison on an effect size of f = 0.25, a power of 80%, assuming α = 0.05. Consequently, a minimum target sample was set at N = 125.

Measures

Emotion dysregulation

Difficulties in Emotion Regulation Scale-16 (DERS-16) [76]. The DERS-16 is a brief form of the 36-item DERS [77]. The DERS-16 is a self-report scale measuring emotion regulation difficulties. It consists of 16 items grouped into 5 dimensions: (a) lack of emotional clarity (Clarity), (b) difficulty engaging in goal-directed behaviour when distressed (Goals), (c) impulse control difficulties when distressed (Impulse), (d) limited access to strategies for regulation (Strategies), and (e) non-acceptance of emotional responses (Non-acceptance). Items are rated on a 5-point Likert scale (1= “almost never” to 5= “almost always”). Higher scores reflect greater emotion regulation difficulties. There is no validated French version of the DERS-16. However, we took the corresponding 16 items from the French-validated 36-item version by Dan-Glauser & Scherer [78]. The scale showed adequate internal consistency (Cronbach's α = .92) among a college sample [76]. In the present study, the internal consistency for the total DERS-16 was overall excellent for the total scale (Cronbach’s α = .95) and for the subscales (Cronbach’s α between .82 and .95).

ASD and BPD traits importance

Autism Spectrum Quotient Short Version (AQ-Short) [79]. The AQ-Short is a brief version of the AQ-50 [80]. The AQ-Short is a self-report scale composed of 28 items that assess core autistic traits in adults. The scale comprises 4 subscales: (a) social skills (b) routine, (c) attention switching, and (d) imagination. Items are rated on a 4-point Likert scale (1= “definitely agree” to 4= “definitely disagree”). Higher scores indicate higher level of autistic traits. There is no validated French version of the AQ-Short. However, we took the 28 corresponding items from the complete French version validated by Lepage et al. [81]. The AQ-Short total score internal consistency has been reported to be good (Cronbach’s α ranging between .77 and .86) [79]. In the present study, the AQ-Short internal consistency was very good (Cronbach’s α = .87).

Short form of the Borderline Symptom List (BSL-23) [82], French validation by Nicastro et al. [83]. BSL-23 is a short version of the BSL-95 [84]. BSL-23 is a self-report scale constituted of 23 items assesses the severity of BPD symptoms and behaviour. Each item is answered on a 5-point Likert scale (0= “not at all” to 4 = 1 “very strong”): 0–1 point refers to no BPD symptoms, 2–23 points to mild, 24–69 points to moderate, and 70–92 points to severe BPD symptoms. The BSL-23 internal consistency is excellent (Cronbach’s α = .94) [82]. In the present study, the internal consistency of the BSL-23 was excellent (Cronbach’s α = .97).

Co-occurring disorders

ADHD Self-Report Scale v1.1 screener (ASRS v1.1 Screener) [85], French validation by Caci et al. [86], is a 6-item self-report measure of attention deficit hyperactivity disorder (ADHD) symptoms in adults. In this study, the ASRS v1.1 was used as a measure of inattention and impulsivity, potentially involved in the biological component of the biosocial model. Items are rated on a 5-point Likert scale (0= “never” to 4= “very often”). Respondents who endorsed at least four out of six items are considered at “elevated” risk for ADHD. The ASRS v1.1 Screener has demonstrated good psychometric properties in studies with adults (Cronbach’s α between .63 and .72) [85, 87]. In the present study, the ASRS v1.1 Screener’s internal consistency was good (Cronbach’s α = .79).

Depression, Anxiety and Stress Scales (DASS-21) [88], French validation by Nahaboo [89] is a shortened version of the DASS-42 [90]. The DASS-21 is a self-report scale that assesses through 21 items the full range of core symptoms of three affective states: depression, anxiety and stress. Items are rated on a 4–point Likert scale (0= “did not apply to me at all” to 3= “applied to me very much”). Scores are summed and multiplied by 2 to create separate seven-item subscales for each dimension. The subscales’ internal consistency ranged between excellent and good in the validation study (Depression α = .91; Anxiety α = .84; Stress α = .90) [88]. In the present study, the DASS-21’s internal consistency was excellent (Depression α = .95) and ranged from very good to excellent for the subscales (Depression α = .93, Anxiety α = .86, Stress α = .93).

Emotional vulnerability

Emotional Vulnerability-Child Scale self-report (EV-Child) [91] is a self-report scale that retrospectively assesses emotional vulnerability during childhood. The scale includes items addressing Linehan’s [33] concept of slow return to emotional baseline. The scale encompasses 22 items rated on a 6-point Likert scale (1= “never” to 6= “always”). Higher scores indicate higher level of emotional vulnerability during childhood. The scale demonstrated excellent internal consistency in the validation study (Cronbach’s α = .92). In the present study, internal consistency was excellent (Cronbach’s α = .94).

ASD-related factors contributing to ED

Camouflaging Autistic Traits Questionnaire (CAT-Q) [92] is a 25-item, self-report questionnaire that assesses social camouflaging behaviour. Items are rated on a 7-point Likert scale (1= “Strongly disagree” to 7= “Strongly agree”). The scale contains 3 subscales: (a) Assimilation, i.e., strategies used to blend in during social situations, (b) Compensation, i.e., strategies to compensate for ASD-related communication and social difficulties, and (c) Masking, i.e., strategies to appear ‘non-autistic’ in social contexts. Higher scores indicated greater camouflaging. There is no French validation of the scale. However, we have used the scale translated and back translated by our team in an ongoing French validation study (Bureau et al., submitted). The CAT-Q has shown excellent internal consistency (Cronbach’s α = .94) [92]. In the current sample the internal consistency for the total scale was very good (Cronbach’s α = .86), good for the Compensation subscale (Cronbach’s α = .76), acceptable for Masking subscale (Cronbach’s α = .58) and poor for the Assimilation subscale (Cronbach’s α = .45).

Sensory Processing Sensitivity Questionnaire - Sensory Sensitivity subscale (SPSQ SS) [93] is 16-item, self-report scale that measures high sensitivity to various stimuli. The tool encompasses two subscales: (a) Sensory sensitivity (SS) which assesses sensory processing sensitivity) and (b) Other sensitivity (OS) which assesses sensitivity to emotions and various life experiences. Items are rated on an 11-point Likert scale (0= “compared to others” to 10= “much more sensitive than the people around me”). Higher scores indicated greater is the sensitivity. In the current study, we only used the SS subscale. The subscale has excellent internal consistency (Cronbach’s α = .89) [93]. In the current sample the internal consistency of the SS was acceptable (Cronbach's α = .67).

Eight‑item General Alexithymia Factor Score (GAFS-8) [94] is a self-report unidimensional scale that measures alexithymia using 8 items derived from the Toronto Alexithymia Scale (TAS-20) [95]. The selection of items (TAS-20 items: 1, 2, 6, 9, 11, 12, 13, and 14) has been found to be a more robust measure of alexithymia in both autistic and non-autistic samples [94]. Items are rated on a 5-point Likert scale (1= “Strongly Disagree” to 5= “Strongly A²gree”). Higher scores indicate higher level of alexithymia. As the measure is recent and has not yet been validated in French, we referred to the corresponding items in the French version of the TAS-20 [96]. In the present sample, the internal consistency of the GAFS-8 was excellent (Cronbach's α = .92).

Invalidation and adverse experiences

Childhood Trauma Questionnaire-Short Form (CTQ-SF) [97], French validation by Paquette et al. [98], is a 28-item, self-report questionnaire designed to assess five types of maltreatment during childhood: (a) physical abuse, (b) sexual abuse, (c) emotional abuse, (d) physical neglect, and (e) emotional neglect; with 5 items per scale (and 3 additional “minimization” items, which were not used in the present study). Each item is scored on a 5-point Likert (1= “never true” to 5= “very often true”). The total score is calculated as the sum of all items, with higher scores reflecting higher rates of childhood traumatic events. The internal reliability of the CTQ-SF is excellent for the total score (Cronbach’s α = .95), good to excellent for four dimensions (Cronbach’s α respectively .81– .86, .84– .89, .92– .95, and .88– .91) and acceptable for physical neglect (Cronbach’s alpha ranging from .61– .78) [97]. In the present sample, internal consistency for the CTQ-SF total scale was very good (Cronbach's α = .92), very good to excellent for 4 of the subscales (Cronbach's α ranging between .85 and .92) and moderate for Physical neglect (Cronbach's α = .68).

Assessment of Bullying Experiences (ABE) [99] is a questionnaire assessing bullying experiences including those unique of autistic and neurodivergent youth (e.g., verbal teasing about social differences) grouped in 4 subscales: (a) Verbal, (b) Physical, (c) Relational, and (d) Cyber bullying. The ABE comprises 22 items that ask parents to rate the frequency with which their child has experienced specific bullying behaviour on a 6-point Likert scale (0= “has never happened” to 5= “weekly or more”). In the current study we used the ABE as a self-report questionnaire and were interested only in the total score. In our sample, the internal consistency of the ABE was excellent (Cronbach's α = .94).

Socio-demographic and clinical data

The demographic characteristics of participants collected in this study were gender (female/male/other), age (years), country of residency, marital status, professional status, educational status, and living situation. The clinical data collected were the neurodevelopmental and/or psychiatric diagnoses received, whether the person had current psychotropic medication or not, as well as ED severity indicators: the presence or not of self-harming behaviour in the previous year, the presence or not of suicidal ideation, history of at least a suicide attempt, and history of hospitalization in psychiatry services.

Statistical analysis

Descriptive and comparative analyses

Descriptive and comparative statistical analyses were conducted using Jamovi 2.6.23 [100]. Descriptive results were expressed in means (M) and standard deviations (SD) for continuous variables and in numbers (n) and percentages for ordinal variables.

To investigate whether a significant interaction exists between the groups (ASD/BPD/NC) and each severity indicator of ED, we used the chi-square test of association [101]. The same test was used to investigate the association between self-harm and suicidal behaviour.

To analyse the predictive value of the DERS-16 total score for self-harm and suicidal behaviour in the ASD and BPD groups, we performed a bimodal logistic regression [102] with two categories (individuals with self-harming and/or suicidal behaviour and those without self-harming and/or suicidal behaviour).

As our data were not normally distributed, we used the Kruskal-Wallis one-way test [103], a non-parametric alternative to the one-way ANOVA, followed by the Steel-Dwass-Chritchlow-Fligner [104] post-hoc test, for pairwise comparisons to assess whether there were significant differences between groups on the scales. Epsilon squared (ε²) was used to calculate the effect size [105]. ε² ≤ 0.05 being considered as small effect size, between 0.06–0.13 as moderate effect size, and ≥ 0.14 as large effect size [106].

Linear regression analyses

To identify ED predictors in each group (ASD/BPD/NC), we conducted linear regression analyses. The following variables were included in the model: ASD, BPD and ADHD traits (assessing impulsivity), emotional vulnerability in childhood, autistic camouflaging, sensory processing particularities, alexithymia, childhood trauma, school bullying and gender for the autistic group (Woman/man). Only women and men were considered due to the small number of participants of other genders. The ordinary least squares (OLS) method was used to estimate the regression coefficients and determine the best-fit line that minimized the sum of squared residuals. We then extracted the model fits statistics, the estimated coefficients for each predictor, and information about the residuals that help interpret the regression analysis.

Machine learning models

To study the importance of the ED predictors, we trained a powerful ML model XGBoost [107] using each one of the three groups. We first separated the data into subsets of 80% for training, 10% for validation, and 20% for testing to prepare it for training. We then normalized the numerical attributes using standard scaling to ensure that all features contribute equally to the model's learning process.

We also encoded the categorical attributes using one-hot encoding to effectively represent them in a numerical format. To optimize the performance of the XGBoost model, we conducted hyperparameter tuning using a GridSearch [108]. This involved systematically exploring various combinations of hyperparameters and evaluating the model's performance using a suitable evaluation metric to find the optimal ones.

To assess the generalizability of the trained model, we employed cross-validation [109]. Cross-validation involves dividing the dataset into multiple subsets, or folds, and training the model on a subset while evaluating it on the remaining folds. By averaging the performance across all folds, we obtained a more robust estimate of the model's performance. Furthermore, we used SHapley Additive exPlanations (SHAP) [110] values to assess the significance of each feature in the XGBoost model. By calculating the relative contributions of each feature to the expected result, SHAP values offer a consistent measure of feature relevance. This study gives us insights into the underlying relationships and influences within the dataset and helps understand the relative impact of features on the model's predictions.

Statistical significance was set at p-values ≤ 0.05 and the trend towards significance threshold was set at 0.08.

Sample description

A total of 2151 individuals began the study and 1049 (49%) provided complete data. 724 (69%) met the inclusion criteria for one of the three groups (Fig. 1). 154 (21%) constituted the ASD group (mean age = 32.62 ± 12.05, range from 18 to 61, 58% female), 111 (15%) constituted the BPD group (mean age = 28.64 ± 8.97, range from 18 to 62, 78% female) and 459 (63%) constituted the NC group (mean age = 25.92 ± 9.43, range from 18 to 65, 75% female) (Table 1). Seventy-nine (11%) participants declared also having a formal diagnosis of ADHD (52 (34%) in the ASD group and 27 (24%) in the BPD group. Ninety-nine % of participants were from France. The remaining were from other French-speaking countries (2 from Belgium, 4 from Canada and 3 from Switzerland).

Table 1

Sample description

ED severity indicators

The BPD group had the highest occurrence of all ED severity indicators (Table 2). Indeed, the prevalence of self-harming behaviour (over the year prior to their participation in the study) in the BPD group was 82% compared to 51% in the ASD group and 35% in the NC group, with these differences being significant (Χ² p_BPD−ASD < 0.001, Χ² p_BPD−NC < 0.001); the lifetime occurrence of suicide attempts in the BPD group (77%) was significantly higher than in the ASD group (39%) and the NC group (12%) (Χ² p_BPD−ASD <0.001, Χ² p_BPD−NC < 0.001); the suicidal ideation occurrence was also higher in the BPD group (91%) compared to the ASD group (55%) and the NC group (36%) (Χ² p_BPD−ASD <0.001, Χ² p_BPD−NC < 0.001); and the history of hospitalization in psychiatric services was more frequent in the BPD group (64%) than in the ASD group (22%) and the NC group (3%) (Χ² p_BPD−ASD <0.001, Χ² p_BPD−NC < 0.001).The ASD group had a higher occurrence of all the ED severity indicators compared to the NC group (p_ASD−NC<0.001 for all indicators).

Table 2

ED severity indicators occurrence in each group and chi-square comparison of participant’s distribution between each pair of groups

In addition, a significant association between self-harming behaviour and suicidal behaviour was found in the three groups (Χ² p_BPD = 0.008; Χ² p_ASD = 0.007; Χ² p_NC <0.001) (Table 3).

Table 3

Chi-square test of association between self-harm and suicidal behaviour in the sample and in each group

Link between ED and self-harm and/or suicidal behaviour

The bimodal logistic regression analysis showed that the DERS-16 mean score significantly predicted the presence of self-harm and/or suicidal behaviour in both the ASD group (z = 3.700, p < 0.011) and the BPD group (z = 2.208, p = 0.027).

The DERS-16 prediction accuracy rates were 71% for the ASD group and 92% in the BPD group. The prediction accuracy rate of the DERS-16 is especially high for the presence of self-harm and or suicidal behaviour (90% for ASD and 100% for BPD) (Table 4).

Table 4

Association between DERS-16 total and self-harm and suicidal behaviours using bimodal logistic regression

Questionnaire scores’ comparisons between groups

Emotion dysregulation

The BPD group had significantly higher scores on the DERS-16 total scale and on its subscales than the ASD group, with large effect sizes, except for Clarity and Goals subscales (W = 11.75, p < 0.001 for the total; W = 5.56, p = 0.005 for Clarity; W = 12.84, p < 0.001 for Impulse; W = 9.72, p < 0.001 for Non-acceptance; W = 5.23, p = 0.011 for Goals; W = 11.00, p < 0.001 for Strategies) and the NC group (W = 17.03, p < 0.001 for the total; W = 13.29, p < 0.001 for Clarity; W = 17.25, p < 0.001 for Impulse; W = 13.69, p < 0.001 for Non-acceptance; W = 11.70, p < 0.001 for Goals; W = 16.75, p < 0.001 for Strategies). The DERS-16 total score was significantly higher in the ASD group compared to the NC group (W = 7.35, p < 0.001). This was also the case for the DERS-16 subscales (W = 9.10, p < 0.001 for Clarity; W = 5.70, p < 0.001 for Impulse; W = 3.48, p = 0.037 for Non-acceptance; W = 7.10, p < 0.001 for Goals; W = 6.86, p = 0.002 for Strategies) (Table 5). See additional material for the correlation analyses between the scales for each group [Additional file 1].

ASD and BPD traits

As expected, on the AQ-Short, the score of the ASD group was significantly higher than the BPD group, with large effect sizes (W= -12.68, p < 0.001) and the NC group (W = 20.26, p < 0.001), while the BPD group scored higher than the NC group (W = 6.20, p < 0.001). On the BSL-23, the BPD group score was significantly higher than the ASD group (W = 13.48, p < 0.001) and the N.C group (W = 17.72, p < 0.001) with large effect sizes, while the ASD group scored significantly higher than the NC group (W = 5.48, p < 0.001).

Co-occurring disorders

On the ASRS v1.1, there was no significant difference between the ASD and BPD groups (W = 1.43, p = 0.568), both scoring significantly higher than the NC group (W_ASD−NC= 7.14, p < 0.001; W_BPD−NC= 8.70, p < 0.001).

On the DASS-21, the score of the BPD group was significantly higher than that of the ASD group (W = 10.36, p < 0.001) and the NC group (W = 15.99, p < 0.001) with large effect sizes except. Specifically, the BPD group scored significantly higher on the Stress and Depression subscales (W = 9.96, p < 0.001 and W = 9.45, p < 0.001, respectively) compared to the ASD group, and significantly higher on all subscales compared to the NC group. The ASD group scored significantly higher than the NC group (W = 6.87, p < 0.001).

Emotional vulnerability

On the EV-Child, the score of the BPD group was significantly higher than the ASD group (W = 5.41, p < 0.001) and the NC group (W = 14.11, p < 0.001), respectively, with small and large effect sizes. The ASD group scored significantly higher than the NC group (W = 11.29, p < 0.001), with a moderate effect size.

ASD-related factors contributing to ED

On the CAT-Q (camouflaging), the ASD group had significantly higher scores than the BPD group (W= -5.35, p < 0.001) and the NC group (W = 13.44, p < 0.001), while the BPD group had a significantly higher score than the NC group (W = 5.86, p < 0.001). This was also the case for the Compensation subscale (W_ASD−BPD= -5.94, p < 0.001; W_ASD−NC= 14.41, p < 0.001; W_BPD−NC= 6.51, p < 0.001). On the Masking subscale, the ASD group had significantly higher scores than the BPD group (W= -5.25, p < 0.001) and the NC group (W = 9.15, p < 0.001), but no significant difference was found between the BPD and NC groups (W = 1.51, p = 0.536). There was no significant difference between the three groups on the Assimilation subscale.

On the SPSQ SS (sensory sensitivity) subscale score there was no significant difference between the ASD and BPD groups (W= -1.69, p = 0.456), both scoring significantly higher than the NC group (W_ASD−NC= -6.87, p < 0.001; W_BPD−NC= 4.25, p = 0.007).

Similarly, on the GAFS-8 (alexithymia), there was no significant difference between the ASD and BPD groups (W= -2.38, p = 0.211), both scoring significantly higher than the NC group (W_ASD−NC= 11.34, p < 0.001; W_BPD−NC= 7.31, p < 0.001).

Invalidation and adverse experiences

On the CTQ-SF, the score was significantly higher in the BPD group compared to the ASD group (W = 5.99, p < 0.001) and the NC group (W = 13.78, p < 0.001). This was also the case for all the CTQ-SF subscales except for Emotional neglect, although a trend towards significance was observed (W = 3.14, p = 0.068). The ASD group scored significantly higher than the NC group on the CTQ-SF total (W = 8.41, p < 0.001) and subscales.

On the ABE, there was no significant difference between the ASD and BPD groups (W= -0.21, p = 0.988), both scoring significantly higher than the NC group (W_ASD−NC= 12.93, p < 0.001; W_BPD−NC= 10.47, p < 0.001).

Table 5

Scales mean scores pairwise comparisons between the groups using the Kruskal-Wallis one-way test

Gender differences

When considering gender differences (Table 6), we did not include the non-binary (i.e., “other”) sub-group because of the small number of participants (n = 19).

Autistic women had a significantly higher DASS-21 score than the autistic men (W = -3.90, p = 0.016) and the BPD group (W= 8.32, p <0.001), especially regarding the Stress (W _ASD-BPD = -5.22, p <0.001 and W _ASD-NC= 7.01, p <0.001) and Anxiety subscales (W_ASD-BPD = -4.42, p= 0.005 and W _ASD-NC= 7.02, p <0.001).

On the EV-Child (emotional vulnerability), the score tended to be higher in autistic women compared to men (W = -3.27, p = 0.054). No significant difference was found relative to the BPD group, but there was a trend for a higher score in the BPD group compared to autistic women (W = 3.29, p = 0.053).

On the CAT-Q (measuring total camouflaging and Compensation subscale), autistic women scored significantly higher than ASD men (W = -4.01, p = 0.013) and the BPD group (W= -5.83, p<0.001), while no differences were found between ASD men and the BPD group (W= 1.63, p = 0.482). On the Assimilation subscale, ASD women scored significantly higher than ASD men (W= -3.97, p = 0.014) but similar to the BPD group (W= -1.49, p = 0.544).

Autistic women had a significantly higher SPSQ SS (sensory processing) scores than ASD men (W= -4.31, p = 0.006) and the BPD group (W= -3.91, p =0.016). There was no significant difference between autistic men and the BPD group on this scale (W= -1.22, p =0.663).

On the CTQ-SF (childhood maltreatment), the Sexual abuse subscale score was significantly higher in autistic women than autistic men (W = -5.00, p = 0.001) and comparable to the BPD group, although they tended to be lower than the latter group (W= 1.46, p = 0.056).

Predictors of ED

Linear regression models

Linear regression analyses showed three significant predictors of ED in the ASD group, with the regression model accounting for 54.6% of the variance (Table 7): BPD traits measured by the BSL-23 (t= 7.348, p <0.001), emotional vulnerability measured by the EV-Child (t= 3.682, p <0.001), and alexithymia measured by the GAFS-8 (t= 3.805, p <0.001). A trend towards significance was found on the SPSQ SS subscale measuring sensory sensitivity (t= 1.916, p= 0.057).

In the BPD group, the same three variables were found to be strong predictors of ED, with the regression model accounting for 39.6% of the variance: BPD traits (t= 3.900, p <0.001), emotional vulnerability (t= 2.386, p= 0.019), and alexithymia (t= 2.256, p= 0.026). A trend towards significance was found on the ASRS v1.1 Screener measuring ADHD symptoms (t= 1.851, p= 0.067).

In the NC group, six significant predictors of ED were found, with the regression model accounting for 66.8% of the variance: BPD traits (t= 13.383, p <0.001), ADHD symptoms (t= 3.054, p= 0.002), emotional vulnerability (t= 6.091, p <0.001), autistic camouflaging measured by the CAT-Q (t= -2.297, p= 0.022), sensory hypersensitivity (t= 2.603, p= 0.010), and alexithymia (t= 8.017, p <0.001).

ED Predictors rankings

Figure 2 presents the SHAP summary plots that display the importance of the 11 variables selected, the magnitude of their impact (i.e., the effect size) in each group, and the direction of a specific feature’s association with ED.

For the ASD group, BPD traits (BSL-23), alexithymia (GAFS-8), emotional vulnerability (EV-Child), sensory sensitivity (SPSQ SS), and autistic camouflaging (CAT-Q), respectively carried most of the general model’s predictive power with an absolute mean SHAP value of 0.37, 0.14, 0.13, 0.11 and 0.09 respectively.

For the BPD group, BPD traits, emotional vulnerability, alexithymia and ADHD traits (ASRS v1.1 Screener) respectively carried most of the general model’s predictive power with an absolute mean SHAP value of 0.28, 0.15, 0.13 and 0.10 respectively.

For the NC group, BPD traits, alexithymia, emotional vulnerability, ADHD traits, and sensory sensitivity respectively carried most of the general model’s predictive power with an absolute mean SHAP value of 0.40, 0.21, 0.15, 0.10, and 0.08 respectively.

For all groups, gender (Woman/Man) ranked at the bottom of the model, indicating a very low predictive value on ED scores.

Furthermore, the SHAP value range for the BSL-23 was wider in the ASD group compared to the BPD group, which means that the BPD traits predicted more strongly ED in the ASD group than in the BPD group.

Our study is the first to provide an investigation of ED in autistic adults compared to BPD and nonclinical controls. Our results suggest that ED scores and its behavioural correlates (i.e., suicidality, self-harming behaviour, and hospitalizations) are increased in ASD compared to ND. Nevertheless, people with BPD had the highest scores of ED and the greatest occurrences of its behavioural correlates in our sample. Interestingly, and contrary to our hypotheses, the same three dimensions, i.e., BPD traits, emotional vulnerability, and alexithymia, significantly predicted ED scores in both the ASD and BPD groups. Consistently, these three dimensions ranked as the greatest ED predictors in the ML models for both the ASD and BPD groups, whereas sensory sensitivity and autistic camouflaging were associated with ED in ASD, and ADHD symptoms with ED in BPD.

First of all, in line with our hypothesis (H4), we found that ED was higher in the BPD group than the ASD group, with both clinical groups scoring higher than the nonclinical controls. This result is not surprising given that ED is a core feature of BPD [2, 3], and that ED is typically considered as a co-occurring difficulty in autistic people [49]. Yet, unlike our results, recent findings by Weiner et al. [27] suggested that ED was heightened in autistic adults, particularly autistic women, compared to women with BPD. However, their sample of autistic people was recruited from a DBT waiting list, hence their results could not be generalized to autistic adults in other contexts. This is not the case for our results, as both the BPD and the ASD groups were recruited from the general population.

Relatedly, consistent with our hypothesis (H1), the prevalence of suicidal behaviour, use of psychiatry services and self-harm was higher in the BPD group compared to the ASD group, while these behaviours were more frequent in both clinical groups compared to the NC group. Interestingly, the rates of suicidal and self-harming behaviour found in the BPD and ASD groups are consistent with previous findings. Indeed, in their review, Oumaya et al. [22] reported that self-harm rates ranged between 50–80% and suicidality rates ranged between 40–85% in BPD, which match the 82% prevalence of self-harm and 77% of suicidal behaviour in our BPD group. In our ASD group, the rate of 41% of self-harm matches the 50% reported by Maddox et al. [19], and the rate of 39% of suicide attempts is congruent with the 35% rate found by Cassidy et al. [111]. These results add to previous research showing that the rates of self-harm and suicidal behaviour in autistic adults are higher than that of non-autistic people [13, 14]. Nevertheless, our study is the first to show that the prevalence of these behavioural ED correlates is decreased in ASD compared to BPD. These results are congruent with the ED levels found in our groups. Indeed, while ED scores are higher in the ASD group compared to the NC group, the BPD group displayed the highest ED scores in our sample. Moreover, consistent with our hypotheses (H2), self-harm and suicidal behaviour were strongly associated in each group and ED strongly predicted these behaviours in both clinical groups (H3). This is in line with numerous studies suggesting that there is a strong link between self-harm and suicidality [14, 23, 112], and that ED predicts these behaviours in both ASD and BPD [113, 114].

In terms of gender differences, contrary to our hypothesis (H5), autistic women did not show a heightened ED compared to autistic men and to the BPD group. Indeed, the BPD group scored the highest compared to both autistic women and men. This result is inconsistent with findings supporting a heightened ED in autistic women compared to both autistic men [24, 25, 27] and women with BPD [27]. However, studies comparing ED in autistic women and people with BPD are scarce [27] and those comparing ED across genders in autistic adults have focused on individuals attending psychiatric facilities, unlike our study [25, 27, 52]. Thus, compared to these findings, our results are probably more representative of ED levels across genders in the adult autistic population irrespective of their co-occurring disorders. It is noteworthy, however, that autistic women had higher levels of anxiety and stress compared to autistic men, but lower scores than the BPD group. This is in line with findings supporting an increased vulnerability to anxiety in autistic women compared to autistic men, that may be attributed to psychosocial (e.g., increased use of camouflaging) and biological (e.g., heightened sensory sensitivity) factors [24, 115, 116].

As expected, the analyses between groups showed that the ASD group had the highest ASD traits (measured by the AQ-Short), while the BPD group showed the highest BPD traits (measured by the BSL-23). However, the ASD group scored higher than the NC group on the BSL-23 and the BPD group scored higher than the NC group on the AQ-Short, which might reflect the overlapping features between ASD and BPD [117, 118]. ADHD traits were equivalent between the ASD and BPD groups, but the ADHD co-occurrence was higher in ASD compared to BPD (34% in the ASD group and 24% in the BPD group). This could be due to an increased awareness among clinicians of the high co-occurrence between ADHD and ASD and their shared neurodevelopmental nature, which might increase the likelihood of screening for ADHD in autistic people [119, 120]. Moreover, alexithymia scores were found to be similar in our clinical groups, but also across genders in the ASD group. Hence, although alexithymia has been strongly linked to ASD traits [121], it can be rather seen as a transdiagnostic feature also found in BPD [122].

Regarding factors specifically related to Linehan's biosocial model, the BPD group showed higher scores on the scales assessing emotional vulnerability and invalidating experiences compared to the ASD group. This is in line with the fact that ED is central in BPD, and the transaction between these components is recognized as predictive of ED in this disorder [33, 22]. Nevertheless, autistic women showed similar emotional vulnerability scores compared to the BPD group and tended to present with higher levels of emotional vulnerability relative to autistic men. The rates of sexual abuse in autistic women were also found to be equivalent to the BPD group in our study, and both were higher than those reported by autistic men and the NC group. These results add to findings pointing to an increased prevalence of sexual abuse among autistic women [123, 124, 125]. In fact, Cazalis et al. [123] reported a 2 to 3-fold increase of victimization in autistic women compared to non-autistic women, as well as high rates of revictimization. This, in turn, is associated with a higher risk to develop mental health issues, including anxiety disorders and PTSD [123, 126]. Additionally, autistic women showed higher scores on both the camouflaging and the sensory processing scales compared to both autistic men and the BPD group. However, no significant difference was found in sensory processing between autistic men and the BPD group. Given that sensory sensitivity has been linked to daily psychophysiological arousal and increased anxiety, this might partially explain why autistic women tended to present with higher emotional vulnerability and anxiety scores relative to autistic men in our study [23, 127].

Interestingly, while autistic women presented with increased emotional vulnerability and sensory sensitivity (biological component), and higher rates of sexual abuse and camouflaging (psychosocial component) than autistic men, it is worth noting that ED scores were equivalent between the two genders. This can be explained by the fact that autistic women may have increased emotion regulation abilities relative to autistic men which might compensate for their heightened exposure to ED risk factors and to anxiety. Consistent with this hypothesis, autistic women have been reported to have increased social skills [128] and greater use of autistic camouflaging [67, 68] than autistic men, which could underlie a greater ability to engage in goal-directed behaviour that may be useful for emotion regulation although the selected goals, e.g., camouflaging, might be detrimental to their well-being on the long run [67, 68]. Previous data has reported an increased use of emotion regulation strategies in NC women compared to men, while ED is equally associated with psychopathology in both genders [129].

Considering the predictors of ED, we found that BPD traits, emotional vulnerability, and alexithymia predicted ED in both clinical groups. Consistently, the SHAP value plots showed that the BPD traits were the main ED predictor across groups (ASD/BPD/NC), followed by emotional vulnerability and alexithymia. These results suggest that ED is central to BPD as conceptualized by Linehan’s theory [33]. Additionally, they indicate that BPD traits, emotional vulnerability and alexithymia might be central to ED, irrespective of the diagnosis. Interestingly, BPD traits were a stronger predictor of ED in the ASD group compared to the BPD group with a wider range of SHAP value. This can be explained by the fact that the BSL-23 (measuring BPD traits) might lack discriminative and predictive power in the BPD group as BPD traits are intrinsic to BPD. Conversely, in ASD, it is likely that BPD traits may refer to the overlapping difficulties between ASD and BPD (e.g., social communication peculiarities, disturbed sense of self) [117, 118]. Moreover, emotional vulnerability was identified as one of the three strongest ED predictor in both groups, suggesting a strong biological basis underlying ED regardless of diagnosis. Alexithymia ranked second in the SHAP value plots for ASD, while emotional vulnerability ranked second for BPD. Thus, although alexithymia has been pinpointed by several studies as a transdiagnostic process involved in ED [e.g., 27,130], this suggests that alexithymia might be particularly central to ED in ASD. ASD-related factors, on the other hand, did not predict ED in the ASD group compared to the BPD group, contrary to our hypotheses (H6 and H7). However, sensory sensitivity showed a trend towards significance in the linear regression model in the ASD group and ranked fourth in the SHAP value plot. This suggest that sensory processing particularities may be the ASD-related factor that contributes the most to ED in ASD, which is congruent with previous findings in autistic youth [20, 131]. It is noteworthy that autistic camouflaging ranked as the fifth strongest ED predictor in the ASD group. In the BPD group, ADHD symptoms ranked as the fourth predictor in the SHAP value plot, which is consistent with the fact that impulsivity is a core feature of BPD [132].

Furthermore, contrary to our hypotheses (H6 and H7), the childhood invalidation measures, including early trauma (CTQ-SF) and bullying (ABE), did not emerge as significant ED predictors in the clinical groups, which is inconsistent with Linehan’s model [33]. Another study also failed to identify parental invalidation as BPD predictors and suggested that it was irrelevant for the model [36], while Keng & Soh [40] suggested that maternal invalidation contributed to BPD. IT is noteworthy that Linehan [33] considered that invalidation included several forms of emotional invalidation (i.e., minimization, punishment, ignoring the emotional experience) with a focus on parental invalidation that was not tackled by the measures we used (i.e., CTQ-SF and the ABE). Moreover, in Linehan’s theory it is rather the transaction between the two components that is key to ED, and this was not assessed in our study [33].

Our study has a number of limitations. First, the sample size (N = 724) was limited, especially since the participants were divided into three groups. Using a larger sample would have also allowed to optimize our predictive models for ED, especially those based on ML as large data sets are supported to improve the accuracy of such models [133, 134]. Second, as the study was conducted online, the ASD and BPD diagnoses were self-reported, which did not allow us to check their accuracy. However, it was stated in the online form that a formal diagnosis provided by a psychiatrist was required. Third, we assessed self-harming and did not distinguish skin-cutting from other forms of self-harm. Given that skin-cutting in particular is has been associated with severe ED [20] and with suicidal behaviour [135], it would be relevant to distinguish between different types of self-harming behaviour in future studies. Fourth, all the study measures were self-reported. Indeed, self-report scales have numerous limitations, including relying on the participant’s ability to accurately self-assess their functioning [136]. This might be particularly challenging in studies on ED since it might be accompanied by high levels of alexithymia, which was the case in our study. Fifth, although we have opted for a short version for each scale, we had many questionnaires, which probably explains the high dropout rate in our study (51%). Reducing the number of measures could lighten the burden on participants, reduce the dropout rate and potentially improve the quality of the data collected [137]. Sixth, in our study we did not explore the transactional and the potential moderating/mediating relationships between the model’s components and factors. Indeed, this study aimed to provide initial results of ED comparing ED correlates between BPD, ASD and a nonclinical group. These results may pave the way to further investigate how the different ED correlates identified for ASD interact. Seventh, our cross-sectional findings prevent any conclusions regarding the direction and causality of relationships between ED and its correlates; longitudinal research is warranted to address the latter. Finally, we measured the psychosocial factors of the model through the CTQ-SF (childhood trauma) and the ABE (school bullying), whereas there are tools that may be more appropriate to measure invalidation as conceptualized by Linehan [33], such as the Socialization of Emotion Scale [91].

To conclude, ED scores are higher in ASD compared to nonclinical controls, but milder than in BPD. While gender did not predict ED scores in our sample, autistic women had increased risk factors to ED relative to autistic men (i.e., emotional vulnerability, sexual abuse, sensory particularities, autistic camouflaging and anxiety). This suggests that it is crucial to consider gender-related factors potentially involved in ED in ASD in future studies. Importantly, the same three dimensions, i.e., BPD traits, alexithymia and emotional vulnerability, seem to be involved in ED across the clinical groups, suggesting that they might be key to ED irrespective of diagnosis. However, sensory processing and autistic camouflaging may be more specific to ED in ASD, while ADHD symptoms may play a specific role in ED in BPD. Given that the same three main contributors to ED were found in ASD and BPD, our results outline that DBT, built upon Linehan’s biosocial model, is likely to be as relevant to treat ED in autistic adults as it is in BPD.

ABE	Assessment of bullying experiences
ADHD	Attention deficit hyperactivity disorder
AQ-Short	Autism spectrum quotient short version
ASD	Autism spectrum disorder
ASRS v1.1 Screener	ADHD self-report scale v1.1 screener
BPD	Borderline personality disorder
BSL-23	Short form of the borderline symptom list
CAT-Q	Camouflaging autistic traits questionnaire
CBT	Cognitive behavioural therapy
cPTSD	Complex post-traumatic stress disorder
CTQ-SF	Childhood trauma questionnaire-Short form
DASS-21	Depression, anxiety and stress scales
DBT	Dialectical behaviour therapy
DERS-16	Difficulties in emotion regulation scale-16
EV-Child	Emotional vulnerability-Child scale
GAFS-8	Eight‑item general alexithymia factor score
M	Mean
NC	Nonclinical controls
ML	Machine learning
OLS	Ordinary least square
SD	Standard deviation
SPSQ SS	Sensory processing sensitivity questionnaire - Sensory sensitivity subscale
SHAP	Shapley additive explanation

Ethics approval and consent to participate

The study was approved by the Research Ethics Committee of the University of Strasbourg (Ethical approval number: Unistra/CER/2022-35). The participants provided informed consent to participate by checking the box intended for this purpose prior to accessing the questionnaires. All study procedures complied with the ethics code outlined in the Declaration of Helsinki.

The study has been preregistered prior to data analysis in the Open Science Framework (OSF; registration DOI: https://doi.org/10.17605/OSF.IO/U8WY6).

Consent for publication

Not applicable.

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare that they have no competing interests.

Funding

This study is part of the first author’s PhD research funded by Fonds de Dotation Neuroglia.

Author contributions

DB performed conceptualization of the project, data collection and analysis, and writing of manuscript. AL performed and analysed linear regression and machine learning predictions, and wrote the corresponding methods and results. LW performed conceptualization of the project, provided supervision, reviewed and edited the drafts and the final manuscript. All authors read and approved the final manuscript.

Acknowledgements

We thank all the participants who took part in the study. We also thank Fonds de Dotation Neuroglia for funding the first author’s PhD work.

Beauchaine TP. Future Directions in Emotion Dysregulation and Youth Psychopathology. J Clin Child Adolesc Psychol. 2015;44(5):875-896.
Daros AR, Williams GE. A Meta-analysis and Systematic Review of Emotion-Regulation Strategies in Borderline Personality Disorder. Harv Rev Psychiatry. 2019;27(4):217-232.
Carpenter RW, Trull TJ. Components of emotion dysregulation in borderline personality disorder: a review. Curr Psychiatry Rep. 2013;15(1):335.
Paulus FW, Ohmann S, Möhler E, Plener P, Popow C. Emotional Dysregulation in Children and Adolescents With Psychiatric Disorders. A Narrative Review. Front Psychiatry. 2021;12:628252.
Weissman DG, Bitran D, Miller AB, Schaefer JD, Sheridan MA, McLaughlin KA. Difficulties with emotion regulation as a transdiagnostic mechanism linking child maltreatment with the emergence of psychopathology. Dev Psychopathol. 2019;31(3):899-915.
Sloan E, Hall K, Moulding R, Bryce S, Mildred H, Staiger PK. Emotion regulation as a transdiagnostic treatment construct across anxiety, depression, substance, eating and borderline personality disorders: A systematic review. Clin Psychol Rev. 2017;57:141-163.
Dvir Y, Ford JD, Hill M, Frazier JA. Childhood maltreatment, emotional dysregulation, and psychiatric comorbidities. Harv Rev Psychiatry. 2014;22(3):149-161.
McLaughlin KA, Hatzenbuehler ML, Mennin DS, Nolen-Hoeksema S. Emotion dysregulation and adolescent psychopathology: a prospective study. Behav Res Ther. 2011;49(9):544-554.
Conner CM, Golt J, Shaffer R, Righi G, Siegel M, Mazefsky CA. Emotion Dysregulation is Substantially Elevated in Autism Compared to the General Population: Impact on Psychiatric Services. Autism Res. 2021;14(1):169-181.
Cai RY, Richdale AL, Uljarević M, Dissanayake C, Samson AC. Emotion regulation in autism spectrum disorder: Where we are and where we need to go. Autism Res. 2018;11(7):962-978.
Samson AC, Phillips JM, Parker KJ, Shah S, Gross JJ, Hardan AY. Emotion dysregulation and the core features of autism spectrum disorder. J Autism Dev Disord. 2014;44(7):1766-1772.
Swain D, Scarpa A, White S, Laugeson E. Emotion Dysregulation and Anxiety in Adults with ASD: Does Social Motivation Play a Role?. J Autism Dev Disord. 2015;45(12):3971-3977.
Licence L, Oliver C, Moss J, Richards C. Prevalence and Risk-Markers of Self-Harm in Autistic Children and Adults. J Autism Dev Disord. 2020;50(10):3561-3574.
Moseley RL, Gregory NJ, Smith P, Allison C, Baron-Cohen S. Links between self-injury and suicidality in autism. Mol Autism. 2020;11(1):14.
Steenfeldt-Kristensen C, Jones CA, Richards C. The Prevalence of Self-injurious Behaviour in Autism: A Meta-analytic Study. J Autism Dev Disord. 2020;50(11):3857-3873.
Richa S, Fahed M, Khoury E, Mishara B. Suicide in autism spectrum disorders. Arch Suicide Res. 2014;18(4):327-339.
Hedley D, Uljarević M. Systematic Review of Suicide in Autism Spectrum Disorder: Current Trends and Implications. Curr Dev Disord Rep. 2018;5:65-76.
Dell'Osso L, Carpita B, Muti D, et al. Mood symptoms and suicidality across the autism spectrum. Compr Psychiatry. 2019;91:34-38.
Maddox BB, Trubanova A, White SW. Untended wounds: Non-suicidal self-injury in adults with autism spectrum disorder. Autism. 2017;21(4):412-422.
Moseley RL, Gregory NJ, Smith P, Allison C, Baron-Cohen S. A 'choice', an 'addiction', a way 'out of the lost': exploring self-injury in autistic people without intellectual disability. Mol Autism. 2019;10:18.
Conner CM, Golt J, Righi G, Shaffer R, Siegel M, Mazefsky CA. A Comparative Study of Suicidality and Its Association with Emotion Regulation Impairment in Large ASD and US Census-Matched Samples. J Autism Dev Disord. 2020;50(10):3545-3560.
Oumaya M, Friedman S, Pham A, Abou Abdallah T, Guelfi JD, Rouillon F. Personnalité borderline, automutilations et suicide : revue de la littérature [Borderline personality disorder, self-mutilation and suicide: literature review]. Encephale. 2008;34(5):452-458.
Moseley RL, Gregory NJ, Smith P, Allison C, Cassidy S, Baron-Cohen S. Non-suicidal self-injury and its relation to suicide through acquired capability: investigating this causal mechanism in a mainly late-diagnosed autistic sample. Mol Autism. 2022;13(1):45.
Sáez-Suanes GP, García-Villamisar D, Pozo Armentia AD. Does the gender matter?: Anxiety symptoms and emotion dysregulation in adults with autism and intellectual disabilities. Autism Res. 2023;16(1):113-121.
Wieckowski AT, Luallin S, Pan Z, Righi G, Gabriels RL, Mazefsky C. Gender Differences in Emotion Dysregulation in an Autism Inpatient Psychiatric Sample. Autism Res. 2020;13(8):1343-1348.
Hirvikoski T, Mittendorfer-Rutz E, Boman M, Larsson H, Lichtenstein P, Bölte S. Premature mortality in autism spectrum disorder. Br J Psychiatry. 2016;208(3):232-238.
Weiner L, Costache ME, Bemmouna D, et al. Emotion dysregulation is heightened in autistic females: A comparison with autistic males and borderline personality disorder. Womens Health (Lond). 2023;19:17455057231174763. doi:10.1177/17455057231174763
Salazar de Pablo G, Pastor Jordá C, Vaquerizo-Serrano J, et al. Systematic Review and Meta-analysis: Efficacy of Pharmacological Interventions for Irritability and Emotional Dysregulation in Autism Spectrum Disorder and Predictors of Response. J Am Acad Child Adolesc Psychiatry. 2023;62(2):151-168.
Beck KB, Conner CM, Breitenfeldt KE, Northrup JB, White SW, Mazefsky CA. Assessment and Treatment of Emotion Regulation Impairment in Autism Spectrum Disorder Across the Life Span: Current State of the Science and Future Directions. Child Adolesc Psychiatr Clin N Am. 2020;29(3):527-542.
Reyes NM, Pickard K, Reaven J. Emotion regulation: A treatment target for autism spectrum disorder. Bull Menninger Clin. 2019;83(3):205-234.
Bemmouna D, Coutelle R, Weibel S, Weiner L. Feasibility, Acceptability and Preliminary Efficacy of Dialectical Behaviour Therapy for Autistic Adults without Intellectual Disability: A Mixed Methods Study. J Autism Dev Disord. 2022;52(10):4337-4354.
Ritschel LA, Guy L, Maddox BB. A pilot study of dialectical behaviour therapy skills training for autistic adults. Behav Cogn Psychother. 2022;50(2):187-202.
Linehan, MM. Cognitive-behavioural treatment of borderline personality disorder. Guilford Press; 1993.
Crowell SE, Beauchaine TP, Linehan MM. A biosocial developmental model of borderline personality: Elaborating and extending Linehan's theory. Psychol Bull. 2009;135(3):495-510.
Brereton A, McGlinchey E. Self-harm, Emotion Regulation, and Experiential Avoidance: A Systematic Review. Arch Suicide Res. 2020;24(sup1):1-24.
Reeves M, James LM, Pizzarello SM, Taylor JE. Support for Linehan's biosocial theory from a nonclinical sample. J Pers Disord. 2010;24(3):312-326.
Gill D, Warburton W. An investigation of the biosocial model of borderline personality disorder. J Clin Psychol. 2014;70(9):866-873.
Kuo JR, Linehan MM. Disentangling emotion processes in borderline personality disorder: physiological and self-reported assessment of biological vulnerability, baseline intensity, and reactivity to emotionally evocative stimuli. J Abnorm Psychol. 2009;118(3):531-544.
Bortolla R., Cavicchioli M., Galli M. et al. A comprehensive evaluation of emotional responsiveness in borderline personality disorder: a support for hypersensitivity hypothesis. Bord personal disord emot dysregul. 2019;6:8.
Keng SL, Soh C. Association between childhood invalidation and borderline personality symptoms: self-construal and conformity as moderating factors. Bord personal disord emot dysregul. 2018;5:19.
Neacsiu AD, Eberle JW, Kramer R, Wiesmann T, Linehan MM. Dialectical behaviour therapy skills for transdiagnostic emotion dysregulation: a pilot randomized controlled trial. Behav Res Ther. 2014;59:40-51. doi:10.1016/j.brat.2014.05.005
Safer DL, Telch CF, Agras WS. Dialectical behaviour therapy for bulimia nervosa. Am J Psychiatry. 2001;158(4):632-634. doi:10.1176/appi.ajp.158.4.632
Fleming AP, McMahon RJ, Moran LR, Peterson AP, Dreessen A. Pilot randomized controlled trial of dialectical behaviour therapy group skills training for ADHD among college students. J Atten Disord. 2015;19(3):260–271. doi: 10.1177/1087054714535951
Calkins SD, Dollar JM, Wideman L. Temperamental vulnerability to emotion dysregulation and risk for mental and physical health challenges. Dev Psychopathol. 2019;31(3):957-970.
Beauchaine TP. Physiological Markers of Emotional and Behavioural Dysregulation in Externalizing Psychopathology. Monogr Soc Res Child Dev. 2012;77(2):79-86.
Rüfenacht E, Pham E, Nicastro R, et al. Link between History of Childhood Maltreatment and Emotion Dysregulation in Adults Suffering from Attention Deficit/Hyperactivity Disorder or Borderline Personality Disorder. Biomedicines. 2021;9(10):1469. Published 2021 Oct 14.
Gruhn MA, Compas BE. Effects of maltreatment on coping and emotion regulation in childhood and adolescence: A meta-analytic review. Child Abuse Negl. 2020;103:104446.
Guérin-Marion C, Bureau JF, Lafontaine MF, Gaudreau P, Martin J. Profiles of Emotion Dysregulation Among University Students Who Self-Injure: Associations with Parent-Child Relationships and Non-Suicidal Self-Injury Characteristics. J Youth Adolesc. 2021;50(4):767-787.
Mazefsky CA, White SW. Emotion regulation: concepts & practice in autism spectrum disorder. Child Adolesc Psychiatr Clin N Am. 2014;23(1):15-24.
Mazefsky CA, Herrington J, Siegel M, et al. The role of emotion regulation in autism spectrum disorder. J Am Acad Child Adolesc Psychiatry. 2013;52(7):679-688.
Keluskar J, Reicher D, Gorecki A, Mazefsky C, Crowell JA. Understanding, Assessing, and Intervening with Emotion Dysregulation in Autism Spectrum Disorder: A Developmental Perspective. Child Adolesc Psychiatr Clin N Am. 2021;30(2):335-348.
Northrup JB, Patterson MT, Mazefsky CA. Predictors of Severity and Change in Emotion Dysregulation among Children and Adolescents with ASD. J Clin Child Adolesc Psychol. 2021;50(6):708-729.
McDonnell CG, Boan AD, Bradley CC, Seay KD, Charles JM, Carpenter LA. Child maltreatment in autism spectrum disorder and intellectual disability: results from a population-based sample. J Child Psychol Psychiatry. 2019;60(5):576-584.
Mandell DS, Walrath CM, Manteuffel B, Sgro G, Pinto-Martin JA. The prevalence and correlates of abuse among children with autism served in comprehensive community-based mental health settings. Child Abuse Negl. 2005;29(12):1359-1372.
Ting V, Weiss JA. Emotion Regulation and Parent Co-Regulation in Children with Autism Spectrum Disorder. J Autism Dev Disord. 2017;47(3):680-689.
Bemmouna D. Weiner L. Linehan’s Biosocial Model applied to Emotion Dysregulation in Autistic Adults: A Narrative Review of the Literature and an Illustrative Case Conceptualization. [Manuscript submitted for publication]
White SW, Mazefsky CA, Dichter GS, Chiu PH, Richey JA, Ollendick TH. Social-cognitive, physiological, and neural mechanisms underlying emotion regulation impairments: understanding anxiety in autism spectrum disorder. Int J Dev Neurosci. 2014;39:22-36. doi:10.1016/j.ijdevneu.2014.05.012
Normansell-Mossa KM, Top DN Jr, Russell N, Freeston M, Rodgers J, South M. Sensory Sensitivity and Intolerance of Uncertainty Influence Anxiety in Autistic Adults. Front Psychol. 2021;12:731753. Published 2021 Nov 16. doi:10.3389/fpsyg.2021.731753
Warrier V, Baron-Cohen S. Childhood trauma, life-time self-harm, and suicidal behaviour and ideation are associated with polygenic scores for autism. Mol Psychiatry. 2021;26(5):1670-1684. doi:10.1038/s41380-019-0550-x
Hellström L. A Systematic Review of Polyvictimization among Children with Attention Deficit Hyperactivity or Autism Spectrum Disorder. Int J Environ Res Public Health. 2019;16(13):2280.
Fuld S. Autism Spectrum Disorder: The Impact of Stressful and Traumatic Life Events and Implications for Clinical Practice. Clin Soc Work J. 2018;46(3):210-219.
Fisher MH, Epstein RA, Urbano RC, Vehorn A, Cull MJ, Warren Z. A population-based examination of maltreatment referrals and substantiation for children with autism spectrum disorder. Autism. 2019;23(5):1335-1340. doi:10.1177/1362361318813998
Bargiela S, Steward R, Mandy W. The Experiences of Late-diagnosed Women with Autism Spectrum Conditions: An Investigation of the Female Autism Phenotype. J Autism Dev Disord. 2016;46(10):3281-3294. doi:10.1007/s10803-016-2872-8
Fenning RM, Baker JK, Moffitt J. Intrinsic and Extrinsic Predictors of Emotion Regulation in Children with Autism Spectrum Disorder. J Autism Dev Disord. 2018;48(11):3858-3870.
Baker JK, Fenning RM, Crnic KA, Baker BL, Blacher J. Prediction of social skills in 6-year-old children with and without developmental delays: contributions of early regulation and maternal scaffolding. Am J Ment Retard. 2007;112(5):375-391.
Cassidy SA, Gould K, Townsend E, Pelton M, Robertson AE, Rodgers J. Is Camouflaging Autistic Traits Associated with Suicidal Thoughts and Behaviour? Expanding the Interpersonal Psychological Theory of Suicide in an Undergraduate Student Sample. J Autism Dev Disord. 2020;50(10):3638-3648.
Beck JS, Lundwall RA, Gabrielsen T, Cox JC, South M. Looking good but feeling bad: "Camouflaging" behaviour and mental health in women with autistic traits. Autism. 2020;24(4):809-821.
Lai MC, Lombardo MV, Ruigrok AN, et al. Quantifying and exploring camouflaging in men and women with autism. Autism. 2017;21(6):690-702.
McQuaid GA, Lee NR, Wallace GL. Camouflaging in autism spectrum disorder: Examining the roles of sex, gender identity, and diagnostic timing. Autism. 2022;26(2):552-559. doi:10.1177/13623613211042131
Chapman AL. Borderline personality disorder and emotion dysregulation. Dev Psychopathol. 2019;31(3):1143-1156.
Dell'Osso L, Cremone IM, Carpita B, et al. Correlates of autistic traits among patients with borderline personality disorder. Compr Psychiatry. 2018;83:7-11.
Brown S, Shankar R, Smith, K. Borderline personality disorder and sensory processing impairment. Prog Neurol Psychiatry. 2009;13(4):10-16.
Vegni N, D'Ardia C, Torregiani G. Empathy, Mentalization, and Theory of Mind in Borderline Personality Disorder: Possible Overlap With Autism Spectrum Disorders. Front Psychol. 2021;12:626353.
Myruski S, Dennis-Tiwary TA. Observed parental spontaneous scaffolding predicts neurocognitive signatures of child emotion regulation. Int J Psychophysiol. 2022;177:111-121.
Faul F, Erdfelder E, Buchner A, Lang AG. Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behav Res Methods. 2009;41(4):1149-1160.
Bjureberg J, Ljótsson B, Tull MT, et al. Development and Validation of a Brief Version of the Difficulties in Emotion Regulation Scale: The DERS-16. J Psychopathol Behav Assess. 2016;38(2):284-296.
Gratz KL, Roemer L.. Multidimensional assessment of emotion regulation and dysregulation: Development, factor structure, and initial validation of the difficulties in emotion regulation scale. J Psychopathol Behav Assess. 2004;26:41-54.
Dan-Glauser ES, Scherer KR. The difficulties in emotion regulation scale (DERS): factor structure and consistency of a French translation. Swiss J Psychol Z Für Psychol Suisse Psychol. 2013;72:5-11.
Hoekstra RA, Vinkhuyzen AA, Wheelwright S, et al. The construction and validation of an abridged version of the autism-spectrum quotient (AQ-Short). J Autism Dev Disord. 2011;41(5):589-596.
Baron-Cohen S, Wheelwright S, Skinner R, Martin J, Clubley E. The autism-spectrum quotient (AQ): evidence from Asperger syndrome/high-functioning autism, males and females, scientists and mathematicians [published correction appears in J Autism Dev Disord 2001 Dec;31(6):603]. J Autism Dev Disord. 2001;31(1):5-17.
Lepage JF, Lortie M, Taschereau-Dumouchel V, Théoret H. Validation of French-Canadian versions of the empathy quotient and autism spectrum quotient. Can J Behav Sci. 2009;41(4):272–276.
Bohus M, Kleindienst N, Limberger MF, et al. The short version of the Borderline Symptom List (BSL-23): development and initial data on psychometric properties. Psychopathology. 2009;42(1):32-39.
Nicastro R, Prada P, Kung AL, et al. Psychometric properties of the French borderline symptom list, short form (BSL-23). Borderline Personal Disord Emot Dysregul. 2016;3:4.
Bohus M, Limberger MF, Frank U, Chapman AL, Kühler T, Stieglitz RD. Psychometric properties of the Borderline Symptom List (BSL). Psychopathology. 2007;40(2):126-132.
Kessler RC, Adler L, Ames M, et al. The World Health Organization Adult ADHD Self-Report Scale (ASRS): a short screening scale for use in the general population. Psychol Med. 2005;35(2):245-256.
Caci H, Didier C, Wynchank D. Validation and bifactor structure of the French Adult ADHD Symptoms Rating Scale v1.1 (ASRS) [published online ahead of print, 2023 Jan 12]. Encephale. 2023;S0013-7006(22)00274-3.
Adler LA, Spencer T, Faraone SV, et al. Validity of pilot Adult ADHD Self- Report Scale (ASRS) to Rate Adult ADHD symptoms. Ann Clin Psychiatry. 2006;18(3):145-148.
Lovibond PF, Lovibond SH. Manual for the Depression Anxiety Stress Scales. Sydney: Psychology Foundation; 1995.
Ramasawmy S. Hicks R. Gilles, P-Y. Développement des échelles de dépression, anxiété et stress (DASS 21) pour la population mauricienne. In P-Y Gilles, & Carlier, M. (dir.), Vive(nt) les différences. Psychologie différentielle fondamentale et applications (pp.). Aix-en-Provence : Presses universitaires de Provence; 2013.
Lovibond PF, Lovibond SH. The structure of negative emotional states: comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and Anxiety Inventories. Behav Res Ther. 1995;33(3):335-343.
Sauer SE, Baer RA. Validation of measures of biosocial precursors to borderline personality disorder: childhood emotional vulnerability and environmental invalidation. Assessment. 2010;17(4):454-466.
Hull L, Mandy W, Lai MC, et al. Development and Validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q). J Autism Dev Disord. 2019;49(3):819-833.
Malinakova K, Novak L, Trnka R, Tavel P. Sensory Processing Sensitivity Questionnaire: A Psychometric Evaluation and Associations with Experiencing the COVID-19 Pandemic. Int J Environ Res Public Health. 2021;18(24):12962.
Williams ZJ, Gotham KO. Improving the measurement of alexithymia in autistic adults: a psychometric investigation of the 20-item Toronto Alexithymia Scale and generation of a general alexithymia factor score using item response theory. Mol Autism. 2021;12(1):56.
Bagby RM, Parker JD, Taylor GJ. The twenty-item Toronto Alexithymia Scale--I. Item selection and cross-validation of the factor structure. J Psychosom Res. 1994;38(1):23-32.
mLoas G, Otmani O, Verrier A, Fremaux D, Marchand MP. Factor analysis of the French version of the 20-Item Toronto Alexithymia Scale (TAS-20). Psychopathology. 1996;29(2):139-144.
Bernstein DP, Stein JA, Newcomb MD, et al. Development and validation of a brief screening version of the Childhood Trauma Questionnaire. Child Abuse Negl. 2003;27(2):169-190.
Paquette D, Laporte L, Bigras M, Zoccolillo M. Validation de la version française du CTQ et prévalence de l'histoire de maltraitance [Validation of the French version of the CTQ and prevalence of the history of maltreatment]. Sante Ment Que. 2004;29(1):201-220.
Morton HE, Gillis JM, Zale EL, Brimhall KC, Romanczyk RG. Development and Validation of the Assessment of Bullying Experiences Questionnaire for Neurodivergent Youth. J Autism Dev Disord. 2022;52(11):4651-4664.
The Jamovi Project Jamovi. (Version 1.1) 2019. https://www.jamovi.org Retrieved from. (computer software)
Pearson K. X. on the criterion that a given system of deviations from the probable in the case of a correlated system of variables is such that it can be reasonably supposed to have arisen from random sampling. London Edinburgh Dublin Philos. Mag. J. Sci. 1900;50(302):157-175.
Cox DR. The regression analysis of binary sequences. Journal of the Royal Statistical Society. Series B (Methodological). 1958;20(2):215-242.
Kruskal WH, Wallis WA. Use of ranks in one-criterion variance analysis. J Am Stat Assoc. 1952;47:583–621.
Crichtlow DE, Fligner MA. On distribution-free multiple comparisons in the one-way analysis of variance. Communications in Statistics—Theory and Methods. 1991;20(1):127.
Kelley TL. An Unbiased Correlation Ratio Measure. Proc Natl Acad Sci U S A. 1935;21(9):554-559. doi:10.1073/pnas.21.9.554
Cohen J. A power primer. Psychol Bull. 1992;112(1):155-159. doi: 10.1037//0033-2909.112.1.155
Chen T, Guestrin C. XGBoost: A Scalable Tree Boosting System. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD ’16), San Francisco, CA, August 13–17, 2016; Association for Computing Machinery: New York, NY, U.S.A., 2016; pp 785–794. doi: 10.1145/2939672.2939785
Bergstra J, Bengio Y. Random search for hyper-parameter optimization. Journal of MachineLearning, 2012;13:281–305.
Kohavi R. A study of cross-validation and bootstrap for accuracy estimation and model selection. In Proc. 14th Int. Joint Conf. on AI, 1995, pp. 1137–1145.
Lundberg SM, Lee Su-In. "A unified approach to interpreting model predictions", In Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS'17). Red Hook: Curran Associates Inc.; 2017. p. 4768–77.
Cassidy S, Bradley P, Robinson J, Allison C, McHugh M, Baron-Cohen S. Suicidal ideation and suicide plans or attempts in adults with Asperger's syndrome attending a specialist diagnostic clinic: a clinical cohort study. Lancet Psychiatry. 2014;1(2):142-147.
Alberdi-Páramo I, Díaz-Marsá M, Sáiz González MD, Carrasco Perera JL. Association between non-suicidal self-injury and suicidal behaviour in Borderline Personality Disorder: a retrospective study. Actas Esp Psiquiatr. 2021;49(5):199-204.
Wolff JC, Thompson E, Thomas SA, et al. Emotion dysregulation and non-suicidal self-injury: A systematic review and meta-analysis. Eur Psychiatry. 2019;59:25-36. doi:10.1016/j.eurpsy.2019.03.004
Gratz KL, Kiel EJ, Mann AJD, Tull MT. The prospective relation between borderline personality disorder symptoms and suicide risk: The mediating roles of emotion regulation difficulties and perceived burdensomeness. J Affect Disord. 2022;313:186-195. doi:10.1016/j.jad.2022.06.066
Murray C, Kovshoff H, Brown A, Abbott P, Hadwin JA. Exploring the anxiety and depression profile in individuals diagnosed with an autism spectrum disorder in adulthood. Research in Autism Spectrum Disorders. 2019;58:1–8. doi:10.1016/j.rasd.2018.11.002
Uljarević M, Hedley D, Rose-Foley K, et al. Anxiety and Depression from Adolescence to Old Age in Autism Spectrum Disorder. J Autism Dev Disord. 2020;50(9):3155-3165. doi:10.1007/s10803-019-04084-z
May T, Pilkington PD, Younan R, Williams K. Overlap of autism spectrum disorder and borderline personality disorder: A systematic review and meta-analysis. Autism Res. 2021;14(12):2688-2710. doi:10.1002/aur.2619
Dudas RB, Lovejoy C, Cassidy S, Allison C, Smith P, Baron-Cohen S. The overlap between autistic spectrum conditions and borderline personality disorder [published correction appears in PLoS One. 2018 Jan 2;13(1):e0190727]. PLoS One. 2017;12(9):e0184447. Published 2017 Sep 8. doi:10.1371/journal.pone.0184447
Hours C, Recasens C, Baleyte JM. ASD and ADHD Comorbidity: What Are We Talking About?. Front Psychiatry. 2022;13:837424. Published 2022 Feb 28. doi:10.3389/fpsyt.2022.837424
Bemmouna D, Weibel S, Kosel M, Hasler R, Weiner L, Perroud N. The utility of the autism-spectrum quotient to screen for autism spectrum disorder in adults with attention deficit/hyperactivity disorder. Psychiatry Res. 2022;312:114580. doi:10.1016/j.psychres.2022.114580
Poquérusse J, Pastore L, Dellantonio S, Esposito G. Alexithymia and Autism Spectrum Disorder: A Complex Relationship [published correction appears in Front Psychol. 2018 Sep 05;9:1638]. Front Psychol. 2018;9:1196. Published 2018 Jul 17. doi:10.3389/fpsyg.2018.01196
Edwards ER, Rose NLJ, Gromatsky M, et al. Alexithymia, Affective Lability, Impulsivity, and Childhood Adversity in Borderline Personality Disorder. J Pers Disord. 2021;35(Suppl A):114-131. doi:10.1521/pedi_2021_35_513
Cazalis F, Reyes E, Leduc S, Gourion D. Evidence That Nine Autistic Women Out of Ten Have Been Victims of Sexual Violence. Front Behav Neurosci. 2022;16:852203. Published 2022 Apr 26. doi:10.3389/fnbeh.2022.852203
Roberts AL, Koenen KC, Lyall K, Robinson EB, Weisskopf MG. Association of autistic traits in adulthood with childhood abuse, interpersonal victimization, and posttraumatic stress. Child Abuse Negl. 2015;45:135-142. doi:10.1016/j.chiabu.2015.04.010
Ohlsson Gotby V, Lichtenstein P, Långström N, Pettersson E. Childhood neurodevelopmental disorders and risk of coercive sexual victimization in childhood and adolescence - a population-based prospective twin study. J Child Psychol Psychiatry. 2018;59(9):957-965. doi:10.1111/jcpp.12884
Hailes HP, Yu R, Danese A, Fazel S. Long-term outcomes of childhood sexual abuse: an umbrella review. Lancet Psychiatry. 2019;6(10):830-839. doi:10.1016/S2215-0366(19)30286-X
Kuiper MWM, Verhoeven EWM, Geurts HM. Stop Making Noise! Auditory Sensitivity in Adults with an Autism Spectrum Disorder Diagnosis: Physiological Habituation and Subjective Detection Thresholds. J Autism Dev Disord. 2019;49(5):2116-2128. doi:10.1007/s10803-019-03890-9
Head AM, McGillivray JA, Stokes MA. Gender differences in emotionality and sociability in children with autism spectrum disorders. Mol Autism. 2014;5(1):19. Published 2014 Feb 28. doi:10.1186/2040-2392-5-19
Nolen-Hoeksema S. Emotion regulation and psychopathology: the role of gender. Annu Rev Clin Psychol. 2012;8:161-187. doi:10.1146/annurev-clinpsy-032511-143109
Derks YPMJ, Westerhof GJ, Bohlmeijer ET. A Meta-analysis on the Association Between Emotional Awareness and Borderline Personality Pathology. J Pers Disord. 2017;31(3):362-384. doi:10.1521/pedi_2016_30_257
Duerden EG, Oatley HK, Mak-Fan KM, et al. Risk factors associated with self-injurious behaviour in children and adolescents with autism spectrum disorders. J Autism Dev Disord. 2012;42(11):2460-2470. doi:10.1007/s10803-012-1497-9
Linhartová P, Látalová A, Barteček R, et al. Impulsivity in patients with borderline personality disorder: a comprehensive profile compared with healthy people and patients with ADHD. Psychol Med. 2020;50(11):1829-1838. doi:10.1017/S0033291719001892
Goretzko D, Bühner M. One model to rule them all? Using machine learning algorithms to determine the number of factors in exploratory factor analysis. Psychol Methods. 2020;25(6):776-786. doi:10.1037/met0000262
Jacobucci R, Grimm KJ. Machine Learning and Psychological Research: The Unexplored Effect of Measurement. Perspect Psychol Sci. 2020;15(3):809-816. doi:10.1177/1745691620902467
Carroll R, Thomas KH, Bramley K, et al. Self-cutting and risk of subsequent suicide. J Affect Disord. 2016;192:8-10. doi:10.1016/j.jad.2015.12.007
Yates N, Gough S, Brazil V. Self-assessment: With all its limitations, why are we still measuring and teaching it? Lessons from a scoping review. Med Teach. 2022;44(11):1296-1302. doi:10.1080/0142159X.2022.2093704
Aiyegbusi OL, Roydhouse J, Rivera SC, et al. Key considerations to reduce or address respondent burden in patient-reported outcome (PRO) data collection. Nat Commun. 2022;13(1):6026. Published 2022 Oct 12. doi:10.1038/s41467-022-33826-4

Tables are available in Supplementary Files section.

No competing interests reported.

Download PDF

Journal Publication

published 18 Dec, 2023

Read the published version in Molecular Autism →

Editorial decision: Major revision
04 Sep, 2023
Reviews received at journal
11 Aug, 2023
Reviewers agreed at journal
18 Jul, 2023
Reviewers invited by journal
14 Jul, 2023
Editor assigned by journal
06 Jul, 2023
Submission checks completed at journal
03 Jul, 2023
First submitted to journal
30 Jun, 2023

You are reading this latest preprint version

The biosocial correlates and predictors of emotion dysregulation in autistic adults compared to borderline personality disorder and nonclinical controls

Status:

Journal Publication

Version 1

Abstract

Background

Methods

Results

Limitations:

Conclusions

Figures

Background

Methods

Participants’ recruitment and study sample

Measures

Emotion dysregulation

ASD and BPD traits importance

Co-occurring disorders

Emotional vulnerability

ASD-related factors contributing to ED

Invalidation and adverse experiences

Socio-demographic and clinical data

Statistical analysis

Descriptive and comparative analyses

Linear regression analyses

Machine learning models

Results

Sample description

Sample description

ED severity indicators

Link between ED and self-harm and/or suicidal behaviour

Association between DERS-16 total and self-harm and suicidal behaviours using bimodal logistic regression

Questionnaire scores’ comparisons between groups

Emotion dysregulation

ASD and BPD traits

Co-occurring disorders

Emotional vulnerability

ASD-related factors contributing to ED

Invalidation and adverse experiences

Scales mean scores pairwise comparisons between the groups using the Kruskal-Wallis one-way test

Discussion

Limitations

Conclusions

Abbreviations

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1