In a pilot randomized controlled trial evaluating specialized body image interventions for young psychiatric inpatients, adding intensive mirror exposure therapy to standard inpatient care failed to produce greater improvements in global body image disturbance or overall eating disorder pathology compared to standard care alone. The investigation, conducted within an inpatient child and adolescent psychiatry department, examined whether twelve structured mirror confrontation sessions delivered over four weeks could provide measurable incremental benefits beyond standard inpatient treatment. Although participants across both study arms demonstrated marked reductions in overall body image disturbance over hospitalization, the intervention group showed no statistically significant advantage on standardized global outcome measures. Eye-tracking assessments further revealed that a pronounced cognitive bias toward self-defined unattractive body regions persisted unchanged following treatment.

Body image disturbance represents a defining maintaining factor in anorexia nervosa, contributing directly to disease progression, resistance to clinical intervention, and elevated rates of post-discharge relapse. While mirror exposure has accumulated empirical support for reducing body-related distress among adult cohorts, empirical investigations addressing adolescent populations remain exceedingly scarce. Most previous evaluations of exposure-based techniques in younger demographics have been embedded within multimodal treatment packages, leaving the specific therapeutic contribution of mirror confrontation unisolated and poorly understood. Moreover, anorexia nervosa typically emerges during adolescence and carries severe somatic complications alongside exceptionally high mortality rates. The research team conducted this pilot trial to determine whether standalone mirror exposure could feasibly be integrated into intensive inpatient schedules and whether it could meaningfully modify maintaining cognitive biases in youth.

Recruitment occurred through the Department of Child and Adolescent Psychiatry at Freiburg University Medical Center between 2018 and 2023, identifying adolescent patients aged 12 to 21 with an International Classification of Diseases tenth edition diagnosis of anorexia nervosa or atypical anorexia nervosa. Clinical protocols mandated that participants achieve a body mass index age percentile of at least the tenth percentile prior to study entry, ensuring sufficient medical stability and cognitive capacity to engage with exposure exercises. Out of sixty-three eligible individuals contacted, twenty-eight provided informed consent, reflecting a relatively low consent rate attributed to the clinical burden of the population and the demanding nature of the intervention. Two participants presented with baseline values marginally below the cutoff but were retained because their measurements were not statistical outliers.

The final analyzed sample comprised twenty-four female adolescent inpatients, randomly allocated to either the experimental mirror exposure condition consisting of eleven patients or the treatment as usual control condition comprising thirteen patients. Within the cohort, nineteen individuals met diagnostic criteria for restrictive-type anorexia nervosa, four met criteria for purging-type anorexia nervosa, and one met criteria for atypical anorexia nervosa. Diagnostic screening identified substantial psychiatric comorbidity across the sample, reflecting profiles commonly encountered in adolescent eating disorder wards. Social phobia was documented in fourteen participants, depressive episodes in six, and generalized anxiety disorder in four. Baseline statistical comparisons revealed no significant pre-treatment differences between cohorts regarding chronological age, body mass index percentiles, or the accumulated duration of inpatient psychiatric treatment.

All participants engaged continuously in the hospital’s standard treatment as usual regimen, which maintained a primary clinical emphasis on medical weight restoration, nutritional normalization, and scheduled individual psychotherapy sessions without structured body image modules. Patients randomized to the active experimental arm received their baseline treatment supplemented by twelve individual sixty-minute mirror exposure sessions conducted over four weeks according to a standardized protocol. To ensure therapeutic rigor, all exposure sessions were administered by trained post-graduate psychologists who received ongoing clinical supervision from an experienced licensed psychotherapist throughout the trial. Ethical oversight was granted by the ethics committee of the Albert-Ludwigs University Freiburg, and the complete trial was formally preregistered in the German Clinical Trials Register under registration number DRKS00019104.

To assess whether mirror exposure influenced underlying cognitive maintaining mechanisms, the research protocol incorporated advanced computerized eye-tracking technology using specialized iViewXTM2 software and BeGaze analysis tools to capture visual attention patterns. Attentional biases were quantified across two distinct laboratory paradigms: an exogenous cueing task and a dot-probe visual presentation task. In the exogenous cueing paradigm, patients viewed photographic images depicting their own physical body or a body mass index-matched control body for 3000 milliseconds before responding to an attention-orienting cue. Investigators computed normalized fixation duration indices and relative fixation frequencies directed toward three personally identified most attractive and unattractive anatomical areas of interest. In the dot-probe paradigm, stimulus pairings displaying self-body, control body, and vase images were shown for 3000 milliseconds, recording initial fixation orientations.

The primary analytical framework evaluated treatment outcomes through mixed-design multivariate analysis of variance and mixed analyses of variance with group assignment and assessment time points serving as primary factors. Primary endpoints encompassed standardized global measures of body image disturbance, including body dissatisfaction, checking habits, and avoidance, alongside general eating disorder pathology. Baseline directional gaze biases were initially scrutinized through paired-samples t-tests contrasting fixation duration across self-perceived unattractive and attractive anatomical zones against matched control bodies. Attentional trajectories across the intervention period were modeled using four-way mixed analyses of variance incorporating group allocation alongside within-subject factors of time, evaluative valence, and body referent. In addition, the investigators conducted secondary correlational analyses utilizing Spearman’s rank correlation coefficients, while exploratory post-hoc Huber regressions evaluated potential predictors of clinical improvement.

In the primary multivariate analysis of variance on global body image disturbance, researchers detected no significant group-by-time interaction, with an F-statistic of 0.87, p = 0.361, partial eta squared of 0.038, and Wilks’ lambda of 0.962. Similarly, the mixed analysis of variance examining general eating disorder pathology showed no significant group-by-time interaction, yielding an F-statistic of 0.979, p = 0.333, and partial eta squared of 0.043. In contrast, the analysis demonstrated a statistically significant main effect of time on body image disturbance, with an F-statistic of 284.06, p < 0.001, and partial eta squared of 0.928. These findings confirmed that while adolescents in both cohorts experienced substantial improvements over time, mirror exposure conferred no additive clinical advantage. Furthermore, changes in specific body image dimensions correlated strongly with reductions in overall eating disorder pathology.

Baseline eye-tracking data verified the presence of a pronounced cognitive attentional bias, as participants exhibited significantly longer fixation durations directed toward self-defined unattractive areas of their own bodies compared to equivalent regions on control bodies, yielding a t-statistic of 3.04 with a p-value of 0.006. No corresponding attentional bias emerged when inspecting self-defined attractive physical features, producing a non-significant t-statistic of negative 1.38 and a p-value of 0.182. While a four-way interaction reached statistical significance for fixation duration, researchers noted that this effect violated statistical assumptions and primarily reflected viewing shifts toward control bodies rather than self-directed gaze changes. In the dot-probe task, an apparent group-by-time interaction for self versus control bodies was compromised by significant pre-intervention baseline differences between groups, while correlations between bias shifts and symptom changes were nonsignificant.

Despite the absence of divergence on global clinical scales, exploratory post-hoc analyses targeting specific subscales and individual questionnaire items revealed nuanced behavioral adjustments within the mirror exposure cohort that did not emerge among control patients. Repeated-measures analyses revealed significant group-by-time interaction effects on measures capturing specific checking, avoidance, and weighing behaviors, most notably within the grooming and weighing subscale of the Body Image Avoidance Questionnaire. Item-level examinations identified localized reductions in the frequency of compulsive mirror checking, inspecting physical reflections in alternative surfaces, and persistent distress surrounding self-weighing routines. In contrast, exploratory repeated-measures analyses assessing momentary affect and evaluative body appraisal failed to demonstrate any significant main effects or interaction patterns across treatment groups. Investigators cautioned that because these post-hoc analyses were exploratory and involved multiple testing, findings require cautious interpretation.

Investigation into clinical predictors of treatment response revealed that participant age and baseline disease severity significantly influenced the magnitude of psychological change achieved during inpatient care. Bivariate correlations established that reductions in body dissatisfaction were positively associated with chronological age, yielding a Spearman’s rho of 0.433 with a p-value of 0.039, as well as baseline eating disorder severity, which produced a correlation coefficient of 0.430 and a p-value of 0.036. Robust Huber regression models demonstrated that older adolescent patients achieved greater reductions in body dissatisfaction regardless of whether they received mirror exposure or standard inpatient care, showing a regression coefficient of 4.42 in the exposure cohort and 11.09 in the control group. An interaction regression model confirmed the absence of a significant age-by-group interaction, suggesting that older adolescents might possess greater cognitive maturity and introspective reflective capacity.

Regarding implementation feasibility, the twelve-session mirror exposure regimen demonstrated high treatment acceptability among participating adolescents, with ninety-one percent stating they would recommend the intervention, eighty-two percent evaluating it as personally helpful, and seventy-three percent expressing satisfaction with session frequency, even though all participants acknowledged the exposure exercises were challenging. Nevertheless, authors detailed substantial methodological limitations, foremost being the small final sample of twenty-four patients, which severely restricted statistical power. Additional constraints included the absence of formal treatment integrity assessments and potential treatment contamination resulting from unblinded group allocation and informal peer interactions within the shared inpatient ward environment. Furthermore, planned three-month post-discharge follow-up evaluations suffered from a low response rate of only forty-one percent, precluding longitudinal analysis and highlighting logistical challenges in this population.

Subject of Research: Feasibility and clinical effects of mirror exposure therapy on body image disturbance and attentional bias in adolescent inpatients with anorexia nervosa

Article Title: Mirror exposure in adolescents with anorexia nervosa: a feasibility study on body image disturbance and attentional bias

Article References: Pelzer, M., Werthmann, J., Svaldi, J., Fleischhaker, C., Haack-Dees, B., Renner, F., & Tuschen-Caffier, B. (2026). Mirror exposure in adolescents with anorexia nervosa: a feasibility study on body image disturbance and attentional bias. Eating and Weight Disorders – Studies on Anorexia, Bulimia and Obesity, 31(1), Article 54. https://doi.org/10.1007/s40519-026-01880-2

Image Credits: AI Generated

DOI: 10.1007/s40519-026-01880-2

Keywords: anorexia nervosa, mirror exposure therapy, adolescent inpatients, body image disturbance, attentional bias, eye-tracking, treatment as usual, randomized controlled trial, body dissatisfaction, body checking, avoidance behavior, eating disorder pathology, exogenous cueing task, dot-probe task, weight restoration, child and adolescent psychiatry, clinical feasibility, grooming and weighing subscale