See discussions, stats, and author profiles for this publication at: [Link]

net/publication/320016678

Psychophysiological Changes during Pure vs Guided Mirror Exposure

Therapies in Women with High Body Dissatisfaction: What Are They Learning
about Their Bodies?

Article in European Eating Disorders Review · September 2017

DOI: 10.1002/erv.2546

CITATIONS READS

22 550

5 authors, including:

Sandra Díaz-Ferrer Sonia Rodríguez-Ruiz

University of Granada University of Granada
10 PUBLICATIONS 62 CITATIONS 70 PUBLICATIONS 1,951 CITATIONS

SEE PROFILE SEE PROFILE

Blanca Ortega-Roldán Jose Luis Mata

University of Granada University of Granada
9 PUBLICATIONS 109 CITATIONS 91 PUBLICATIONS 730 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Sonia Rodríguez-Ruiz on 24 April 2018.

The user has requested enhancement of the downloaded file.

RESEARCH ARTICLE
Psychophysiological Changes during Pure vs Guided Mirror
Exposure Therapies in Women with High Body Dissatisfaction:
What Are They Learning about Their Bodies?
Sandra Díaz-Ferrer* , Sonia Rodríguez-Ruiz, Blanca Ortega-Roldán, José Luis Mata-Martín &
M. Carmen Fernández-Santaella
Department of Personality, Assessment and Psychological Treatment, Faculty of Psychology, University of Granada, Spain

Abstract
This study aimed to examine the psychophysiological changes resulting from two mirror exposure treatments that are effective at reduc-
ing body dissatisfaction. Thirty-five university women with body dissatisfaction and subclinical eating disorders were randomly assigned
to one of two groups: pure (n = 17) or guided exposure (n = 18). The participants received six sessions of treatment. Their thoughts,
feelings and avoidance behaviours were assessed after each session. Their subjective discomfort, heart rate and skin conductance were
assessed within the sessions. Both groups showed improvement in cognitive-affective and avoidance behaviour symptoms. Nevertheless,
the pure exposure group showed faster habituation of subjective discomfort and a greater physiological response than the guided expo-
sure group. These findings suggest that both procedures are effective interventions for improving body image disturbances, although psy-
chophysiological changes observed within session suggest that each technique would act through different processes. Copyright © 2017
John Wiley & Sons, Ltd and Eating Disorders Association.
Received 5 April 2017; Revised 5 August 2017; Accepted 14 August 2017

Keywords
body dissatisfaction; body discomfort; heart rate; skin conductance

*Correspondence
Sandra Díaz-Ferrer, PhD, Department of Personality, Assessment and Psychological Treatment, Faculty of Psychology, University of Granada, Campus de la Cartuja
s/n, 18071, Granada, Spain. Tel: (+34) 958 24 42 51; Fax: (+34) 958 24 37 49.
Email: sdiaz_1@[Link]

Published online in Wiley Online Library ([Link]) DOI: 10.1002/erv.2546

Introduction a therapist to describe their physical appearance (head, torso, legs,

arms and overall appearance) in detail as accurately and as neu-
Body image disturbances, particularly cognitive-affective ones, are trally as possible while observing themselves in a full-length mir-
one of the main risk factors for developing eating disorder (ED) ror. Recently, Moreno-Domínguez et al. (2012) evaluated whether
symptoms; their severity predicts treatment outcome, and their this neutral and accurate description is crucial for reducing body
persistence increases the risk of relapse (Keel, Dorer, Franko, dissatisfaction, proposing a new approach in the treatment of
Jackson, & Herzog, 2005; Farrell, Shafran, & Lee, 2006). Several body image disturbances called pure exposure. In this technique,
studies yield evidence of the effectiveness of different mirror ex- patients also observe their bodies in a full-length mirror but
posure techniques (e.g. cognitive dissonance-based mirror expo- without the therapist’s guidance through a body neutral descrip-
sure and mindfulness-based mirror exposure) to reduce body tion. Instead, the therapist helps patients to express their
image disturbances in women with EDs (Delinsky & Wilson, thoughts and emotions elicited by their bodies. This new ap-
2010; Jansen et al., 2016) and body-dissatisfied women (Luethcke, proach was first tested in a non-clinical sample comparing the ef-
McDaniel, & Becker, 2011; Moreno-Domínguez, Rodríguez-Ruiz, fectiveness of pure mirror exposure with guided mirror exposure
Fernández-Santaella, Jansen, & Tuschen-Caffier, 2012). Despite (Moreno-Domínguez et al., 2012). Body dissatisfaction, negative
this evidence, mirror exposure techniques in this field are in con- thoughts and feelings of ugliness were equally reduced in both
tinuous development to improve their effectiveness. treatments.
A common approach in the treatment of body image distur- In spite of these positive results, empirical evidence on the pro-
bances is guided mirror exposure. This technique has been dem- cesses by which guided and pure mirror exposure improves body
onstrated to be effective in reducing body dissatisfaction and satisfaction is still insufficient. With this aim, Trentowska et al.
related symptoms in individuals with EDs (Hilbert, Tuschen- (2013) examined the habituation process and the cognitive
Caffier, & Vögele, 2002; Trentowska, Bender, & Tuschen-Caffier, changes during guided mirror exposure in women with bulimic
2013; Trentowska, Svaldi, & Tuschen-Caffier, 2014; Tuschen- symptoms (BN) and in those without EDs. Over the course of
Caffier & Florin, 2002). In this approach, patients are guided by the treatment, negative thoughts and emotions decreased,

Eur. Eat. Disorders Rev. (2017) Copyright © 2017 John Wiley & Sons, Ltd and Eating Disorders Association.
Pure vs Guided Exposure S. Díaz-Ferrer et al.

whereas positive emotions increased, becoming significantly more the physiological responses should show the expected activation
numerous in women with BN. In all sessions, subjective discom- and reduction within and between sessions. For its part, the inhib-
fort experienced was greater in BN women. Between sessions, itory learning model assumes that feared associations are elimi-
subjective discomfort was characterized by a general tendency to- nated by the development of new non-threatening associations
wards habituation in both groups. Within session, subjective dis- that inhibit the original fear structure without removing or mod-
comfort increased in relation to the torso, legs and overall ifying it (Craske et al., 2008; Craske, Treanor, Conway, Zbozinek,
appearance and decreased in relation to the participants’ arms. & Vervliet, 2014). Accordingly, this model does not assume the
The authors suggest that these results are consistent with a classi- reduction of subjective fear and its associated physiological re-
cal habituation process (Foa & Kozak, 1986), although they pro- sponses within and between sessions.
pose further investigation to determine the precise mechanism The primary aim of this study was to replicate the results of
involved. Díaz-Ferrer et al. (2015) concerning the effectiveness of pure mir-
In line with the previous study, a recent study compared guided ror exposure compared with guided mirror exposure to reduce
mirror exposure with pure mirror exposure in university women body dissatisfaction, thoughts and avoidance behaviours in six
with BN through psychological indices recorded within and be- treatment sessions in female university students with high body
tween sessions (Díaz-Ferrer, Rodríguez-Ruiz, Ortega-Roldán, dissatisfaction and subclinical ED symptoms. The secondary aim
Moreno-Domínguez, & Fernández-Santaella, 2015). Over the was to examine the physiological (heart rate and skin conductance)
course of the treatment, both exposure procedures reduced nega- and emotional changes that occur in pure and guided mirror expo-
tive thoughts and increased positive thoughts. Nevertheless, body sure sessions. Consistent with previous studies (Díaz-Ferrer et al.,
dissatisfaction decreased more in the pure exposure group. Be- 2015; Moreno-Domínguez et al., 2012), we hypothesized that (i)
tween sessions, subjective discomfort decreased in both groups. both techniques will improve positive thoughts and body satisfac-
However, within sessions, in the pure mirror exposure, subjective tion and decrease negative thoughts and associated avoidance be-
discomfort gradually decreased, whereas in the guided mirror ex- haviours, (ii) the pattern of emotional and physiological changes
posure, subjective discomfort showed an initial decrease followed observed within the sessions will differ between the two exposure
by a progressive increase towards the middle of the session, simi- techniques and (iii) the pure exposure technique will elicit
lar to the results described by Trentowska et al. (2013). This dif- greater body discomfort and greater heart rate and skin conduc-
ferent pattern of changes in subjective discomfort between both tance responses within the sessions than the guided exposure
groups was also reported by Moreno-Domínguez et al. (2012) in technique.
a non-clinical sample. Based on these findings, pure mirror expo-
sure was identified as a simpler and easier technique for improv-
ing positive emotions and reducing body dissatisfaction among Method
women with BN and with non-clinical symptoms. The authors
interpret the results as evidence that different processes may be Participants
involved within each technique, and the question arises whether Thirty-five college women with high body dissatisfaction and sub-
such differences may be interpreted in terms of different habitua- clinical ED symptoms voluntarily participated in the study. These
tion processes. participants were screened from a sample of 623 women who
Physiological changes that occur within and between mirror responded voluntarily to the Body Shape Questionnaire (Cooper,
exposure sessions may help to clarify this issue. To our knowl- Taylor, Cooper, & Fairburm, 1987) and the Eating Attitude Test-
edge, only one study has examined the heart rate and skin con- 40 (EAT; Garner & Garfinkel, 1979; Peláez-Fernández, Ruiz-
ductance reactions during one single session of guided mirror Lázaro, Labrador, & Raich, 2014) during class in different
exposure in women with EDs (Vocks, Legenbauer, Wächter, departments at the University of Granada (Spain). The inclusion
Wucherer, & Kosfelder, 2007). In this study, no significant heart criteria were as follows: (i) a total Body Shape Questionnaire score
rate changes were found; only an increase in skin conductance ≥105 (to select participants with high body dissatisfaction), (ii) a
was found from baseline to initial mirror exposure. Recently, a total EAT-40 score ≥20 or ≤40 (to select women without clinical
non-treatment study in a group of women with low and high eating-related pathologies), (iii) a body mass index (BMI) be-
body dissatisfaction examined heart rate and skin conductance re- tween 18 and 29 and (iv) an age range of 18–30 years (to select
sponses caused by confronting different body parts (Servián- university women with similar age characteristics). After pre-
Franco, Moreno-Domínguez, & Reyes del Paso, 2015). Greater screening, potential participants (n = 60) were provided an indi-
physiological responses were found in both groups when focusing vidual appointment by telephone to participate in an assessment
on specific body parts, such as the thighs, buttocks and belly. interview. Participants were excluded if they (i) had symptoms
These studies did not examine habituation of the physiological re- of substance abuse or addiction, (ii) were under psychological
sponses between sessions. and/or psychiatric treatment for EDs or other mental disorders,
Two models have been proposed to explain the improvement (iii) were receiving or had received body image treatment, (iv)
in body satisfaction and in positive emotions and thoughts after were currently following a weight loss programme or (v) had his-
mirror exposure treatment: habituation and inhibitory learning. tory of a cardiovascular disease/disorder. Eligible participants
The habituation model assumes that mirror exposure to one’s (n = 36) were randomly assigned to one of the two treatment
own body would lead to the initial activation of the fear structure groups (pure exposure, n = 18, and guided exposure, n = 18).
related to the body, followed by habituation within and between One participant abandoned the pure group because of health
sessions (Trentowska et al., 2013; Vocks et al., 2007). Accordingly, problems during the first session, leaving a sample of 35

Eur. Eat. Disorders Rev. (2017) Copyright © 2017 John Wiley & Sons, Ltd and Eating Disorders Association.
S. Díaz-Ferrer et al. Pure vs Guided Exposure

individuals seeking treatment (see patient flow in Figure S1). Each attend to the thoughts and feelings that arose while observing
participant was compensated with course credit for her their bodies without resisting them. To check for avoidance, dur-
participation. ing the session, the participants were asked to freely and continu-
ously indicate the different areas of their bodies that they were
Procedure focusing on and to verbalize their thoughts and feelings. The ther-
In the first appointment, a trained clinical psychologist conducted apist gave these instructions to help patients before starting the
an individual semi-structured investigator-based interview that session:
assessed whether the inclusion and exclusion criteria were met.
Participants who met the criteria and agreed to take part in the ‘I’m going to ask you to stand in front of the mirror, watching your
study completed the baseline psychological measures to assess whole body; while you are doing it, let your thoughts and emotions
the following: (i) positive/negative thoughts, (ii) behavioural flow, do nothing to counter them, and keep talking to me. You can
avoidance and (iii) feelings of satisfaction with one’s body. Finally, tell yourself how you feel and what you are experiencing...do not
worry if throughout the session you experience thoughts and emo-
the participants were weighed and measured, and they signed an
tions; share them with me. For example, if you are looking at the
informed consent form according to the University’s Ethical shape of your ear, you can tell me what you are looking at, how
Committee. After this, all participants attended six individual ses- you see it, and what thoughts come to your head when you are
sions of 40- to 45-min duration twice a week for over 3 weeks. watching’.
The participants were provided a set of beige underwear (T-shirt
and shorts), which they were required to wear during the expo- Moreover, every 5 min, the therapist inquired about body dis-
sure session. For both treatments, participants were instructed comfort and asked what the participants were thinking and the
to remain in front of the mirror in a stable position and were areas of their body they focused on. In this procedure, the partic-
asked to limit movements during exposure. The overall exposure ipants could observe the body areas in the order that they wanted,
session was subdivided into nine exposure trials of 5-min dura- with no guidance from the therapist. The body areas reported by
tion to evaluate subjective discomfort changes and physiological participants during the nine exposure trials were classified into
reactions. At the beginning of each exposure session and in all the same categories used in the guide exposure group. General in-
nine exposure trials, subjective discomfort was evaluated using a structions for each group were identical in each session.
10-point subjective discomfort scale. In the first and last treat-
ment sessions, skin conductance and heart rate were continuously
recorded during the nine exposure trials. The trials of physiolog- Measures
ical measures began every time that the participants were asked
about their subjective level of discomfort. Because the goal was Psychological measures
to capture the initial reaction to this question, only the first 37 sec- The Body Image Automatic Thoughts Questionnaire (BIATQ;
onds of the 5-min trial exposures were considered (because of Cash, Lewis, & Keeton, 1987) is a 52-item questionnaire that as-
preliminary analyses comparing different latency windows). The sesses the frequency with which an individual has had negative
participants were asked to rate positive and negative thoughts, be- or positive thoughts regarding his or her body based on a Likert
havioural avoidance and feelings of satisfaction with one’s body scale, ranging from never (0) to very often (4). It has two subscales:
after the first and last treatment session. All treatment sessions positive thoughts (15 items) and negative thoughts (37 items).
were conducted by the same therapist at the time scheduled with This instrument was validated in a Spanish population (Perpiñá
each participant (10:00, 12:00, 17:00 or 19:00 h) and were et al., 2003). The overall alphas for the positive and negative
matched in both groups. BIATQ in this study were 0.81 and 0.94, respectively.
The Body Image Avoidance Questionnaire (BIAQ; Rosen, Srebnik,
Treatments Saltzberg, & Wendt, 1991) is a 19-item questionnaire that assesses be-
Participants were required to observe their bodies in a mirror that havioural manifestations of body image disturbances based on a
was sufficiently large to view themselves from all angles. In the Likert scale, ranging from never (0) to always (5). In accordance with
guided exposure group, the participants had to describe their Vocks, Kosfelder, Wucherer, and Wächter (2008), we developed a
body in the most neutral and objective manner possible, in accor- general score to assess body-related avoidance behaviours that
dance with the manual by Tuschen-Caffier and Florin (2002). consisted of 13 items from the BIAQ. This score reflects the tendency
With the aid of this manual, the therapist guided the participants to disguise or hide appearance with specific clothes and to avoid social
to begin focusing on various areas of the body, always in the same situations in which weight or appearance could become a focus of
order in every session: head, torso, lower extremities and upper attention. The BIAQ overall alpha in this study was 0.71.
extremities, ending with a body overview. These five general body The Visual Analogue Scale was used to measure the body
areas were divided across the nine trials as follows: head 1 (eyes, dissatisfaction-satisfaction dimension. The word dissatisfaction
eyebrows, nose, chin, mouth and teeth), head 2 (ears, cheeks, face appeared on the left side of the visual scale, and the satisfaction
skin, hair and overall head), torso 1 (neck, shoulders and breasts), appeared on the right side. The participants rated her satisfaction
torso 2 (waist, hips and belly), back (back and buttocks), lower by marking a vertical mark on the 100-mm line. A score of zero
extremities 1 (legs), lower extremities 2 (feet and overall legs), up- was assigned to the endpoint of satisfaction, and a score of 100
per extremities (arms) and overview. After each 5-min body area was assigned to the endpoint of dissatisfaction.
description, the therapist asked the participant to rate subjective The Subjective Discomfort Scale was used to measure the level
discomfort. In the pure exposure group, the participants had to of subjective discomfort within each treatment session. The

Eur. Eat. Disorders Rev. (2017) Copyright © 2017 John Wiley & Sons, Ltd and Eating Disorders Association.
Pure vs Guided Exposure S. Díaz-Ferrer et al.

participants reported their level of discomfort using a 10-point Results

scale (1 = minimum level of discomfort and 10 = maximum level of
discomfort). Participant characteristics and clinical symptoms at
baseline
The groups did not differ in age (F(1, 33) = 2.31, p < .13,
Physiological measures
ηp2 = .06) or BMI (F(1, 33) = .77, p < .38, ηp2 = .02). There were
Skin conductance, expressed in microsiemens, was measured no significant differences in measures of positive (F(1, 33) = .79,
using two Ag–Ag Cl electrodes placed on the hypothenar area of p < .38, ηp2 = .02) and negative thoughts (BIAQT; F(1, 33) = .62,
the left hand using a Biopac GSR100C amplifier (Biopac Systems, p < .43, ηp2 = .01), body satisfaction (Visual Analogue Scale;
Inc., Goleta, CA, USA). Heart rate, expressed in beats per min, was F(1, 33) = .28, p < .59), ηp2 = .00) and avoidance behaviours
measured through electrocardiography using three disposable elec- (BIAQ; F(1, 33) = .87, p < .35, ηp2 = .02) at baseline (Table 1).
trodes that were placed according to the lead II configuration and
connected to a Biopac ECG100C amplifier. The continuous record- Psychological measures between exposure
ing of both variables was analysed using ACQKNOWLEDGE 4.2 soft- sessions
ware (Biopac Systems, Inc., Goleta, CA, USA). Physiological values
second by second were transformed into differential scores with The 2 × 2 ANOVA [Group (pure exposure, guided exposure) ×
respect to the baseline of each trial (the second prior to the trial). Session (Session 1, Session 6)] results for positive and negative
thoughts (BIAQT) indicate significant effects for the Session fac-
tor (positive thoughts: F(1, 33) = 5.85, p < .021, ηp2 = .15; negative
Statistical analyses thoughts: F(1, 33) = 32.98, p < .0001, ηp2 = .50). The Group factor
Differences in age, BMI and psychological measures between the and the interactions were not significant in any case. There was an
two groups at baseline were analysed using univariate analyses increase in positive thoughts and a decrease in negative thoughts
of variance (ANOVAs). Changes in psychological measures be- in the last session in both groups (Table 1).
tween the first and last treatment sessions were analysed using 2 The 2 × 2 ANOVA [Group (pure exposure, guided exposure) ×
(Group) × 2 (Session) repeated-measures ANOVAs. Psychological Session (Session 1, Session 6)] results for feelings of satisfaction
measures within sessions were analysed using 2 (Group) × 2 (Visual Analogue Scale) exhibited significant effects for the Ses-
(Session) × 10 (one baseline and nine trials) repeated-measures sion factor (F(1, 33) = 46.85, p < .0001, ηp2 = .58). The Group fac-
ANOVAs. Physiological measures within sessions were analysed tor and the interactions were not significant. There was an
using 2 (Group) × 2 (Session) × 9 (Trials) × 37 (Time) repeated- increase in feelings of body satisfaction in the last session in both
measures ANOVAs. Because of the presence of artefacts in the re- groups (Table 1).
cordings, one participant (pure exposure) was excluded only from The 2 × 2 ANOVA [Group (pure exposure, guided exposure) ×
the statistical analysis of physiological measures. The body parts Session (Session 1, Session 6)] results for avoidance behaviours
more frequently informed by the participants were analysed using (BIAQ) revealed significant effects for the Session factor (F(1,
frequency analysis. The Greenhouse-Geisser correction was ap- 33) = 35.04, p < .0001, ηp2 = .51). The Group factor and the inter-
plied to the repeated-measures factors to control for sphericity vi- actions were not significant. There was a decrease in avoidance
olation. The results are reported with the original degrees of behaviours in the last session in both groups (Table 1).
freedom, the corrected p-values and the effect sizes (ηp2). Orthog-
onal analyses were carried out using tests of contrasts within sub- Psychological measures within exposure sessions
jects. Post hoc analyses were performed using the Bonferroni test. The 2 × 2 × 10 ANOVA [Group (pure exposure, guided exposure)
The significance level was set at .05 for all analyses. × Session (Session 1, Session 6) × Trials (baseline, T1, T2, T4, T5,

Table 1 Descriptive statistics (mean and standard deviation) for baseline sample characteristics (i.e. age, BMI and clinical symptoms) and psychological variables
assessed after the first (session 1) and last (session 6) exposure session in both treatment groups

Pure exposure (n = 17) Guided exposure (n = 18)

Exposure Exposure Exposure Exposure

BL Session 1 Session 6 BL Session 1 Session 6
Variable M (SD) M (SD) M (SD) M (SD) M (SD) M (SD)

BIAQT-POS 19.17 (7.56) 17.70 (7.63) 22.52 (8.92) 21.88 (10.20) 19.16 (7.48) 21.83 (10.68)
BIAQT-NEG 71.35 (19.02) 83.00 (21.21) 58.94 (21.80) 64.83 (28.61) 78.27 (27.87) 60.72 (29.26)
BIAQ 10.32 (4.40) 12.35 (4.18) 9.41 (4.87) 9.16 (2.79) 10.63 (3.49) 8.00 (4.61)
VAS 77.70 (22.48) 82.94 (15.04) 54.05 (19.71) 81.22 (15.88) 81.33 (17.91) 57.44 (20.36)
Age (years) 19.35 (1.65) 20.33 (2.11)
2
BMI (kg/m ) 24.09 (3.22) 23.22 (2.62)

Note: BMI, body mass index; BIATQ-POS, positive thoughts; BIATQ-NEG, negative thoughts; BIAQ, avoidance behaviours; VAS, feelings of dissatisfaction.

Eur. Eat. Disorders Rev. (2017) Copyright © 2017 John Wiley & Sons, Ltd and Eating Disorders Association.
S. Díaz-Ferrer et al. Pure vs Guided Exposure

T6, T7, T8, T9)] results for subjective discomfort within sessions Physiological measures within exposure sessions
and between the first and last sessions indicated significant effects The 2 × 2 × 9 × 37 ANOVA [Group (pure exposure, guided expo-
for the Session factor (F(1, 33) = 81.71, p < .0001, ηp2 = .71), sure) × Session (Session 1, Session 6) × Trials (exposure trial 1,
the Trial factor (F (9, 297) = 2.54, p < .042, ηp2 = .07), the 2, 3, 4, 5, 6, 7, 8, 9) × Time (seconds 1, 2, 3, 4, 5, 6, 7, 8, 9, 10,
Trial × Group (F (9, 297) = 12.01, p < .0001, ηp2 = .26) and …, 37)] results for skin conductance revealed significant effects
Session × Trial (F(9, 297) = 3.57, p < .012, ηp2 = .09) interac- for the Time factor (F(36, 1152) = 7.34, p < .0001, ηp2 = .18) and
tions. The Group factor, the Session × Group and the Trial × Time × Group interaction (F(288, 9216) = 2.01,
Session × Trial × Group interactions were not significant. A de- p < .039, ηp2 = .05). The Group factor presented a marginally sig-
crease in subjective discomfort from the first to the last session nificant effect (F(1, 32) = 3.72, p < .063, ηp2 = .10). There was no
was observed in both groups. Changes in subjective discomfort significant main effect for the Session and Trial factors. Analysis
were also present within both sessions. However, orthogonal anal- of the triple Trial × Time × Group interaction revealed that the ac-
ysis of the Session × Trial interaction revealed that the pattern of tivation pattern of skin conductance during the session differed
changes within each session was different [significant differences between the two groups (Figure 2). Orthogonal analyses showed
in the linear (p < .05) and quadratic (p < .0001) trends]. In con- that the pure exposure group showed a greater increase of skin
trast, analysis of the Trial × Group interaction revealed a different conductance within (throughout the 37 s) and between trials
pattern of change for each group during both sessions (Figure 1). (across the nine trials) compared with the guided exposure group
Orthogonal analyses showed that the pure exposure group pre- [significant differences in the linear trend (p < .041)]. Post hoc
sented greater discomfort than the guided exposure group at the analyses indicated that the pure exposure group maintained a sig-
beginning of the session, and this discomfort progressively de- nificantly higher level of skin conductance than the guided expo-
creased. Nevertheless, the guided exposure group presented lower sure group at the end of exposure trial 1 (seconds 25 and 31 to
discomfort than the pure exposure group at the beginning of the 36), trial 2 (seconds 19 and 35), trial 3 (seconds 12, 14, 19 and
session, followed by greater discomfort towards the end of the ses- 24 to 37), trial 5 (seconds 28 to 30 and 35 to 37) and trial 8 (sec-
sion [significant differences in the linear (p < .0001) and quadratic onds 13 to 37). No differences were found between the groups in
(p < .006) trends]. A comparison of both groups over the course of the other trials. The other interactions were not significant.
nine trials in the first and last sessions revealed significantly greater The 2 × 2 × 9 × 37 ANOVA [Group (pure exposure, guided ex-
discomfort for the pure exposure group in exposure trials 1 (T1) posure) × Session (Session 1, Session 6) × Trials (exposure trial 1,
and 2 (T2) in the first session and significantly greater discomfort 2, 3, 4, 5, 6, 7, 8, 9) × Time (seconds 1, 2, 3, 4, 5, 6, 7, 8, 9, 10,
for the guided exposure group in exposure trials 6 (T6), 7 (T7) and …,37)] results for heart rate exhibited significant effects for the
9 (T9) in the last session. Trial factor (F(8, 256) = 2.41, p < .029, ηp2 = .07), the Time factor
(F(36, 1152) = 6.34, p < .0001, ηp2 = .16) and the Time × Group
interaction (F(36, 1152) = 3.45, p < .001, ηp2 = .10). Analysis of
Body parts informed within exposure session the Time × Group interaction revealed faster heart rate activation
Within the first and the last sessions, the body parts more in the pure exposure group than in the guided exposure group
frequently informed by the participants in the pure exposure (Figure 3). Orthogonal analyses showed that both groups first
group along the nine exposure trials were the torso (Session showed a deceleration, an acceleration and a final deceleration;
1 = 18.08%; Session 6 = 15.25%) and legs (Session 1 = 18.06%; however, the pure exposure group showed a faster cubic pattern
Session 6 = 16.16%; see Table S1). Body parts looked at and de- compared with the guided exposure group [significant differences
scribed in the guided exposure group were head 1 (T1), head 2 in the cubic trend (p < .004)]. Post hoc analyses indicated that the
(T2), torso 1 (T3), torso 2 (T4), back (T5), lower extremities 1 significant differences between the two groups appeared in sec-
(T6), lower extremities 2 (T7), upper extremities (T8) and over- onds 7 to 10, with the pure exposure group presenting a higher
view (T9). heart rate than the guided exposure group, and seconds 18 to

Figure 1. Subjective responses to discomfort during exposure in the first and last treatment sessions for both treatment groups: EXP (pure exposure) and G-EXP
(guided exposure). Low levels indicate less subjective discomfort levels (1 = minimum level of discomfort and 10 = maximum level of discomfort). Standard errors
are represented in the figure by the error bars attached to each line

Eur. Eat. Disorders Rev. (2017) Copyright © 2017 John Wiley & Sons, Ltd and Eating Disorders Association.
Pure vs Guided Exposure S. Díaz-Ferrer et al.

Figure 2. Differential scores for changes in skin electrical conductance in both treatment groups during exposure: EXP (pure exposure) and G-EXP (guided exposure).
Negative scores are decreases in skin electrical conductance

20, with the guided exposure group presenting a higher heart rate
than the pure exposure group. Other significant main effects or
interactions were not observed.

Discussion
This study aimed to evaluate the effectiveness of pure and guided
mirror exposure to improve body image disturbances and the
psychophysiological reactions during individual treatment ses-
sions. Overall, our results confirm that both exposure techniques
improved cognitive-affective and avoidance behaviour symptoms
towards one’s body. Nevertheless, we noted that each technique
caused different patterns of emotional and physiological activa-
tion within exposure.
In line with our hypothesis, both techniques decreased negative
thoughts and increased positive thoughts and feelings of satisfac-
tion throughout the sessions, as found in previous research
(Díaz-Ferrer et al., 2015; Trentowska et al., 2013; Vocks et al.,
Figure 3. Differential scores for changes in heart rate in both treatment groups 2007). Self-reported avoidance behaviours also improved consid-
during exposure: EXP (pure exposure) and G-EXP (guided exposure). Negative erably, although treatment was not explicitly directed towards
scores are decreases in heart rate avoidance behaviours, also replicating previous findings

Eur. Eat. Disorders Rev. (2017) Copyright © 2017 John Wiley & Sons, Ltd and Eating Disorders Association.
S. Díaz-Ferrer et al. Pure vs Guided Exposure

(Luethcke et al., 2011; Vocks et al., 2008). Avoidance behaviours subjective discomfort. Pure mirror exposure produces higher skin
are aimed at decreasing negative emotions and critical assess- conductance responses than guided exposure. This maintained
ments of one’s body (Shafran, Fairburn, Robinson, & Lask, arousal pattern may occur because participants undergoing pure
2004; Shafran, Lee, Payne, & Fairburn, 2007). Consequently, the exposure focus on the ‘problem areas’ without a programmed se-
improvement in such behaviours may reflect improvement in quence. According to the inhibitory learning model (Craske et al.,
body satisfaction and a decrease in negative thoughts. 2014), for example, conducting exposure therapy by selecting
Our findings confirm the prediction concerning subjective items of the hierarchy out of order (unlike traditional exposure
discomfort. In the first session, the subjective discomfort was procedures) produces variability during the session. Such variabil-
maintained in both groups at a high rate without evidence of a ity promotes higher levels of physiological arousal and anxiety dur-
within-session habituation process. However, in the last session, ing exposure that fails to habituate (Kircanski et al., 2012). As
the discomfort decreased, which can be explained as a between- mentioned previously, pure mirror procedure promotes variability
session habituation process. From a theoretical perspective, this during the session (although not exactly the same variability pro-
outcome seems more consistent with the assumption of the inhib- posed by the inhibitory learning model), which could explain the
itory learning model (Abramowitz, 2013; Craske et al., 2008, higher physiological reactivity. In contrast, guided mirror expo-
2014), namely, that fear reduction within an exposure session is sure did not maintain a high skin conductance response. One pos-
not a critical index of therapeutic change, as proposed by the clas- sible reason is that participants in guided exposure have less
sical habituation model (Foa & Kozak, 1986). A similar finding variability during the session because they follow the same se-
was reported by Baker et al. (2010) in patients with anxiety disor- quence of neutral body descriptions in every session. Interestingly,
ders. While fear reduction between sessions was predictive of we did observe a simultaneous increase in skin conductance and
short-term therapeutic change, no relationship was found be- subjective discomfort during guided exposure in trials in which
tween fear reduction within sessions and treatment outcomes. typical dissatisfying body areas are described—for example, in trial
Nevertheless, the pattern of subjective discomfort changes 4 (waist, hips and belly). It is important to note that this coinci-
within sessions was different for each group, in line with findings dence represents the first objective data reflecting the different
in previous studies (Díaz-Ferrer et al., 2015; Moreno-Domínguez pattern of changes in subjective discomfort between the two
et al., 2012). In pure mirror exposure, the pattern is characterized techniques observed in previous studies (Díaz-Ferrer et al.,
by a linear decrease. A possible explanation is that pure exposure 2015; Moreno-Domínguez et al., 2012).
allows the verbalization of negative emotions and thoughts, focus- In relation to heart rate, pure exposure elicited faster heart rate
ing on the most dissatisfying body parts from the beginning of each activation than guided exposure, in line with the skin conduc-
session, which can explain the initial strong subjective discomfort. tance results. Pure mirror exposure focuses attention on more
According to various studies, women with body dissatisfaction and dissatisfying body areas, accompanied by the verbalization of neg-
patients with EDs are characterized by an increased selective visual ative emotions and thoughts experienced while observing the
attention for the most dissatisfying body areas (Hewig et al., 2008; body, which could explain the observed rapid heart rate accelera-
Jansen, Nederkoorn, & Mulkens, 2005; Tuschen-Caffier et al., tion. Instead, guided exposure focuses attention on the cognitive
2015). Pure exposure does not follow an orderly method of expo- neutral reprocessing of various body parts, starting with the less
sure, and it could explain why they tend to direct their attention to dissatisfying body areas, which could explain the delayed heart
these areas. This interesting finding raises the question of whether rate acceleration in this group. Taken together, the physiological
it is better in mirror exposure to guide the focus of attention to- and emotional reactivity helps to illuminate the therapeutic pro-
wards positive or negative body parts. Recently, Jansen et al. cess underlying each technique.
(2016) observed that focusing on the most unattractive parts, de-
spite being emotionally more disturbing for patients, led to a better
appraisal of those parts. These results suggest that prolonged atten- Limitations and future research
tion to the most loathed body parts can change their negative va- Several limitations and future research directions must be men-
lence to become more positive, leading to a greater acceptance of tioned. First, the measures presented in this paper have been
their negative emotions and thoughts at the end of the treatment. recorded before, during and immediately after exposure. Long-
In contrast, in guided mirror exposure, the pattern of subjective term follow-up measures are necessary to confirm that the stron-
discomfort changes is characterized by an initial decrease coincid- ger and faster physiological arousal elicited by pure exposure does
ing with a description of the head, followed by an increase towards not affect the overall improvement of patients. Second, our study
the middle of the session while looking at and describing the torso was restricted to a population without clinical EDs. Future re-
(e.g. hips, waist and belly) and lower extremities (e.g. legs). This search should be expanded to include individuals with different
pattern could be explained by the fact that the head is a less EDs. Third, our study did not include direct physiological mea-
dissatisfying area for many women compared with the torso and sures of attention (e.g. the eye tracker) to determine whether there
legs (‘problem areas’; Hewig et al., 2008; Jansen et al., 2005; was any correlation between attention towards different body
Tuschen-Caffier et al., 2015). Thus, this procedure might have in- parts and emotional and physiological activation patterns. Fourth,
terfered with the expected initial activation of the negative emo- both treatments followed a different protocol from traditional ex-
tional responses, which, as suggested by Foa and Kozak’s (1986) posure; for example, the duration of mirror exposure was fixed
model, would have facilitated their subsequent habituation. and did not continue until habituation occurred. Specific strate-
Regarding the physiological changes, our results also reveal dif- gies to enhance inhibitory learning were not proposed. Finally, fu-
ferences in the pattern of reactivity between both groups, similar to ture studies should control for time exposure to evaluate the

Eur. Eat. Disorders Rev. (2017) Copyright © 2017 John Wiley & Sons, Ltd and Eating Disorders Association.
Pure vs Guided Exposure S. Díaz-Ferrer et al.

habituation process and change the protocols to introduce strate- programmes for body image disturbances because it may facilitate
gies to enhance inhibitory learning. new learning associations by focusing on the body and emotional
expression.
Conclusions
Acknowledgements
This study found further evidence of the beneficial effects of pure
mirror exposure as a therapeutic strategy for improving body im- The authors would like to thank all students who assisted with
age disturbances. Our findings also highlight the positive effects of this study. This research was supported by grants from the
treating body image problems early in women who are at risk for Spanish Ministry of Economy and Competitiveness (ref.
developing EDs in order to prevent them. Pure mirror exposure PSI2012-31395 and PSI2009-08417) and the Spanish Ministry of
could be an effective tool as a single component in treatment Education (FPU grant ref. AP2009-3078).

REFERENCES Hewig, J., Cooper, S., Trippe, R. H., Hecht, H., Straube, T., & Body Image Automatic Thoughts Questionnaire (BIATQ) in a
Miltner, W. H. (2008). Drive for thinness and attention Spanish population. Poster session presented at the 37th Annual
Abramowitz, J. S. (2013). The practice of exposure therapy: Rele- toward specific body parts in a nonclinical sample. Psychoso- Convention of Association for Advancement of Behavior Ther-
vance of cognitive-behavioral theory and extinction theory. Be- matic Medicine, 70(6), 729–736. [Link] apy, Boston, MA.
havior Therapy, 44(4), 548–558. [Link] PSY.0b013e31817e41d3. Rosen, J. C., Srebnik, D., Saltzberg, E., & Wendt, S. (1991). Develop-
beth.2013.03.003. Hilbert, A., Tuschen-Caffier, B., & Vögele, C. (2002). Effects of ment of a body image avoidance questionnaire. Psychological As-
Baker, A., Mystkowski, J., Culver, N., Yi, R., Mortazavi, A., & Craske, prolonged and repeated body image exposure in binge eating dis- sessment: A Journal of Consulting and Clinical Psychology, 3(1),
M. G. (2010). Does habituation matter? Emotional processing order. Journal of Psychosomatic Research, 52(3), 137–144. 32–37. [Link]
theory and exposure therapy for acrophobia. Behaviour Research Jansen, A., Nederkoorn, C., & Mulkens, S. (2005). Selective visual at- Servián-Franco, F., Moreno-Domínguez, S., & Reyes del Paso, G. A.
and Therapy, 48(11), 1139–1143. [Link] tention for ugly and beautiful body parts in eating disorders. Be- (2015). Body dissatisfaction and mirror exposure: Evidence for a
brat.2010.07.009. haviour Research and Therapy, 43(2), 183–196. [Link] dissociation between self-report and physiological responses in
Cash, T. F., Lewis, R. J., & Keeton, P. (1987, March). Development 10.1016/[Link].2004.01.003. highly body-dissatisfied women. PloS One, 10(4), e0122737.
and validation of body image automatic thoughts questionnaire: A Jansen, A., Voorwinde, V., Hoebink, Y., Rekkers, M., Martijn, M., & doi: [Link]
measure of body-related cognitions. Paper presented at the Meet- Mulkens, S. (2016). Mirror exposure to increase body satisfac- Shafran, R., Fairburn, C. G., Robinson, P., & Lask, B. (2004). Body
ing of the Southeastern Psychological Association. GA: Atlanta. tion: Should we guide the focus of attention towards positively checking and its avoidance in eating disorders. International Jour-
Cooper, P. J., Taylor, M., Cooper, Z., & Fairburm, C. (1987). The de- or negatively evaluated body parts? Journal of Behavior Therapy nal of Eating Disorders, 35(1), 93–101. [Link]
velopment and validation of the body shape questionnaire. Inter- and Experimental Psychiatry, 50, 90–96. [Link] eat.10228.
national Journal of Eating Disorders, 6(4), 485–494. [Link] [Link].2015.06.002. Shafran, R., Lee, M., Payne, E., & Fairburn, C. G. (2007). An exper-
org/10.1002/1098-108X(198707)6. Keel, P. K., Dorer, D. J., Franko, D. L., Jackson, S. C., & Herzog, D. imental analysis of body checking. Behaviour Research and Ther-
Craske, M. G., Kircanski, K., Zelikowsky, M., Mystkowski, J., B. (2005). Postremission predictors of relapse in women with apy, 45(1), 113–121. [Link]
Chowdhury, N., & Baker, A. (2008). Optimizing inhibitory eating disorders. American Journal of Psychiatry, 162(12), Trentowska, M., Bender, C., & Tuschen-Caffier, B. (2013). Mirror
learning during exposure therapy. Behaviour Research and Ther- 2263–2268. [Link] exposure in women with bulimic symptoms: How do thoughts
apy, 46(1), 5–27. [Link] Kircanski, K., Mortazavi, A., Castriotta, N., Baker, A. S., and emotions change in body image treatment? Behaviour Re-
Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, Mystkowski, J. L., Yi, R., et al. (2012). Challenges to the tradi- search and Therapy, 51(1), 1–6. [Link]
B. (2014). Maximizing exposure therapy: An inhibitory learning tional exposure paradigm: Variability in exposure therapy for brat.2012.03.012.
approach. Behaviour Research and Therapy, 58, 10–23. https:// contamination fears. Journal of Behavior Therapy and Experi- Trentowska, M., Svaldi, J., & Tuschen-Caffier, B. (2014). Efficacy
[Link]/10.1016/[Link].2014.04.006. mental Psychiatry, 43(2), 745–751. [Link] of body exposure as treatment component for patients with
Delinsky, S. S., & Wilson, G. T. (2010). Cognitive behavior therapy jbtep.2011.10.010. eating disorders. Journal of Behavior Therapy and Experimental
with body image exposure for bulimia nervosa: A case example. Luethcke, C. A., McDaniel, L., & Becker, C. B. (2011). A comparison Psychiatry, 45(1), 178–185. [Link]
Cognitive and Behavioral Practice, 17, 270–277. [Link] of mindfulness, nonjudgmental, and cognitive dissonance-based 09.010.
10.1016/[Link].2010.02.004. approaches to mirror exposure. Body Image, 8(3), 251–258. Tuschen-Caffier, B., Bender, C., Caffier, D., Klenner, K., Braks, K., &
Díaz-Ferrer, S., Rodríguez-Ruiz, S., Ortega-Roldán, B., Moreno- [Link] Svaldi, J. (2015). Selective visual attention during mirror expo-
Domínguez, S., & Fernández-Santaella, M. C. (2015). Testing Moreno-Domínguez, S., Rodríguez-Ruiz, S., Fernández-Santaella, sure in anorexia and bulimia nervosa. PloS One, 10(12), 1–18.
the efficacy of pure versus guided mirror exposure in women with M. C., Jansen, A., & Tuschen-Caffier, B. (2012). Pure versus [Link]
bulimia nervosa: A combination of neuroendocrine and psycho- guided mirror exposure to reduce body dissatisfaction: A prelim- Tuschen-Caffier, B., & Florin, I. (2002). Teufelskreis Bulimie. Ein
logical indices. Journal of Behavior Therapy and Experimental Psy- inary study with university women. Body Image, 9(2), 285–288. Manual zur psychologischen Therapie. Hogrefe Verlag: Göttingen.
chiatry, 48, 1–8. [Link] [Link] Vocks, S., Kosfelder, J., Wucherer, M., & Wächter, A. (2008). Does
Farrell, C., Shafran, R., & Lee, M. (2006). Empirically evaluated treat- Peláez-Fernández, M. A., Ruiz-Lázaro, P. M., Labrador, F. J., & habitual body avoidance and checking behavior influence the de-
ments for body image disturbance: A review. European Eating Dis- Raich, R. M. (2014). Validación del cuestionario Eating Attitudes crease of negative emotions during body exposure in eating dis-
orders Review, 14(5), 289–300. [Link] Test (EAT) como prueba de cribado de trastornos de la conducta orders? Psychotherapy Research, 18(4), 412–419. [Link]
Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: Ex- alimentaria en la población general [validation of the Eating At- 10.1080/10503300701797008.
posure to corrective information. Psychological Bulletin, 99(1), titudes Test as a screening instrument for eating disorders in gen- Vocks, S., Legenbauer, T., Wächter, A., Wucherer, M., & Kosfelder,
20–35. [Link] eral population]. Medicina Clínica, 142(4), 153–155. [Link] J. (2007). What happens in the course of body exposure? Emo-
Garner, D. M., & Garfinkel, P. E. (1979). The Eating Attitudes Test: An org/10.1157/13119489. tional, cognitive, and physiological reactions to mirror confron-
index of the symptoms of anorexia nervosa. Psychological Medi- Perpiñá, C., Borra, C., Baños, R., Botella, C., Quero, S., & tation in eating disorders. Journal of Psychosomatic Research,
cine, 9(2), 273–279. [Link] Jorquera, M. (2003, November). Psychometric properties of the 62(2), 231–239. [Link]

Supporting information

Additional Supporting Information may be found online in the supporting information tab for this article.

Eur. Eat. Disorders Rev. (2017) Copyright © 2017 John Wiley & Sons, Ltd and Eating Disorders Association.

View publication stats