Table 3 Details of studies that used tactile and tactile-less embodiment methods in ED populations

Eshkevari et al. [26]

Do people with EDs (AN, BN, or an eating disorder not otherwise specified [EDNOS]) report a stronger experience of the RHI than healthy controls (HCs)?

N = 139 females

(AN = 36; Age: M = 23.00 years; BMI: M = 16.10)

(EDNOS = 20; Age: M = 27.50 years; BMI: M = 19.70)

(BN = 22; Age: M = 22.50 years; BMI: M = 20.90)

(HC = 61; Age: M = 24.00 years; BMI: M = 21.5)

Location: UK

Participants saw a rubber hand being stroked synchronously or asynchronously with their own unseen hand

Embodiment measured via a subjective embodiment questionnaire and objective proprioceptive drift

Participants with EDs reported a significantly stronger experience of the RHI on both subjective and objective measures. Note: no comparison between ED groups reported by authors

Eshkevari et al. [1]

Do people who have recovered from an ED show differences in embodiment (RHI) when compared to current ED and HC participants?

N = 167 females

(current ED and HC groups as per Eshkevari et al. [26]

(Recovered ED = 28; Age: M = 25.5 years; BMI: M = 20.7)

Location: UK

As per Eshkevari et al. [26]

Recovered and current ED participants experienced similar levels of subjective embodiment; both groups showed higher levels of subjective embodiment compared to HCs. For the objective measure, no difference was reported between any groups

Note: there was a failure to replicate the finding between ED and HC groups for the objective measure, despite being the same sample as Eshkevari et al. [26] (likely due to insufficient power with the addition of the recovered ED sample, although rounded up scores differed between studies)

Keizer et al. [27]

Does the experience of ownership over a rubber hand change body size perception in AN participants?

N = 60 females

(AN = 30; Age: M = 26.37 years; BMI: M = 17.5 and HC = 30; Age: M = 21.80 years; BMI: M = 21.19)

Location: Netherlands

Participants saw a rubber hand being stroked synchronously or asynchronously with their own unseen hand

Embodiment measured via a subjective embodiment questionnaire and objective proprioceptive drift. Hand size over-estimation was used as a measure for the effect of embodiment

AN participants showed stronger RHI than HCs when measured subjectively, but only in the synchronous condition. For proprioceptive drift, no difference between groups

Reduced hand size over-estimation occurred in both synchronous and asynchronous conditions (compared to pre-embodiment) in AN participants only

Keizer et al. [28]

Does experiencing a FBI change the estimation of body parts that are more emotionally salient than the hand?

N = 59 females

(AN = 30; Age: M = 22.03 years; BMI: M = 18.11 and HC = 29; Age: M = 21.07 years; BMI: M = 20.77)

Location: Netherlands

Through VR goggles, participants observed a virtual avatar’s body (from a first-person perspective) being stroked in synchrony or asynchrony with their own unseen abdomen

Embodiment measured via a subjective embodiment questionnaire. Objective body size estimation task (circumference and width of abdomen, hips, shoulders; body height) was used as a measure for the effect of embodiment

AN and HC groups did not differ in the strength of subjective embodiment. Note: no analyses were performed using the total subjective questionnaire; only with each subscale

Experiencing a FBI decreased both AN participants’ and HCs’ over-estimation of all body parts, except for abdomen circumference (only AN participants showed a decrease in abdomen circumference over-estimates)

Over-estimation changes from pre to follow-up (2 h and 45 min after the FBI) were larger in AN participants than HCs in shoulder width and circumference, and hip circumference estimations

At follow-up after the FBI, AN participants shoulder size estimation was fully normalised, although a larger percentage of body size overestimations for the abdomen and hips were still present when compared to HCs

Malighetti et al. [40]

Can a VR induced FBI improve body satisfaction and body estimation accuracy, and decrease body image related concerns?

N = 7 females with AN (Age: M = 17 years; BMI: M = 15.95)

Location: Italy

Participants experienced tactile-less visuo-motor synchronization with an avatar in the VR setting from first- and third-person perspectives. The BMI of the avatar corresponded to the participant’s real BMI at the beginning, but increased over successive sessions until it reached a healthy BMI

Embodiment measured by a subjective embodiment questionnaire. Body size estimation task required participants to verbally specify how to change the avatar’s body to match it with their ideal and perceived real body sizes. Body satisfaction and image related concerns were assessed by questionnaires and scales pre and post VR intervention

Participants reported a trend of reducing body dissatisfaction and body image related concerns after VR intervention

After the intervention, they preferred a body with a closer to normal BMI

Porras-Garcia et al. [29]

Can a VR induced FBI added to treatment as usual (TAU; including nutritional rehabilitation, group counseling, and cognitive-behavioural therapy) decrease the fear of gaining weight (FGW) and other body-related disturbances?

N = 35 females with AN (Age: M = 18.73 years; BMI: M = 17.42)

Location: Spain

Participants were exposed to a virtual body with their real size and BMI, which slightly increased over following sessions until their healthy BMI target was reached. Both tactile-less visuo-motor and visuo-tactile stimulations were used

Self-reported FBI and FGW levels were evaluated in the VR setting using a one item subjective analogue scale for the intensity of the FBI and FGW, respectively. Body image disturbances were assessed by several questionnaires pre and post intervention, and at follow-up

The experimental group who received both TAU and VR intervention had significantly lower levels of FGW and body image disturbances than the control group who only received TAU at follow-up (3 months) as well as after the intervention

Porras-Garcia et al. [30]

Can a VR induced FBI effectively measure key body-related cognitive and emotional responses in AN?

N = 73 females

(AN = 30; Age: M = 17.73 years; BMI: M = 17.55 and HC = 43; Age: M = 21.12 years; BMI: M = 21.94)

Location: Spain

Through a head-mounted display, participants observed a virtual avatar’s body (from a first-person perspective) being stroked in synchrony/asynchrony with their own unseen abdomen

Embodiment measured via a one item subjective visual analogue scale for the intensity of the FBI

The level of FBI experience was significantly lower in AN participants than HCs. There was a significant negative relationship between FBI and body image disturbances (i.e., in people with AN, as disturbances increased, susceptibility to FBI decreased)

Provenzano et al. [2]

Does the experience of a VR induced embodiment illusion reduce body dissatisfaction in AN participants?

Do AN and HC groups differ with respect to emotional reactions to experiencing embodiment?

N = 40 females

(AN = 20; Age: M = 23.30 years; BMI: M = 15.87 and HC = 20; Age: M = 23.85 years; BMI: M = 18.98)

Location: Italy

Through a head-mounted display, participants observed three virtual avatar’s bodies (from a first-person perspective) being stroked in synchrony or asynchrony with their own unseen abdomen. These avatars varied in relation to body size (lower than real BMI, real BMI, and larger than BMI). Each avatar was experienced one at a time

Embodiment measured via subjective visual analogue scales. Affective reactions were measured using visual analogue scales in response to the experience of embodiment with each avatar. Dissatisfaction was measured using a perceived vs ideal body task (i.e., out of a series of avatars, participants chose those which they felt related to (i) their real size and (ii) their ideal size)

AN participants showed more negative response to the larger avatar used for embodiment compared to HCs. HCs preferred the larger rather than smaller avatar, whilst AN participants preferred the smaller rather than larger avatar

AN and HC groups did not show any change in body dissatisfaction regardless of avatar body size

Serino et al. [39]

Does the illusion of owning a virtual body change body size estimation in AN participants?

N = 23 females suffering from AN; Age: M = 22.76 years; BMI: M = 15.50

Location: Italy

Participants watched a virtual body (abdomen) being stroked synchronously or asynchronously with their own abdomen through a head-mounted display. The illusion was delivered in two sessions, one before (T1) and one after (T2) experiencing treatment provided by a centre of excellence

The effect of embodiment was measured via an objective body size (abdomen, shoulders, and hips) estimation task

There were significant decreases in the circumference estimation of the abdomen and hips only following the illusion delivered after the treatment (T2)

Serino et al. [38]

Can VR-based embodiment illusions reduce body image distortions in an AN participant?

One female with AN; Age = 30 s (no exact information for age provided); BMI = 13.69

Location: Italy

The participant observed a virtual abdomen of a healthy-weight woman being stroked in synchrony or asynchrony with her own abdomen through a head-mounted display. The illusion was delivered in three sessions over the course of her outpatient treatment (i.e., start, end, and 1 year follow up)

Embodiment measured via subjective embodiment questionnaire. The effect of embodiment was assessed by an objective body size estimation task (abdomen, hips, and shoulders)

At the beginning of the treatment, she reported high scores for all three subscales of the embodiment questionnaire. No evidence for embodiment at end of treatment or 1 year follow-up, although she reported a high score for the asynchronous condition at follow up for the ownership subscale

Body size estimations for width were reduced at the start of treatment for synchronous compared to asynchronous stroking (it should be noted her general tendency to overestimate body width even before embodiment happened). At the end of the treatment, experiencing embodiment did not reduce estimations involving width (i.e., she still showed a general tendency for overestimating width). At 1 year follow-up, experiencing embodiment generally did not influence width estimations (although her general estimations were already accurate), although she over-estimated the width of her hips more after asynchronous embodiment. Estimations for body circumferences showed similar patterns to those for width across time points

In general, ‘more emotionally-laden sites’ such as the abdomen and hips were more susceptible to overestimation than shoulders throughout the procedure

Zopf et al. [31]

Do AN and HC groups differ in susceptibility to RHI?

N = 46 females

(AN = 23; Age: M = 21.87 years; BMI: M = 15.82 and HC = 23; Age: M = 21.48 years; BMI: M = 21.16)

Location: Australia

Participants saw a rubber hand being stroked synchronously or asynchronously with their own unseen hand

Embodiment measured via a modified subjective RHI questionnaire and objective reaching task

AN participants experienced a stronger RHI than HCs when measured subjectively and objectively