Introduction: The purpose of this research is to develop a new component for cognitive-behavioral treatment (CBT), empirically validated, for binge eating behavior in patients with bulimia nervosa (BN) and binge eating disorder (BED), by means of cue-exposure therapy (CET) with virtual reality. While CBT is an effective treatment for these eating disorders, an important percentage of patients do not improve despite treatment. It is necessary to explore how to enhance the effectivity of usual treatments with the incorporation of new technologies and procedures. The innovation of this investigation is based on the use of VR for cue-exposure therapy, which has been found effective with patients resistant to conventional treatments (CBT and pharmacological intervention), and seeks to enhance this efficacy by increasing possibilities of generalization and providing more useful tools to clinicians, diminishing the logistic complications of the exposure to real cues (food).

Participants: 60 patients with a diagnosis of BN or BED according the DSM-5 who had been proposed for participation in the study by their referral mental health's professional after conducting unsuccessfully a first level treatment (CBT structured intervention) at the clinical sites involved will participate in the study.

Procedure: Participants with active episodes of binge eating (and purging, in the case of BN) during the last 2 weeks of the structured intervention of CBT will be provided with detailed information about the study. All patients who agree to participate in the study and who sign the informed consent form will be randomly assigned to one of the two second-level treatment conditions: virtual reality based cue-exposure therapy (VR-CET) or additional cognitive-behavioral treatment (A-CBT). Both second-level treatment conditions consist of six 60-minute sessions that are held two times per week over a period of three weeks.

In an interview setting, all of the participants will be administered several questionnaires (BN, DT and BD subscales of EDI-3, FCQ-T/S, and STAI-Y) and participants assigned to the VR-CET will also be exposed to various types of virtual foods in different virtual environments (kitchen, dining-room, bedroom, and café) where they will be asked about their food cravings and anxiety experienced in each situation to develop a proper exposure hierarchy according to the specific characteristics and needs of each patient.

Sample Sixty patients who have been diagnosed with BN or BED according to the Diagnostic and Statistical Manual of Mental Disorders Fifth Edition (DSM-5) and who have been found to be resistant to an initial structured program of CBT (Fairburn, Marcus, & Wilson, 1993) will participate in the study. Participants will be recruited from collaborating hospitals and health centers including Hospital de Bellvitge (Barcelona, Spain), Adult Mental Health Center of the Consorcio Sanitario de la Anoia (Igualada, Spain), Hospital Universitari Joan XXIII (Tarragona, Spain), Centro ABB (Tarragona, Spain) and the Instituto Auxologico Italiano (Milano, Italy). Suicidal ideation and diagnoses of other severe mental disorders (substance use disorders, bipolar disorder, psychosis) are exclusion criteria. Patients with active episodes of binge eating (and purging, in case of BN) during the last two weeks of the structured program of CBT will be considered for participation in the study. All prospective participants will be comprehensively informed about the characteristics of the study and will be required to sign an informed consent form. Authorization of the research ethics committees of the applicant organization (University of Barcelona) and the collaborating institutions will be required.

Patients who agree to participate in the study will be randomly assigned to one of two second-level treatment conditions: the VR-CET group will receive cue-exposure sessions using VR environments (six 60-minute (maximum) sessions over the course of three weeks; two per week) and the A-CBT group will receive six additional sessions of the usual cognitive behavioral treatment. Consistent with the American Psychological Association and the CONSORT guideline recommendations regarding sample size in this type of research, 30 patients will be included in each group. If necessary, dropout cases will be replaced.

Procedure At the end of the first level of treatment with the structured program of CBT, all patients will be assessed relative to their status and will complete several questionnaires (BN, DT and BD subscales of EDI-3, FCQ-T/S, and STAI-Y) (pre-randomization assessment).

Those with active behavioral symptoms who agree to participate in the intervention study and who have signed the informed consent form will be randomly assigned to one of the two second-level treatment conditions: VR-CET or A-CBT sessions. Also, participants assigned to the VR-CET condition will be assessed through a validated VR-based cue exposure software (Gutiérrez-Maldonado, Pla-Sanjuanelo et al., 2016) in order to construct the exposure hierarchy of 3D interactive (environment and food) situations, used in the VR-CET sessions. For that, participants assigned to the VR-CET condition:

1. Will be exposed to 30 graphic representations of different foods. The level of craving elicited by each food will be assessed by means of a visual analog scale (VAS) that ranges from 0 (not at all) to 100 (extreme).
2. These participants will also be exposed to four graphic representations of different environments (kitchen, dining room, bedroom, and café). The level of craving elicited by each environment will be assessed by means of a VAS that ranges from 0 (not at all) to 100 (extreme).
3. They will be individually exposed to the ten different virtual foods assessed as eliciting the highest levels of craving in each of the four virtual environments (kitchen, dining room, bedroom, and café) for a total of 40 virtual situations from lowest to highest craving scores as previously registered in phases 1 and 2. Once in the virtual situation, the participant will be asked to sit at a virtual table where the food is placed. Once the participant sits, the time is measured. During the first 20 seconds, the patient can handle the virtual food by means of the laptop's mouse but cannot eat the food. When time runs out, the levels of craving and anxiety will be assessed by means of visual analog scales that range from 0 (not at all) to 100 (extreme). Virtual environments will be displayed by means of a 3D laptop with specific software that is capable of creating a stereoscopic effect that is duly processed by polarized glasses. Headphones will also be used. This assessment will be conducted in a quiet darkened room to increase the participant's immersion in the virtual scenarios. The therapist will remain behind the patient during the assessment so as not to interfere with the task but to be available in the event a participant requires assistance.
4. Anxiety scores obtained from each of the 40 situations will be used to develop the exposure hierarchy that will be used in the VR-CET sessions.

Virtual reality based cue-exposure therapy (VR-CET) sessions: There will be six 60-minute (maximum) exposure sessions (two sessions per week for three weeks). Participants (n=30) will be exposed to different VR environments related to binge behaviors, according to the previously constructed hierarchy. During exposure, patients will face high risk situations to diminish or extinguish the conditioned response of anxiety when exposed to food related cues. In each session, the patient will be exposed to the corresponding steps of the hierarchy.

Additional cognitive-behavioral treatment (A-CBT) sessions: Participants (n=30) in this condition will receive six CBT booster sessions (twice weekly) over the course of three weeks to improve the output of treatment. CBT sessions are based on the approach described by Eldredge and colleagues (Eldredge et al., 1997). In this treatment, patients are trained to self-monitor their food patterns and to identify and manage thoughts, emotions, and environmental factors related to disrupted eating behaviors.

VR based cue-exposure booster sessions The cue-exposure procedure used in this study is an adaptation of the cue-exposure treatment for BN proposed by Martinez-Mallen et al. (2007). Exposure sessions will be conducted individually in a quiet, darkened room. In each session, the participant will sit in front of a 3D laptop with software that is capable of creating a stereoscopic effect that is duly processed by polarized glasses. Headphones with active noise cancellation will also be used to expose the participant to the auditory stimuli in the virtual environments and, at the same time, to isolate the patient from external noises. Participants can interact within the virtual environments in real time. They can move around the scenarios, sit at a table, and they can handle foods displayed using the keyboard and the laptop's mouse.

During the first minutes of the session, patients will be reminded of the purpose of the exposure, that is, to place them in contact with food despite the high levels of anxiety they experience. Doubts and questions raised by the participants will be attended to at this time. The participants will be exposed to the corresponding virtual environments and foods based on the previously established hierarchy. Along the VR-CET sessions, craving and anxiety levels were assessed periodically. Exposure will terminate when subjective anxiety has decreased a 40% in relation to the level registered at the initiation of the exposure session or after 60 minutes. Once the anxiety level has decreased significantly, the patient will be exposed to the next level in the hierarchy. In the event the session terminates before the anxiety level has been sufficiently reduced in the virtual situations, the patients will be exposed to the same situation in the following session. Subjective craving and anxiety will be recorded using visual analogic scales (VAS) that range from 0 (not at all) to 100 (extreme) at the beginning, every minute, and at the end of the virtual scenario exposures. During exposure, the therapist will remain behind the patient and will only intervene if instruction is needed or if the patient needs assistance.

Assessment During each VR-CET session, subjective craving and anxiety will be assessed using VAS (range from 0 to 100) and will be recorded at the beginning, every minute, and at the end of the exposure.

The following questionnaires will be administered to the participants of the two groups before initiating treatment (pre-randomization phase), at the end of the second level-level treatment conditions, and at the six-month follow-up: STAI-Y (State-Trait Anxiety Inventory-Form Y), BN, DT and BD subscales of EDI-3 (Eating Disorders Inventory-3) and FCQ-T/S (Food Craving Questionnaire-Trait/State).

Finally, outcome assessment for core behavioural features included frequency (i.e., number) of binge eating episodes and purging during the two weeks prior to beginning the second-level treatment conditions, during the two weeks after the end of the intervention, and during two weeks after six-month follow-up.

Data analyses Mixed between-within subject analysis of variance will be conducted. Between-subject variable refers to both conditions, the VR-CET versus A-CBT. The within-subject variable refers to the assessment time: prior to beginning the second-level treatment conditions (both VR-CET and A-CBT conditions), at the end of the treatment, and at the six-month follow-up. Thus, comparisons between results obtained by both groups before treatment, after treatment and at follow-up for each of the assessed measures will be conducted. Using a similar analysis, anxiety measures registered during each exposure session will also be compared.

A significant interaction between condition (VR-CET vs. A-CBT) and time is expected in the first block of analyses. This would indicate that the results of the questionnaires as well as the number of binges and purges would remain stable before the second-level treatment conditions (pre-randomization phase), after the treatment and at follow-up in the control group (A-CBT), whereas these measures would be reduced in the experimental group (VR-CET) after treatment and at follow-up. In the second analysis block, it is expected that measures of subjective anxiety will decrease as each exposure session progresses.

Inclusion Criteria:

• Diagnosis of bulimia nervosa or binge eating disorder according the DSM-5
• Written and informed consent to participate

Exclusion Criteria:

• Other concurrent severe mental disorders (substance use disorders, bipolar disorder, psychosis)
• Suicidal ideation