Self-reported interoceptive sensibility across eating disorder symptoms: associations with sociodemographic factors

Abstract

Background

Interoception is the capacity to perceive, interpret, and respond to internal bodily signals and is increasingly recognised as an important factor in the development and maintenance of eating disorders (EDs). Altered interoception contributes to disrupted hunger and satiety, body image disturbances, and difficulties with emotional awareness and responsiveness, all of which are central to ED psychopathology. Despite this growing theoretical and clinical interest, empirical research examining associations between interoception and diverse ED symptoms remains limited, particularly with respect to the potential moderating role of sociodemographic factors.

Objective

This study used self-report instruments to investigate interoceptive sensibility across ED symptoms.

Methods

A community sample of 221 Australian adults aged 18 or above completed validated self-report measures of interoceptive sensibility (Interoception Sensory Questionnaire, ISQ) and ED symptoms (e.g., Eating Disorder Questionnaire, EDE-Q and Nine Item ARFID Screener, NIAS). Correlation analyses were conducted to assess associations between ISQ scores and ED measures. In addition, we undertook moderation analyses to assess whether sociodemographic factors (i.e., racial background, gender identity, age, employment status, living situation, and sex assigned at birth) influenced ISQ scores and their associations with ED measures.

Results

Statistically significant correlations between ISQ scores and all ED measures were found, with the strongest associations observed with the NIAS and the EDE-Q. Some sociodemographic factors (i.e., sexual orientation, racial background, and sex assigned at birth) also influenced ISQ scores and their association with ED measures.

Conclusions

Difficulties with interoceptive sensibility are evident across a wide range of self-reported ED symptoms, suggesting that altered interoceptive sensibility may be a transdiagnostic feature of EDs, though the statistical strength of associations between ISQ scores and ED measures varied. Integrating interoceptive sensibility into ED assessment and treatment may enhance the individualisation of care.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40337-026-01530-z.

Plain language summary

People with eating disorders often report difficulties with interoception, which is the ability to notice, interpret, and act on signals from within the body such as hunger, fullness, pain, or tiredness. Previous research has typically focused on single eating disorder diagnoses, limiting understanding of how interoceptive difficulties relate to a broader range of eating disorder symptoms. This study examined associations between interoceptive sensibility and eating disorder symptoms in a community sample of adults. Participants completed an anonymous online survey assessing interoceptive sensibility as well as eating related behaviours and attitudes. Results indicated that eating disorder symptoms including food restriction or avoidance, body image concerns such as fear of weight gain or body dissatisfaction, binge eating, and purging were associated with poorer interoceptive sensibility. These associations were evident across symptom types and were influenced by sociocultural factors including sex assigned at birth and racial background. These findings suggest that interoceptive difficulties are common across diverse eating disorder presentations and highlight the potential value of more personalised, sensory informed approaches to treatment that support individuals in reconnecting with their bodies in ways that align with their lived experiences and support needs.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40337-026-01530-z.

Background

Eating disorders (EDs) are complex mental health conditions characterised by distorted thoughts and behaviours related to eating and self-image [1]. These disorders occur along a spectrum of phenomenology and can manifest through restrictive eating, binge eating, or compensatory behaviours such as self-induced vomiting, laxative misuse, or compulsive exercise. EDs are associated with severe physiological consequences, including electrolyte imbalances, cardiomyopathy, hypoglycaemia, and renal impairment [1], and they significantly impact psychological wellbeing [2, 3].

Estimates suggest lifetime ED prevalence ranges from 0.74 to 8.4% [4], with recent studies documenting a troubling rise in prevalence [5, 6]. However, the pathophysiological mechanisms underlying EDs remain poorly understood [7–9]. Recovery rates remain low, with fewer than half of affected individuals achieving sustained remission following evidence-based treatment [10]. High relapse and mortality rates have further underscored the need for more personalised treatment approaches [11–16]. As Knight et al. [17] argue, “adaptable and flexible service models are essential for effective care.”

One promising direction for treatment enhancement involves addressing interoception, defined as the perception, interpretation, and integration of internal bodily cues, such as hunger, satiety, thirst, pain, discomfort, fatigue, and temperature [18]. Interoception is foundational to adaptive behaviour, including eating and drinking [19], body image and self-awareness [20–22], and emotional regulation [23, 24]. Disruptions in these domains are central to ED pathology [25–28].

Garfinkel et al.’s [29] multidimensional model of interoception provides a critical framework for understanding why subjective reports of bodily experience often diverge from objective performance on interoceptive tasks (e.g., heartbeat or respiratory rate tracking). By distinguishing interoceptive accuracy, interoceptive sensibility, and interoceptive awareness, the model clarifies that these components are dissociable. In particular, interoceptive sensibility reflects individuals’ beliefs, interpretations, and self-attributions regarding internal bodily signals, rather than the precision with which those signals are detected (i.e., interoceptive accuracy), thereby accounting for discrepancies between self-report and behavioural or task-based measures [29].

Subsequent theoretical syntheses have reinforced the relevance of this distinction for mental health research. Integrative frameworks emphasise that interoceptive sensibility is influenced by higher-order cognitive processes, including attentional focus and prior beliefs about the body, which can bias the interpretation of afferent signals independently of physiological accuracy [30, 31]. As a result, elevated interoceptive sensibility does not necessarily indicate heightened interoceptive accuracy and may instead reflect altered weighting or amplification of bodily sensations.

Within this conceptual context, the Interoception Sensory Questionnaire (ISQ) maps directly onto the interoceptive sensibility dimension articulated by Garfinkel et al. [29]. The measure captures subjective sensory experiences, aligning with contemporary models that position self-reported interoceptive sensibility as a clinically meaningful construct distinct from both objective detection performance (i.e., interoceptive accuracy) and metacognitive insight (i.e., interoceptive awareness). This distinction is particularly relevant in psychopathology research, where altered bodily beliefs and sensory interpretations may contribute to symptom development and maintenance even in the absence of measurable disruptions in interoceptive accuracy [29, 31].

Beyond its conceptual decomposition, interoception is increasingly understood as a multilevel process involving the sensing, interpretation, and integration of internal bodily signals across peripheral, subcortical, and cortical systems [32]. Contemporary neurobiological models emphasise that interoceptive signals ascend via spinal and vagal afferent pathways to brainstem nuclei and thalamic relays, before being integrated within higher-order cortical regions, particularly the insular cortex and anterior cingulate cortex, where they contribute to subjective feeling states and regulatory control [32]. Importantly, these neural representations support both conscious and non-conscious interoceptive processing, underscoring that self-report measures capture only one dimension of a broader interoceptive system [32].

Moreover, sociodemographic context, including cultural, ethnic, and racial background, may be associated with differences in bodily awareness and the subjective interpretation of internal signals, suggesting that interoception-informed ED interventions should be culturally sensitive and socially responsive. Cross-cultural research indicates that interoception is not attended to or expressed uniformly across cultural groups, as reflected in differences in language use, idiomatic expressions, and culturally embedded models of mind and body when compared with non-Western cultural contexts [33]. Some non-Western cultural groups report heightened interoceptive sensibility on self-report measures, yet demonstrate lower or altered interoceptive accuracy on behavioural tasks such as heartbeat perception when compared with Western samples [33]. This dissociation between interoceptive sensibility and interoceptive accuracy underscores that these constructs capture distinct aspects of interoception. Cultural differences may therefore reflect variation in attentional focus, interpretation, and meaning-making of bodily signals. Cultural schemas, epistemologies, and explanatory models of the body are therefore likely to play a role in shaping how bodily sensations are noticed, labelled, and evaluated [33–35].

Contemporary theoretical frameworks further highlight the relevance of interoception to EDs. Interoceptive inference models propose that the brain constructs predictions about internal bodily states and continuously updates these based on interoceptive input [29]. When prediction errors occur, such as failing to accurately interpret signals of hunger or satiety, maladaptive, non-homeostatic, behaviours may emerge, such as restrictive or binge eating. Within this framework, interoceptive sensibility alterations in EDs may reflect persistent mismatches between expected and actual internal states, leading to both physiological and affective disturbances.

Embodied cognition perspectives emphasise that the experience of the body and self arises from dynamic multisensory integration involving interoceptive, exteroceptive, and proprioceptive dimensions [36]. Alterations in interoceptive sensibility may therefore undermine embodiment and body trust,1 contributing to body image disturbances and low self-esteem commonly reported in EDs [37]. These theoretical perspectives provide a compelling rationale for examining interoceptive sensibility in relation to EDs.

Emerging research suggests that altered interoception may contribute to ED symptoms through multiple pathways. Impaired detection or interpretation of hunger and fullness cues can drive restrictive behaviours (e.g., early satiety) or binge eating episodes (e.g., delayed satiety) [25]. Disrupted interoceptive sensibility is also linked to executive functioning difficulties, including remembering to eat, meal planning, and maintaining regular eating routines [12, 25, 36]. For instance, in anorexia nervosa (AN), reduced hunger detection may contribute to restriction and reinforce physiological disconnection and emotional dysregulation [38–40]. In bulimia nervosa (BN), impaired interoceptive sensibility has been linked to difficulty detecting satiety and regulating binge urges [41, 42]. Disrupted interoceptive sensibility is further implicated in body image disturbances and mistrust of bodily sensations in both AN and BN [37]. In binge eating disorder (BED), confusion between emotional and physiological states may contribute to non-homeostatic eating and loss-of-control eating episodes [43].

Avoidant/restrictive food intake disorder (ARFID) has also been associated with altered interoceptive sensibility, particularly in subtypes involving reduced appetite or low interest in eating [44, 45]. In such cases, altered detection of hunger or discomfort may underlie restrictive intake. Orthorexia nervosa (ON), a clinical construct that is not currently recognised as a standalone diagnosis in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5-TR; 88), is characterised by an obsessive fixation on healthy eating [46]. ON has been associated with heightened anxiety about internal bodily cues, such as gastrointestinal discomfort, which may be misinterpreted as indicators of food impurity or toxicity [47, 48]. Individuals with ON also report lower embodiment and intuitive eating, as well as reduced body appreciation and functionality awareness [49, 50].

Interoception-focused treatments, such as interoceptive exposure or training, have already been integrated into some ED interventions and show promise in improving emotion regulation and reducing ED symptom severity [45, 51]. However, it is important to prioritise person-centred care and avoid practices that are applied in a relatively uniform, ‘one size fits all’ manner [52], with limited consideration of specific ED presentations and psychosocial factors that may influence individual interoceptive experiences and related support needs.

Lack of treatment individualisation can result in unintended iatrogenic harm. The long-standing failure of mental health research to assess and report treatment-related harm in clinical trials, systematic reviews, and meta-analyses is increasingly recognised as a serious methodological and ethical gap [53–56]. Despite ongoing discussions of treatment safety, quantitative evidence addressing psychotherapy-related harm remains limited, while qualitative research consistently documents clients’ negative and harmful treatment experiences [54, 55].

Although iatrogenic harm is rarely considered or acknowledged in clinical trials and systematic reviews, it has been documented in qualitative ED research [7], where mis-attuned, standardised protocols such as family-based therapy (FBT) and CBT have been associated with feeling misunderstood, ED exacerbation, mental health deterioration, and treatment-related trauma for Autistic individuals [57]. Lived experience-led research further highlights how the lack of person-centred or individually adapted care contributes to worsening ED symptoms, disengagement from services, and prolonged illness trajectories [58–61]. These accounts describe how invalidating clinical conceptualisations and resistance to addressing treatment-related harm can reinforce ED pathology and entrench systemic harm within mental healthcare services.

In order to mitigate the documented risks associated with indiscriminately universalised treatment models, attention to sociodemographic factors and individualised support needs is therefore important when considering interoception-informed treatment frameworks for EDs. For example, emerging evidence suggests that while interoceptive or body-focused practices can support emotional regulation and adaptive eating behaviours, they may also exacerbate gender dysphoria in transgender and gender diverse individuals when exercises involve sustained attention to incongruent body parts [62, 63]. Broader gender-affirming and trauma-informed care guidelines also recommend adapting body-based interventions by using individualised language for body parts and interoceptive sensations and offering alternatives or opt-outs to avoid triggering gender dysphoria [62]. Identifying associations between difficulties in interoceptive sensibility and a broad range of self-reported ED symptoms, as well as the potential moderating influence of sociodemographic factors on these relationships, may provide insights that support the personalisation and optimisation of interoception-informed treatment approaches. Such knowledge may help ensure that ED care is better aligned with the unique, multidimensional, and contextually shaped support needs of individuals.

Although research on the role of interoception in EDs is expanding, few studies have compared interoceptive sensibility across a wide range of self-reported ED symptoms using validated psychometric instruments within the same sample. Moreover, much of the existing literature overlooks potential variations in interoceptive sensibility across intersectional sociodemographic factors such as racial background, gender identity, and sex assigned at birth [64]. Addressing these gaps may help clarify associations between interoceptive sensibility and self-reported ED symptoms and support the development of more individualised ED treatment approaches. Investigating interoceptive sensibility across multiple self-reported ED measures that capture a diverse range of ED symptoms, including food restriction, distress related to weight or shape, orthorexia, and binge eating, within the same sample, and examining associations with sociodemographic factors, therefore adds to the existing literature.

The present study examined associations between self-reported interoceptive sensibility and diverse ED symptoms. We also investigated whether sociodemographic factors, including racial background, gender identity, age, employment status, living situation, and sex assigned at birth, were associated with interoceptive sensibility or moderated its relationships with ED symptoms. By adopting this approach, the study contributes to a growing body of research on interoception in EDs and provides insights relevant to the development of more personalised, interoception-informed ED treatment strategies.

Methods

Recruitment and data collection

Participants were eligible to take part in the study if they were aged 18 years or older, resided in Australia, and were proficient in reading English. Data were collected between April and July 2025 via an anonymous online survey administered through Qualtrics. Recruitment was conducted through multiple social media platforms, including Facebook, Instagram, X (formerly Twitter), and LinkedIn (Appendix 1). In addition, several organisations, such as the Butterfly Foundation, the National Eating Disorders Collaboration, the Australian and New Zealand Academy for Eating Disorders, and Eating Disorders Neurodiversity Australia supported recruitment by sharing the survey through their newsletters and social media channels. To incentivise participation, a prize draw was conducted in April 2025, offering ten Coles vouchers valued at $20 AUD each. A detailed overview of the study procedures, including potential risks of discomfort, was made available online to prospective participants. By choosing to complete the anonymous online survey, participants provided implied consent. The first author (LC) also disseminated the survey through personal and professional networks.

Survey components and characteristics

Participants self-reported lifetime ED diagnoses using the question: “Have you ever been diagnosed with an eating disorder? (Select all that apply).” Response options included: “Yes, anorexia nervosa”; “Yes, bulimia nervosa”; “Yes, binge eating disorder”; “Yes, other specified feeding or eating disorder or eating disorder not otherwise specified (e.g., atypical anorexia)”; “Yes, avoidant/restrictive food intake disorder”; and “No, I have never been diagnosed with an eating disorder.” This question referred to lifetime history rather than current diagnostic status and allowed participants to self-report more than one ED diagnosis.

Lifetime ED history was collected for descriptive purposes only and was not used analytically in the correlational or moderation analyses. All statistical analyses focused on current self-reported ED symptoms and interoceptive sensibility measured at the time of survey completion. As such, findings should be interpreted as reflecting associations between current symptom dimensions and interoceptive sensibility, rather than effects attributable to specific ED diagnoses or illness phases. This distinction is important given the potential mismatch between lifetime diagnostic history and current symptom presentation, recovery status, or illness chronicity.

Allowing participants to endorse multiple ED diagnoses reflected the well-documented diagnostic crossover and fluctuation observed across ED illness trajectories. Longitudinal research demonstrates frequent transitions between AN, BN, BED, and other specified feeding or eating disorder (OSFED; particularly atypical AN), as well as between AN subtypes (restricting and binge-purge), underscoring the dimensional and fluid nature of EDs [65–68].

The Qualtrics survey platform incorporated several built-in data quality and engagement controls. Automated checks flagged questionnaires completed more than two standard deviations faster than the median duration, indicating potential inattention. The integrated reCAPTCHA system identified likely automated (‘bot’) responses. Duplicate detection flagged multiple submissions originating from the same internet protocol (IP) address, and straightlining detection identified cases where identical responses were selected across all items in a matrix, suggesting non-differentiated responding. These embedded procedures supported data integrity by identifying inattentive or invalid response patterns prior to analysis. Of all participants, two were flagged as ‘unreliable’ by Qualtrics quality checks. These surveys were not included in the analyses. Although the Qualtrics platform includes embedded data quality features, such as detection of unusually rapid completion times, duplicate IP addresses, straightlining, and automated ‘bot’ screening, these procedures are not equivalent to validated attention check items. As such, while these embedded checks supported basic data integrity, they cannot fully rule out inattentive responding. In addition, we also excluded participants who did not fully complete the survey (i.e., survey completion less than 100%).

Eating disorder questionnaire (EDE-Q)

The EDE-Q (version 6) is a 28-item self-report measure adapted from the Eating Disorder Examination interview [69, 70]. It assesses ED behaviours and attitudes over the past 28 days and comprises four subscales: Restraint, Eating Concern, Shape Concern, and Weight Concern. Items are rated using a seven-point forced-choice scale. The EDE-Q demonstrates good internal consistency, with Cronbach’s alphas of 0.84 (Restraint), 0.93 (Shape Concern), 0.89 (Weight Concern), and 0.78 (Eating Concern) [71]. It also shows strong diagnostic sensitivity [72]. In addition, the Cronbach’s alphas for this study’s sample were 0.97 (EDE-Q total), 0.90 (Restraint), 0.94 (Shape Concern), 0.88 (Eating Concern), and 0.89 (Weight Concern).

Nine item ARFID screen (NIAS)

The NIAS is a nine-item self-report scale assessing behaviours related to ARFID [73]. It includes three subscales: Picky Eating, Appetite, and Fear (e.g., fear of aversive consequences from eating), each consisting of three items rated on a six-point Likert scale from “strongly disagree” to “strongly agree.”

The NIAS demonstrates high internal consistency (α = 0.90) and strong test–retest reliability [73]. Subscale reliabilities have also been confirmed: Picky Eating (0.86), Appetite (0.91), and Fear (0.90) [45]. The Cronbach’s alphas for this study’s sample were 0.90 (NIAS total), 0.92 (Picky Eating), 0.88 (Appetite), and 0.89 (Fear).

Düsseldorf orthorexia scale (English version; E-DOS)

The English version of the Düsseldorf Orthorexia Scale (E-DOS; 74) was also administered; as this version is already validated in English, no translation was required. The Düsseldorf Orthorexia Scale (DOS) is a German 10-item self-report questionnaire designed to measure orthorexic eating behaviours [75]. Responses are rated on a four-point Likert scale, ranging from 1 (does not correspond to my behaviour at all) to 4 (corresponds well to my behaviour). The English version shows good internal consistency (α = 0.88) and strong construct validity, with E-DOS scores correlating with the Eating Habits Questionnaire (r = 0.76, p < 0.001) [74]. The Cronbach’s alpha for this study’s sample was 0.90.

Binge eating disorder screener (BEDS-7)

The BEDS-7 is a seven-item screener developed to identify binge eating behaviours [76]. It assesses the presence and frequency of binge eating episodes, perceived loss of control, and post-binge distress, with items rated on a Likert scale from “never/rarely” to “always.” The BEDS-7 has demonstrated good internal consistency in non-clinical samples, with Cronbach’s alpha values above 0.82 [77, 78]. The Cronbach’s alpha for this study’s sample was 0.91.

Interoception sensory questionnaire (ISQ)

The ISQ is a 20-item self-report measure designed to assess interoceptive sensibility, specifically the subjective ability to sense and interpret internal bodily states [79]. The ISQ has a single-factor structure, and each item is rated on a seven-point Likert scale, ranging from "not true at all of me" to "very true of me." A higher score indicates difficulties with the recognition, interpretation, and expression of somatic sensations, also referred to as alexisomia [80]. The ISQ demonstrates excellent internal consistency (α = 0.96) across multiple studies [79, 81]. The Cronbach’s alpha for this study’s sample was 0.97.

Data analysis

Spearman’s rank-order correlations were used to assess associations between ISQ scores and ED measures. Moderation analyses were then conducted to examine whether sociodemographic factors, specifically sex assigned at birth, racial background, gender identity, sexual orientation, age, and living situation, influenced the relationship between ISQ scores and ED measures (Appendix 2). All moderation analyses were conducted using Hayes’ PROCESS macro for SPSS [82]. The suitability of linear models was evaluated by inspecting Q-Q plots and residual-versus-predicted scatterplots.

Although Cohen’s [83] guidelines (0.10 = small, 0.30 = medium, 0.50 = large) remain the most frequently cited reference for interpreting correlation magnitudes, they were intended as heuristics rather than universal standards. Subsequent methodological work has refined these conventions to better reflect the empirical distribution of correlations in behavioural and psychological science. For example, Schober, Boer, and Schwarte [84] recommend interpreting correlations of less than 0.20 as weak or small, 0.20 to 0.39 as moderate, and 0.40 or greater as strong or large, emphasising that these cut-points are context dependent rather than absolute [85].

Evidence from Gignac and Szodorai [85] further supports this perspective. In their analysis of over 700 meta-analytically derived correlations, median observed values clustered around r = 0.19, with upper-quartile values near r = 0.29, indicating that correlations above 0.40 represent comparatively large and uncommon effects in individual-differences research [86]. In line with these converging lines of evidence, adopting correlations of 0.20 or less as small, values up to 0.40 as moderate, and values greater than 0.40 as large provides a balanced, empirically informed, and interpretable framework for effect size interpretation that aligns with contemporary methodological recommendations and empirical realities [87].

Results

Participants

The sample (N = 221) predominantly comprised white (67%) or Asian (22.6%) individuals assigned female at birth (89.1%). Most participants identified as women (70.1%), were aged between 18 and 44 years (82.8%), and had a high level of education (84.6% had post-secondary education). There was diversity regarding sexual orientation: while nearly half identified as heterosexual (49.8%), large minorities identified as bisexual (16.7%), gay/lesbian (12.2%), or pansexual (10.9%). An ED diagnosis was self-reported by 47.5% of participants; most commonly AN (24.9%), followed by OSFED (14.5%), BED (11.8%), BN (10.9%), and ARFID (10.9%). In addition, 43 participants self-reported two or more ED diagnoses.

The high prevalence of post-secondary education and diversity regarding sexual orientation in the sample may be attributable to some participants having been recruited directly by the first author (LC), who is a member of the LGBTQIA + community (Lesbian, Gay, Bisexual, Transgender, Queer/Questioning, Intersex, Asexual), followed by subsequent recruitment through these participants’ social networks, consistent with a snowball sampling effect (Table 1).

Table 1.

Demographics

Category	N (total = 221)	%
Age range
18–24 years	80	36.2
25–34 years	65	29.4
35–44 years	47	21.3
45–59 years	24	10.9
60–69 years	5	2.3
Gender identity
Woman	155	70.1
Non-binary	30	13.6
Man	27	12.2
Agender	9	4.1
Sex assigned at birth
Female	197	89.1
Male	24	10.9
Racial background
White	148	67
Asian	50	22.6
Multiracial	12	5.4
Middle Eastern/North African	8	3.6
Education level
Bachelor’s degree	74	33.5
Master’s degree	54	24.4
Secondary school or below	52	23.6
Certificate/Diploma/Associate’s degree	36	16.3
Doctorate	5	2.3
Employment status
Not employed	86	38.9
Part-time	82	37.1
Full-time	53	24
Sexual orientation
Straight (heterosexual)	110	49.8
Bisexual	37	16.7
Gay/Lesbian	27	12.2
Pansexual	24	10.9
Asexual	23	10.4
Self-reported ED diagnosis
No ED diagnosis	116	52.5
More than 1 ED diagnosis	43	15.8
Any ED diagnosis	105	47.5
AN	55	24.9
OSFED	32	14.5
BED	26	11.8
BN	24	10.9
ARFID	24	10.9

Note. AN = Anorexia Nervosa; BN = Bulimia Nervosa; BED = Binge Eating Disorder; OSFED = Other Specified Feeding or Eating Disorder; ARFID = Avoidant/Restrictive Food Intake Disorder; ED = Eating Disorder; N = Number of participants; % = Percentage of total sample

Correlational analysis between ISQ, EDE-Q, NIAS, BEDS-7, and E-DOS

All p-values are based on two-tailed tests and are statistically significant at p < 0.001.

Spearman correlations presented in Table 2 revealed statistically significant associations between ISQ scores2 and all ED measures, including total and subscale scores. The strongest correlations were found with the NIAS and EDE-Q. Weaker but statistically significant correlations were also observed between the ISQ, the BEDS-7, and the E-DOS.

Table 2.

Correlations between ISQ scores and ED measures

	ISQ total
	rho	95% CI	p
EDE-Q total	0.43	0.31 to 0.55	< 0.001
Restraint	0.33	0.21 to 0.43	< 0.001
Shape Concern	0.40	0.28 to 0.51	< 0.001
Eating Concern	0.49	0.37 to 0.59	< 0.001
Weight Concern	0.40	0.27 to 0.52	< 0.001
NIAS total	0.58	0.49 to 0.67	< 0.001
Picky Eating	0.46	0.35 to 0.57	< 0.001
Appetite	0.55	0.45 to 0.64	< 0.001
Fear	0.40	0.29 to 0.50	< 0.001
BEDS-7 total	0.26	0.14 to 0.40	< 0.001
E-DOS total	0.28	0.15 to 0.41	< 0.001

Note. ISQ = Interoception Sensory Questionnaire; EDE-Q = Eating Disorder Examination Questionnaire; NIAS = Nine Item ARFID Screen; BEDS-7 = Binge Eating Disorder Screener; E-DOS = Düsseldorf Orthorexia Scale (English version); rho = Spearman’s rank correlation coefficient; CI = Confidence Interval

Moderator analysis

In the first part of the moderator analysis, sociodemographic factors (i.e., racial background, gender identity, age, employment status, living situation, and sex assigned at birth) were used to predict ISQ scores (Appendix 2). Several sociodemographic factors emerged as statistically significant predictors. Participants who identified as heterosexual or pansexual tended to report lower ISQ scores compared to those of other sexual orientations. Additionally, racial background influenced ISQ scores, with white participants tending to report higher ISQ scores than participants from other racial backgrounds.

The second part of the analysis investigated whether these sociodemographic factors moderated the relationships between ISQ scores and ED measures, including the EDE-Q, NIAS, BEDS-7, and E-DOS. No significant moderation effects were found in the relationships between ISQ and either NIAS or E-DOS scores. However, several sociodemographic factors did influence the relationship between ISQ and EDE-Q scores; specifically, racial background and sex assigned at birth. The association between ISQ and EDE-Q scores was stronger among participants not identifying as white. Furthermore, the relationship between ISQ and EDE-Q scores was statistically significant only among individuals assigned female at birth, though the lack of statistical power among participants assigned male at birth may have limited interpretability. In addition, several statistically significant moderation effects emerged when examining the relationship between ISQ and BEDS-7 scores. Racial background once again played a role, with the ISQ and BEDS-7 relationship being stronger for individuals not identifying as white. Finally, sex assigned at birth moderated this relationship as well, with an association observed among individuals assigned female at birth, while the association was negative and non-significant among individuals assigned male at birth.

Discussion

This study examined self-reported interoceptive sensibility, as measured by the ISQ, across different self-reported ED symptoms, as measured by the EDE-Q, BEDS-7, NIAS, and E-DOS. ISQ scores showed strong positive correlations with the NIAS, moderate correlations with the EDE-Q, and weaker but statistically significant correlations with the BEDS-7 and E-DOS. In addition, several sociodemographic factors influenced ISQ scores and their associations with ED measures. These associations indicate that heightened difficulties related to interoceptive sensibility correspond with ED symptoms and support the conceptualisation of altered interoceptive sensibility as a transdiagnostic feature of EDs, observable across diverse symptoms [17, 21, 25, 41, 42].

The correlations observed between ISQ scores and all ED measures indicate that greater interoceptive sensibility difficulties are associated with ED symptoms across domains. Moreover, ISQ scores showed the strongest associations with the NIAS Appetite and Picky Eating subscales, moderate associations with the EDE-Q Restraint, Shape Concern, and Eating Concern subscales, and smaller but statistically significant correlations with the BEDS-7 and E-DOS total scores. These dimensional relationships suggest that different facets of ED symptomatology are linked to varying degrees of interoceptive sensibility disruption, which may help clinicians identify individual interoceptive sensibility targets for intervention. Collectively, these associations suggest that poor interoceptive sensibility contributes to both behavioural (e.g., food restriction, binge eating) and cognitive-affective features (e.g., body image disturbances) of EDs [12, 18, 25].

Notably, the NIAS, which primarily measures ED features consistent with ARFID, showed the strongest associations with ISQ scores. These findings align with research indicating that difficulties with interoception in ARFID may underlie food avoidance behaviours, driven by a lack of hunger detection or fear of somatic discomfort [44, 73]. For example, Datta and Lock [44] noted that adolescents with ARFID show reduced accuracy on heartbeat perception tasks (interoceptive accuracy) and report difficulties interpreting interoceptive signals (interoceptive sensibility), which may contribute to symptom presentation. In contrast, adolescents with AN demonstrate higher heartbeat perception accuracy (interoceptive accuracy) but describe limited trust in bodily cues (interoceptive sensibility) [44]. Similarly, Zickgraf and Ellis [73] identified low appetite as a core feature of ARFID. Burton Murray et al. [45] further supported this association by showing that ARFID symptom dimensions (e.g., reduced interest in eating, fear of discomfort) were closely linked to interoception. Altogether, this study’s results align with the DSM-5-TR diagnostic criteria for ARFID, which highlight that disordered eating may be underpinned by alterations in multisensory processing and integration [88].

Moderator analyses added further nuance to this study. Sexual orientation was shown to influence interoceptive sensibility. Participants identifying as heterosexual reported less interoceptive sensibility difficulties than those identifying as bisexual, gay, or lesbian. One possible explanation for this finding involves differences in body awareness linked to minority stress. Sexual and gender minority individuals often experience unique forms of stigma, discrimination, and body image ideals and associated pressures that can heighten awareness and self-consciousness related to bodily cues and internal states [89]. Minority stress theories suggest that these chronic psychosocial stressors may increase attentional focus on bodily experiences associated with anxiety, vigilance, or self-presentation. By contrast, heterosexual individuals, who experience fewer minority stressors, may engage in less habitual self-monitoring of internal states, resulting in lower self-reported interoceptive sensibility difficulties.

Our moderator analyses also show that the association between ISQ and EDE-Q scores was statistically significant only among individuals assigned female at birth. Research has highlighted that people assigned female at birth tend to report higher interoceptive sensibility but lower interoceptive accuracy than those assigned male at birth, potentially reflecting both hormonal and neurobiological differences [90, 91]. However, focusing solely on sex-based binaries risks overlooking the diverse interoceptive experiences shaped by gender identity and dysphoria. Although the lack of statistical power for trans and gender diverse individuals in this study may have limited interpretability, emerging evidence highlights that transgender and gender diverse people may experience interoceptive signals differently, particularly in the context of gender incongruence. For instance, Frietchen et al. [92] found that trans and gender diverse adults reported distress or detachment from internal bodily signals due to a mismatch between their felt gender and physical characteristics or bodily sensations. This highlights the need for gender-affirming approaches in interoception-informed ED interventions, particularly for individuals whose interoceptive challenges are compounded by experiences of gender dysphoria, systemic oppression, or minority stress.

Moreover, the strength of association between the ISQ scores and ED measures varied based on racial background. These findings suggest that cultural and ethnicity-related factors may shape not only how interoception is experienced but also how bodily signals are interpreted, understood, and acted upon, highlighting that interoception is not merely neurobiological but is also socially and culturally constructed. This underscores the importance of integrating cultural dimensions into interoception-informed ED treatments, ensuring clinical practices are responsive to the lived experiences and support needs of diverse populations outside of Western contexts. This aligns with embodied cultural cognition perspectives, which propose that bodily self-awareness and internal state perception are influenced by cultural schemas, practices, and lived experience [33, 93, 94]. Consequently, an intersectional understanding of interoceptive sensibility in the context of EDs, one that acknowledges the influence of embodied identity, sociodemographic context, and marginalisation, is essential to fully appreciate how embodiment is differentially shaped across diverse populations.

Clinically, these results suggest that routine screening for interoceptive sensibility could be incorporated into standard ED assessment to identify individuals experiencing significant disconnection from bodily states and difficulties interpreting and responding to bodily cues. For example, those with restrictive EDs may benefit from interventions targeting appetite and hunger cues, while those experiencing binge eating or purging behaviours might benefit from strategies to improve the interpretation of satiety and discomfort signals.

In addition, our moderator analyses’ findings support movement away from standardised intervention assumptions and toward approaches that work responsively with each person’s sensory, emotional, environmental, and cultural relationship to their bodily cues. Rather than applying a single model of interoceptive-informed care, ED treatment could be adapted in dialogue with how individuals describe, interpret, and inhabit bodily sensations.

Emerging treatment approaches seek to strengthen embodied awareness and self-regulation capacities. For example, Mindful Awareness in Body-Oriented Therapy (MABT) and intuitive eating interventions have been shown to enhance interoceptive sensibility, support adaptive responses to hunger and satiety, and improve body trust [18, 95]. Such approaches extend beyond exposure to interoceptive cues by actively cultivating curiosity, tolerance, and self-compassion toward bodily signals. However, it should be noted that some interoceptive training practices may cause iatrogenic harm for some populations. For example, interoceptive training can worsen body-related distress in trans and gender diverse populations if the focus is placed on gender incongruent bodily parts or sensations [62, 63]. Therefore, caution is warranted when using interoceptive training methods to ensure the treatment targets are safe and culturally valid.

Furthermore, difficulties with interoceptive sensibility influence executive functioning challenges, including problems with planning, task initiation, decision-making, organisation, and working memory [12, 25, 36]. These difficulties are themselves associated with ED symptomatology, suggesting that interventions designed to support executive functioning, such as structured routines, external aids, prompts and reminders, body doubling, or coaching, may be particularly beneficial when paired with interoception-focused strategies [12]. Together, these avenues underscore the potential of multimodal, personalised interventions that combine interoceptive training with executive function support to enhance treatment outcomes and foster self-efficacy and self-compassion.

The implications for future research are threefold. First, longitudinal studies are needed to clarify the directionality of the relationship between interoceptive sensibility and ED symptoms. It remains unclear whether such difficulties precede the onset of disordered eating, emerge as a consequence of it, or reflect a bidirectional relationship in which interoceptive sensibility disruptions increase vulnerability to EDs while disordered eating further exacerbates interoceptive sensibility alterations. Although emerging evidence suggests that difficulties with interoceptive sensibility may persist after remission (64), their role as predisposing factors requires further investigation. Second, neurobiological research should investigate how interoceptive signalling within regions such as the thalamus and insula differs across ED presentations compare to healthy controls and how these neural patterns may change with treatment. Finally, future research should prioritise inclusive and culturally responsive approaches, as the moderator findings from this study indicate that interoceptive experiences are also influenced by sociodemographic factors.

Limitations and strengths

While this study provides valuable insights into the relationships linking interoceptive sensibility and EDs, there are several limitations. The cross-sectional design of this study prevents causal inferences from being drawn regarding the relationship between interoceptive sensibility and ED symptoms. The sample comprised participants located exclusively in Australia, limiting the generalisability of findings to other cultural or geographic contexts. The study also relied on self-reported questionnaires. Although self-report allows for efficient data collection in community-based samples, self-reported ED diagnoses or symptoms may introduce bias through inaccurate recall, limiting the generalisability of our findings to clinical populations. Future research should therefore replicate these findings using diagnostic interviews or clinician-administered assessments. There is also a possibility of selection bias, given that recruitment relied on social media outreach and snowball sampling. Consequently, findings should be interpreted with caution when generalising to broader demographic populations. Furthermore, there was no assessment of attentive responding. The small number of identified invalid responses (n = 2) may be an underestimate as all the internal consistencies in this study are higher than other published studies, which is potentially consistent with inattentive responding.

Despite these limitations, several methodological strengths enhance the credibility and interpretability of the findings. The sample size provided sufficient statistical power to detect associations between self-reported interoceptive sensibility and ED symptoms. In addition, the use of validated and psychometrically robust instruments, such as the ISQ, EDE-Q, NIAS, BEDS-7, and E-DOS, ensured that the constructs were assessed reliably. Our findings not only replicate prior work suggesting transdiagnostic interoceptive sensibility difficulties in EDs but also offer novel insights by including a broader range of ED measures as well as investigating the moderating effects of sociodemographic factors. Importantly, this sample was substantially more diverse than those typically represented in ED research, including participants across a wide range of racial backgrounds, gender identities, and sexual orientations. This study also included under-researched ED symptoms, such as that characteristic of ON (measured with the E-DOS) and ARFID (measured with the NIAS). This diversity strengthens the ecological validity of the findings and provides valuable insight within under-represented populations in ED research.

Conclusions

This study provides evidence that self-reported interoceptive sensibility difficulties are associated with a broad range of self-reported ED symptoms. Correlational analyses indicated that higher ISQ scores were associated with ED symptoms across multiple measures, including the EDE-Q, NIAS, BEDS-7, and E-DOS. The strongest associations were observed with the NIAS Appetite and Picky Eating subscales, followed by the EDE-Q Eating Concern and Shape Concern subscales, supporting the view that interoceptive sensibility difficulties are linked to both behavioural and cognitive-affective features of EDs.

Taken together, these findings suggest that altered interoceptive sensibility represents a transdiagnostic characteristic of EDs, with variation in the strength of associations across symptom domains. Although BEDS-7 and E-DOS scores showed statistically significant associations with ISQ scores, these relationships were weaker than those observed for the NIAS and EDE-Q. This pattern indicates that interoceptive sensibility alterations are not confined to a single ED diagnostic construct but constitute a shared feature across ED presentations. Moreover, the stronger correlations observed for the NIAS Appetite and Picky Eating subscales, alongside moderate correlations for the EDE-Q Eating Concern and Shape Concern subscales, suggest that interoceptive sensibility difficulties may be more closely linked to specific behavioural and cognitive symptom domains than others, providing insight into which aspects of ED symptomatology may be most influenced by disrupted interoceptive sensibility.

Our study also evidenced that interoceptive sensibility is influenced by sociodemographic factors, particularly racial background, sex assigned at birth, and sexual orientation. This is an important consideration for future investigations and individualising of interoception-informed ED care that is culturally adapted and socially responsive.

Therefore, these findings support further investigation into the incorporation of interoceptive sensibility screening and interoception-informed interventions within ED clinical practice. Interventions aimed at improving interoceptive sensibility, or at developing individualised support strategies to reduce the impact of interoceptive difficulties on eating behaviours, may facilitate the re-establishment of a more compassionate and trusting relationship with the body. Future research should examine whether altered interoceptive sensibility precedes ED onset, identify its neurobiological correlates, and evaluate culturally informed approaches to enhance the clinical relevance and applicability of interoception-informed ED care.

Abbreviations

α

Cronbach’s alpha (measure of internal consistency)

AN

Anorexia Nervosa

ANZAED

Australia and New Zealand Academy for Eating Disorders

ARFID

Avoidant/Restrictive Food Intake Disorder

BED

Binge Eating Disorder

BEDS-7

Binge Eating Disorder Screener

BN

Bulimia Nervosa

DOS

Düsseldorf Orthorexia Scale (E-DOS: English version)

DSM

Diagnostic and Statistical Manual of Mental Disorders

ED

Eating Disorder

EDs

Eating Disorders

EDE-Q

Eating Disorder Examination Questionnaire

EDNA

Eating Disorders Neurodiversity Australia

EHQ

Eating Habits Questionnaire

IP

Internet Protocol

ISQ

Interoception Sensory Questionnaire

LGBTQIA

Lesbian, Gay, Bisexual, Transgender, Queer/Questioning, Intersex, Asexual

MABT

Mindful Awareness in Body-Oriented Therapy

N

Sample Size

NIAS

Nine Item ARFID Screener

ON

Orthorexia Nervosa

OSFED

Other Specified Feeding or Eating Disorder

p

P-value (probability value indicating statistical significance)

rho

Correlation coefficient (Spearman’s or rank-biserial)

Author contributions

Laurence Cobbaert: Conceptualisation; data curation; formal analysis; investigation; methodology; project administration; resources; writing—original draft; writing—review and editing. Phillipa Hay: Methodology; supervision; writing—review and editing. Philip B. Mitchell: Methodology; supervision; writing—review and editing. ain Perkes: Methodology; supervision; writing—review and editing.

Funding

No funding was obtained for this research. The prize draw was self-funded by the first author (LC).

Data availability

The data for this research is available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

A detailed overview of the study and its goals, including potential risks of discomfort, was made available online to prospective participants. By choosing to complete the anonymous online survey, participants provided implied consent.

Ethics approval for this study was obtained from the University of New South Wales Human Research Ethics Committee (Reference Number: iRECS7680).

Competing interests

Laurence Cobbaert is on the Board of Directors of not-for-profit organisation Eating Disorders Neurodiversity Australia (EDNA). They receive no direct financial remuneration from their Directorship position with EDNA. Laurence Cobbaert is also the Founder and Chair of the neurodiversity special interest group at the Australia and New Zealand Academy for Eating Disorders (ANZAED). Laurence Cobbaert is a PhD candidate and receives a PhD stipend from the University of New South Wales, Sydney, Australia. Phillipa Hay is the Editor-in-Chief for the Journal of Eating Disorders. However, she was not involved in the peer review process regarding this manuscript.