Alexithymia and eating disorder symptoms: the mediating role of emotion regulation

ABSTRACT

Objective

Preliminary research has found evidence to suggest alexithymia acts as a risk factor for the development of psychopathology symptoms, due to its impairing role on emotion regulation ability. Eating disorder symptoms have been extensively linked to high levels of alexithymia and emotion regulation difficulties, yet little is known about the precise mechanisms behind these interactions. The current study investigates whether emotion regulation difficulties drive the association between alexithymia and eating disorder symptoms.

Method

An online questionnaire battery consisting of psychometric self-report measures was administered to 255 undergraduate students. Measures included the Perth Alexithymia Questionnaire (PAQ), Perth Emotion Regulation Competency Inventory (PERCI), and the Eating Disorder Examination Questionnaire (EDE-Q).

Results

A parallel mediation analysis was conducted, finding that the relationship between alexithymia and eating disorder symptoms was partially mediated by difficulties regulating positive emotions.

Conclusions

These findings support contemporary theory, which recognises alexithymia as an important risk factor for psychopathology symptoms due to its impairing effect on emotion regulation ability. These results highlight the importance of considering the relationship between alexithymia and emotion regulation when conceptualising cases and planning eating disorder treatment and prevention measures.

Key Points

What is already known about the topic:

1. The attention-appraisal model posits that alexithymia is associated with emotion-based psychopathology symptoms due to alexithymia’s impairing effect on emotion regulation ability.

2. Alexithymia is extensively linked with eating disorders.

3. Emotion regulation difficulties are a transdiagnostic feature among eating disorders.

What this topic adds:

1. Difficulty regulating positive emotions appears to drive, at least in part, the relationship between alexithymia and eating disorder symptoms.

2. Our results highlight the importance of considering both alexithymia and emotion regulation when understanding eating disorder symptoms.

3. Alexithymia may interfere with eating disorder treatment due to its detrimental effect on ability to regulate positive emotions.

KEYWORDS:

• Alexithymia
• emotion regulation
• eating disorders

Introduction

Alexithymia is a multidimensional personality trait that is normally distributed throughout the population, whereby individuals high in the trait experience emotions in a less nuanced or differentiated manner compared to those low in alexithymia (Taylor et al., 1997). Alexithymia is comprised of at least three core components: (1) difficulty identifying one’s own feelings (DIF), (2) difficulty describing feelings (DDF), (3) and an externally orientated thinking style (EOT), which is an exaggerated focus on external events or stimuli as opposed to internal emotional states and sensations (Nemiah & Sifneos, 1970; Preece, Becerra, Robinson, Allan et al., 2020). Evidence suggests the emotional deficits associated with alexithymia place individuals at a greater risk of developing clinical symptoms (Leweke et al., 2011; Taylor et al., 1997), thus explaining its presence as a transdiagnostic feature across a variety of clinical populations including patients with depression (Bamonti et al., 2010; Honkalampi et al., 2001), personality disorders (Nicolò et al., 2011), substance abuse disorders (Hamidi et al., 2010; Thorberg et al., 2008), and eating disorders (Nowakowski et al., 2013; Westwood et al., 2017). Although the association between alexithymia and clinical disorders has been well established in previous literature, few studies have attempted to investigate the precise mechanisms underlying these relationships.

The current study investigates the relationship between alexithymia and eating disorder symptoms due to their high prevalence and relatively poor therapeutic outcomes in contrast to other disorders (Wilson et al., 2007). The aim of our study is to help address some of the abovementioned gaps, by testing a recently introduced hypothesis that in emotion-based psychopathologies, like eating disorders, alexithymia may be a risk factor for psychopathology because it impairs emotion regulation abilities (Preece, Becerra, Allan, Robinson & Dandy 2017; Preece, Mehta et al., 2022).

Theoretical framework

The attention-appraisal model of alexithymia posits that clinical symptoms may manifest in cases of high alexithymia due, at least in part, to the impairing effect of alexithymia on emotion regulation ability (Preece, Becerra, Allan, Robinson & Dandy 2017; Preece, Mehta et al., 2022). The model is situated within broader emotion regulation theory, specifically Gross’s (1998, 2015) process model of emotion regulation which explains the process whereby individuals alter the dynamics of their emotions.

According to Gross (2015), acts of emotion regulation occur via a series of dynamic and interacting valuation systems that can be subdivided into four stages: situation, attention, appraisal, and response. Emotions are regulated when an emotion is present (situation stage), one then focuses attention on the emotion (attention stage), appraises what the emotion is and what it means for their goals (appraisal stage), and based on that appraisal, one might activate a goal to try to regulate the unfolding emotion (response stage; i.e., emotion regulation). The attention-appraisal model of alexithymia proposes that alexithymia involves deficits at the attention and appraisal stages of the emotion regulation process (see Figure 1). More specifically, externally orientated thinking affects a person’s ability to focus attention on their emotional states (attention stage), and difficulties identifying and describing emotions affect a person’s ability to accurately appraise their emotions (appraisal stage) (Preece, Becerra, Allan, Robinson & Dandy 2017).

Figure 1. Visual representation of attention-appraisal model of alexithymia.

Components of alexithymia: EOT (externally orientated thinking), DIF (difficulty identifying feelings), and DDF (difficulty describing feelings) mapped onto the attention and appraisal stages of Gross’s (2015) process model of emotion regulation.

By disrupting these fundamental attention and appraisal stages of the emotion regulation process, alexithymia should therefore impair the effectiveness of subsequent emotion regulation decisions at the response stage (Preece, Mehta et al., 2022). Indeed, there is a growing body of literature robustly linking alexithymia to emotion regulation difficulties, such as reliance on less effective or avoidant emotion regulation strategies (Preece, Becerra, Robinson & Gross 2020; Stasiewicz et al., 2011; Swart et al., 2009; Venta et al., 2013). Given the close established relationships between emotion dysregulation and a wide variety of psychopathology categories, these links between alexithymia and emotion regulation are therefore of substantial interest. In line with the above theorising, a recent mediation study found that the link between alexithymia and affective disorder symptoms (depression and anxiety) was fully mediated by emotion regulation abilities (Preece, Mehta et al., 2022).

Alexithymia, emotion regulation, and eating disorders

Alexithymia levels are substantially elevated in eating disorder populations compared to healthy controls, a pattern observed across many eating disorder subtypes including anorexia nervosa, bulimia nervosa, and binge-eating disorder (for reviews see, Nowakowski et al., 2013; Westwood et al., 2017). This pattern also extends to non-clinical populations, in which individuals with higher levels of eating disorder symptomatology tend to report higher levels of alexithymia (De Berardis et al., 2007). Furthermore, higher alexithymia levels have been associated with less successful therapeutic outcomes and eating disorder recovery rates (Pinna et al., 2014).

Similarly, it is well established that affective disturbance is commonly present in eating disorder presentations. Emotion regulation difficulties, such as a lack of problem-solving and reappraisal strategies, have been recognised as a transdiagnostic feature across eating disorder subtypes (see for review, Oldershaw et al., 2015; Prefit et al., 2019). In addition, prominent theoretical models conceptualise eating disorder behaviours (e.g., purging, binge eating, restricted food intake, excessive exercise) as maladaptive regulatory techniques that are employed to manage uncontrollable or unpleasant emotional states (Fairburn et al., 2003; Taylor et al., 1997). Some theorists also suggest individuals with eating disorders hold beliefs that emotions are unacceptable, leading to secondary emotions of guilt or shame which consequently manifest as eating disorder symptoms (Corstorphine, 2006; Goss & Allan, 2009).

However, to date, few studies have specifically investigated how alexithymia and emotion regulation might jointly interact in eating disorder populations. Sfärlea et al. (2019) found that within an adolescent sample of girls diagnosed with eating disorders, alexithymia levels predicted the use of more maladaptive emotion regulation strategies and less frequent use of adaptive emotion regulation strategies. Furthermore, Brown et al. (2018) documented that within an adult sample of anorexic patients, alexithymia predicted levels of emotion regulation difficulties at discharge, suggesting that alexithymia may interfere with the efficacy of treatment programs. Within non-clinical populations, similar findings have been established (e.g., Whiteside et al., 2007). A recent inpatient study demonstrated that despite remaining clinically significant, reduced alexithymia levels were associated with reductions in emotion regulation difficulties and eating disorder symptoms following treatment (Meneguzzo et al., 2022). The findings of these studies appear to conceptually align with the attention-appraisal model, whereby alexithymia is theorised to disrupt the ability to successfully regulate emotions, which may then put people at risk of clinical symptoms characterised by affective disturbance (Preece, Mehta et al., 2022).

While there exists an expansive body of literature documenting the relationships between eating disorders, alexithymia, and emotion regulation, few studies have directly tested how these variables interact in a single model, and whether the link between alexithymia and eating disorder symptoms might be explained via the impairing effect of alexithymia on emotion regulation. So far, two studies to date have found evidence to suggest that the relationship between alexithymia and eating disorder symptoms may indeed be mediated by emotion regulation difficulties (Goetz et al., 2020; Khodabakhsh et al., 2017). However, these studies have some key limitations, including Khodabakhsh et al. (2017) use of the Difficulties in Emotion Regulation Scale (DERS: Gratz & Roemer, 2008) to operationalise emotion regulation difficulties. Use of the DERS in the alexithymia field is somewhat problematic because it is based on a broad definition of emotion dysregulation that also includes alexithymia components (labelled as the clarity and awareness subscales in the DERS) within its emotion regulation score. Thus, alexithymia is effectively represented twice in the model in such studies, once under the label of alexithymia, and once combined under the label of emotion regulation, confounding the capacity to cleanly examine these relationships. Additionally, the scales used in the abovementioned studies do not distinguish between positive (i.e., happiness and joy) and negatively valenced emotions (i.e., sadness and anger) when assessing emotion regulation ability.

Research in this field to date overwhelmingly focuses on the role of only negative emotions. However, different psychopathologies are often characterised by valence specific deficits; for example, in bipolar disorder, positive emotion dysregulation is a major risk factor for manic episodes (Gruber, 2011). Despite the known benefits of positive psychology, only a small range of studies have investigated the role of positive emotions in the context of disordered eating (Tchanturia et al., 2015). Although limited, these studies have reported interesting findings which suggest positive emotions may play a role in the production and maintenance of eating disorder behaviours. For example, Gruber et al. (2011) have noted that high levels of positive emotion can lead to an increased engagement in riskier behaviour in eating disorder and substance abuse populations. Furthermore, Selby et al. (2015) found that anorexic individuals with higher levels of both positive and negative emotional instability reported more frequent weight loss behaviours, such as laxative misuse and frequent weigh-ins. Another study conducted by Selby et al. (2014) found that a low ability to differentiate between emotions predicted increased weight loss behaviours in an anorexic sample. Considering these findings, it could be proposed that during states of heightened or unregulated positive emotions, individuals may struggle to control or inhibit disordered eating behaviours. Selby et al. (2014) suggest that positive emotions may reinforce eating disorder behaviours, when they provide individuals with a sense of weight loss success or control. It is apparent that further research into positive emotions may enhance the current understanding of eating disorders, and provide valuable treatment insight (Tchanturia et al., 2015)

The current study

The aim of the current study was to investigate whether emotion regulation difficulties (assessed across both negative and positive emotions) mediate the association between alexithymia and eating disorder symptoms (see Figure 2 for specified mediation model). We predicted that alexithymia would impair an individual’s ability to successfully regulate their emotions, which would in turn predict a higher level of eating disorder symptoms (i.e., a significant indirect effect).

Figure 2. β coefficients for PAQ mediation model where c’ represents direct effect between PAQ and EDE-Q scores.

***p < .001, **p < .01, *p < .05

Method

Participants and procedure

Ethics approval was granted by the University of Western Australia Human Research Ethics Committee (approval number: 2021/ET000282). A total of 255 undergraduate psychology students completed a series of psychometric measures in an online survey using Qualtrics software. The survey was advertised to students enrolled in undergraduate psychology units via SONA system, an online system used to manage and schedule experiments. Participants were reimbursed with unit credit in exchange for their voluntary completion of the survey. The questionnaire battery contained standardised self-report measures of alexithymia (Perth Alexithymia Questionnaire), emotion regulation ability (Perth Emotion Regulation Competency Inventory), and eating disorder symptomatology (Eating Disorder Examination Questionnaire), as well as demographic questions assessing age, gender, and ethnicity. Height and weight estimates provided from the Eating Disorder Examination Questionnaire were used to calculate body mass index (BMI: kg/m²). Participant age ranged from 18–62 years (M = 20.71, SD = 5.23), with 69.4% of the sample identifying as female and 30.6% as male. A majority of the sample reported their ethnicity as either White/Caucasian (50.2%), or Asian (41.6%). The average BMI of the sample was 23.13 (SD = 4.88).

Materials

Perth Alexithymia Questionnaire (PAQ)

The PAQ (D. Preece et al., 2018) is a 24-item self-report measure designed to assess alexithymia. Items include statements such as, “When I’m feeling bad, I can’t find the right words to describe those feelings”, which were responded to on a 7-point Likert scale ranging from “strongly disagree” to “strongly agree”. Several subscales can be derived from the PAQ, including DIF, DDF, and EOT composite scores which correspond to the three facets of alexithymia. The current study used the total scale score as an overall marker of alexithymia, for which higher scores are indicative of higher levels of alexithymia. The scale has previously demonstrated good reliability and validity (D. Preece et al., 2018, Preece, Becerra, Allan et al., 2020), and the internal consistency of the total scale score was excellent in this sample (α = .95).

Perth Emotion Regulation Competency Inventory (PERCI)

The PERCI (D. A. Preece et al., 2018) is a 32-item self-report questionnaire designed to assess emotion regulation ability across both positive and negative emotions. Items include statements such as, “When I’m feeling bad, I have no control over the strength and duration of that feeling” and “When I’m feeling good, I don’t have many strategies (e.g., activities or techniques) to increase the strength of that feeling”., which are responded to on a 7-point Likert scale ranging from “strongly disagree” to “strongly agree”. The PERCI is comprised of eight subscales which can be split via valence – half concern the regulation of positive emotions, the other half concern the regulation of negative emotions. The subscales measure (1) difficulties altering the experiential manifestations of emotions (e.g., down-regulating negative emotions and up-regulating positive emotions), (2) difficulties controlling the behavioural manifestations of emotions (e.g., inhibiting behavioural responses to emotions and activating non-dominant behaviours), and (3) difficulties activating emotion regulation goals when appropriate (e.g., ability to tolerate emotions). The subscales can combine into two theoretically meaningful composites, which are used in the current study: a Difficulty Regulating Positive Emotions composite score and a Difficulty Regulating Negative Emotions composite score. Higher scores on each indicate greater emotion regulation difficulties in that valence domain (or lower emotion regulation ability). The PERCI has previously demonstrated good psychometric properties (D. A. Preece et al., 2018; Preece, Mehta et al., 2022) and had strong reliability in this sample (α = .93 for both composite scores).

Eating Disorder Examination Questionnaire (EDE-Q)

The EDE-Q (Fairburn & Beglin, 1994) is a self-report measure adapted from the Eating Disorder Examination (Fairburn et al., 1993), widely regarded as a gold-standard interview-based diagnostic tool. The EDE-Q provides a comprehensive assessment of the attitudes and behaviours that are central to eating disorders by assessing their frequency across the previous 28 days. Items include questions such as, “Have you been deliberately trying to limit the amount of food you eat to influence your shape or weight (whether or not you have succeeded)?”, which are measured on a 7-point Likert scale ranging from “no days” to “every day”. The EDE-Q is composed of four subscales, including Restraint, Weight Concern, Shape Concern, and Eating Concern. A global score can be obtained by summing the four subscale scores and dividing the total by four. Higher global scores are indicative of higher levels of eating disorder symptoms. The EDE-Q has demonstrated excellent internal consistency and overall psychometric adequacy (Berg et al., 2012; Luce & Crowther, 1999; Mond et al., 2004) and performed well in this sample (α = .97).

Analytic strategy

The hypothesis that emotion regulation difficulties would mediate the relationship between alexithymia and eating disorder symptoms, was tested using a parallel mediation analysis design, in which difficulty regulating negative emotions and difficulty regulating positive emotions were inputted as mediators. Analyses were conducted using IBM SPSS 26, with Hayes’ (2017) PROCESS package for mediation analysis. Scores from the PAQ, PERCI and EDE-Q were used as markers of their associated constructs in each analysis.

Pearson correlations were conducted to examine the associations between all variables prior to mediation analyses, and all variable scores were converted into z-scores to generate standardised values. A mediation analysis was conducted (see Figure 2 for a visual representation of the tested model) with a specified bootstrapped sample of 5000 with 95% confidence intervals to test for direct and indirect effects. When testing indirect effects, bootstrapped confidence intervals that did not cross zero were recognised as significant (Hayes, 2017). Age and participant gender (coded: male = 0, female = 1) were included in the mediation analysis as covariates to control for potential demographic effects.

Results

Data screening

Timing-checks were embedded throughout the Qualtrics survey, which informed the removal of 17 participants. Participants that completed each questionnaire with an average of less than 2 seconds an item were removed to control for insufficient effort or inattentive responding (Huang et al., 2012). Two additional participants were identified as outliers and removed from the sample according to the Outlier Labelling Rule, with a specified g value of 2.2 (Hoaglin & Iglewicz, 1987; Hoaglin et al., 1986). These data cleaning procedures consequently reduced the sample for analysis down to 236 participants. Updated demographic information for the final sample is as follows: age ranged from 18–62 years (M = 20.76, SD = 5.37), 70.3% female and 29.7% as male, 52.1% White/Caucasian and 39.8% Asian, and an average BMI of 23.13 (SD = 4.92).¹

Scores on all measures were normally distributed according to Field’s (2013) guidelines, which specify skew and kurtosis cut-offs as less than |2.0| and |7.0| respectively. Descriptive statistics for each variable are displayed in Table 1. Of note, the mean global EDE-Q score for the sample is relatively higher than available community norms (Fairburn & Beglin, 1994; Mond et al., 2006), suggesting the sample had higher than average levels of eating disorder symptomatology.

Table 1. Descriptive statistics for PAQ, PERCI, and EDE-Q composites.

Variables	M(SD)	Range
PAQ total	78.41(27.16)	24–154
PERCI Pos	41.55(15.36)	16–97
PERCI Neg	59.60(18.73)	16–105
EDE-Q	1.91(1.49)	.00–5.65

PAQ total = Alexithymia total score, PERCI Pos = Difficulty regulating positive emotions, PERCI Neg = Difficulty regulating negative emotions, EDE-Q = Eating disorder symptoms.

As anticipated, Pearson correlations (see Table 2 for correlation matrix) demonstrated that alexithymia (PAQ total score) was moderately positively correlated with eating disorder symptoms (EDE-Q total score), and strongly positively correlated to both difficulty regulating positive emotions and difficulty regulating negative emotions scores (PERCI composites). Furthermore, both difficulty regulating positive emotions and difficulty regulating negative emotions were moderately positively correlated with eating disorder symptoms. Results from this correlation analysis are in accordance with previous literature, and confirm that each variable is appropriate for use in a mediation analysis.

Table 2. Correlation matrix for all variables in mediation analysis.

Variables	1.	2.	3.	4.	5.
1. PAQ total
2. PERCI Pos	.56**
3. PERCI Neg	.58**	.52**
4. EDE-Q	.34**	.32**	.33**
5. Age	−.22**	−.16*	−.19**	−.08
6. Gender	.09	.03	.14*	.38**	−.02

PAQ total = Alexithymia total score, PERCI Pos = Difficulty regulating positive emotions, PERCI Neg = Difficulty regulating negative emotions, EDE-Q = Eating disorder symptoms. **p < .01, *p < .05.

Mediation analysis

The analysis revealed that the overall model accounted for 26.7% of variance in the level of eating disorder symptoms, F(5, 230) = 16.75, p < .001, R² = .27. The covariate gender was found to significantly predict eating disorder symptoms (β = .35, SE = .06, 95% CI [.23, .46], p < .001), meaning on average females scored higher on the EDE-Q than males; which is unsurprising as eating disorders are more prevalent in females (Hoek, 2006).

Alexithymia had a significant total effect on eating disorder symptoms (β = .30, SE = .06, 95% CI [.19, .42], p < .001). There was no significant indirect effect via difficulty regulating negative emotions (β = .06, SE = .05, 95% CI [−.03, .16]), however, there was a significant indirect effect via difficulty regulating positive emotions (β = .09, SE = .04, 95% CI [.01, .18]). The mediation effect was partial, as indicated by the continued significance of the direct effect between alexithymia and eating disorder symptoms after controlling for indirect pathways (β = .16, SE = .08, 95% CI [.01, .30], p = .042). This suggests that alexithymia deficits predict difficulties regulating positive emotions, which in turn partially predicts the presence of greater eating disorder attitudes and behaviours. See Figure 2 and Table 3 for all model coefficients.²

Table 3. Coefficients for mediation model.

Variable	β (95% CI)	SE	t	p
Outcome variable: eating disorder symptoms
Alexithymia	.16(.01, .30)	.08	2.05	.042*
Difficulty regulating positive emotions	.17(.02, .31)	.07	2.31	.022*
Difficulty regulating negative emotions	.10(−.04, .25)	.07	1.37	.171
Age	.01(−.11, .12)	.06	.09	.932
Gender	.35(.23, .46)	.06	6.04	<.001***
Outcome variable: difficulty regulating positive emotions
Alexithymia	.56(.45, .67)	.06	10.01	<.001***
Age	−.04(−.15, .07)	.06	−.75	.452
Gender	−.02(−.13, .08)	.05	−.42	.674
Outcome variable: difficulty regulating negative emotions
Alexithymia	.56(.45, .67)	.05	10.28	<.001***
Age	−.07(−.17, .04)	.05	−1.20	.231
Gender	.09(−.01, .20)	.05	1.70	.090

***p < .001, **p < .01, *p < .05.

To supplement these analyses, we also conducted additional mediation analyses split by gender, which revealed a different pattern of results in females vs males (see Table S3 and S4 of supplementary materials). No significant mediation effects were observed in the male sample, in contrast, the association between alexithymia and eating disorder symptoms within the female sample was fully mediated by difficulty regulating positive emotions. This suggests that female participants were primarily driving the effect observed in the combined sample. However, the low male sample size and related power concerns limits our ability in this study to confidently infer that the mediation effect is limited to females. Our sample was not designed to look at gender differences; as such, we consider our analysis on the total sample to be our primary analysis, and future work should look at potential gender differences in more detail.

Discussion

The aim of the study was to investigate whether emotion regulation difficulties function as a link underlying the association between alexithymia and eating disorder symptoms. Overall, we found evidence for such a pathway, with our results highlighting the importance of considering valence, as difficulty regulating positive emotions was found to partially mediate the relationship between alexithymia and eating disorder symptoms.

These findings are broadly consistent with two previous studies investigating a similar mediation model, therefore providing further support for the importance of considering the interaction between alexithymia and emotion regulation in relation to eating disorder symptoms (Goetz et al., 2020; Khodabakhsh et al., 2017). It is also consistent with Preece, Mehta et al. (2022) finding based upon a similar hypothesis, where overall emotion regulation ability was found to mediate the association between alexithymia and depression and anxiety symptoms. Taken together, this pattern of findings is therefore in line with theoretical models specifying that, as emotional awareness provides an important foundation for effective emotion regulation (Preece, Becerra, Allan, Robinson & Dandy 2017), and eating disorder psychopathologies are characterised by emotion dysregulation (Oldershaw et al., 2015), that alexithymia and emotion regulation may interact in predisposing people to eating disorder symptoms.

The current study expanded upon previous literature through use of a more sophisticated parallel mediation model, that split emotion regulation by valence. This addition was necessary as emotion regulation refers to the process of altering the trajectory of all emotions across the valence spectrum, not just the negative valence domain (Gross, 2015). This approach revealed the association between alexithymia and eating disorders was partially mediated by difficulties regulating positive emotions specifically. In fact, regression coefficients indicated that when accounting for difficulty regulating negative emotions and positive emotions in the same model, the negative emotion domain did not significantly predict eating disorder symptoms, despite their strong positive correlation in terms of raw associations.

This finding is a novel contribution to the literature, as positively valenced emotions are often overlooked when investigating emotion regulation in eating disorder populations. For example, cognitive models of eating disorders (especially bulimia nervosa) often conceptualise engagement in eating disorder behaviours as a response to negative emotional states, such as sadness and anger (Cooper et al., 2004). As a result, many studies administer emotion regulation measures focused only on negative emotions. More recently in the broader psychopathology field, research has begun to recognise the important role of positive emotions in the maintenance and production of symptoms in emotional disorders, such as bipolar affective disorder and substance abuse disorders (see for review, Carl et al., 2013). We have drawn upon this literature to inform the findings of the current study, especially considering high instances of comorbidity and similarities in emotion dysregulation profiles have been documented between eating disorder and substance abuse disorder populations (Holderness et al., 1994; Morie & Ridout, 2018). Gruber (2011) has shown that despite popular belief, not all forms of positive emotions are necessarily adaptive. For example, happiness can be maladaptive if it is pursued in the wrong ways, at the wrong time, or in excess amounts; making it an important area of scientific inquiry (Gruber et al., 2011).

There are several possible explanations for why difficulty regulating positive emotions was found to drive, at least in part, the association between alexithymia and eating disorder symptoms in the current study. First, individuals high in eating disorder attitudes and behaviours may struggle down-regulating their experience of positive emotions, particularly if they lack a foundation of good emotional awareness (i.e., they have high alexithymia). One component of successful emotion regulation is the ability to inhibit dominant behavioural responses when experiencing emotions (D. A. Preece et al., 2018). This is assessed by PERCI items such as, “When I’m feeling good, my behaviour becomes out of control”. Individuals with alexithymia and eating disorder symptoms may struggle to control impulses to engage in eating disorder behaviours such as binge eating and purging, when experiencing heightened or unregulated states of positive emotion. The theory that individuals with eating disorder symptoms struggle to down-regulate their positive emotions is supported by the study conducted by Overton et al. (2005), which found elevated levels of positive emotions in individuals with eating disorders compared to healthy controls. Furthermore, this theory aligns with substance abuse literature, which has found that heightened states of positive emotions may reduce an individual’s capacity to resist the urges to engage in abuse of substances such as drugs and alcohol (Weiss et al., 2018). Unregulated levels of positive emotionality may also act to reinforce maladaptive eating disorder attitudes and behaviours. Overton et al. (2005) found that individuals with eating disorders consistently associated eating disorder themes with positive emotions rather than negative emotions. Anecdotally, these positive associations have been linked to a sense of control and achievement that is gained from attaining weight loss goals (Panza et al., 2019; Selby et al., 2015). An excess of positive emotionality caused by an inability to down-regulate, may allow positive emotions to reinforce eating disorder attitudes and behaviours. Positive emotions may not only trigger engagement in riskier impulsive eating disorder behaviour, but aid in maintaining symptoms via reinforcement (Overton et al., 2005; Selby et al., 2015).

A second potential explanation for the observed effect is that individuals high in alexithymia and eating disorder symptomatology may also struggle up-regulating their positive emotions. This difficulty may be influenced by an experience of positive emotional states as particularly aversive, due to underlying feelings of shame, a common feature in the emotion profiles of individuals with eating disorders (Goss & Allan, 2009; Troop et al., 2008). Individuals high in eating disorder symptomatology may feel as though they are unworthy of positive emotions, particularly if they do not understand the cause and nature of these states in cases of high alexithymia.

The findings of the current study in combination with recent clinical findings (Meneguzzo et al., 2022), highlight the importance of considering the interacting role of alexithymia and emotion regulation in eating disorder case formulations. Further clinical research of this nature may indicate that targeting alexithymia and emotion regulation elements across positive and negative domains, should be an essential consideration in the assessment and treatment of eating disorder symptoms. Techniques from CBT programs like the Unified Protocol for the Transdiagnostic Treatment of Emotional Disorders (Barlow et al., 2017), which contain modules designed to target both alexithymia and emotion regulation, may therefore be appropriate.

Limitations and directions for future research

We believe our study makes a useful preliminary contribution, however several limitations should be noted. First, our study exclusively used self-report measures, as they were practical for our research design and are currently the most widely used and validated type of measure in the alexithymia and emotion regulation field. Future studies may benefit from the incorporation of observer-rated tools and psychophysiological measures to capture other channels of emotional experience. Secondly, causal claims cannot be made based on this study due to its cross-sectional design. Although the study design is grounded in a robust body of literature and built upon key theoretical assumptions, longitudinal designs are necessary before directionality can be confidently inferred. Thirdly, our participants were all from a university in a single Western country (Australia), therefore future work would be necessary to examine the cross-cultural generalisability of our findings. Finally, treatment implications cannot be made until replicability in achieved in specialised clinical samples. Relatedly, we examined eating disorder symptoms generally, rather than those specific to a specific type of eating disorder. Future work might usefully examine similarities and differences in emotional functioning across eating disorder categories, as different mechanisms relating to positive emotions may be at work for different eating disorders.

Conclusion

The current study supports that emotion regulation ability may be a mechanism driving the link between alexithymia and eating disorder symptoms. These findings further reinforce the transdiagnostic relevance of these emotional constructs, and highlight the importance of considering the positive valence domain in the eating disorder field. Clinical research disentangling these constructs further, may assist in the conceptualisation, assessment, and treatment of eating disorder symptoms.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author, XM, upon reasonable request.

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/00050067.2023.2236280.

Notes

1. Descriptive statistics for males and females can be found in supplementary materials.

2. See supplementary materials for hierarchical regression analysis predicting eating disorder symptoms.