Vol

ume19,Year2021
I
SSN:1554-
8716

Vi
rt
ualReal
i
tyTherapyi
ntheMet
aver
se:
Mergi
ngVRf ort
heOutsi
de
wi
thVRfortheInsi
de

Edi
tor
s:

Br
endaK.Wi
eder
hol
d,Ph.
D.,MBA,BCB,BCN

Gi
useppeRi
va,Ph.
D.,M.
S.,M.
A.

ht
tp:
//
www.
arc
tt.
inf
o

19
ANNUAL REVIEW OF CYBERTHERAPY
AND TELEMEDICINE 2021
Annual Review of Cybertherapy
And Telemedicine 2021
Virtual Reality Therapy in the Metaverse:
Merging VR for the Outside with VR for the Inside

Edited by

Brenda K. Wiederhold
Interactive Media Institute, San Diego, CA, USA
Virtual Reality Medical Institute, Brussels, Belgium

Giuseppe Riva
Catholic University of Milano, Milan, Italy
Istituto Auxologico Italiano, Milan, Italy
Annual Review of CyberTherapy and Telemedicine, Volume 19

Annual Review of CyberTherapy and Telemedicine

Copyright © 2021 Interactive Media Institute

6540 Lusk Boulevard, Suite C115
San Diego, CA 92121

ISBN: 1554-8716

All rights reserved.

Printed in the United States of America

Journal Website: http://www.arctt.info

Interactive Media Institute Website: http://www.interactivemediainstitute.com

LEGAL NOTICE
The publisher is not responsible for the use which might be made of the following
information.
 Editors-in-Chief

Brenda K. Wiederhold, Ph.D., BCB, BCN

Interactive Media Institute
Virtual Reality Medical Center

Giuseppe Riva, Ph.D., M.S., M.A.

Istituto Auxologico Italiano
Università Cattolica del Sacro Cuore

Managing Editor

Silvia Serino, Ph.D.

Università Cattolica del Sacro Cuore

Assistant Editors

Clelia Malighetti, Ph.D

Università Cattolica del Sacro Cuore

Kathryn Phan, B.S.

University of North Texas

Local Chair Conference Coordinator

Pietro Cipresso
Istituto Auxologico Italiano
Università Cattolica del Sacro Cuore

Claudia Repetto
Università Cattolica del Sacro Cuore

Daniela Villani
Università Cattolica del Sacro Cuore
 Editorial Board

Nathan Appel, PhD Cristina Botella, PhD Willem-Paul Brinkman, PhD

National Institute on Drug Abuse

Universitat Jaume, Delft University of Technology
(NIDA)
Castelló de la Plana Castellón, Spain Delft, Netherlands
Bethesda, Maryland, USA

Stéphane Bouchard, PhD Andrew Campbell, PhD Cosimo Tuena, PhD Cand.

Université du Québec en Outaouais The University of Sydney Università Cattolica del Sacro Cuore di
Gatineau, Canada Sydney, Australia Milano Italy

Scott Debb, EdD Chris Fullwood, PhD Andrea Gaggioli, PhD

Università Cattolica del Sacro Cuore

Norfolk State University Norfolk, University of Wolverhampton
di Milano
Virginia, USA Wolverhampton, United Kingdom
Italy

Luciano Gamberini, PhD Pedro Gamito, PhD Karen Holmes, PhD

Universidade Lusófona de
University of Padova Norfolk State University Norfolk,
Humanidades e Tecnologia Lisbon,
Padova, Italy Virginia, USA
Portugal

Thomas Parsons, PhD

Stella Gelli, MD José Gutierrez Maldonado, PhD

University of Barcelona University of North Texas

Università Cattolica del Sacro Cuore
Barcelona, Spain Denton, Texas, USA
di Milano Italy

Susan Persky, PhD Wendy Powell, PhD Olivia Realdon, PhD

National Institutes of Health (NIH),

Tilburg University University of Milano-Bicocca
Bethesda
Tilburg, Netherlands Milan, Italy
Maryland, USA

Michael Roy, M.D., MPH, FACP,

Claudia Rodella, PhD Dan Romer, PhD
COL (Ret.)

Università Cattolica del Sacro Cuore University of Pennsylvania

Uniformed Services University,
di Milano Philadelphia
Bethesda Maryland, USA
Italy Pennsylvania, USA

Daniele Di Lernia, PhD Stefan Stieger, PhD Darlene Colson, PhD

Università Cattolica del Sacro Cuore Karl Landsteiner University Norfolk State University Norfolk,
di Milano Italy Krems an der Donau, Austria Virginia, USA

Claudia Carissoli, PhD Wendy Powell, PhD Marta Matamala Gomez, PhD
Università degli Studi di Milano Tilburg University
Università degli Studi Milano-Bicocca
Statale Tilburg, Netherlands
 General Information

Annual Review of CyberTherapy and Telemedicine (ARCTT – ISSN: 1554-8716) is published

annually (once per year) by the Interactive Media Institute (IMI), a 501c3 non-profit organization,
dedicated to incorporating interdisciplinary researchers from around the world to create, test, and
develop clinical protocols for the medical and psychological community. IMI realizes that the mind
and body work in concert to affect quality of life in individuals and works to develop technology
that can be effectively used to improve the standards and reduce the cost of healthcare delivery.

Interactive Media Institute

6540 Lusk Boulevard, Suite C115
San Diego, CA 92121 USA
Telephone:+1 858 642 0267
E-mail: [email protected]
Web site: http://www.interactivemediainstitute.com
Journal Web site: http://www.arctt.info

Copyright © 2020 by Interactive Media Institute. Printed in the United States of America.

About the Journal

ARCTT is a peer-reviewed all-purpose journal covering a wide variety of topics of interest to the
mental health, neuroscience, and rehabilitation communities. The mission of ARCTT is to provide
systematic, periodic examinations of scholarly advances in the field of CyberTherapy and
Telemedicine through original investigations in the Telemedicine and CyberTherapy areas, novel
experimental clinical studies, and critical authoritative reviews. It is directed to healthcare
providers and researchers who are interested in the applications of advanced media for improving
the delivery and efficacy of mental healthcare and rehabilitative services.

Manuscript Proposal and Submission

Because Annual Review papers examine either novel therapeutic methods and trials or a specific
clinical application in depth, they are written by experienced researchers upon invitation from our
Editorial Board. The editors nevertheless welcome suggestions from our readers. Questions or
comments about editorial content or policies should be directed to the editors only.

Manuscript Preparation
Manuscripts should be submitted in electronic format on CD-Rom or floppy disks as well as on
81⁄2 x 11-in. paper (three copies), double-spaced format. Authors should prepare manuscripts
according to the Publication Manual of the American Psychological Association (5th Ed.).
Original, camera-ready artwork for figures is required. Original color figures can be printed in
color at the editors' discretion and provided the author agrees to pay in full the associated
production costs; an estimate of these costs is available from the ARCTT production office on
request. ARCTT policy prohibits an author from submitting the same manuscript for concurrent
consideration by two or more publications. Authors have an obligation to consult journal editors
concerning prior publication of any data upon which their article depends. As this journal is a
primary journal that publishes original material only, ARCTT policy also prohibits publication of
any manuscript that has already been published in whole or substantial part elsewhere unless
authorized by the journal editors.

Disclaimer
All the published papers, editorial news and comments, opinions, findings, conclusions, or
recommendations in ARCTT are those of the author(s), and do not necessarily reflect or constitute
the opinions of the Journal, its Publisher, and its editorial staff.
 REVIEW BOARD
We would like to extend a warm and heartfelt thank you to all members of the review board whose help made this
year’s publication possible:

A. Gaggioli, G. Riva, C. Malighetti, D. Di Lernia, S. Serino, C. Repetto, D. Villani, E. Pedroli, D.

Colombo, A. Chirico, I. M. A. Benzi
Brenda K. Wiederhold Giuseppe Riva
PRINCIPLES of the ANNUAL REVIEW
Ethical Standards.
Authors published in the Annual Review must uphold and respect ethical standards in the preparation
and realization of their research, as well as in the writing style of the papers (avoid prejudice, protect
confidentiality of the patients, etc.). Research involving human subjects must have been conducted with
respect to current ethical practices and after participants expressed their free and informed consent (e.g.,
with a signed consent form or other appropriate method). Ethical standards also apply to research that is
not conducted with humans (e.g., animal protection protocol), and to publishing issues (e.g., plagiarism,
research fraud, authorship misappropriation, duplication of publications).

Conflicts of Interest.
It is the position of the Annual Review that potential conflicts of interest must be made available to the
readers since an author’s economic and commercial interests may, even involuntarily, bias objectivity.
Economic and commercial interests do not necessarily constitute a conflict of interest, and conflicts of
interest do not necessarily imply biased research. But the best practice is to disclose activities and
relationships that, if known to others, might be viewed as a conflict of interest.
Potential conflicts of interest include, but are not limited to: (a) funding or remuneration, including
salaries and equipment, from organizations that may gain or lose financially through the publication of
the paper, (b) personal financial interests, including receiving royalties or holding stocks and shares in
companies that may gain or lose financially from publication, (c) holding patent and patent applications
whose financial value may be affected, and (d) employment by an organization that may gain or lose
from publication of the paper. All contributing authors are expected to provide the Editor with a signed
presenter disclosure form, and all contributing authors and reviewers are encouraged to contact the
Editor at any stage in the manuscript review process if they believe that a potential conflict of interest
needs to be examined.

Upholding the Annual Review’s Standards.

Our publication pays careful attention to the protection of a patient’s anonymity in case reports and
elsewhere.
Identifying information such as names, initials, and hospital numbers must be avoided. Also, authors
should disguise identifying information when discussing patients’ characteristics and personal history.
 Preface
ARCTT is a peer-reviewed all-purpose journal covering a wide variety of topics of
interest to the mental health, neuroscience, and rehabilitation communities. The
mission of ARCTT is to provide systematic, periodic examinations of scholarly
advances in the field of Cybertherapy and Telemedicine through original
investigations in the telemedicine and cybertherapy areas, novel experimental
clinical studies, and critical authoritative reviews.
Healthcare delivery systems have been evolving to rely more heavily on
technology. There has been a shift in care diagnosis and treatment that has
decreased the importance of traditional methods of care delivery. Technology has
not only helped to extend our lifespan, but it has improved the quality of life for all
citizens.
We have put a great deal of effort into the definition of the structure of the
volume and in the sequence of the contributions, so that those in search of a
specific reading path will be rewarded. To this end, we have divided the different
chapters into six main sections:
1. Editorial: This introductory text expresses the position of the Editors –
Brenda K. Wiederhold Giuseppe Riva - about the focus of this year’s
issue;
2. Critical Reviews: These chapters summarize and evaluate emerging
cybertherapy topics including technology, enhanced rehabilitation,
Interreality, and Intersubjectivity;
3. Evaluation Studies: These chapters are generally undertaken to solve
some specific practical problems and yield decisions about the value of
cybertherapy interventions;
4. Original Research: These chapters’ research studies address new
cybertherapy methods or approaches;
5. Clinical Observations: These chapters include case studies or research
protocols with long-term potential;
6. Work in Progress: These chapters include papers describing a future
research work;
7. Brief Communications: These chapters include brief papers reporting
preliminary data on-going research work and/or new developments.
For both health professionals and patients, the selected contents will play an
important role in ensuring that the necessary skills and familiarity with the tools are
available, as well as a fair understanding of the context of interaction in which they
operate.
In conclusion, this volume underlines how cybertherapy has started to make
progress in treating a variety of disorders. However, there is more work to be done
in a number of areas, including the development of easy-to-use and more affordable
hardware and software, the development of objective measurement tools, the need
to address potential side effects, and the implementation of more controlled studies
to evaluate the strength of cybertherapy in comparison to traditional therapies.
We are grateful to Silvia Serino and Ian T. Miller for their work in collecting
and coordinating chapters for this volume.
We sincerely hope that you will find this year’s volume to be a fascinating and
intellectually stimulating read. We continue to believe that together, we can change the face
of healthcare.
Brenda K. Wiederhold
Giuseppe Riva

v
 Contents
Preface v
Brenda K. Wiederhold and Giuseppe Riva

Section I. Editorial

Virtual Reality Therapy in the Metaverse:

Merging VR for the Outside with VR for the Inside 3
Giuseppe Riva, Daniele Di Lernia,Elena Sajno, Maria Sansoni, Sabrina Bartolotta,
Silvia Serino, Andrea Gaggioli and Brenda K. Wiederhold

Section II. Critical Reviews

1. Implicit Measures of Perceived Realness in Virtual Reality 11

Marius Rubo, Jasin Sahraoui and Simone Munsch
2. Between benevolent lies and harmful deception: Ethical analysis of deceptive
practices in dementia care technology 15
Ans Tummers, Rens Brankaert and Wijnand Ijsselsteijn
3. Normative Affordances: Utilising the constraint of context-specific expectation
in simulated environments 21
John Francis Leader

Section III. Evaluation Studies

4. Empathic interactions in online treatment: experiences of mental healthcare

practitioners 29
Milou Feijt, Yvonne De Kort, Joyce Westerink and Wijnand Ijsselsteijn
5. Further Validation of Russian Video Games Addiction Scale (VGAS) 35
Vitalii Epishin, Nataliya Bogacheva and Diana Medakovskaya
6. A Motion Controlled Virtual Reality Paradigm for Ostracism Research 41
Patrick Mulvaney, Brendan Rooney and Brendan Rooney
7. A qualitative and quantitative virtual reality usability study for the early
assessment of ASD children 47
Maria Eleonora Minissi, Irene Alice Chicchi Giglioli, Fabrizia Mantovani,
Marian Sirera, Luis Abad and Mariano Alcañiz
8. A voice recognition application for the semantic and prosodic analysis of ASD
caregivers 53
Irene Alice Chicchi Giglioli, Luna Maddalon, Lucía Gómez-Zaragozá, Maria
 Eleonora Minissi, Javier Marín Morales, Marian Sirera, Luis Abad and
Mariano Alcañiz

Section IV. Original Research

9. Exploratory Factor Analysis of the Virtual Reality Stroop Task 61

Justin Asbee and Thomas Parsons
10. Cognitive stimulation using non-immersive virtual reality tasks in children with
learning disabilities 67
Lucileide Santos, Jorge Oliveira and Pedro Gamito
11. Mind your time: The implications of prolonged Instagram use and drive for
thinness in university students 73
Alexandra Ghita, Teuntje Elfrink, Adriana Bülter, Svenja Gabriel, Marie Geise,
Greta Grewe and Gerben Westerhof
12. Influence of Photorealism and Non-Photorealism on Connection in Social VR 79
Nienke Bierhuizen, Wendy Powell, Tina Mioch, Omar Niamut and Hans
Stokking
13. Idealization on dating apps: Seeing fewer photos of the potential partner leads
to expectancy violation and lower attraction 85
Simona Sciara, Clelia Malighetti, Giorgia Martini, Giuseppe Riva and Camillo
Regalia

Section V. Clinical Observations

14. Going beyond body exposure therapy. Presenting an innovative Virtual Reality
and Eye-Tracking body-related attentional bias task. 93
Bruno Porras Garcia, Alana Singh, Marta Ferrer-Garcia, Helena Miquel,
Gemma Taña-Velasco, Natalia Briseño-Oloriz, Jesus Fleta, Emma Iglesias and
José Gutiérrez-Maldonado
15. The way we look at our own body really matters! Body-related attentional bias
as a predictor of worse clinical outcomes after a virtual reality body exposure
therapy 99
Marta Ferrer-Garcia, Bruno Porras-Garcia, Helena Miquel, Eduardo Serrano-
Troncoso, Marta Carulla-Roig and José Gutiérrez-Maldonado
16. A Virtual Reality tool using embodiment and body swapping techniques for the
treatment of obesity: A usability study 105
Dimitra Anastasiadou, Bernhard Spanlang, Mel Slater, Julia Váquez-De
Sebastian, Josep Antoni Ramos-Quiroga, Gemma Parramon Puig, Andreea
Ciudin, Marta Comas and Pilar Lusilla-Palacios
17. Aphasia360°: A virtual reality intervention for anomia rehabilitation in post-
stroke patients 111
Claudia Repetto, Alice Cancer, Claudia Rodella, Marta Campagna and
Alessandra Maietti
18. Rescripting emotional eating with virtual reality: a single case study 177
Clelia Malighetti, Ciara Schnitzer, Georgia Potter, Katherine Nameth, Theresa
Brown, Emily Vogel, Giuseppe Riva, Cristin Runfola and Debra Safer
Section VI. Work In Progress

19. The immersive 3D objects’ library for applying non-invasive brain stimulation
in research on the motor control and the mirror neurons system: a call for
collaboration 125
Andrey Vlasov, Fanir Kilmukhametov and Matteo Feurra
20 Meta Cognition on the Internet: Expected Accuracy of Human and AI Virtual
Assistants’ First Impressions about Us Online 129
Elena Tsankova and Ergyul Tair
21. Virtual reality for relaxation in a pediatric hospital setting: an interventional
study with a mixed-methods design 133
Sylvie Bernaerts, Bert Bonroy, Jo Daems, Romy Sels, Dieter Struyf and Wessel
van de Veerdonk

Subject Index 136

Author Index 137
SECTION I

EDITORIAL

This introductory text expresses the

position of the editors – Brenda K.
Wiederhold and Giuseppe Riva – the
focus of this year’s issue.

Brenda K. Wiederhold and Giuseppe Riva

 Annual Review of Cybertherapy and Telemedicine 2021 3

Virtual Reality Therapy in the Metaverse:

Merging VR for the Outside with VR for
the Inside
Giuseppe RIVA1 a,b, Daniele DI LERNIA a,
Elena SAJNO a, Maria SANSONI a, Sabrina BARTOLOTTA a,c, Silvia
SERINO a, Andrea GAGGIOLI a,b,c, Brenda K. WIEDERHOLD d,e
a
Humane Technology Lab., Università Cattolica del Sacro Cuore, Milan, Italy
b
Applied Technology for Neuro-Psychology Lab. Istituto Auxologico Italiano, Milan, Italy
c
ExperienceLab, Università Cattolica del Sacro Cuore, Milan, Italy
c
Interactive Media Institute, San Diego, CA, USA
d
Virtual Reality Medical Center, La Jolla, CA, USA
Abstract. The Metaverse can be defined as a hybrid (digital/physical)
environment offering places for rich user interaction. In this view, the main feature of
the Metaverse is a twofold link between the virtual and physical worlds: (a) behavior
in the physical world influences the experience in the virtual one and, (b) behavior in
the virtual world influences the experience in the real one. Furthermore, any change in
the physical world is mirrored in its digital representation (the digital twin), and
feedback is sent in the opposite direction (i.e., if the avatar is touched, haptic feedback
is provided to the physical body). Currently, this is achieved through 3D shared XR
worlds, biosensors and activity sensors (from the real to the virtual world), two-way
Internet of Things (IoT) object connections, social media, and wearable devices
including smartphones (from the virtual world to the real one). Our view is that
bridging technologies that simulate both the external world and the internal world (our
bodily experience) will allow the simulated, cognitive, and embodied dimension of the
Metaverse to merge, thereby transforming it into the ultimate clinical technology. In
particular, it will allow for the emergence of Regenerative VR: the integration of
external and internal simulated technologies to rewrite a faulty bodily experience and
to regenerate the wellbeing of an individual.

Keywords. Metaverse, Interreality, Virtual Reality, Virtual Reality Therapy, Mixed

Reality, Interoceptive Technologies, Regenerative VR

1. Introduction

In his classical 1992 novel Snow Crash, Neal Stephenson described the Metaverse
as a new digital experience in which virtual spaces offer the same possibilities and
opportunities as the real world. In the novel, Stephenson depicted the Metaverse as a
digital escape from a physical and less interesting world. However, the actual vision
driving the research and development work of many technological companies is instead
trying to seamlessly connect the physical and the digital domains [1]. An example of this
vision is digital twins [2]: digital representations of real-world entities – an object,
system, or process – that are synchronized with the real world.
Following this vision, the emerging Metaverse can be defined as a hybrid
(digital/physical) environment offering enhanced places for rich user interaction. In this
view, the main feature of the Metaverse is a two-way link between the virtual and
physical worlds: (a) behavior in the physical world influences the experience in the
virtual world and, (b) behavior in the virtual world influences the experience in the real
world. Furthermore, any change in the physical world is mirrored in its digital
representation (the digital twin), and feedback is sent in the opposite direction (i.e., if the
avatar is touched, haptic feedback is provided to the physical body). This will be
achieved through the merging and interaction between different digital technologies: 3D
1
Corresponding author: [email protected]
Riva et al./ Virtual Reality Therapy in the Metaverse: Merging VR for the Outside with VR for the Inside
4

shared XR worlds, biosensors and activity sensors (from the real to the virtual world),
two-way Internet of Things (IoT) object connections, social media, and wearable devices
including smartphones (from the virtual world to the real one).
Currently, XR technologies – virtual reality and augmented reality – have been
successfully used for the assessment and treatment of different mental health disorders
including anxiety disorders, stress-related disorders, obesity and eating disorders, acute
and chronic pain, addictions, and schizophrenia [3-5].
However, the Metaverse aims to become the most advanced form of human-computer
interaction allowing individuals to act, communicate, and become present in digital and
digitally-enhanced physical environments. Even if these features will further improve
the clinical potential of XR technologies, we suggest that there is a significant missing
piece to this equation: our physical body.
Our view is that bridging technologies that simulate both the external world and the
internal world (our bodily experience) will allow the simulated, cognitive, and embodied
dimension of the Metaverse to merge, thereby transforming it into the ultimate clinical
technology.

2. Virtual Reality Therapy Today

Virtual Reality (VR) is usually referred to as an interactive 3D visualization system

(e.g., a computer, gaming console, or smartphone) supported by one or more position
trackers and a head-mounted display. VR is, however, more than a collection of
technologies. The term "virtual reality" is made up of two words: "virtual" (nearly as
described) and "reality" (the actual state of things). As a result, the term "virtual reality"
might be defined as "almost reality" or "near reality," implying that VR is a type of reality
simulation. Experiential learning has a long history as a therapeutic practice, and virtual
reality's simulation strength makes it the ideal tool for this technique.
Virtual reality allows patients to learn by reflecting on their actions. "One reason it
is so difficult to get people to update their assumptions is that change typically needs a
prior step – identifying the distinction between an assumption and a perception,"
according to Glantz and colleagues [6]. "Assumptions shape the reality until they are
proven to be false; they appear to be perceptions, and as long as they do, they are resistant
to change" (p. 96). Using the VR experience, the therapist can more easily illustrate to
the patient that what appears to be a fact is actually a result of his or her thinking.
The simulation potential of VR also transforms it into an advanced imaginal system
– a type of experiencing imagery that is as effective as reality at eliciting emotional
reactions [7; 8]. Numerous studies have indicated that VR can increase subjectively
reported anxiety in phobic participants confronted with a dangerous virtual setting,
comparable to the effects seen in in vivo conditions (Powers and Emmelkamp, 2008;
Opris et al., 2012). As a result, as indicated by two meta-analyses [9; 10], VR is an
effective and equal medium for exposure therapy. In fact, exposure treatment (virtual
reality exposure therapy, or VRE) is the most popular clinical application of VR. It is
used to replicate an external reality. In other words, virtual reality is employed in clinical
settings to make patients believe that something that isn't there is "real."

3. The Body: From the Brain to the Metaverse

"Predictive coding" [11-13] is a common neuroscience idea that suggests our brain
actively develops an internal model (simulation) of the body and its surroundings. This
model is used to make predictions about the sensory information that will be received
and to reduce the number of prediction errors (or "surprise"). Specifically, our brains
build an embodied simulation of the body that represents its expected future states to
successfully interact with the world (intentions and emotions). This simulation has two
distinct properties [14; 15].
It is, first and foremost, a simulation of sensory-motor experiences using
visceral/autonomic (interoceptive), motor (proprioceptive), and sensory (e.g., visual,
aural) information as sources. Second, embodied simulations are based on the subject's
expectations and reactivate multimodal brain networks that caused the
simulated/expected result earlier.
Riva et al./ Virtual Reality Therapy in the Metaverse: Merging VR for the Outside with VR for the Inside 5

One of the most important goals of this process is to reduce the average of surprise
(i.e., the gap between intentions and the consequences of enacting them) across all
representations and to learn how to model and forecast incoming content.
Virtual reality works similarly; it uses technology to create a virtual experience that
people can manipulate and explore as if they were there. In other words, VR technology
tries to predict the sensory consequences of users' actions by displaying the same
outcome that our brains expect in the real world. As explained by Riva and colleagues
[14]: “To achieve it, like the brain, the VR system maintains a model (simulation) of the
body and the space around it. This prediction is then used to provide the expected sensory
input using the VR hardware. Obviously, to be realistic, the VR model tries to mimic the
brain model as much as possible: the more the VR model is similar to the brain model,
the more the individual feels present in the VR world” (p. 89).
Up until now, VR has been used clinically to make people believe that something
that is not present is "real." However, VR simulations of our body can also fool the
predictive coding mechanisms that regulate bodily experience, making people feel "real"
in situations that are not.
In fact, recent key discoveries in neuroscience are outlining a new conceptual
framework suggesting that mental health disorders are linked to the processing of
multisensory bodily signals [15; 16]. As recently explained by Paulus and colleagues
[17]: “these conceptual models suggest that mental disorders can be broadly
characterized by a dysfunction in the way the brain computes and integrates
representations of the inner and outer worlds of the body across time. According to this
view, changes in mood and anxiety are a by-product of the brain’s biased translation of
what it expects will happen versus what is actually happening in these worlds, producing
a persistent discrepancy/error signal when outcomes are observed.” (p. 99).
In this scenario, VR can also be used to structure, augment, and/or replace the body's
experience for clinical purposes. Various researchers have used advanced technologies,
including VR, to alter bodily perceptions in clinical and non-clinical populations since
the discovery of the rubber hand illusion [18] and the emergence of non-invasive brain
stimulation methodologies [19].
Brain stimulation techniques, including transcranial direct current stimulation (tDCS),
transcranial magnetic stimulation (TMS), and vagus nerve and galvanic vestibular
stimulations, have been successfully used to modify both bottom-up [20] and top-down
[21] bodily signals. More recently, the use of VR allowed the emergence of a new
approach: virtual embodiment. This approach uses virtual reality technologies to trick
the predictive coding mechanisms of the brain, thereby inducing in users a sense of
ownership over a virtual body. Some researchers are beginning to use virtual
embodiment – the use of VR to replace multisensory bodily contents with synthetic ones
– for chronic pain and eating disorders therapy [22].
The rationale behind this approach is to use VR's embodiment potential to correct a
dysfunctional representation of the affected body/body part. For example, virtual
embodiment is currently being used to treat phantom limb pain, which is caused by
dysfunctional changes in amputees' representations of their bodies [23]. Additionally,
Matamala-Gomez and colleagues used virtual embodiment to obtain pain relief in
chronic pain patients [24; 25]. Finally, Serino and colleagues [26] used a VR-based body
swapping illusion to correct the dysfunctional representation of the body in anorexia
nervosa.
Although randomized controlled trials are not yet available, the above case studies
show that VR interventions have a high potential for modifying the experience of the
body.

4. Simulating the Inner Body

Existing research, however, suggests that the above-mentioned approaches' effects

on higher cognitive processes are only temporary, even in non-pathological individuals.
According to Freeman and colleagues [27], the longest follow-up in studies with virtual
bodily illusions for correcting body perception in participants with eating disorders is
only 2 hours. Why?
Riva et al./ Virtual Reality Therapy in the Metaverse: Merging VR for the Outside with VR for the Inside
6

Regardless of the success of virtual embodiment, what distinguishes our body from
other physical objects is that, unlike other physical objects, we not only perceive it
through external senses (exteroception), but we also have internal access to it via inner
(interoceptive, proprioceptive, and vestibular) signals [28].
In this view, interoceptive technologies [29] that modulate interoceptive signals can
play a critical role in simulating our inner bodily experience in the Metaverse. They
include technologies that produce direct modulation of interoceptive signals (for
example, c-fiber stimulation [30] or sonoception [28]) as well as technologies that
produce illusions by providing false feedback of individuals' physiological states [31].

Figure 1. The different components of Regenerative VR (Adapted from [32])

The integration of technologies able to simulate/stimulate both the external body

and the internal one will allow for a new clinical intervention – “Regenerative Virtual
Therapy” (RVT) [32] – which uses technology-based somatic modification techniques
to restructure the maladaptive bodily representations behind a pathological condition. To
achieve this goal, we will combine two different strategies (see Figure 1). First, a
technological intervention merging brain stimulation techniques with a high-rewarding
and novel synthetic multisensory bodily experience (i.e., a virtual reality full-body
illusion in sync with a low predictability interoceptive modulation). Second, a
psychological intervention including mindful attention and cognitive reappraisal.
Riva et al./ Virtual Reality Therapy in the Metaverse: Merging VR for the Outside with VR for the Inside 7

5. Conclusions

The Metaverse can be defined as a hybrid (digital/physical) environment providing

places for rich user interaction. In this view, the main feature of the Metaverse is a
twofold link between the virtual and physical worlds: (a) behavior in the physical world
influences the experience in the virtual one and, (b) behavior in the virtual world
influences the experience in the real one. Moreover, any change in the physical world is
reflected in the digital representation (the twin) and feedback gets sent in the other
direction.
XR technologies – virtual reality and augmented reality – are currently being used to
assess and treat a variety of mental health disorders including anxiety disorders, stress-
related disorders, obesity and eating disorders, acute and chronic pain, addictions, and
schizophrenia [3-5].
However, the Metaverse aims to become the most advanced form of human-computer
interaction, allowing individuals to act, communicate, and become present in digital and
digitally-enhanced physical environments. Even if these features will further improve
the clinical potential of XR technologies, we suggest that there is a significant missing
piece to this equation: our physical body.
Our view is that adding technologies that simulate both the external world and the
internal world (our bodily experience) will allow the simulated, cognitive, and embodied
dimension of the Metaverse to merge, transforming it into the ultimate clinical
technology. A critical role will be played by the emergence of interoceptive technologies
that produce a direct modulation of interoceptive signals or generate interoceptive
illusions by providing false feedback of individuals' physiological states.
The integration of technologies able to simulate/stimulate both the external body
and the internal one will allow the emergence of Regenerative VR: the integration of
external and internal simulated technologies to rewrite a faulty bodily experience and to
regenerate the wellbeing of an individual.

References
[1] B.K. Wiederhold, Ready (or Not) Player One: Initial Musings on the Metaverse, Cyberpsychology,
Behavior, and Social Networking 25 (2021), 1-2.
[2] F. Tao and Q. Qi, Make more digital twins, Nature (2019), 490-491.
[3] G. Riva, Virtual Reality, in: The Palgrave Encyclopedia of the Possible, Springer International
Publishing, Cham, 2020, pp. 1-10.
[4] L. Valmaggia, The use of virtual reality in psychosis research and treatment, World Psychiatry 16 (2017),
246-247.
[5] S.W. Jerdan, M. Grindle, H.C. van Woerden, and M.N. Kamel Boulos, Head-Mounted Virtual Reality
and Mental Health: Critical Review of Current Research, JMIR Serious Games 6 (2018), e14.
[6] K. Glantz, N.I. Durlach, R.C. Barnett, and W.A. Aviles, Virtual reality (VR) and psychotherapy:
Opportunities and challenges, Presence, Teleoperators, and Virtual Environments 6 (1997), 87-105.
[7] F. Vincelli, From imagination to virtual reality: the future of clinical psychology, Cyberpsychology &
Behavior 2 (1999), 241-248.
[8] F. Vincelli, E. Molinari, and G. Riva, Virtual reality as clinical tool: immersion and three-dimensionality
in the relationship between patient and therapist, Studies in Health Technology and Informatics 81
(2001), 551-553.
[9] M.B. Powers and P.M. Emmelkamp, Virtual reality exposure therapy for anxiety disorders: A meta-
analysis, J Anxiety Disord 22 (2008), 561-569.
[10] D. Opris, S. Pintea, A. Garcia-Palacios, C. Botella, S. Szamoskozi, and D. David, Virtual reality
exposure therapy in anxiety disorders: a quantitative meta-analysis, Depress Anxiety 29 (2012), 85-93.
[11] K.J. Friston, The free-energy principle: a unified brain theory?, Nat Rev Neurosci 11 (2010), 127-138.
[12] K.J. Friston, Embodied inference and spatial cognition, Cognitive Processing 13 Suppl 1 (2012), S171-
177.
[13] K.J. Friston, Does predictive coding have a future?, Nat Neurosci 21 (2018), 1019-1021.
[14] G. Riva, B.K. Wiederhold, and F. Mantovani, Neuroscience of Virtual Reality: From Virtual Exposure
to Embodied Medicine, Cyberpsychol Behav Soc Netw 22 (2019), 82-96.
[15] G. Riva, The Neuroscience of Body Memory: from the Self through the Space to the Others, Cortex 104
(2018), 241-260.
[16] O. Blanke, M. Slater, and A. Serino, Behavioral, Neural, and Computational Principles of Bodily Self-
Consciousness, Neuron 88 (2015), 145-166.
[17] M.P. Paulus, J.S. Feinstein, and S.S. Khalsa, An active inference approach to interoceptive
psychopathology, Annual Review of Clinical Psychology 15 (2019), 97-122.
[18] M. Botvinick and J. Cohen, Rubber hands 'feel' touch that eyes see, Nature 391 (1998), 756.
[19] E. Tatti, S. Rossi, I. Innocenti, A. Rossi, and E. Santarnecchi, Non-invasive brain stimulation of the
aging brain: State of the art and future perspectives, Ageing Research Reviews 29 (2016), 66-89.
Riva et al./ Virtual Reality Therapy in the Metaverse: Merging VR for the Outside with VR for the Inside
8

[20] O. Pollatos, B.M. Herbert, S. Mai, and T. Kammer, Changes in interoceptive processes following brain
stimulation, Philosophical Transactions of the Royal Society B: Biological Sciences 371 (2016),
20160016.
[21] A. Marotta, A. Re, M. Zampini, and M. Fiorio, Bodily self-perception during voluntary actions: The
causal contribution of premotor cortex and cerebellum, Cortex 142 (2021), 1-14.
[22] M. Matamala-Gomez, A. Maselli, C. Malighetti, O. Realdon, F. Mantovani, and G. Riva, Virtual Body
Ownership Illusions for Mental Health: A Narrative Review, Journal of clinical medicine 10 (2021),
139.
[23] J. Dunn, E. Yeo, P. Moghaddampour, B. Chau, and S. Humbert, Virtual and augmented reality in the
treatment of phantom limb pain: A literature review, NeuroRehabilitation 40 (2017), 595-601.
[24] M. Matamala-Gomez, A.M. Diaz Gonzalez, M. Slater, and M.V. Sanchez-Vives, Decreasing Pain
Ratings in Chronic Arm Pain Through Changing a Virtual Body: Different Strategies for Different Pain
Types, J Pain 20 (2019), 685-697.
[25] M. Matamala-Gomez, T. Donegan, S. Bottiroli, G. Sandrini, M.V. Sanchez-Vives, and C. Tassorelli,
Immersive Virtual Reality and Virtual Embodiment for Pain Relief, Front Hum Neurosci 13 (2019), 279.
[26] S. Serino, N. Polli, and G. Riva, From avatars to body swapping: The use of virtual reality for assessing
and treating body-size distortion in individuals with anorexia, Journal of Clinical Psychology 75 (2019),
313-322.
[27] D. Freeman, S. Reeve, A. Robinson, A. Ehlers, D. Clark, B. Spanlang, and M. Slater, Virtual reality in
the assessment, understanding, and treatment of mental health disorders, Psychological Medicine (2017),
1-8.
[28] G. Riva, S. Serino, D. Di Lernia, E.F. Pavone, and A. Dakanalis, Embodied Medicine: Mens Sana in
Corpore Virtuale Sano, Frontiers in Human Neuroscience 11 (2017).
[29] F. Schoeller, A.J.H. Haar, A. Jain, and P. Maes, Enhancing human emotions with interoceptive
technologies, Physics of life reviews (2019).
[30] D. Di Lernia, M. Lacerenza, V. Ainley, and G. Riva, Altered Interoceptive Perception and the Effects
of Interoceptive Analgesia in Musculoskeletal, Primary, and Neuropathic Chronic Pain Conditions, J
Pers Med 10 (2020).
[31] P. Iodice, G. Porciello, I. Bufalari, L. Barca, and G. Pezzulo, An interoceptive illusion of effort induced
by false heart-rate feedback, Proc Natl Acad Sci U S A 116 (2019), 13897-13902.
[32] G. Riva , S. Serino, D. Di Lernia, and F. Pagnini, Regenerative Virtual Therapy: The use of Multisensory
Technologies and Mindful Attention for Updating the Altered Representations of the Bodily Self,
Frontiers in System Neuroscience (2021), 749268.
SECTION II

CRITICAL REVIEWS

In general, there are two reasons why

cybertherapy is used: either because there is
no alternative, or because it is in some sense
better than traditional medicine.

In this sense, telehealth has been used

very successfully for optimizing health
services delivery to people who are isolated
due to social and physical boundaries and
limitations.

Nevertheless, the benefits of

cybertherapy, due to the variety of its
applications and uneven development, are not
self-evident.

However, the emergence of cybertherapy

is supporting the cost-effectiveness of certain
applications such as assessment,
rehabilitation, and therapy in clinical
psychology and neuroscience.

Wiederhold & Riva, 20

 Annual Review of Cybertherapy and Telemedicine 2021 11

Implicit Measures of Perceived Realness

in Virtual Reality
Marius RUBOa,1, Jasin SAHRAOUI b and Simone MUNSCHa
a
Clinical Psychology and Psychotherapy, Fribourg University, Switzerland
b
University of Würzburg, Germany

Abstract. Virtual Reality (VR) provides an interesting form of stimulus presentation

due to its capability to elicit lifelike experiences. People often feel present (i.e. as if they were
there) in a virtual scenario, but they can additionally obtain a sense of owning a virtual body
and experience virtual conspecifics as socially present. These experiences can be assessed
explicitly using questionnaires, but several studies have also employed behavioral and
physiological measures to trace when people experience a virtual situation as real. This article
gives a brief introduction to how implicit measures were used to explore these distinct but
related facets of perceived realness in VR and how they may be exploited to better understand
mental health conditions.

Keywords. Virtual Reality, Presence, Virtual Body Ownership, Social Presence, Implicit
Measures

1. Introduction

Early on in VR research, it was observed that physiological reactions in VR (such

as changes in heart rate in response to a threatening virtual situation [8]) can correspond
well to those expected in a real-life situation, and it was argued that such measures may
be interpreted as objective surrogates of presence. Later research additionally indicated
that a weaker fronto-central Mismatch Negativity component (MMN; an event-related
component) to task-irrelevant beeping tones in participants’ EEG was indicative of their
experienced presence [9]. However, [10] concluded that physiological correlates of the
sense of presence are not yet sufficiently robust overall. Moreover, since the evocation
of physiological reactions often required an induction of specific states (e.g. stress) or
additional stimuli (beeping tones), it was argued that behavioral measures could be more
suitable to unobtrusively monitored presence instead [2].
In a study by [11], participants were asked to point towards a radio which they
had seen both in the real laboratory room and a virtual scene but in different locations.
Participants more strongly pointed in the direction of the virtual radio as compared to the
real radio when the virtual scene was depicted at a higher visual acuity, substantiating
the idea that this behavioral measure may mirror experiences of presence. In a study by
[12], participants walked more slowly across a virtual steel girder when exposed to
height in VR, following a safety strategy which can be expected in a real-life situation.
Another indirect measure of presence may lie in analyzing how experiences in
VR are stored in memory. Here, the observation that treatment effects from virtual reality
exposure therapy can transfer well to real life [13] may point to high levels of presence
during confrontation with a feared stimulus. Indeed, object learning was found to be
facilitated in VR compared to the viewing of a computer monitor [14]. As a more direct
test for the realness of VR experiences in memory, our own study tested how well people
can distinguish real from virtual experiences in hindsight [15]. Participants viewed 3D
models either in reality, VR, or on a computer screen and were later asked to recall where
they had seen each model. Participants more frequently misattributed VR and reality
compared to VR and the computer monitor as sources of a memory trace, possibly
revealing a mechanism driving the good generalization of VR treatment effects. Future
research may investigate if a tendency for such source memory errors is predictive of
therapeutic success in virtual reality exposure therapy.

1
Corresponding Author: [email protected].
Rubo et al./ Implicit Measures of Perceived Realness in Virtual Reality
12

Overall, it was argued that, while a plethora of behavioral markers for presence
can be envisioned in various virtual scenarios, the field is still understudied [2], and
potential approaches to implicitly monitor presence remain to be exploited. Note that the
approaches described here do not allow for a continuous monitoring of presence; they
either require the implementation of specific events (e.g. exposure to height, asking
participants to point to an object) or are only assessed after the experience (in memory
tasks).

2. Implicit Measures of Body Ownership

A large number of studies have documented that people can develop a sense of
ownership over a virtual body or body parts. The illusion is facilitated by spatial overlap
between the two bodies or body parts, human-like visual representation, visuo-tactile
congruency, and visuo-motor congruency [16, 17, 4]. A common indirect measure for
the ownership over a virtual or fake hand is a proprioceptive drift [18]. After taking
ownership over a hand model which is located at a slightly different position compared
to their real hand, participants are asked to indicate its felt position by pointing towards
it using their other hand. Here, pointing more towards the location of the fake hand is
interpreted as sign of a successful ownership illusion (although it may only index an
ownership illusion in situations where visuo-tactile congruency is present [19]). A
parallel measure was introduced for assessing the experience of taking ownership over a
whole fake body. In a seminal study [20], participants saw their own back from behind
by wearing a head-mounted display (HMD) which displayed the live-streamed video
recorded by a camera positioned 2 meters behind them. After participants’ backs were
stroked for one minute (which participants could view on their own backs in front of
them), they were passively displaced while being blind-folded and asked to return to
their initial position. Participants moved to a position in space closer to where their
virtual own body was located, indicating a shift in self-location and thus an experience
of the illusory situation as real.
In studies where participants took ownership over a virtual body in VR, the body
illusion was furthermore found to be associated with heart rate deceleration [21] and skin
conductance reactions [22, 23] following a threat towards the virtual body. In our own
research, we investigated continuous walking behavior as a proxy for a body illusion
[24]. Participants took ownership over a more corpulent virtual avatar and either
experienced intact or deteriorated visuo-tactile congruency when touching their own
hips. Participants who experienced intact visuo-tactile congruency more consistently
walked in the laboratory as if they were actually more corpulent than before, keeping
distances to obstacles in line with being more corpulent.
Additionally, several studies documented how successful ownership over a virtual
body different from one’s own can alter social cognition [25]. For instance, [26] and [27]
observed a shift towards associating oneself with child-like concepts in an Implicit
Association Test (IAT) after taking ownership over a childlike virtual body. In the future,
these approaches may help to more closely trace body schema disturbances in mental
disorders such as anorexia [7] or explore self-associations in mood disorders [26].

3. Implicit Measures of Social Presence

Implicit markers of social presence are reactions towards virtual agents which are
comparable to those observed towards real humans. An interesting demonstration for
such an effect made use of the social inhibition of return. This effect, where a stimulus
is reacted to with additional delay if it was previously reacted to by a conspecific, is
commonly observed when interacting with a real counterpart, and was also – albeit in a
smaller magnitude – observed when interacting with a virtual agent shown in VR [28].
Another study demonstrated that virtual avatars can induce social stress [29]: when
forming a committee in a Trier Social Stress Test (TSST), participants reacted with
similarly increased stress markers (both in explicit reports and on a physiological level)
as towards a real TSST committee. In order to directly compare the influence of VR on
an implicit measure of social presence, we compared gaze behavior towards a virtual
 Rubo et al./ Implicit Measures of Perceived Realness in Virtual Reality 13

agent seen in VR with that of the same scene displayed on a computer monitor [30]. Here,
participants in the VR condition more strongly reciprocated the virtual agent’s social
gaze, pointing to a more natural reaction to this (artificial) social situation and thus to a
stronger social presence. Such findings may be helpful in better monitoring and
interpreting social avoidance behavior in the course of VR therapy programs for treating
social anxiety disorder [31].

4. Conclusion

A large body of research investigated how experiences in VR can involve a sense of

presence [2], illusory body ownership [4], and social presence [5]. Common future use
cases of VR such as sharing a virtual space in embodied social encounters may
incorporate all of these three facets of experiencing a VR situation as real. All three
phenomena have seen a thorough investigation of boundary conditions as well as a
systematic advancement of explicit measures. By contrast, although there exist examples
for implicit measures, such developments currently remain more scarce and fragmented,
especially regarding behavioral markers. The development of sensitive and reliable
implicit measures of the perceived realness could, especially if they can be assessed in a
continuous and nonintrusive manner [2], help to deepen our knowledge on how people
experience situations in VR, and allow us to monitor problems and progress during VR
therapy programs in a variety of mental disorders.

References

[1] Riva G. Virtual Reality in Clinical Psychology. Reference Module in Neuroscience and Biobehavioral
Psychology. 2022:B978-0-12-818697-8.00006-6.
[2] Skarbez R, Brooks, Jr. FP, Whitton MC. A survey of presence and related concepts. ACM Comput Surv.
2018;50(6):1–39.
[3] Gutiérrez M, Vexo F, Thalmann D. Stepping into Virtual Reality. 1st ed. London: Springer; 2008. 2-3 p.
[4] Maselli A, Slater M. The building blocks of the full body ownership illusion. Front Hum Neurosci. 2013
Mar 21;7(38).
[5] Oh CS, Bailenson JN, Welch GF. A systematic review of social presence: Definition, antecedents, and
implications. Front Robot AI. 2018 Sep 11;5(114).
[6] Roth D, Lugrin JL, Latoschik ME, Huber S. Alpha IVBO-construction of a scale to measure the illusion
of virtual body ownership. In: Proceedings of the 2017 CHI Conference Extended Abstracts on Human
Factors in Computing Systems; 2017 May 6-11; Denver, USA. New York (NY): Association for
Computing Machinery. p. 2875-83.
[7] Bailenson, JN, Blascovich J, Beall AC, Loomis JM. Interpersonal distance in immersive virtual
environments. Pers Soc Psychol Bull. 2003 Jul 1;29(7):819-33.
[8] Meehan M, Insko B, Whitton M, Brooks Jr FP. Physiological measures of presence in stressful virtual
environments. ACM Trans Graph. 2002 Jul;21(3):645-52.
[9] Terkildsen T, Makransky G. Measuring presence in video games: An investigation of the potential use of
physiological measures as indicators of presence. Int J Hum Comput Stud. 2019 Jun;126:64-80.
[10] Grassini S, Laumann K. Questionnaire measures and physiological correlates of presence: A systematic
review. Front Psychol. 2020 Mar 19;11(349).
[11] Slater M, Usoh M, Chrysanthou Y. The influence of dynamic shadows on presence in immersive virtual
environments. In: Göbel M, editors. Virutal environments’95. Eurographics. Vienna: Springer; c1995. p.
8–21.
[12] Kisker J, Gruber T, Schöne B. Behavioral realism and lifelike psychophysiological responses in virtual
reality by the example of a height exposure. Psychol Res. 2021 Feb;85(1):68-81.
[13] Morina N, Ijntema H, Meyerbröker K, Emmelkamp PM. Can virtual reality exposure therapy gains be
generalized to real-life? A meta-analysis of studies applying behavioral assessments. Behav Res Ther,
2015 Nov;74:18-24.
[14] Krokos E, Plaisant C, Varshney A. Virtual memory palaces: immersion aids recall. Virtual Real. 2019
May 16;23(1):1-15.
[15] Rubo M, Messerli N, Munsch S. The human source memory system struggles to distinguish virtual reality
and reality. Comput Hum Behav Rep. 2021 Aug-Sep;4(100111).
[16] Botvinick M, Cohen J. Rubber hands ‘feel’touch that eyes see. Nature. 1998 Feb 19;391(6669):756-756.
[17] Slater M, Pérez Marcos D, Ehrsson H, Sanchez-Vives MV. Inducing illusory ownership of a virtual
body. Front Neurosci, 2009 Sep 15;3(2):214-220.
[18] Kalckert A, Ehrsson HH. The moving rubber hand illusion revisited: Comparing movements and
visuotactile stimulation to induce illusory ownership. Conscious Cog. 2014 May;26:117-132.
[19] Rohde M, Di Luca M, Ernst MO. The rubber hand illusion: feeling of ownership and proprioceptive drift
do not go hand in hand. PloS One. 2011 Jun 28;6(6):e21659.
[20] Lenggenhager B, Tadi T, Metzinger T, Blanke O. Video ergo sum: manipulating bodily self-
consciousness. Science. 2007 Aug 24;317(5841):1096-1099.
Rubo et al./ Implicit Measures of Perceived Realness in Virtual Reality
14

[21] Slater M, Spanlang B, Sanchez-Vives MV, Blanke O. First person experience of body transfer in virtual
reality. PloS One. 2010 May12;5(5):e10564.
[22] Kilteni K, Normand JM, Sanchez-Vives MV, Slater M. Extending body space in immersive virtual
reality: A very long arm illusion. PloS One. 2012 July 19;7(7): e40867.
[23] Petkova VI, Ehrsson HH. If i were you: Perceptual illusion of body swapping. PloS One. 2008 December
3;3(12):e3832.
[24] Rubo M, Gamer M. Visuo-tactile congruency influences the body schema during full body ownership
illusion, Conscious. Cog. 2019 Aug;73(102758).
[25] Maister L, Slater M, Sanchez-Vives MV, Tsakiris M. Changing bodies changes minds: Owning another
body affects social cognition. Trends Cogn Sci. 2015 Jan;19(1):6–12.
[26] Banakou D, Groten R, Slater M. Illusory ownership of a virtual child body causes overestimation of
object sizes and implicit attitude. Proc Natl Acad of Sci U S A, 2013 Jul 30;110(31):12846–12851.
[27] Tajadura-Jiménez A, Banakou D, Bianchi-Berthouze N, Slater, M. Embodiment in a child-like talking
virtual body influences object size perception, self-identification, and subsequent real speaking. Sci Rep.
2017 Aug 29;7(1).
[28] Wienrich C, Gross R, Kretschmer F, Müller-Plath G. Developing and proving a framework for reaction
time experiments in VR to objectively measure social interaction with virtual agents. In: 2018 IEEE
conference on virtual reality and 3D user interfaces: 2018 Mar 18-22; Reutlingen, DE. IEEE; 2018. p.
191-198.
[29] Zimmer P, Buttlar B, Halbeisen G, Walther E, Domes G. (2019). Virtually stressed? A refined virtual
reality adaptation of the Trier Social Stress Test (TSST) induces robust endocrine
responses. Psychoneuroendocrinol. 2019 Mar;101:186-192.
[30] Rubo M, Gamer M. Stronger reactivity to social gaze in virtual reality compared to a classical laboratory
environment. Br J Psychol. 2021 Jun 2;112:301-314.
[31] Emmelkamp PM, Meyerbröker K, Morina N. Virtual reality therapy in social anxiety disorder. Curr
Psychiatry Rep. 2020;22(7)
 Annual Review of Cybertherapy and Telemedicine 2021 15

Between benevolent lies and harmful

deception: Reflecting on ethical
challengesin dementia care technology
Ans I.M. TUMMERS-HEEMELS1a, Rens G.A.
BRANKAERTa, Wijnand A. IJSSELSTEIJNa
a
Eindhoven University of Technology, Eindhoven, Netherlands

Abstract. In the context of dementia care, deception is a common yet controversial practice,
generating substantial attention from scholars. Though complicated, consensus has seemed to
emerge that, whereas lying is generally frowned upon, benevolent (white) lies can be acceptable
if the aim is to improve the life of the recipient. However, with the increasing omnipresence of
technology as a means of improving quality of life and care efficiency, many technologies,
implicitly or explicitly, embody deceptive practices. In the current paper, we expand our ethical
analysis and understanding of deceptive practices to include technological designs and human-
technology relations in dementia care settings, by reviewing current literature and exploring
relevant case studies. With our analysis, we hope to create awareness and proactive engagement
of technology developers, interaction designers, as well as care professionals, who want to
ethically develop and deploy care technologies containing benevolent deceptive elements.

Keywords. Dementia, Care Technology, Deception, Ethics

1. Introduction

Imagine, if you will, the following four scenarios:

i. An elderly lady with middle stage dementia loses her pet dog.
After weeks of intense grief, she is given an interactive robot
cat, to which she immediately develops a deep attachment.
She cares for and caresses the robot continuously, and her
grief over the lost dog is significantly lessened. She calls the
robot cat “her dog” and uses the name of her deceased dog.
When the batteries of the catrun low, she is deeply distressed
and calls her informal carer, telling him “the dog is dying”.
ii. An elderly lady in the later stages of dementia occasionally
shows intermittent episodes of significant restlessness and
emotional distress. The nursing staff, responding to her calls,
hand her what looks like an old-fashioned dial telephone that
connects the lady to the prerecorded voice of her son. The
system responds through scripted questions and answers,
where AI-based language recognition and voice stress
analysis allow for some level of flexibility and tuning of the
conversation. Believing she is talking to her son, the
conversation has a soothing effect on the elderly lady who
ends the conversation by asking when her son iscoming over
to visit her again. The computer responds, in the voice of her
son:“I’ll be over this evening” – an answer that puts a big
smile on her face.

1
Corresponding author a.i.m.tummers- [email protected]
Tummers et al./ Between benevolent lies and harmful deception: Reflecting on ethical challenges
in dementia care technology
16
iii. The garden of a nursing home for people with middle to late-
stage dementia hasits own bus stop. It is designed with all
the familiar bus stop signs, timetables, a booth, and a bench
to sit on. However, no bus will ever arrive at this stop. It is a
fake bus stop erected with the express purpose to attract
people with dementia prone to wandering around or off the
nursing home grounds – a significant source of stress for
caregivers and care facilities. Here, they sit and wait for the
bus.
iv. An ambient assisted living facility of a senior couple has
been outfitted with a new, state-of-the-art, dynamic lighting
system. The system is designed to detect the mood of the
residents and subsequently adjust the lighting in the room to
either calm or activate the residents. The system uses sensors
and affective computing algorithms to measure and interpret
the couple’s mood and controlsthe LED-based light settings
using a pre-constructed mood model. As the “active” mood
lighting kicks in at around 9:00 a.m., the senior couple feel
thatit must be a nice bright day outside and plan to go for a
walk.

All four scenarios are based on actual, existing technological interventions – some
are based on high-tech engineering including AI or robotics, others include physical
redesign of familiar environmental characteristics. They all share the aim of wanting to
improve the quality of life for people with dementia (PwD) and/or alleviate the care
burden for informal and professional carers. They also share the use of deceptive
practices to reach that aim. Whereas in recent literature on nursing and dementia care,
attention has been growing regarding the use, boundary conditions, and ethical
implications of people using benevolent (white) lies as part of caring for PwD, there has
been scant attention to the implications of technology embodying deceptive practices in
dementia care. The current paper aims to address this urgent issue.

2. Deception, and its use in dementia care practice

In this paper, our contribution will focus on deception in dementia care, and in
particular the role that technology can play as a conduit of deception. Deception is “the
act of causing someone to accept as true or valid what is false or invalid” (Merriam –
Webster). Deception implies intent, as well as an intentional agent (the deceiver), and is
typically situated in the context of social interactions between human beings. In most
definitions of deception, there is an element of personal gain on the part of the deceiver.
Deceptive acts may include acts or statements that mislead or promote a falsehood.
Deceptive acts may also wittingly withhold or otherwise hide information required to
access or appreciate the truth.
Deception is generally seen as morally reprehensible for a variety of reasons. From
some philosophical (e.g., deontological) and religious perspectives, it is seen as a
fundamental moral wrong – a universal denouncement of any form of deception.
Instrumentalist or consequentialist views of deception attempt to weigh the costs of
deception against its potential benefits, creating a more situational or contextualized
moral view on deception rather than an absolute denouncement. In this view, there are a
variety of cases of deception that are not regarded as clear moral transgressions, such as
lying under threat of harm. In general though, deception impinges on the rights of the
person being deceived to make a free and informed decision, effectively diminishing that
person to a means to achieve the deceiver’s purpose, thus negatively impacting their
autonomy.
It is important to note that not all lies (or other forms of deception) are necessarily
done with malicious intent or for personal gain. The best-known example of this is the
white lie, which is a well-intentioned untruth, generally meant to avoid hurting
someone’s feelings. White lies are frequent, rather trivial occurrences in day-to-day
interpersonal relationships, wherein they typically help grease the social wheels and
make people feel good about themselves (e.g., “you haven’t aged a bit”).
Tummers et al./ Between benevolent lies and harmful deception: Reflecting on ethical challenges in dementia
care technology 17

When reviewing recent literature on nursing and dementia care, we note that white
lies are a commonly used strategy in this context, generally with an aim to serve the best
interest of a person with dementia [4]. Moreover, the rationale for deception appears to
be more about managing a person’s behavior and emotional distress than it is about
controlling information. For example, in the more progressed stages of dementia,
confabulation is frequently encountered, where a person with dementia may be totally
convinced of a parallel but untrue reality, frequently associated with events or relations
from one’s past life. If this happens, caregivers can reach someone by carefully listening
to, and not contradicting, the confabulated stories, essentially validating the person’s
altered sense of reality [5]. As another example, people with dementia may be reliving
an emotionally charged episode from their past (e.g., a stressful episode at work, a sick
child, or a dying mother) where they may be highly distressed and may think they
urgently need to act on a perceived responsibility. These “time-shifted” stressful
experiences can lead to distress and restless behavior. In these cases, it is common for
staff to use white lies – such as “your mom is fine” – to comfort and distract the person
with dementia [7]. In a recent cross-sectional survey study, Cantone et al. [1] reported
on attitudes and behaviours of geriatric and psychiatric nurses towards the use of lies in
dementia care. Only a few of the respondents (10 to 12%) stated that they never used lies
or that it would never be acceptable to use lies in care practice. Attitudes differed
depending on the situation; more respondents were prone to tell a lie “to prevent or
reduce aggressive behaviors” than “to avoid wasting time giving explanations.” In all, it
appears that views on deception and truth-telling in dementia care practice are strongly
instrumentalist in nature. Rather than being an intrinsic wrong, deception is evaluated
and used in a contextualized fashion, serving the interests of the patient and/or the care
provider.
In recent years, a number of guidelines and communication strategies have been
published, based on qualitative research amongst care professionals and PwD, on how
to deal with deception in dementia care settings. One example is the work by James et
al.[4] formulating 12 guidelines, including that a lie should only be told in the best
interest of the PwD (e.g., to ease distress), that consideration should be given to an
individual’s ability to discern truth from falsehood, that family consent should be
obtained, the lies should be documented, and that an individualized and flexible approach
should be adopted towards each case, weighing the benefits and costs relating to a lie.
In addition to such guidelines, the same group [8] developed a four-stage
communication strategy where telling the truth is the first option to meet a resident’s
need (e.g., wish to see one’s husband), and subsequent stages include addressing the
underlying need (e.g., comforting in response to perceived loneliness), offering
distraction (e.g., ask the PwD to help to set the table), and finally, telling a therapeutic
lie. However, in practice, health care professionals often struggle with the interpretation
of these guidelines and strategies and may strongly vary in their use of therapeutic lies.
Many professionals are also conflicted in striking the right balance between the virtues
of being an honest person and doing the right thing given the circumstances. All this is
suggestive of a situated and complementary mix of care ethics and virtue ethics – of
responding with empathy and integrity on a case-by-case basis where the therapeutic lie
should be used sparingly and only when it is used to protect and prevent distress of the
person with dementia, and alternative strategies are not available or don’t work [2, 6].
Indeed, a recent study using a sample of people with dementia and their carers stresses
that in addition to having good intentions, “the carer telling the lie had to really know the
person and be cognizant of family preferences” [2].

3. Considering deceptive practices mediated by dementia care technologies

The introduction of care technologies brings new challenges and complexities to the
fore in the way they can, intentionally or unintentionally, be embodiments or conduits of
deceptive practices. The four scenarios mentioned at the start of this paper – (i) the robot
cat, (ii) the AI phone conversation, (iii) the fake bus stop, and (iv) the mood-sensitive
lighting – all offer illustrative cases of technologies that are being used, or proposed to
be used, in current dementia care practice. They are used here to help reflect on the
Tummers et al./ Between benevolent lies and harmful deception: Reflecting on ethical challenges
in dementia care technology
18
ethical challenges and tradeoffs that the introduction of technology in dementia care may
imply, and specifically where they touch upon the use of deception.
As a consequence of their cognitive vulnerabilities, PwD, especially in more
advanced stages of the disease, are more likely than healthy seniors to be unable to
distinguish simulated or mediated reality from actual reality. This “illusion of non-
mediation” [6] can pertain to social simulations as in cases (i) and (ii), as well as
reproductions of physical environments (as in (iii), but also seen, for example, in VR-
based bike rides “to work”, or in life-size pictures of one’s former front door attached to
the door of the resident’s current apartment). Such simulations/reproductions are
frequently meant to encourage connection to socially significant others, stimulate
reminiscence, or elicit a sense of familiarity of surroundings. However, little data is
available on the extent to which these experiences may also confuse, defamiliarize, and
disconnect PwD from their current lived reality. Expectations (e.g., of talking to one’s
son, of waiting for the bus to arrive) may be violated, and such “reality breakdowns”
may reinforce the confusion and alienation that many PwD already experience. Relatedly,
the inner complexities of some of the technologies in use may lead to an inaccurate
mental model of the workings of the technology, and their derived meaning. Scenario
(iv) is but one example of a host of “smart home” innovations where the causal chain of
sensing, interpretation, and actuation may remain entirely opaque to the resident. As a
consequence, s/he may engage in unproductive interactions in an attempt to exert control
or may interpret experiences in ways which are unrelated to physical reality (see also 3).
In relation to our discussion of deception in dementia care, one can generally
recognize and acknowledge the positive intentions with which technologies are being
developed and introduced. The rationale for these technologies typically includes
improving the quality of life of PwD and supporting care efficiency and effectiveness.
However, a number of considerations that are part of the guidelines and situated ethics
around deception by human care staff and family can also be valuable in considering the
ethical cost-benefit tradeoffs when technologies are used as part of deception.
Despite their increased sophistication, technologies generally lack the intelligence
(and empathy) required to really know a person, their context, life history, and family
preferences. Many technologies have been developed with only bare-bones adaptivity or
personalization features that supports the technology’s functionality. Most technologies
currently in use are typically not receptive to surprise, confusion, anger or
disappointment on the part of the PwD, nor can these technologies respond adequately
to such responses if they occur. Importantly, even though care technologies have clear
moral implications, they are not moral agents in themselves; they cannot take an “all
things considered” approach that is required to support ethical decision making. In some
cases, the nursing staff may play an active role in choosing to deploy the technology,
assessing the ethical pros and cons, and managing or ameliorating any unforeseen,
potentially negative effects. In other cases, the technology is omnipresent or always-on,
and may impact the PwD - positively or negatively - whether they are being presently
attended to by staff or family members or not.
As our review of deception in dementia care demonstrated, deceptive practices by
nursing staff are deeply contextualized and adapted to the situation and person at hand.
Similarly, this should hold true for the technologies we develop and deploy. In addition
to good intentions, technologies need to be used in ways that are adaptive to the situation
at hand: to only use deception when other interaction strategies are ineffective, and then
to use deception only sparingly and with integrity and restraint. Deception is not intended
to create entirely new realities, but rather to defuse potentially harmful or stressful
situations. Secondly, when using deception, it is necessary to only do this with a constant
sensitivity to a person’s needs and responses, the particularities of the context, and family
preferences. To echo Casey et al. [2], whoever is doing the deceiving has to really know
the person. This requires technology to be developed using person-centered co-design
methods to ensure it is genuinely serving the needs of the PwD (the Warm Technology
approach, 3). Moreover, after its introduction, care technology really requires constant
human involvement to verify and ensure that the technology is still appropriate given
changing circumstances and human needs – an explicit and continuous assessment of its
fitness-for-purpose and its associated ethical cost-benefit tradeoffs. The complexities of
introducing technologies that contain benevolent deceptive elements into care practice
Tummers et al./ Between benevolent lies and harmful deception: Reflecting on ethical challenges in dementia
care technology 19

thus emphasizes the fact that care technologies should always be designed to supplement
rather than replace human care.

References

[1] Cantone, D, Attena, F., Cerrone, S, Fabozzi, A, Rossiello, R, et al. Lying to patients with dementia:
Attitudes versus behaviours in nurses. Nursing Ethics, 2019, 26(4): 984-992.
[2] Casey D., Lynch U., Murphy K, Cooney A, Gannon M, Houghton C, et al. Telling a ‘good or white
lie’: The views of people living with dementia and their carers. Dementia, 2020 Nov;19(8):2582-600.
[3] IJsselsteijn W, Tummers-Heemels A, Brankaert R. Warm Technology: A novel perspective on design
for and with people living with dementia. In: R. Brankaert & G. Kenning (eds). HCI and Design in the
Context of Dementia 2020 (pp. 33-47). Springer, Cham.
[4] James, I. A, Wood‐Mitchell, AJ, Waterworth, AM, Mackenzie, LE, & Cunningham, J. Lying to
people with dementia: Developing ethical guidelines for care settings. International Journal of
Geriatric Psychiatry, 2006, 21(8): 800-801.
[5] Lindholm, C. Parallel realities: The interactional management of confabulation in dementia care
encounters. Research on Language and Social Interaction, 2015, 48(2): 176-199.
[6] Lombard M, Ditton T. At the heart of it all: The concept of presence. Journal of Computer-Mediated
Communication. 1997 Sep 1;3(2):JCMC321.
[7] Turner, A, Eccles, F, Keady, J, Simpson, J, & Elvish, R. The use of the truth and deception in
dementia care amongst general hospital staff. Aging & Mental Health, 2017, 21(8): 862-869.
[8] Wood-Mitchell A, Waterworth A, Stephenson M, James I. Lying to people with dementia: sparking
the debate. Journal of Dementia Care, 2006, 14(6):30-1.
 Annual Review of Cybertherapy and Telemedicine 2021 21

Normative Affordances: Utilising the

constraint of context-specific expectation in
simulated environments
John Francis LEADERa,1
a
Media & Entertainment Psychology Lab, University College Dublin, Ireland

Abstract. Affordances are the interactional opportunities that exist between us and
our environment. The design processes of simulations (using technologies such as
virtual reality and mixed reality) can benefit from considering three distinct forms
of affordances. Possible affordances are interactional opportunities that exist but are
not perceived. Perceived affordances appear to offer interaction possibilities but in
fact do not. Normative affordances are those actions that are consistent with the
socially-constrained behavioural expectations of a given setting. By considering
these three types of affordances and, in particular, utilising normative affordances
in the creation of affordance arrays, it is argued that more compelling narrative
experiences can be created with comparatively minimal resources.

Keywords: Affordances, Normative Affordances, Simulated environments

1. Introduction

By virtue of our embodiment, different aspects of our environment become more

or less salient. Individual differences between human beings—and between species—
create variations in what aspects of the environment are significant. This difference of
umwelt, or our subjective sense of our own environment as per the work of Jakob von
Uexküll (1864–1944), means that even the same environment can tell a very different
story for different participants [1]. The eyesight of a bird, a dog’s senses of hearing and
of smell, and a bat’s echolocation system all offer radically different ways of engaging
with potentially the same environment [2]. Just as importantly, the diverse needs of
organisms lead to a filtering of the environment in different ways. The endurance of this
meaningfulness can vary, with something like oxygen in the air being persistently
meaningful to those who have lungs, while an interest in food would be circumstantially
meaningful depending on recent eating patterns. James J. Gibson (1904–1979), in his
work in ecological psychology, described affordances as follows:
“The affordances of the environment are what it offers the animal, what it provides or
furnishes, either for good or ill. The verb to afford is found in the dictionary, the noun
affordance is not. I have made it up. I mean by it something that refers to both the
environment and the animal in a way that no existing term does. It implies the
complementarity of the animal and the environment.” [3]

Affordances, in Gibson’s view, exist relationally. That is, a chair will be a rather
different thing to an adult human than to a child, and to other species like cats or ants.
While some things—like oxygen—may be afforded automatically to an organism by its
environment, in other cases—like food—resources must be actively sought. For some
organisms with simpler needs, there are fewer degrees of abstraction [1]. For human
beings, the parameters are substantially more complex. Base needs, as something like
the bottom level of Abraham Maslow’s [4] hierarchy of human needs, are in and of
themselves more involved: varied nutritional needs, shelter, warmth, etc. However, even
more complex are the variety of psychological and existential needs that human beings

1
Corresponding Author, e-mail: [email protected]
 Leader et al./ Normative Affordances: Utilising the constraint of context-specific expectation in simulated
environmenta 22

possess, and which motivate action. Through life experience, these inherent needs
become mediated through patterns of interaction, and we learn effective and ineffective
ways of achieving things. We want food, and there may be competition, so eating it first
means we’re more likely to get it. But, if we take it away from others who want it, they
may be less likely to share with us in the future. In this way, natural biological tendencies,
as they are enacted in the context of an environment, are shaped and directed.
The role of experiential therapy and training then, whether conducted physically,
imaginally, or through simulative tools such as Virtual Reality (VR) and Mixed Reality
(MR), is to help develop these relations to ensure the most appropriate affordances are
prioritised, and that the strategies for achieving them are sustainable [5]. In recognising
this meaningful relationship between the organism and their environment, the distributed
nature of the subject becomes more evident, and, for behavioural change purposes,
various potential modes of intervention start to stand out. Rather than just focusing on
how the person makes sense of the world, can the world be changed, manipulated, or
presented in ways that change the person? Fundamentally, we know that the answer is
yes. Techniques of behavioural shaping have demonstrated this not just in therapies and
trainings, but in advertising, marketing, and shopping reward cards [6]. However, while
basic principles such as reward and punishment are effective for certain well-defined
behavioural outcomes, it is possible to work at an even finer grain, and to not just leave
the cognitive system as a black box in the middle of a variety of environmental cues.
Instead, the aim is to meet its nuances, understand them, and scaffold them via
environmental supports.

2. Possible and perceived affordances

Gibson’s use of the term affordance was directed at the physical environment, but
how does such a concept transfer to considering the virtual? In a Human Computer
Interaction (HCI) context, a field where both affordances and virtuality have been
considered significant, Donald Norman [7] describes affordances as follows:
…“the term affordance refers to the perceived and actual properties of the thing,
primarily those fundamental properties that determine just how the thing could possibly
be used. […] Affordances provide strong clues to the operations of things. Plates are for
pushing. Knobs are for turning. Slots are for inserting things into. Balls are for throwing
or bouncing. When affordances are taken advantage of, the user knows what to do just
by looking: no picture, label, or instruction needed.” [7]

For Norman, an affordance—contrary to Gibson’s definition—is the perception of

interaction afforded rather than the actuality of that possibility when we attempt to push
the plate, turn the knob, insert something into the slot, or throw the ball. Virtual versions
of the above which lack such actual interactivity would, therefore, be classed as
affordances by Norman but not by Gibson. From Gibson’s particularly realist position,
an affordance would exist even if it was never detected by an organism.

Figure 1. Perceived and possible affordances

In William Gaver’s diagram above, affordances may exist and be perceptible (top right)
or may not exist nor appear to (bottom left). They may exist and not be perceptible
(bottom right) or may appear to exist but in fact not (top left) [8].
Leader et al./ Normative Affordances: Utilising the constraint of context-specific expectation in simulated
environments 23

William Gaver [8] draws together perceived and actual affordances yielding four
categories, as shown in Figure 1 above. On the top row of the grid are circumstances
where there appears to be an affordance, which may turn out to be a false affordance or
a perceptible affordance that offers actual interactive possibilities. On the bottom row
are circumstances where there is no apparent affordance, which may lead to correct
rejection if there is in fact no affordance or—if there is—then this is a hidden affordance.
For more complex interactions, Gaver considers groups of affordances that are sequential
in time or nested in space, leading to the possibility of a combination of the above
categories in a given task. For Gibson, only the categories of perceptible and hidden
affordances come under the formal definition of affordances. Norman, while not denying
the existence of hidden affordances, is interested in his definition in perceived
affordances, whether false or not.
In carefully considering these distinctions, it becomes clear that it would be a
mistake to conflate the virtual and the digital. A physical door—complete with a handle
that does not move and that leads to nowhere—would be an entirely physical object yet
it would offer false affordances. A digital door in a virtual simulation may function
perfectly to enable progress into another room, albeit also virtual. Affordances are of
interest in training, assessment, and therapeutic contexts because as they occur
relationally, they allow an insight into psychological phenomena that may not be
otherwise accessible. In introducing an affordance, or arrays of affordances for more
substantial scenarios, consideration is needed as to what level of interaction is required.
For example, in working with a phobia of heights, a perceived sense of height is all that
is needed, and an actual fall in no way adds value to the simulation. Whether we make
use of physical, virtual, or imagined affordances is likely to be immaterial—all other
things being equal—provided we can include the interaction possibilities relevant to the
scenario [9].

3. Normative affordances

Figure 2 below shows three notional ways of viewing affordances in the context of
narrative scenarios created for therapeutic, training, and assessment purposes.

Figure 2. Normative affordances chart

Possible affordances, shown in the bottom left circle of the Venn diagram above, exist
but are not perceived. In the bottom right circle, perceived affordances appear to offer
interaction possibilities but in fact do not. In the top circle, normative affordances are
those actions that are consistent with the socially-constrained behavioural expectations
of a given setting. The intersections show combinations: PPN, PUN, IPN and PPA.
Leader et al./ Normative Affordances: Utilising the constraint of context-specific expectation in simulated
environments
24

The bottom left circle of Figure 2 above, possible affordances, relates to Gibson’s
notion of affordances as existing whether an agent is aware of them or not. The bottom
right circle, perceived affordances, relates to Norman’s view of affordances as being
cues to action, whether or not they are acted on and whether or not such interaction
possibilities in fact exist. The upper circle, normative affordances, refers to a kind of
sociocultural barrier that serves to constrain behaviour. For example, to a passenger with
no aviation experience being given a tour of an aircraft cockpit, there are a range of
possible affordances in terms of the aircraft’s equipment that may not be apparent to a
novice, but which are apparent to the pilot. The capacity for such interaction exists and
may be recognised as the pilot demonstrates it, or if the passenger later trains as a pilot.
However, without any introduction, the passenger may notice—or at least is capable of
noticing—a variety of opportunities for interaction—buttons that may be pressed, levers
that may be pulled, cabinets that may be opened. However, just because they can be
interacted with does not mean that they will as, in most cases, sociocultural conditioning
serves to moderate and constrain behaviour to a more limited set of possibilities; these
can be considered normative affordances. One of the functions imagination serves is as
a testing ground for the violation of these norms: “what would happen if I…?” The
results of this thought experiment can be helpful in guiding present behaviour and can
also—in imaginal and virtual worlds—have the therapeutic effect of allowing a person
to stretch beyond usual socially normative behavioural restrictions. This need not
necessarily be because they disagree with them, but sometimes in order to clarify their
boundaries and strengthen them [10]. Many forms of media are based upon this principle.

4. Conclusion

In designing therapeutic, training, and assessment environments—as well as

simulations more generally—a great deal of leverage can be gained by considering the
distinction between perceived and normative affordances. Adding virtual assets when
designing a simulation is typically much easier than facilitating detailed interaction with
them. However, when storytelling is engaged, it is possible to cast the possibilities for
interaction in such a way that interaction will not be expected. When engineered
successfully, perceived affordances in a virtual environment can be more numerous than
possible affordances, and through the use of narrative-led normative affordances, this
never needs to be found out by the participant.
For example, in a rich virtual scene that contains a river: with sufficient presence
there may feel like there is a possibility of getting wet (perceived affordance). However,
in actuality, there is no river and only the visual appearance of one—and no suitable
haptic system to simulate the feeling of interacting with water (possible affordance). If
the participant tries to interact with the water, presence will be broken, and if the river is
removed from the simulation, the benefits of its inclusion are lost. However, if there is a
suitable drop to the river, or if a guiderail is put in place in the simulation, then normative
affordances will contain the expectation of interaction with the river while still allowing
for the sense that it is in fact possible, maintaining the fidelity contract of participant
expectation [11]. This can be further enhanced when a mixed reality approach is taken,
with the digital guiderail mapped onto a physical one in real life as a convincer which
primes subsequent perception.
If their affordances match, then physical, virtual, and imaginal approaches are
functionally equivalent. And, from a narrative design perspective, not all possible
affordances need to match—just those that are actually engaged with.

References

[1] Von Uexküll J. A Stroll Through the Worlds of Animals and Men: A Picture Book of Invisible
Worlds. In: Schiller CH (ed) Instinctive Behavior: The Development of a Modern Concept. New
York, NY: International Universities Press, 1957, pp. 5–80.
[2] Nagel T. What is it like to be a bat? The philosophical review 1974; 83: 435–450.
[3] Gibson JJ. The ecological approach to visual perception: classic edition. New York, NY:
Psychology Press, 2014.
[4] Maslow AH. A theory of human motivation. Psychological Review 1943; 50: 370–396.
Leader et al./ Normative Affordances: Utilising the constraint of context-specific expectation in simulated
environments 25

[5] Leader JF. Mixed Reality Therapy: Presentation & Experience In Applied Psychology. Annual
Review of Cybertherapy and Telemedicine 2016 2016; 14: 215–218.
[6] [Foxall GR. Behavior analysis and consumer psychology. Journal of Economic Psychology 1994;
15: 5–91.
[7] Norman DA. The Psychology of Everyday Things. New York, NY: Basic, 1988.
[8] Gaver WW. Technology affordances. In: Proceedings of the SIGCHI conference on Human
factors in computing systems. New York, NY: Association for Computing Machinery, 1991, pp.
79–84.
[9] Dalgarno B, Lee MJ. What are the learning affordances of 3‐D virtual environments? British
Journal of Educational Technology 2010; 41: 10–32.
[10] Cialdini RB, Trost MR. Social influence: Social norms, conformity and compliance. In: Gilbert
DT, Fiske ST, Lindzey G (eds) The handbook of social psychology. New York, NY: McGraw-
Hill, 1998, pp. 151–192.
[11] Leader JF. Mixed Reality Therapy Clinic Design. In: 2018 IEEE Games, Entertainment, Media
Conference (GEM). IEEE, 2018, pp. 1–9.
.
SECTION III

EVALUATION STUDIES

To date, some cybertherapy applications

have improved the quality of health care, and
later they will probably lead to substantial cost
savings.

However, cybertherapy is not simply a

technology but a complex technological and
relational process.

In this sense, clinicians and health care

providers that want to successfully exploit cy-
bertherapy need to give significant attention
to clinical issues, technology, ergonomics, hu-
man factors, and organizational changes in the
structure of the relevant health service.

Wiederhold & Riva, 2004

 Annual Review of Cybertherapy and Telemedicine 2021 29

Empathic interactions in online treatment:

experiences of mental healthcare
practitioners
Milou A. FEIJT a1, Yvonne A.W. DE KORT a, Joyce H.D.M. WESTERINK a,b,
and Wijnand A. IJSSELSTEIJNa
a
Eindhoven University of Technology, Eindhoven, Netherlands
b
Philips Research, Eindhoven, Netherlands

Abstract. Studies on practitioners' perceptions of eMental Health repeatedly report

experienced difficulties in establishing empathic interactions when communicating
with their clients online – a fact that plays an important role in the adoption, or lack
thereof, of technology-mediated treatments. Especially now, in times when social
distancing and lockdown measures are effectuated globally to combat the COVID-
19 pandemic, the need for high-quality online formats of mental health care has
become more urgent than ever. To increase our understanding of the process of
achieving empathic interactions in technology-mediated communication, the
current paper presents an online survey study using open questions on practitioners'
experiences of establishing empathy in online therapeutic interactions (n = 363).
Responses were analyzed using thematic analysis. From the derived themes, a
conceptual model was built showing several properties of online communication
that influence empathy and ways in which these properties affect the therapeutic
interaction on emotional, conversational, and relational levels. Additionally, our
findings highlight the behavioral strategies practitioners employ to find
workarounds to manage these effects and attempt to (re)connect with their clients.
Our findings thus provide experience-based insights into the process of building
empathy in online treatments – insights that can be used to improve current and
future eMental Health technologies, enabling practitioners to reach the desired level
of empathic understanding during remote therapeutic interactions.

Keywords. eMental Health, Online Psychological Treatment, Empathic Interaction,

Technology-Mediated Communication

1. Introduction

A good empathic interaction and therapeutic alliance are considered important

factors to successful therapeutic outcomes [1]. Even though quantitative studies on
eMental Health show comparable levels of therapeutic alliance and empathy for face-to-
face and online psychological treatments, at least from the clients' perspective [2], reports
of mental healthcare professionals repeatedly indicate that online communication
compromises empathic interactions [3]. These concerns about relational aspects were
found to be among the most prominent obstacles for professionals in their adoption of
eMental Health [3, 4]. In order for online treatment to offer a better experience for
practitioners and clients alike, it is important to specifically focus on technology-
mediated barriers to empathy in online interactions. As a first step to achieve this, we
need a deeper understanding of the empathic process in online therapeutic interactions
and how this is experienced from the perspective of mental healthcare professionals.

1
Corresponding Author:[email protected]
Fejitet al. / Empathic interactions in online treatment: experiences of mental healthcare practitioners
30

Empathy has been defined in a multitude of ways, reflecting the variety of fields in
which this concept has been researched [1]. Despite the differences, there seems to be a
general agreement that empathy refers to an interpersonal process of understanding what
another person is experiencing or trying to express, where empathy consists of three
components: cognitive (i.e., recognizing others' emotions), affective (i.e., emotional
convergence), and behavioral (i.e., emotional responding) [5]. In the field of mental
healthcare, Barret-Lennard's Empathy Cycle [6] describes empathy in therapeutic
interactions as repeated cycles of therapists attuning to expressed emotions and
experiences of clients, followed by empathic responses of therapists that are then
received and responded to by clients. The success of these cycles heavily relies on the
availability of (non-)verbal cues and one's ability to accurately use this information.
In their comprehensive framework, Grondin et al. [7] describe several ways in which
technology-mediated communication affects empathy in online therapeutic interactions.
Their main argument is that technology-mediated communication filters out important
non-verbal cues which constrains the empathic cycle described above, in line with the
so-called “cues filtered out” theories in the field of computer-mediated communication
[8]. While the framework provides helpful starting points towards a deeper
understanding of practitioners' difficulties with online psychotherapy, it primarily
focuses on the negative consequences of technology-mediated communication, whereas
technology also provides unique opportunities; options that are unavailable in face-to-
face interactions [9]. Moreover, some scholars argue that online communication can
actually be more personal and intimate, as people may feel more at ease to self-disclose
in relatively anonymous settings (e.g., [10]).
Despite the potential benefits that online psychological treatment offers, such as
increased accessibility, convenience, and autonomy of clients [3, 11], the perceived
shortcomings in non-verbal communication and empathic attunement keep its potential
from being fulfilled. In addition, the COVID-19 pandemic during Spring 2020 and its
corresponding social distancing measures forced many professionals to transfer their
face-to-face treatments to online means [4, 12], further stressing the need for developing
high-quality remote care. Because, until recently, online treatments were not part of
standard clinical practice, it was challenging to probe the experiences of a broad range
of practitioners and thoroughly explore the source of the reported difficulties. However,
the dramatic increase in online therapy delivery during the COVID-19 pandemic enabled
us to probe a much broader and potentially more representative sample of healthcare
professionals on what exactly practitioners are missing when attempting to establish an
empathic interaction online, and on which part of the empathic process they find to be
most constrained. With this work, we aim to contribute to improving the quality of
remote mental health care by elucidating the most important factors in establishing
empathy online and how these affect the therapeutic interaction.

2. Methods

A sample of 363 mental healthcare professionals (74% female, ages ranging from
18 to 70 years, M = 39.1, SD = 11.5) participated in an online survey. Most frequent
professions were clinical or counselling psychologists (37%), psychiatric nurses (33%),
and social workers (22%). Data was collected from June to September 2020, in the period
after the first lockdown due to the COVID-19 pandemic. The entire survey consisted of
33 questions regarding attitudes, skills, knowledge, and use of eMental Health. It
included two open-ended items on practitioners' experiences in establishing an empathic
online interaction with their clients, which is the focus of the current paper: 1) Could you
indicate to what extent you are (not) able to establish an empathic interaction?; 2) Which
strategies or information do you use to communicate empathically through eHealth?
These items were answered by 350 and 336 respondents respectively, with an average of
18 words per response. The textual responses were analyzed using thematic analysis
through the following steps: familiarization with the data, generating initial codes,
extracting, reviewing, and defining themes, and drawing up the results [13].
 Fejitet al. / Empathic interactions in online treatment: experiences of mental healthcare practitioners 31

3. Results

Sixteen (sub)themes were derived from the responses to the open-ended questions.
These could be divided into four main themes: technology-mediated communication
properties, therapeutic context factors, effects on the therapeutic interaction, and
behavioral strategies, along with twelve subthemes. Furthermore, the expressions of
practitioners indicated how these themes could be related to each other: e.g., "Because
the non-verbal information is lacking, you sometimes miss important parts of the
communication. Therefore, it is harder to reach a deeper emotional level." Based on this,
we integrated these (sub)themes and their interrelations into a conceptual model (see
Figure 1).

Figure 1 Conceptual model of the derived (sub)themes and relations of practitioners' experiences.
The most frequently reported properties of technology-mediated communication
that influence online empathic interactions were technological problems, limited access
to non-verbal cues ("There is literally an increased distance to the client: body language
and facial expressions are not, or much less, perceptible for both parties."), and absence
of situated interaction rituals and social conventions ("The beginning and closure of an
appointment feel incomplete, not being able to give a handshake, walk to the consultation
room together, or offer coffee."). The extent to which these properties affected the
empathic interaction seemed to be influenced by the therapeutic context: client
characteristics, previous (face-to-face) contact, and the topic of therapy ("It goes well
when you discuss more superficial topics, but when it concerns very emotional issues,
discussing them via a screen does not work."). The combination of these factors then
affected the therapeutic interaction on three levels: 1) emotional: it is harder to stay
emotionally attuned with the client ("When you are communicating in real life, it is
easier to feel what something does to someone. This is much more difficult through online
contact"); 2) relational: difficulties to support clients and stay connected; 3)
conversational: communication is more explicit and verbal ("I have to verbalize much
more of what I see and feel, instead of just 'being there"). To manage these effects,
professionals reported to apply various behavioral strategies to establish an empathic
interaction: use more verbalizing techniques ("Ask explicitly after emotions and
experiences and check my observations more often."), exaggerate non-verbal cues ("Use
clearer facial expressions, hum louder than I would normally do in a face-to-face
interaction"), and adjust the therapeutic setting.

4. Discussion

The current study focused on practitioners' perceptions of establishing empathic

interactions during online psychological treatments. By probing actual experiences with
online treatment from a large representative sample, we obtained reports emerging
directly from practice, thus addressing the how and why of these perceptions. According
to practitioners, various properties of technology-mediated communication affect the
therapeutic interaction on the emotional, relational, and conversational level, although
the exact effects depend on the therapeutic context. Remote therapy is especially
complicated in regard to severe mental health issues, emotionally disturbing topics, and
the absence of previous face-to-face contact. Practitioners adopt various strategies to
manage these effects, primarily by being more verbally explicit in their communication.
Fejit et al. / Empathic interactions in online treatment: experiences of mental healthcare practitioners
32

Our findings support the perceptions that practitioners expressed in earlier studies,
before they had to use eMental Health regularly [3, 4]. That is, it seems that their general
expectation that technology-mediated communication would impede the establishment
of an empathic interaction is something they indeed experienced when providing remote
psychotherapy. Therein, it has to be noted that there were large differences between the
respondents: though most participants indicated struggling with online interactions,
others experienced little or even no difference between face-to-face or online modalities,
in line with earlier findings that the experienced drivers and barriers differ depending on
their level of adoption of eMental Health [11]. Analyses of the quantitative survey data
are out of scope for this paper and will be reported elsewhere, but these individual
differences should be kept in mind while interpreting the results. In the current study, it
has become more explicit which parts of the process are specifically considered
troublesome. Our results suggest that particularly the affective and behavioral
components of empathy are affected in technology-mediated communication, reflected
by the difficulties to emotionally tune in with clients and provide comfort and support.
To compensate for this, therapists seem to adopt a more rational communication style by
explicitly formulating their observations and conveying their empathic responses
through verbal statements instead of non-verbal expressions, such as softly humming or
smiling comfortingly. These results are similar to findings of a study that compared
different types of empathy in non-therapeutic digital settings [14].
The presented conceptual model is, to our knowledge, the first to provide links
between characteristics of technology-mediated communication and the therapeutic
context, experienced effects on the therapeutic interaction, and practitioners' behaviors.
Based on our insights, we can derive directions for the development of solutions that can
address practitioners' needs more precisely and facilitate them in achieving desired levels
of empathy in online psychological treatments. One approach is to focus on the
technologies either by developing tools that compensate for the experienced lack of non-
verbal cues, such as eye-gaze correction technologies [15], or by examining how
technologies could bring a unique added value, for example using physiological
feedback to extend the gamut of social and affective cues (cf. [9, 16]). Solutions could
also be sought in enhancing therapists' skills as research indicates that technology-
mediated therapy requires other techniques than face-to-face treatment [17], while only
a minority of practitioners has received some form of training in providing online therapy
[4].
By focusing specifically on the empathic interaction and gauging experience-based
accounts from a broad, representative sample, the current study provides a more detailed
and multi-faceted view on practitioners' lived experience of online empathy and their
workarounds to enable the delivery of effective care. Hopefully, the gathered insights
and directions that come forth from this study can be utilized to improve technology-
mediated therapeutic interactions and contribute to eMental Health becoming a full-
fledged mode of practicing mental healthcare for both professionals and clients.

References

[1] Elliott R, Bohart AC, Watson JC, Murphy D. Therapist empathy and client outcome: An updated
meta-analysis. Psychotherapy. 2018;55(4):399–410.
[2] Berger T. The therapeutic alliance in internet interventions: A narrative review and suggestions for
future research. Psychother Res. 2017;27(5):511–24.
[3] Connolly SL, Miller CJ, Lindsay JA, Bauer MS. A systematic review of providers’ attitudes toward
telemental health via videoconferencing. Clin Psychol Sci Pract. 2020 Jun;27(2):e12311.
[4] De Witte NAJ, Carlbring P, Etzelmueller A, Nordgreen T, Karekla M, Haddouk L, et al. Online
consultations in mental healthcare during the COVID-19 outbreak: An international survey study
on professionals’ motivations and perceived barriers. Internet Interv. 2021 Sep;25:100405.
[5] Decety J, Jackson PL. The functional architecture of human empathy. Behav Cogn Neurosci Rev.
2004 Jun 18;3(2):71–100.
[6] Barrett-Lennard GT. The empathy cycle: Refinement of a nuclear concept. J Couns Psychol.
1981;28(2):91–100.
[7] Grondin F, Lomanowska AM, Jackson PL. Empathy in computer-mediated interactions: A
conceptual framework for research and clinical practice. Clin Psychol Sci Pract. 2019 Dec;26(4):1–
17.
[8] Walther JB, Parks MR. Cues filtered out, cues filtered in: Computer-mediated communication and
relationships. In: Knapp ML, Daly JA, editors. Handbook of interpersonal communication. 3rd ed.
Thousand Oaks, CA: SAGE; 2002. p. 529–63.
Fejitet al. / Empathic interactions in online treatment: experiences of mental healthcare practitioners 33

[9] Comer JS, Timmons AC. The other side of the coin: Computer-mediated interactions may afford
opportunities for enhanced empathy in clinical practice. Clin Psychol Sci Pract. 2019;26(4):e12308.
[10] Walther JB. Computer-Mediated Communication. Communic Res. 1996 Feb 29;23(1):3–43.
[11] Feijt MA, De Kort YAW, Bongers IMB, IJsselsteijn WA. Perceived drivers and barriers to the
adoption of eMental health by psychologists: The construction of the levels of adoption of eMental
health model. J Med Internet Res. 2018 Apr;20(4):e153.
[12] Feijt MA, De Kort YAW, Bongers IMB, Bierbooms JJPA, Westerink JHDM, IJsselsteijn WA.
Mental health care goes online: Practitioners’ experiences of providing mental health care during
the COVID-19 pandemic. Cyberpsychology, Behav Soc Netw. 2020 Dec 1;23(12):860–4.
[13] Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol. 2006;3(2):77–101.
[14] Powell PA, Roberts J. Situational determinants of cognitive, affective, and compassionate empathy
in naturalistic digital interactions. Comput Human Behav. 2017;68:137–48.
[15] Grondin F, Lomanowska AM, Békés V, Jackson PL. A methodology to improve eye contact in
telepsychotherapy via videoconferencing with considerations for psychological distance. Couns
Psychol Q. 2020 Jun 21;1–14.
[16] Feijt MA, De Kort YAW, Westerink JHDM, IJsselsteijn WA. Enhancing empathic interactions in
mental health care: Opportunities offered through social interaction technologies. Annu Rev
CyberTherapy Telemed. 2018;25–31.
[17] Williams R, Bambling M, King R, Abbott Q. In-session processes in online counselling with young
people: An exploratory approach. Couns Psychother Res. 2009 Jun;9(2):93–100.
 Annual Review of Cybertherapy and Telemedicine 2021 35

Further Validation of Russian Video Game

Addiction Scale (VGAS)
Vitalii EPISHINa, Nataliya BOGACHEVAa,1 and Diana MEDAKOVSKAYA a
a
I.M. Sechenov First Moscow State Medical University (Sechenov University), Russia

Abstract. The current study re-examines the psychometric properties of the Russian
Video Game Addiction Scale (VGAS). A new sample of video game players (N =
361; 89.2% male; aged 16-55) was added to the existing data pool (N = 515; 74.6%
male; aged 16-56). Previously found 7-factor structure of VGAS was confirmed by
principal component analysis with varimax rotation, with one highly cross-loaded
item excluded. The alpha-values confirmed VGAS’ good internal consistency.
Three sub-groups based on game-genre preferences were identified in the sample:
shooter players (N = 125), RPG players (N = 104), and video game players who
preferred other genres (N = 132). Shooter and RPG video game players had higher
gaming addiction compared to the “other genres” group. The Dark Triad traits were
measured by the Dirty Dozen questionnaire. Machiavellianism positively correlated
with gaming disorder in all groups except for RPG video game players. Psychopathy
only correlated with gaming addiction in the “other genre” group. Those results
matched the existing data, indirectly supporting VGAS’ construct validity as a new
gaming addiction measurement.

Keywords. Gaming Disorder, Video Game Addiction, Video Game Players, Dark
Triad

1. Introduction

The upcoming 11th revision of the International Classification of Diseases (ICD-11)

introduces Gaming disorder (6C51) as a specific diagnosis under the “disorders due to
addictive behaviors” category. The World Health Organization (WHO) describes
gaming disorder as a pattern of online or offline digital-gaming behavior, characterized
by impaired control over gaming, video games’ increased priority over other activities,
and continuation or escalation of gaming despite negative consequences, accompanied
by significant impairment in social, professional, or educational life [1]. While there are
still ongoing scientific debates over the decision to include gaming disorder in ICD-11
(e.g. [2]), it becomes increasingly important to introduce adequate diagnostic tools to
assess and measure the severity of this condition based on diagnostic criteria proposed
by WHO. While there are several relatively well-known gaming addiction scales, almost
none were ever properly adapted on a Russian sample. Additionally, most of those scales
are based on criteria different from that proposed by WHO (e.g., IGD-20 is based on
nine Internet Gaming Disorder criteria from DSM-5 [3]). At the same time, in some
countries, Russia included, the use of ICD among clinical practitioners is more prevalent.
Thus, we aimed to create a Russian Gaming Disorder questionnaire based on ICD-11
criteria that could be useful for Russian psychologists and clinicians as a research tool
and screening instrument. The first version of Video Game Addiction Scale (VGAS)
included 26 items developed through several stages. First, ten clinical psychology
students belonging to a cyberpsychology-related scientific project were tasked to
produce 20 or more statements each, describing a typical person with gaming addiction.
From the total list of 217 non-repetitive items, 57 most representative ones were
extracted under professional psychologists’ supervision
.

1
Corresponding Author: [email protected].
Epishin et al. / Further Validation of Russian Video Game Addiction Scale (VGAS)
36

An expert group of four psychologists evaluated the list and shortened it to 28 items,
most related to ICD-11 criteria. Two more items were excluded after principal
component analysis due to their low factor loads. This 26-item questionnaire version
showed good internal consistency and significant correlations withinternet addiction,
gaming motivations, and various psychopathological symptoms, indicating its construct
validity (on a sample of 515 participants) [4]. This particular study aimed to explore
VGAS’ psychometric qualities further, and to learn more about video game addiction
concerning specific video game genres such as shooters and RPGs (found to be more
addictive by some studies, e.g. [5]). Recent studies also suggest that the Dark Triad
personality traits play a part in various addictive behaviors including internet and gaming
addiction [6,7]. Thus, we included the Dark Triad traits in the further analysis of our
questionnaire’s construct validity.

2. Method

2.1. Sample

The new sample included 361 regular or occasional video game players (89.2% male)
aged 16-55 (M = 20.79; SD = 4.4; Me = 2) reporting to prefer different video game
genres: shooters (N = 125), RPGs (N = 104), or other games such as strategy, racing, etc.
(N = 132). The questionnaire battery included VGAS (26-item version), and The Dark
Triad Dirty Dozen [8], adapted into Russian by Kornilova et al. [9].
Additionally, for further factor structure and reliability evaluation, the data from our
previous sample was also included. This sample consisted of 515 regular or occasional
video game players (74.6% male) aged 16-56 (M = 24.97; SD = 6.46; Me = 24). For this
study, we only used VGAS data.
All the data in both studies was gathered online. Participation in the research was
anonymous and voluntary, with an informed consent form provided to all the participants.

2.2. Questionnaires

Video Game Addiction Scale (VGAS) is a new Russian gaming disorder assessment
questionnaire, currently in the process of multi-stage validation and psychometric
qualities evaluation. The previous stages of VGAS development were briefly described
in the introduction section of this article and the previous study [4].
This version of VGAS consists of 26-items, with responses rated on a 5-point Likert
scale from 1-strongly disagree to 5-strongly agree. The questionnaire has a 7-factor
structure, which we aimed to check and correct in this study and provides 8
measurements.
The overall VGAS score represents the severity of video game addiction symptoms
in general. At the same time, its individual factors allow us to evaluate symptoms
according to ICD-11 criteria, including various aspects of impaired control over gaming
(factors 2 and 3 – behavioral control over gaming and problems with time-management
respectively), video games’ increased priority over other activities (factor 4 assessed
gaming priority directly while factors 1 and 5 represent emotional aspects such as gaming
being the primary source of positive emotions and gamer’s emotional involvement in
video games), continuation of gaming despite negative consequences represented by
conflicts with others (factor 6), and neglect of one’s health and well-being (factor 7).
The Dark Triad Dirty Dozen (DTDD) [8] is a brief personality inventory used to
assess the so-called “dark triad traits” (named so due to their social undesirability and
connection with possible malevolent behaviors) – Machiavellianism, narcissism, and
psychopathy – in adults without clinically diagnosed personality disorders.
The questionnaire consists of 12 items (as implied by its name), each subscale
including 4 items. In the original DTDD, the responses are based on a 7-point Likert
scale, while the Russian modification [9] uses a 5-point Likert scale instead but keeps
the original structure otherwise.
The measurements obtained through this questionnaire are:
 Epishin et al. / Further Validation of Russian Video Game Addiction Scale (VGAS) 37

Machiavellianism as cold-hearted manipulativeness, exploitative with indifference

to morality and other people’s interests
Narcissism as the sense of grandiosity, self-importance, and superiority
Psychopathy as the lack of empathy, accompanied by high levels of impulsivity and
recklessness
All data was analyzed by using IBM SPSS Statistics for Windows (Version 22.0).

3. Results

3.1. VGAS Factor Structure and Reliability After the Second Study

In the previous study, principal component analysis (PCA) with varimax rotation
suggested a seven factor structure for VGAS. The total percentage of the explained
variance for this solution was 52.53%. To further evaluate VGAS reliability and factor
structure, we merged new sample data (N = 361) with the previously existing data pool
(N = 515), resulting in a total sample of 876. PCA with varimax rotation was again used
to identify groups of intercorrelated variables. The same seven factors with the same
items were obtained, yet one item previously related to factor 2 was excluded due to high
cross loads with factor 4.
The percentage of the explained variance for this solution was 56.29%. Thus, we
updated VGAS to consist of 25-items in seven factors. The names of the factors were
chosen based on their items:
• Factor 1: Positive emotions due to gaming (e.g., translated from Russian:
“My gaming life is more emotional than my life in reality”)
• Factor 2: Impaired gaming control (e.g., “I will continue playing video
games even if I have important unfinished tasks”)
• Factor 3: Problems with time-management while gaming (e.g., “I usually
play longer than planned”)
• Factor 4: Gaming priority (e.g., “There are few things other than gaming
that interest me”)
• Factor 5: Emotional engagement in gaming (e.g., “I tend to rage/panic
when I cannot control the in-game situation”)
• Factor 6: Conflicts with others due to gaming (e.g., “I often argue with my
family because of my gaming habits”)
• Factor 7: Gaming regardless of own well-being (e.g., “I continue playing
even if I feel unwell”)
Table 1 includes Cronbach`s α-values (internal consistency) of the whole scale and
different factors in the previous study, the current merged sample, and the number of
items in previous and current versions of the questionnaire.
Table 1. Reliability test (Cronbach’s α) for VGAS and its seven factors (F.1-7): previous and current samples.
VGAS F.1 F.2 F.3 F.4 F.5 F.6 F.7
Previous sample, N=515 0.856 0.782 0.674 0.584 0.649 0.552 0.635 0.435
Total sample, N=879 0.866 0.808 0.668 0.568 0.631 0.616 0.667 0.456
Items before/now 26/25 7/7 4/3 3/3 3/3 3/3 3/3 3/3

As seen in Table 1, both versions of VGAS have good internal consistency for the whole
scale. Alpha-values of the factors were decent considering a few items in most of them,
with only small changes between the previous and the current studies.

3.2. Gaming Disorder in Relation to Video Game Genres and the Dark Triad

Welch`s ANOVA with Tamhane’s T2 test showed both shooter and RPG game
players to have significantly (p < 0.05) higher VGAS scores compared to the “other
genres” video game players group. Previous studies showed that MMORPG and shooter
game players were more likely to exhibit gaming disorder symptoms (based on DSM-5
criteria for Internet Gaming Disorder) [5]. A similar result in our sample, but with ICD-
11-based VGAS, suggests that gaming disorder might be game-genre specific and
Epishin et al. / Further Validation of Russian Video Game Addiction Scale (VGAS)
38

supports the external validity of our questionnaire. Next, Pearson`s correlations between
VGAS score and The Dark Triad were calculated; results are presented in Table 2.

Table 2. VGAS (general score) correlations with the Dark Triad scales (Pearson’s r), significant only.
Whole sample (N=361) Shooters (N=125) Other genres (N=132)
Machiavellianism 0.216 (p = 0.000) 0.233 (p = 0.009) 0.213 (p = 0.014)
Psychopathy - - 0.184 (p = 0.035)

Machiavellianism was the only Dark Triad trait that showed a significant positive
correlation with VGAS score on the whole sample. While some authors argue
Machiavellianism is the only Dark Triad trait not linked to substance-related and non-
substance-related addictions in general [6], it seems to be important in gaming and
internet addiction studies [7, 10]. Kircaburun and Griffiths [7] found that higher
Machiavellianism was associated with higher online gaming and gambling, among other
online activities.
The authors assumed people with a high Machiavellianism trait might use video
games to fulfill their need for competition and linger towards online communication to
avoid offline social rejection, leading to excessive gaming. Another study [10] also
showed direct effects of Machiavellianism and psychopathy on gaming disorder
symptoms’ severity, while escapism tendencies mediated narcissism’s influence in video
game players. Our current results partially match those findings and suggest some genre-
related specifics of the Dark Triad traits in relation to video game addiction, as RPG
video game players did not share the same correlations for shooter and “other genres”
video game players.

4. Conclusion

Results from the extended sample supported our previous conclusions about our new
ICD-11-based gaming disorder questionnaire VGAS and its 7-factor structure internal
consistency and validity. As a new questionnaire, VGAS will still require further
research and validation as a diagnostic instrument, specifically in comparison with other
existing gaming disorder questionnaires based on different models such as IGD-20 [3].
The links between gaming addiction and the Dark Triad also matched the pre-existing
research findings. Thus, we can conclude that VGAS performs similarly to other video
gaming addiction questionnaires, further supporting its validity.

References

[1] World Health Organization [Internet]. Geneva (Switzerland): World Health Organization; c2022.
Addictive behaviours: Gaming disorder; 2020 Oct 22 [cited 2022 Jan 26]; [about 2 screens]. Available
from: https://www.who.int/news-room/q-a-detail/addictive-behaviours-gaming-disorder
[2] Ferguson CJ, Colwell J. Lack of consensus among scholars on the issue of video game “addiction”.
Psychol Popular Media. 2020;9(3):359–366.
[3] Pontes HM, Király O, Demetrovics Z, Griffiths MD. The conceptualisation and measurement of DSM-5
Internet Gaming Disorder: the development of the IGD-20 Test. PLoS One. 2014 Oct 14;9(10):e110137.
doi: 10.1371/journal.pone.0110137. PMID: 25313515; PMCID: PMC4196957.
[4] Epishin V, Bogacheva N. Development of ICD-11-based Russian Video Games Addiction Scale. In: Riva
G, Wiederhold BK, editors. Abstracts from the 25th Anniversary International CyberPsychology,
CyberTherapy & Social Networking Conference; 2020 Jun 22-24; Milan, Italy. Larchmont (NY): Mary
Ann Liebert, Inc.; c2020. p.72 (Cyberpsychology, Behavior, and Social Networking; vol. 23, no. 10,
suppl.). Available from: https://www.liebertpub.com/doi/suppl/10.1089/cyber.2020.29198.cypsy
[5] Na E, Choi I, Lee TH, Lee H, Rho MJ, Cho H, Jung DJ, Kim DJ. The influence of game genre on Internet
gaming disorder. J Behav Addict. 2017 Jun 29;6(2):1-8. doi: 10.1556/2006.6.2017.033. Epub ahead of
print. PMID: 28658960; PMCID: PMC5520129.
[6] Jauk E, Dieterich R. Addiction and the Dark Triad of Personality. Front Psychiatry. 2019 Sep 17;10:662.
doi: 10.3389/fpsyt.2019.00662. PMID: 31607963; PMCID: PMC6757332.
[7] Kircaburun K, Griffiths MD. The dark side of internet: Preliminary evidence for the associations of dark
personality traits with specific online activities and problematic internet use. J Behav Addict. 2018 Dec
1;7(4):993-1003. doi: 10.1556/2006.7.2018.109. Epub 2018 Nov 14. PMID: 30427212; PMCID:
PMC6376394.
[8] Jonason PK, Webster GD. The dirty dozen: a concise measure of the dark triad. Psychol Assess. 2010
Jun;22(2):420-32. doi: 10.1037/a0019265. PMID: 20528068.
 Epishin et al. / Further Validation of Russian Video Game Addiction Scale (VGAS) 39

[9] Kornilova TV, Kornilov SA, Chumakova MA, Talmach MS. Metodika Diagnostiki Lichnostnyh Chert
“Temnoj Triady”: Aprobacija Oprosnika “Temnaja Djuzhina” [The Dark Triad personality traits
measure: Approbation of the Dirty Dozen Questionnaire]. Psihologicheskij zurnal. 2015;36(2):99-112.
Russian.
[10] Tang WY, Reer F, Quandt T. The interplay of gaming disorder, gaming motivations, and the dark triad.
J Behav Addict. 2020 Apr 4;9(2):491-496. doi: 10.1556/2006.2020.00013. PMID: 32544080.
 Annual Review of Cybertherapy and Telemedicine 2021 41

A Motion Controlled Virtual Reality

Paradigm for Ostracism Research
Pat MULVANEY1, Lauren CHRISTOPHERS, Brendan ROONEY
University College Dublin

Abstract. Despite being a popular field of study, ostracism research paradigms

have been criticised for their abstract nature and lack of mundane reality. The
current paper sets out to address the problem of ecological validity through the
introduction of a newly developed paradigm that makes use of virtual immersive
environments, motion controls, and haptic feedback. Based on the Cyberball-Game
(1), this VR tool places participants in a playground where they are asked to play a
game of catch with two agents. When passed to them, the player can pick up the ball
and throw it to whoever they choose using motion controls. After a period, the other
players will only pass between themselves, inducing the feeling of being ostracised
in the participant. A pilot study (n = 24) was run to assess the effectiveness of the
tool in inducing feelings of ostracism using the Basic Needs Scale (2). The sample
size was too small to report inferential statistics, but descriptive statistics
demonstrate that the tool was successful in inducing an experience of being
ostracised when the control (M = 72.46, SD = 10.64) and ostracism (M = 43.73, SD
= 11.1) conditions are compared. All participants were successfully able to use the
device with no errors. This tool presents significant advantages in comparison to
conventional methodology, with greater ecological validity resulting from improved
presence and embodiment. Greater experimental control and measurement is also
offered when compared to in vivo methods (3).

Keywords. Virtual Reality, Ostracism, Cyberball-Paradigm, Social Stress,

Narrative

1. Introduction

Ostracism is the act of leaving out and ignoring an individual or a group in a social
context. Being ostracised has been found to have significant and often long lasting
adverse psychological and/or physical consequences. The Cyberball-Game (1) is one of
the most widely used experimental methods for studying ostracism. Current ostracism
research paradigms have been criticised for lacking ecological validity due to their
abstract nature and lack of mundane reality (3). The tension between ecological validity
and the desire for experimental control is a long-established concern of psychology
research (3). Virtual reality offers potential to address this divide, allowing for excellent
customisation and detailed data recording while more closely resembling real life
scenarios.
The Cyberball paradigm has previously been adapted into immersive virtual
environments (IMVE) that make use of head mounted displays (HMD) (4). These tools
allow for greater control of proximity and characteristics of avatars but while this
program addressed some identified issues, it was still controlled using a keyboard.
Motion tracking has been identified to significantly contribute to overall immersion (5).
Recent research has also linked embodied and social cognition, with findings suggesting
that sensorimotor experiences can shape perceptions of social situations (6).
To account for the limitations of the previous tool, we developed an immersive
virtual reality tool based upon the Cyberball-Game (1) paradigm that makes use of
motion controls and haptic feedback.

1
Corresponding Author: [email protected]
 Mulvaney et al. / A Motion Controlled Virtual Reality Paradigm for Ostracism Research
42

This tool was developed with the intent of investigating the role of narrative in
ostracism outcomes.

2. Method/Tool

2.1. Design

The pilot study used an independent groups design with an ostracism and control
condition. Self-reported measures of basic needs fulfilment and distress were used as
dependent variables. Basic needs fulfilment was measured using the Basic Needs Scale
(2). Distress was measured using the Subjective Unit of Discomfort Scale (7).

2.2. Participants

Twenty-four participants, 20 of which were from a university campus, volunteered for

this study (17 female, 7 male, mean age = 27.167, SD = 9.187). Participants were
randomly assigned evenly across the two conditions.

2.3. Measures

Basic Human Needs: Participants completed the Basic Needs Scale (1). The 12-item
scale assesses the fulfilment level of four basic human needs. Each need has three items
assessed using a 9-point Likert scale (1-9) of level of agreement. The subscales include
(1) self-esteem, (2) belonging, (3) control, and (4) meaningful existence. This measure
covers the essential human needs that are impacted by the experience of ostracism as
identified in previous research (8).

Distress: The level of distress experienced by participants during use of the VR tool was
also assessed to determine its level of negative impact. The Subjective Unit of Discomfort
Scale (7) was used following the completion of the game. The scale has been found to
be a valid measure of both physical discomfort (7). Participants were asked to rate their
level of distress on a scale that ranged from 0 to 100, with 100 representing the highest
level of distress. In the present study, a visual scale with definitions for the number
values was used.

2.4 VR Cyberball

This tool was developed in Unity by the first author based on the Cyberball paradigm
(1). We designed the immersive environment ostracism experience to be ambiguous by
including no dialogue and neutral expressions, along with using simple cartoon models
for the players and environment (see Figure 1). This was chosen both to moderate
negative impact as well as allow for versatility in study topic/context.
The program situates the player in a bright playground with simple, low polygon
count (level of 3D model detail) objects (see Figure 2). Participants are seated during the
game with background park ambience noises playing through the headphones. When the
participant has the ball thrown to them, it stops a short distance in front of them. From
here they can grab the ball and throw it to whichever avatar they choose using the HTC
Vive’s motion controllers. Participants can throw from a seated or standing position. The
gender and number of avatars you can throw to can also be manipulated.
 Mulvaney et al. / A Motion Controlled Virtual Reality Paradigm for Ostracism Research 43

Figure 1. This image displays the VR tool in motion from the perspective of the participant.

Figure 2. This image displays the park environment the participant is placed in.

2.5 Procedure

Testing was conducted on a one-on-one basis in a room on a university campus. All

measures were filled out and recorded on an online Qualtrics form along with a basic
demographic survey of age, gender, and level of experience with VR. Participants were
not informed of the full nature of the experiment, instead only being told that the study
concerned embodied cognition. Once they had filled out basic demographic information,
they put on the Head-Mounted-Display (HTC Vive, HMD) and were given the chance
to adjust to VR and practice throwing balls at a target in a neutral environment with no
scenery. Participants were seated during all VR sections of the study.
This practice continued until they had successfully thrown the ball several times and
confirmed they were comfortable with the VR setup. This took roughly 5 minutes.
Following this, participants completed the VR playground experience. In all cases, for
the first minute of the game, the ball was evenly (33%) passed to through manipulated
randomisation. In the exclusion condition, following the first minute the other players no
longer passed to the participant for the remainder of the game.
 Mulvaney et al. / A Motion Controlled Virtual Reality Paradigm for Ostracism Research
44

At this point, the participant had no ability to interact with the other players, move
to another location, or get the ball back. In the control condition, the player was evenly
passed to for the full duration of the game. The program closed once 4 minutes had
passed from the first throw. The time of 4 minutes was chosen based on previous research
along with the results from Hartgerink et al.'s (2015) review suggesting that the length
of the game had no clear impact on the mean ostracism effect.
Immediately following the game closing, the participant removed the HMD and
completed a series of measures presented in a randomised order. The study in total lasted
roughly 30 minutes for each participant.

3. Results

While the sample size of the pilot study was too small to report inferential statistics,
descriptive statistics and personal accounts of the experience from the participants
demonstrated that the tool was successful in inducing an experience of being ostracised.
Participants in the control (M = 72.46, SD = 10.64) and ostracism (M = 43.73, SD =
11.1) conditions reported large differences in scores on the basic needs scale.
All participants were successfully able to use the device and tool with no display
or audio errors. The controls for picking up and throwing the ball were intuitive enough
that most participants (>80%) only needed a couple minutes of practice to become
comfortable with consistently throwing the ball. This was true even for participants with
little or no experience in VR, though experienced participants picked it up more quickly.
One participant reported audio issues, but this was determined to be a result of the PC’s
audio settings rather than the tool itself.

4. Discussion

The current paper aimed to build upon and improve traditional ostracism study
paradigms, particularly regarding ecological validity and abstraction. The findings of the
pilot suggest that VR Cyberball was successful in inducing feelings of ostracism.
This tool presents significant advantages in comparison to conventional
methodology, with greater ecological validity resulting from improved presence and
embodiment. Greater experimental control and measurement is also offered when
compared to in vivo methods [3]. When compared to previously developed VR tools,
our tool offers potential for greater immersion and embodiment in the scene thanks to
the act of throwing being physically mirrored with motion controls.
These improvements in turn can allow for ostracism research that better reflects
real life experiences. With ostracism outcomes recently being found to be largely
heterogenous in both the short and long term (9), there remains considerable ground to
be covered in understanding the factors that moderate the impact of social exclusion. The
use of this tool that allows for greater physical embodiment and immersion may lead to
better understanding of the source of individual differences in ostracism outcomes.
One relatively unexplored avenue of interest is the role of narrative. This was
highlighted by Richmand and Leary (10) whose model suggests that long term ostracism
outcomes are moderated by an individuals’ understanding of the event. Linked in with
this, existential meaning, one of the identified basic human needs (1), is closely tied with
personal narrative (11). This tool was developed to better study this topic.
To further validate this tool, as well as determine any differences in experience
compared to traditional paradigms, this method should be utilized in future ostracism
research.

References

[1] Williams KD, Jarvis B. Cyberball: A program for use in research on interpersonal ostracism and
acceptance. Behav Res Methods. 2006 Feb 1;38(1):174–80.
[2] Williams KD. Ostracism: the power of silence [Internet]. Guilford; 2001 [cited 2021 Apr 12].
Available from: https://researchers.mq.edu.au/en/publications/ostracism-the-power-of-silence
Mulvaney et al. / A Motion Controlled Virtual Reality Paradigm for Ostracism Research 45

[3] Parsons TD. Virtual Reality for Enhanced Ecological Validity and Experimental Control in the
Clinical, Affective and Social Neurosciences. Front Hum Neurosci [Internet]. 2015 [cited 2020 Dec
21];9. Available from: https://www.frontiersin.org/articles/10.3389/fnhum.2015.00660/full
[4] Kothgassner OD, Hlavacs H, Beutl L, Glenk LM, Palme R, Felnhofer A. Two Experimental Virtual
Paradigms for Stress Research: Developing Avatar-Based Approaches for Interpersonal and
Evaluative Stressors. In: Entertainment Computing - ICEC 2016 [Internet]. Springer, Cham; 2016
[cited 2021 Apr 12]. p. 51–62. Available from: https://link-springer-
com.ucd.idm.oclc.org/chapter/10.1007/978-3-319-46100-7_5
[5] Cummings JJ, Bailenson JN. How Immersive Is Enough? A Meta-Analysis of the Effect of
Immersive Technology on User Presence. Media Psychol. 2016 Apr 2;19(2):272–309.
[6] Zappa A, Bolger D, Pergandi J-M, Mallet P, Dubarry A-S, Mestre D, et al. Motor resonance during
linguistic processing as shown by EEG in a naturalistic VR environment. Brain Cogn. 2019 Aug
1;134:44–57.
[7] Tanner BA. Validity of global physical and emotional SUDS. Appl Psychophysiol Biofeedback.
2012 Mar;37(1):31–4.
[8] Zadro L, Williams KD, Richardson R. How low can you go? Ostracism by a computer is sufficient
to lower self-reported levels of belonging, control, self-esteem, and meaningful existence. J Exp
Soc Psychol. 2004 Jul 1;40(4):560–7.
[9] Hartgerink CHJ, van Beest I, Wicherts JM, Williams KD. The Ordinal Effects of Ostracism: A
Meta-Analysis of 120 Cyberball Studies. Van Yperen NW, editor. PLOS ONE. 2015 May
29;10(5):e0127002.
[10] Richman LS, Leary MR. Reactions to discrimination, stigmatization, ostracism, and other forms of
interpersonal rejection: A multimotive model. Psychol Rev. 2009;116(2):365–83.
[11] Daiute C. Narrating Possibility. In: Cultural Psychology of Education. 2015. p. 157–72.
 Annual Review of Cybertherapy and Telemedicine 2021 47

A qualitative and quantitative virtual reality

usability study for the early assessment of
ASD children.
Maria Eleonora MINISSIa 1, Irene Alice CHICCHI GIGLIOLIa, Fabrizia
MANTOVANIb, Marian SIRERAc, Luis ABADc, Mariano ALCAÑIZa
a
Instituto de Investigación e Innovación en Bioingeniería (i3B), Universitat
Politécnica de Valencia, Valencia, Spain
b
Centre for Studies in Communication Sciences “Luigi Anolli” (CESCOM) -
Department of Human Sciences for Education ‘‘Riccardo Massa’’, University of
Milano - Bicocca, Milan, Italy
c
Centro de Desarrollo Cognitivo Red Cenit, Valencia, Spain

Abstract. The diagnosis of autism spectrum disorder (ASD) is usually done using
structured and semi-structured interviews directed to children and caregivers. These
procedures are administered by certified clinicians who have expertise in the
assessment of ASD. However, on one side, semi-structured procedures addressed to
children are usually administered in settings requiring ecological validity such as
the laboratory; on the other side, structured interviews to caregivers rely on self-
report that might be affected by psychological response biases. There is the need to
fulfil aforementioned needs, improving ASD assessment procedures through the use
of both ecological settings and objective measures. The present study aims to
investigate the usability of a novel procedure to assess ASD based on virtual reality
(VR) and quantitative measures. 20 children with ASD and 20 children with typical
development (TD) performed four basic tasks in the VR system Cave Assisted
Virtual Environment (CAVE™) while an examiner analysed the usability of the
application as well as children’s user experience. Quantitative behavioural variables
related to children’s performance across tasks were measured. Included tasks
required children to interact in the virtual environment with childlike objects.
Findings demonstrated that VR application was promising for the assessment of
ASD due to good usability in three tasks out of four and positive user experience.
Moreover, quantitative behavioural outcomes revealed differences between groups
on time spent playing and accuracy across tasks. Quantitative and qualitative
usability studies improve the effectiveness of new objective and technology-based
ASD assessment procedures, in particular when children represent the population
target.

Keywords. Autism Spectrum Disorder, Virtual Reality, Usability, Assessment

1. Introduction

Autism spectrum disorder (ASD) is a neurodevelopmental disorder affecting social

skills and interaction, characterized by the presence of restrictive and repetitive patterns
of behaviors and interests [1]. Currently, the ASD assessment occurs in clinical settings
using the Autism Diagnostic Observation Schedule (ADOS) [2] and the Autism
Diagnostic Interview-Revised (ADI-R) [3]. The former utilizes semi-structured
observational tasks aimed to qualitatively evaluate children’s behaviors in different
situations; the latter is a semi-structured interview for caregivers concerning children’s
behaviors in daily life. Despite these procedures being considered the gold standard in
ASD assessment [4], they present some limitations that might affect the objectivity of
the assessment process [5-7].

In particular, ADOS usually takes place in non-ecological settings that are far from
representing real-life environments, while ADI-R is based on caregivers’ reports about
children’s habits and behaviors that may be biased by subjectivity and further
psychological tendencies such as social desirability. To fulfil

1
Corresponding Author: [email protected]
Minissi et al. / A qualitative and quantitative virtual reality usability study for the early assessment of ASD
children.ù
48

aforementioned needs in the current ASD assessment, new objective technology-

based procedures are being developed and tested. In particular, virtual reality (VR) has
the potential to overcome the need for ecological validity in ASD assessment [8]. VR
can reproduce every-day life situations in controlled settings, providing users the sense
of presence as if it was the real world. Moreover, objective measures related to users’
behaviors that are recorded by the VR system can improve assessment procedures,
supplying quantitative behavioral biomarkers of ASD. In the present study, the usability
as well as the user experience in a novel VR procedure to assess ASD early on were
tested. In addition, behavioral differences between ASD and typical developmental (TD)
children were assessed to initially test the procedure’s ability to disentangle ASD. The
Cave Assisted Virtual Environment (CAVE™) was chosen as the VR system due to the
non-intrusiveness and suitability for ASD children [5-6]. The chosen virtual environment
(VE) represented the VE skeleton of a multimodal VR procedure that will be tested in
the near future in attempts to foster the early and objective assessment of ASD.

2. Method

2.1. Participants

Forty children with an age range between 3 and 7 years old were enrolled in the
study. ASD children were recruited from the Development Neurocognitive Centre Red
Cenit in Valencia, Spain, whereas TD children were found through study promotion on
social media. Twenty ASD children (age in months = 53.44 ± 13.22; males = 16, females
= 4) and 20 TD children (age in months = 59.40 ± 11.95; males = 10, females = 10) took
part in the study. ASD children were diagnosed by ADOS-2, and caregivers provided
children’s assessment reports to participate in the study. ADOS-2 mean scores of ASD
children were 12.75 ± 3.68 in social affect, 2.50 ± 1.32 in restricted and repetitive
behaviors, and 15.12 ± 4.05 in total score. TD children’s caregivers answered a short ad
hoc developed questionnaire to ensure the absence of cognitive impairments related to
neurodevelopmental disorders. Before the study, the caregivers were informed about the
procedure and they gave written consent for the participation of children. Study was
approved by the Ethical Committee of the Polytechnic University of Valencia.

2.2. Procedure

ASD and TD children were asked to perform four basic tasks in the VR system
CAVE™. These tasks required them to perform basic movements, which allowed the
interaction in the VE, as well as in the multimodal VR procedure that will be tested in
the near future. The present VE was developed in the Institute for Research and
Innovation in Bioengineering (i3B) at the Polytechnic University of Valencia. The VE
was projected on one surface of the three-surface CAVETM with dimensions of 4 m x 4
m x 3 m. The CAVETM was characterized by three ultra-short lens projectors in the
ceiling, which can project 100° images at 55 cm of distance. Interaction in the VE was
provided by the Azure Kinect DK (Microsoft Corporation, 2019), with depth camera in
resolution mode 640 x 576 at 30 frames per second. The camera depth of field allowed
the user's body tracking over the entire room. The camera was set on a 40 cm high tripod
in front of the central surface of the CAVE™ where the VE was projected. The camera
did not interfere with the user's vision of the VE. Users interacted in the VE through a
non-filled virtual human shape which moved accordingly, mirroring movements of head,
trunk, and limbs. To facilitate the personal identification with the avatar and the
immersion in the VE, at the beginning of the experience users choose between male and
female human shape.
 Minissi et al. / A qualitative and quantitative virtual reality usability study for the early assessment of ASD
children. 49

Figure 1: Screen captions of the four tasks. a. FT; b. BT; c. KT; d. HT.

Participants had to perform four basic tasks requiring basic body movements with
the purpose of interacting and taking virtual actions. At the beginning of each task, the
experimenter instructed children on the goal of the task using basic and standardized
sentences. When participants did not understand task requirements, more in-depth
instructions were provided. Each task was repeated twice, and task order was randomized.
The four tasks were developed to be engaging for children, involving childlike and
colorful objects. In the flower task (FT; see Figure 1a) participants had to pick up a
flower and move it rightward so as to leave it on a bench, repeating the action five times.
In the bubble task (BT; see Figure 1b) children had to move limbs in order to touch thirty
colored bubbles falling down in pairs. Users touched them so as to make them explode.
The first ten bubbles fell down slowly (i.e., slow bubbles), the second ten fell down more
quickly (i.e., moderate bubbles), and the last ten bubbles fell down rapidly (i.e., rapid
bubbles). The kick task (KT; see Figure 1c) required moving the lower limbs in order to
kick a ball presented on the virtual floor five times. The ball appeared red and then turned
green to avoid unintentional kicks. Finally, the hand task (HT; see Figure 1d) required
users to guide a virtual hand in the VE by moving their hand to select three buttons
representing toys.
Usability and user experience of the application were qualitatively assessed by an
expert evaluator who observed children’s behavior across tasks, assigning usability
scores. Children also had to say which tasks they liked most. Finally, the behavioral
performance between groups (i.e., mean time and accuracy) was quantitatively tested.

3. Results

3.1. Qualitative results

The qualitative analysis of children’s interaction in the VE reported effective

usability and good user experience in the majority of tasks, especially in trial 2, where
both groups already learned how to interact. KT and BT were the most enjoyed tasks,
due to intuitive interaction, good usability, and entertainment. FT was perceived as more
challenging by the ASD group since it required more cognitive effort than BT and KT,
and it did not present the same level of entertainment. Finally, HT was the most
challenging task for both groups and it was the least liked due to complex usability
related to technical features. The hard usability in HT yielded bad user experience. TD
children nonetheless were more patient than ASD children in performing the task, and
they manifested less frustration over performance.

3.2. Quantitative results

Data analysis was performed using SPSS Statistics 22 (IBM, 2018). Outliers in age
were checked with the 3 interquartile range method and no subject was excluded from
the analysis. Normality assumption was assessed by Shapiro-Wilk’s test (p > .05), and
homogeneity assumption was tested by Levene's test (p > .05). One-way ANOVA was
used to assess differences between groups on mean time and accuracy in both task trials.
Whether assumptions were violated, the Kruskal-Wallis rank-based non-parametric test
was conducted.
 Group participants were the same age (F(1, 38) = 2.268; p = .140). In trial 1 of KT,
participants spent the same amount of time kicking each ball (p > .05), whereas in trial
2, ASD children were slower than TD children (χ2(1) = 4.093, p = .043; η2 = .084).
Regarding accuracy, participants kicked the same number of balls in both trials (p > .05).
In BT, ASD children were slower than TD children exploding slow bubbles in trial 1
(χ2(1) = 9.677, p = .002; η2 = .228), moderate bubbles in both trials (trial 1: F(1, 38) =
25.013; p = .0001; η2 = .397; trial 2: χ2(1) = 17.129, p = .0001; η2 = .424), and rapid
bubbles in trial 2 (F(1, 38) = 8.531; p = .006; η2 = .183). In the rest of trials, the two
groups acted in the same manner (p > .05). Regarding accuracy in BT, ASD children
exploded less bubbles than TD children in both trials of all bubble types: slow (trial 1:
F(1, 38) = 17.792; p = .0001; η2 = .319; trial 2: F(1, 38) = 7.080; p = .011; η2 = .157),
moderate (trial 1: F(1, 38) = 14.462; p = .001; η2 = .276; trial 2: F(1, 38) = 15.204; p
= .0001; η2 = .286), and rapid (trial 1: F(1, 38) = 44.100; p = .008; η2 = .170; trial 2: F(1,
38) = 12.850; p = .001; η2 = .253). In both trials of FT, ASD children were slower than
TD children in picking up each flower and living it on the bench (trial 1: (χ2(1) = 18.723,
p = .0001; η2 = .479; trial 2: χ2(1) = 15.119, p = .0001; η2 = .392). Regarding accuracy,
there was no difference between groups in trial 1 (p > .05), whereas in trial 2, ASD
children were less accurate since they picked up less flowers than TD children (χ2(1) =
6.829, p = .009; η2 = .153). Finally, in HT, ASD and TD children spent the same amount
of time selecting each virtual button (p > .05). In trial 1 however, ASD children selected
fewer buttons than TD children (F(1, 38) = 8.138; p = .007; η2 = .176), while in trial 2
they selected the same number of buttons (p > .05).

4. Discussion

The aim of the present study was to qualitatively assess usability and user experience
of a virtual application for CAVETM, representing the skeleton of a multimodal VR
procedure for the assessment of ASD. In addition, quantitative behavioral differences in
performance between ASD and TD children were measured.
The qualitative analysis provided evidence in both groups of good usability and
enjoyable user experience in three tasks out of four, and in particular in KT and BT. In
KT, TD children immediately got the task, sometimes without needing instructions. On
the contrary, even though some ASD children got the task immediately, other ASD
children needed either to see an example of how to interact or more than one attempt to
kick the ball. Both groups expressed happiness and enjoyment after each kick: the
majority of the TD group by smiling, raising their hands, or running, whereas the
majority of the ASD group did so by smiling, throwing themselves on the floor, or doing
stereotypies with arms and hands. Regarding quantitative analysis in KT, both groups
kicked the same number of balls and in trial 2, TD children were faster in kicking than
ASD children. This was likely due to the transfer effect between trials in the TD group
but not in the ASD group, who likely needed more time to get used to this type of virtual
interaction. Considering BT, both groups enjoyed the task, which was reported by the
majority of participants as the best one. While TD children showed entertainment staying
focused on the task and trying to do it in the best way, some ASD children expressed fun
and enjoyment either doing stereotypies or staying calm, fascinated by the falling
bubbles. In addition, few ASD children tried to explode bubbles directly on the CAVETM
surface, which might be a consequence of the high-level cognitive load required to take
actions in the VE due to the intangibility of the interaction. Aforementioned qualitative
observations might explain why the ASD group exploded less bubbles than the TD group
in both trials, regardless of bubble type. Regarding time needed to explode each bubble,
with slow bubbles in trial 1, TD children were faster than ASD children, whereas in trial
2, ASD children performed the same as TD children, demonstrating also in the ASD
group transfer effects between trials. Slow bubbles indeed represent a type of game that
ASD children could cope with, even though they were less accurate than TD children.
Regarding the other bubble types, TD children were faster in moderate bubbles of both
trials, and in rapid bubbles of trial 2. Although these bubble types were more challenging
for ASD than TD children, as reflected by their worse accuracy, rapid bubbles in trial 1
were difficult for both groups since they spent the same amount of time to explode them.
However, transfer effect between trials facilitated the TD group but not the ASD group.
 Minissi et al. / A qualitative and quantitative virtual reality usability study for the early assessment of ASD
children. 51

This might be likely due to the presence of high-level cognitive load in the ASD group
due to the higher bubble speed and the interaction intangibility. In FT, the ASD group
strived further to understand the task than the TD group. Some ASD children either
needed to see an example of how to play or touched the flower on the CAVETM surface
instead of interacting with the virtual human shape. In a few cases, the high-level
cognitive effort in ASD caused by mirroring themselves in the virtual human shape was
also evident when they tried to leave the flower on the bench and not with the hand used
to pick it up. Such observations might explain why TD children were faster than ASD
children in both trials of FT. In trial 1, ASD children were as accurate as TD children,
but in trial 2 they picked less flowers, likely due to the tiredness caused by the cognitive
effort required to interact with the VE. Finally, HT was difficult for both groups and
particularly for ASD children who selected fewer virtual buttons in trial 1 than TD group.
Compared to the other tasks, the interaction in HT was in a two-dimensional space rather
than in three dimensions, and the speed of the virtual hand was difficult to control. Due
to poor task usability, the majority of participants got frustrated regardless of group.
However, in trial 1, TD children employed strategies to cope with frustration, which led
them to select more buttons than ASD in the same amount of time. For instance, they
tried to control the virtual hand’s speed by holding with the other hand the one they were
using for the interaction, or they hid the hand behind their back to take it out and start
again. Conversely, some ASD children got frustrated by the disappearance of the virtual
human shape, or they expressed frustration with the task by moving the hand quickly and
in a casual manner. In trial 2, there was no difference between groups in the number of
selected virtual buttons, likely due to more distractibility and tiredness of TD children
who coped with frustration in trial 1, but in trial 2 performed equal to ASD children.

5. Conclusion

Along with the traditional assessment of ASD, ecological procedures based on

technology and quantitative measures similar to the present one could enhance
objectivity in early diagnosis. This usability study can be considered the implementation
outset of a multimodal VR procedure for the assessment of ASD that is being developed
by the authors. Results confirmed that both groups adequately interacted in the VE,
performing basic movements required for taking actions. HT is the only task in which
usability and user interaction are going to be discussed and changed, while performance
weaknesses of ASD children across the other tasks will be considered to quantitatively
disentangle ASD, hence making an objective assessment.

References

[1] American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th ed. 2013.
[2] Lord C, Rutter M, DiLavore PC, Risi SA. Diagnostic Observation Schedule-WPS (ADOS-WPS). Los
Angeles, CA: Western Psychological Services; 1999.
[3] Lord C, Rutter M, Le Couteur A. Autism diagnostic interview revised: a revised version of a diagnostic
interview for caregivers of individuals with possible pervasive developmental disorders. J. Autism Dev.
Disord. 1994; 24:659–685.
[4] Goldstein S, Ozonoff S, Cook A, Clark E. Alternative methods, challenging issues, and best practices in
the assessment of autism spectrum disorders. Assessment of autism spectrum disorders. 2009; 358-372.
[5] Alcañiz Raya M, Marín-Morales J, Minissi ME, Teruel Garcia G, Abad L, Chicchi Giglioli IA. Machine
learning and virtual reality on body movements’ behaviors to classify children with autism spectrum
disorder. Journal of clinical medicine. 2020; 9(5):1260.
[6] Alcañiz Raya M, Chicchi Giglioli IA, Marín-Morales J, Higuera-Trujillo JL, Olmos E, Minissi ME,
Teruel Garcia G, Sirera M, Abad L. Application of Supervised Machine Learning for Behavioral
Biomarkers of Autism Spectrum Disorder Based on Electrodermal Activity and Virtual Reality. Frontiers
in human neuroscience. 2020; 14:90.
[7] Minissi ME, Giglioli IAC, Mantovani F, Raya, MA. Assessment of the Autism Spectrum Disorder Based
on Machine Learning and Social Visual Attention: A Systematic Review. Journal of Autism and
Developmental Disorders. 2021;1-16.
[8] Brahnman S, Jain LC. Advanced computational intelligence paradigms in healthcare 6: virtual reality in
psychotherapy, rehabilitation, and assessment. Springer; 2011. Vol. 337.
 Annual Review of Cybertherapy and Telemedicine 2021 53

A Voice Recognition Application for the

Semantic and Prosodic Analysis of ASD
Caregivers
Irene Alice CHICCHI GIGLIOLIa1, Luna MADDALONa, Lucía GÓMEZ-
ZARAGOZÁa, Maria Eleonora MINISSIa, Javier MARÍN MORALESa, Marian
SIRERAb, Luis ABADb, Mariano ALCAÑIZa
a
Instituto de Investigación e Innovación en Bioingeniería (i3B), Universitat Politécnica
de Valencia, Valencia, Spain
b
Centro de Desarrollo Cognitivo Red Cenit, Valencia, Spain

Abstract. The voice manifests and conveys numerous components of meaning in

addition to words, such as prosody and semantics. Previous studies have found that
parents of children with Autism Spectrum Disorder (ASD) seem to have a delayed
response time compared to parents of children with typical development (TD).
Words and number repetitions, duration of pronunciation, and meaning used by
parents vary by child diagnosis as well. The aim of this project is to demonstrate
that the parent’s voice can be a powerful behavioral biomarker in the diagnosis of
ASD. Parental quality of life may also be a strong predictor of the quality of life of
children with ASD. Given this goal, we propose the creation of a voice analysis
application that through Machine Learning (ML) algorithms, is able to detect
elements of prosody and semantics for investigation purposes. The application is
based on the Autism Diagnostic Interview-Revised (ADI-R) and contains some
personality questionnaires. This article focuses on potential voice metrics to extract
for in-depth voice analysis. Findings outlined semantic and prosodic metrics that
will be implemented in voice recognition analysis of ASD parents. Future studies
are expected to recognize that parents of ASD children have distinct differences in
prosodic and semantic levels compared with parents of control children. The
uniqueness of this study lies in the creation of a tool focused on the voice, through
combined ML and psychological techniques. This application has the potential to
empower the ADI-R methodology by meeting the terms of validity and objectivity.

Keywords. Voice Recognition, Autism Spectrum Disorder, Biomarkers, Autism

Diagnostic Interview-Revised, Machine Learning

1. Introduction

Humans communicate semantic meanings through the medium of voice. In the act
of pronouncing a word along with the linguistic elements, they are intrinsically
associated with the prosodic aspects of intonation, tone, rhythm, and intensity of speech.
Semantics lies in the study of words' meanings. Significance is configured as an
interpretive procedure to explain and give meaning to the events that represent the
content of the experiences object of the communication. Prosody on the other hand,
performs a key function in the organization and interpretation of speech as it conveys
emotional, socio-linguistic, and dialectal information. It thus appears to be a property of
the vocal signal that modulates and enhances its meaning.
Voice is shown to be an effective behavioral biomarker in the diagnosis of ASD [1].
The great interest towards the identification of biomarkers for ASD has led to the
extensive study of linguistic elements related to the disease [2, 3].

Prosodic and semantic elements have also been investigated from the perspective of
caregivers of ASD children. In the same way that children with ASD assimilate parental
communicative input for their vocabulary development [4], parents adapt their speech to

1
Corresponding Author: [email protected]
reflect their children's developmental level [5]. For instance, studies suggest that
compared with caregivers of TD children, parents of ASD children tend to use less causal
talk and fewer desire or cognitive terms [5, 6]. Furthermore, parents of children with
ASD use a greater amount of concrete nouns and active verbs and rarely use abstract
nouns, stative verbs, adjectives, and adverbs compared to caregivers of TD children [7].
Few studies investigated the relationship of verbal (semantic) and paraverbal
(prosodic) communication of parents of ASD children, with their personality and with
the development of the pathology in children [8]. Due to these studies, we currently know
that some characteristics of parents may predispose to the development of ASD in
children. These include personality (i.e., obsessive-compulsive traits, neuroticism
characteristics), poor quality of interpersonal relationships, social support (characterized
by lower emotional regulation), and cases of psychopathology (i.e., depressive and
anxiety symptoms) [9, 10]. Parental quality of life may also be a strong predictor of the
quality of life of ASD children. It is known that caregiving for ASD children affects
parents’ life financially, in combining daily activities or with the presence of depressive
symptoms [10].
Biomarker research aims to improve accuracy of disorder diagnoses. So far, the
diagnosis of ASD is performed through two complementary tools: Autism Diagnostic
Observation Schedule-Second Edition (ADOS-2) which is designed for children older
than 2 years, and ADI-R which is addressed to caregivers. This methodology is used to
estimate the severity of the disease and for planning an educational project. ADI-R has
some limitations. First, being this a qualitative measure, the responses given are
evaluated according to the experience and training of the therapist. The professional’s
interpretation of the data could lead to distortion of the results, thus not making this
survey methodology objective and standardized [11]. Furthermore, since ADI-R is an
interview, responses could be systematically biased according to the principle of social
desirability, therefore influencing the responses given by the caregiver. Finally, this
diagnostic takes a long time to administer (from 1 and a half to about 2 hours).
The ultimate purpose of this project is to investigate the effectiveness of treatments
on ASD children through the analysis of parental voice. Given this goal and to overcome
the limitations of ADI-R, we propose to develop an application that recognizes vocal
feature differences between caregivers through ML algorithms, standardizing the tool
accordingly. In support of our proposal, studies suggest that analysis of speech
production in ASD using ML has the potential to measure biometric data, acoustic
patterns, and supplement traditional clinical assessment [2, 12]. The possibility that vocal
features could be used as a marker of ASD has also been supported by previous
researchers [2]. This article is meant to identify the voice and text metrics that can be
extracted through ML techniques to outline a broad overview of voice analysis.
Parameters of investigation have been defined considering a later implementation in a
wide study. In the following sections, the metrics of semantics and prosody presented
have been identified and extracted from a sample of two subjects.

2. Methods

2.1. The Voice Recognition Application

The application includes two randomized phases, the first containing a compilation
of 11 psychological questionnaires and a sociodemographic questionnaire that work to
obtain a multidimensional profile along with the parent’s quality of life. The second part
has been formulated according to the 8 dimensions of the ADI-R, thereby synthesizing
the original 93 questions into 12 open-ended questions encompassing and satisfying all
investigated dimensions. These 12 significant questions can be identified in 3 dimensions
of analysis: communication and social interaction (6 questions), language (3 questions),
and stereotypies and narrow interests (3 questions). Caregivers are guided by the
application’s instructions in carrying out both phases, hence completing the task
independently.
The GENCAT scale is one of the measures included in the questionnaires that aim
to investigate the caregivers’ quality-of-life [13]. The remaining ten personality
questionnaires are the validated and adapted Spanish versions of the original
Locati et al. / Training Mentalizing Skills In Virtual Reality: An Experimental Treatment For Children 55

questionnaires: State-Trait Anxiety Inventory (STAI) [14], Short Big Five Inventory
(BFI-S) [15], Emotional Expressivity Scale (EES) [16], Ambivalence Over Emotional
Expression Questionnaire (AEQ) [17], Difficulties in Emotion Regulation Scale (DERS)
[18], Duke-Unk Functional Social Support Scale (Duke-UNC-11) [19], Behavioral
Inhibition/Activation Scales (BIS/BAS) [20], General Self-efficacy Scale (GSE) [21],
Perceived Stress Scale (PSS) [22], Symptoms Checklist-90-Revised (SCL-90-R) [23].

2.2. Data Analysis

Inferential statistics will be performed with the results of the questionnaires to gain
control over the characteristics of the sample and parents, but also to observe whether
there are certain patterns that correlate with the child's type of diagnosis.
The voice recorded in the responses will be analyzed through ML algorithms that
identify the semantic and prosodic components. Here, we propose the use of supervised
ML to classify the groups to which parents belong and thereby to diagnose ASD to their
children or not. Programming language Python (version 3.7.4.) and software LIWC with
Spanish dictionary will be used to extract text metrics (i.e., social content, repetitions,
negations, etc.). OpenSMILE toolkit will be used to extract voice metrics (i.e., pitch,
rhythm, duration, amount, and accent). Specifically, the GeMAPS feature package will
be implemented. Finally, the PRAAT (version 6.0.52.) software package will be
implemented for speech analysis in phonetics.

3. Results

One-sample data (a parent dyad) were extracted from a preliminary study. The data
collected, along with the relevant literature, led to the detection and outline of text and
voice metrics relevant for investigative purposes. The semantics metrics listed in Figure
1 were chiefly extracted from LIWC [24] according to ASD literature’s criteria. The
categories included in “marks added in the transcription”, “general” and “spoken
categories” investigate the speech’s organization. Parameters included in “linguistic
processes”, “personal concerns” and “psychological processes” investigate the meaning
of the expressed content. The six semantic categories together create an inclusive profile
of emotional, cognitive, and structural components present in individuals' verbal speech.

TEXT METRICS
Laughing expressions.
Incomplete and unfinished sentences.
Hand-crafted Marks added in the transcription
Incomplete and unfinished words.
Word repetition per question.
Total word count.
General Number of words per question.
Percentage of words captured in the LIWC dictionary. It allows to control the parameters obtained.
Functional words (articles, prepositions, conjunctions and pronouns).
Personal pronouns.
Impersonal pronouns.
Linguistic Processes Past tense verbs.
Present tens verbs.
Future tense verbs.
Conjunctions.
Work.
Achievement.
Leisure.
LIWC Personal concerns
Home.
Money.
Death.
Uncertainty and non-fluent.
Spoken categories
Sentence reconstruction.
Social Processes (family, friends).
Affective Processes (anxiety, anger, sadness).
- Positive emotion.
- Negative emotion.
Psychological Processes
Perceptual Processes (see, hear, feel).
Cognitive Processes.
Relativity (motion, space, time).
Biological Processes (body).
Figure 1. Semantic analysis parameters.

The prosodic metrics shown in Figure 2 explore general constructs such as pauses,
delay, speed, and duration of answers. GeMAPS measures metrics such as rhythm,
accent, and pitch (i.e., fundamental frequency, intensity, phonological duration). These
parameters investigate the speech rate and related emotion. F0, shimmer, and jitter have
been found to be related to stressful situations, trembling, and nervous speech. Moreover,
voice and unvoiced are correlates of confidence and accuracy in speech [25].

VOICE METRICS
Hand-crafted Behavioral metrics Delay. Latency time to press the record button per question
Logarithmic F0 on a semitone frequency scale, starting at 27.5 Hz (semitone 0). For unvoiced frames this parameter is 0.
F0 Men (107–140 Hz); F0 Women (170–240 Hz).
Measures whether the frequency is rising or falling, and whether the frequency is acute or grave.
Frequency related
Jitter. Deviations in individual consecutive F0 period lengths.
Jitter is the deviation from true periodicity of a presumably periodic signal, often in relation to a reference clock signal.
Measures micro-prosodic variations of the length of the fundamental frequency for harmonic sounds.

Shimmer. Difference of the peak amplitudes of consecutive F0 periods.

GeMAPS
Energy/amplitude related Measures micro-prosodic variations of the amplitude of the fundamental frequency for harmonic sounds.

Loudness. Estimate of perceived signal intensity from an auditory spectrum.

Mean Voiced. The mean length and the standard deviation of continuously voiced regions(F0>0).
Measure average speaking time.
Mean Unvoiced. The mean length and the standard deviation of unvoiced regions (F0= 0; approximating pauses).
Temporal features
Measure the mean time of the silences.
Voiced Segment. The number of continuous voiced regions per second (pseudo syllable rate).
Speech rate measurement.
Number of pauses.
PRAAT Pauses, speed, duration
Total audio duration. Measure the duration of the recording per question (with pauses).

Figure 2. Prosody analysis parameters.

4. Conclusions

Findings outlined semantic and prosodic metrics that will be implemented in voice
recognition analysis of ASD caregivers. Future studies foresee to include caregivers
belonging to 3 different groups based on their child’s diagnosis: parents of ASD children
diagnosed more than 12 months beforehand, parents of ASD children diagnosed less than
12 months beforehand, and parents of TD children.
Given the theoretical and methodological assumptions made so far, voice is
expected to be an effective biomarker in the diagnosis of ASD. Confirming what is
measured by the ML algorithms, it is estimated that parents of ASD children (<12
months) will have a delayed reaction/response time in comparison with the other two
groups. It is also presumed that this group may exhibit repetitive verbal behaviors and
long word pronunciation. Regarding the type of child’s diagnosis, we hypothesize that
greater severity and less acceptance of the condition correspond to fewer and negative
words used by the parent. Finally, questions that cause more stress are expected to
produce faster speech than less emotional questions.
Through the metrics now identified, it is expected to detect prosodic and semantic
differences between parents of ASD children and parents of TD children. Accordingly,
via caregivers' psycho-physiological processes, it would be possible to diagnose ASD in
their child.
The present study has the potential to empower the ADI-R methodology by meeting
the terms of validity and objectivity. The application uniqueness is that, unlike traditional
ADI-R methodology, it allows us to perform an objective analysis of both semantics and
prosody through ML techniques. Along with the development of a short and effective
methodology, this project proposes a different perspective of diagnosis of ASD through
analysis of caregivers’ voices. Finally, the application created could have a major impact
for clinics specialized in the disease.

References

[1] Patel SP, Kim JH, Larson CR, Losh M. Mechanisms of voice control related to prosody in autism
spectrum disorder and first‐degree relatives. Autism Res. 2019;12(8):1192-210.
[2] Bonneh YS, Levanon Y, Dean-Pardo O., Lossos L, Adini Y. Abnormal speech spectrum and increased
pitch variability in young autistic children. Front Hum Neurosci. 2011;4: 237.
[3] Fusaroli R, Lambrechts A, Bang D, Bowler DM, Gaigg SB. Is voice a marker for Autism spectrum
disorder? A systematic review and meta‐analysis. Autism Res. 2017;10(3):384-407.
[4] Bang J, Nadig A. Learning language in autism: Maternal linguistic input contributes to later
vocabulary. Autism Res. 2015;8(2):214-23.
Locati et al. / Training Mentalizing Skills In Virtual Reality: An Experimental Treatment For Children 57

[5] Bottema‐Beutel K, Kim SY. A systematic literature review of autism research on caregiver talk. Autism
Res. 2021;14(3):432-449.
[6] Hutchins TL, Deraway C, Prelock P, O’Neill A. Mothers’ and children’s story-telling: A study of dyads
with typically developing children and children with ASD. J Autism Dev Disord. 2017;47(5):1288-304.
[7] Tubul-Lavy G, Jokel A, Leon-Attia O, Gabis LV. Content words in child-directed speech of mothers
toward children with autism spectrum disorder. Am J Speech Lang Pathol. 2020;29(3):1434-47.
[8] Losh M, Klusek J, Martin GE, Sideris J, Parlier M, Piven J. Defining genetically meaningful language
and personality traits in relatives of individuals with fragile X syndrome and relatives of individuals with
autism. Am J Med Genet B Neuropsychiatr Genet. 2012;159(6):660-8.
[9] Özgür BG, Aksu H, Eser E. Factors affecting quality of life of caregivers of children diagnosed with
autism spectrum disorder. Indian J Psychiatry. 2018;60(3):278.
[10] Hoefman R, Payakachat N, van Exel J, Kuhlthau K, Kovacs E, Pyne J, Tilford JM. Caring for a child
with autism spectrum disorder and parents’ quality of life: application of the CarerQol. J Autism Dev
Disord. 2014;44(8):1933-45.
[11] Frigaux A, Evrard R, Lighezzolo-Alnot J. ADI-R and ADOS and the differential diagnosis of autism
spectrum disorders: Interests, limits and openings. Encephale. 2019;45(5):441-8.
[12] Drimalla H, Scheffer T, Landwehr N, Baskow I, Roepke S, Behnia B, Dziobek I. Towards the automatic
detection of social biomarkers in autism spectrum disorder: introducing the simulated interaction task
(SIT). NPJ Digit Med. 2020;3(1):1-10.
[13] Verdugo MA, Schalock RL, Gómez LE, Arias B. Construcción de escalas de calidad de vida
multidimensionales centradas en el contexto: la Escala GENCAT. Siglo Cero. 2007;38(4):57-72.
[14] Seisdedos N. Cuestionario de ansiedad estado-rasgo. Adaptación española. 1988.
[15] Lang FR, John D, Lüdtke O, Schupp J, Wagner GG. Short assessment of the Big Five: Robust across
survey methods except telephone interviewing. Behav Res Methods. 2011;43(2):548-67.
[16] Kring AM, Smith DA, Neale JM. Individual differences in dispositional expressiveness: development
and validation of the Emotional Expressivity Scale. J Pers Soc Psychol. 1994;66(5):934.
[17] Páez D, Adrián JA, Basabe N. Balanza de afectos, dimensiones de la afectividad y emociones: una
aproximación sociopsicológica a la salud mental. Influencias sociales y psicológicas en la salud mental.
Madrid: Siglo XXI. 1992.
[18] Guzmán-González M, Trabucco C, Urzúa A, Garrido L, Leiva J. Validez y confiabilidad de la versión
adaptada al español de la Escala de Dificultades de Regulación Emocional (DERS-E) en población
chilena. Terapia psicológica. 2014;32(1):19-29.
[19] Saameño JAB, Sánchez AD, del Castillo JDL, Claret PL. Validez y fiabilidad del cuestionario de apoyo
social funcional Duke-UNC-11. Aten Primaria. 1996;18(4):153-63.
[20] Rull MAP, Fong LR, Mora AF, Pérez PV. Propiedades psicométricas de la escala BIS/BAS en una
muestra de estudiantes universitarios mexicanos. Enseñanza e Investigación en
Psicología 2016;21(3):300-10.
[21] Suárez PS, García AMP, Moreno JB. Escala de autoeficacia general: datos psicométricos de la adaptación
para población española. Psicothema. 2000;12(Su2):509-13.
[22] Remor E. Psychometric properties of a European Spanish version of the Perceived Stress Scale
(PSS). Span J Psychol. 2006;9(1):86.
[23] González de Rivera JL, Derogatis LR, De las Cuevas C, Gracia Marco R, Rodríguez-Pulido F, Henry-
Benítez M, Monterrey AL. The Spanish version of the SCL-90-R. Normative data in the general
population. 1989.
[24] Ramírez-Esparza N, Pennebaker JW, García FA, Suriá R. La psicología del uso de las palabras: Un
programa de computadora que analiza textos en español. Revista mexicana de psicología. 2007;24(1):
85-99.
Eyben F, Scherer KR, Schuller BW, Sundberg J, André E, Busso C., ... Truong KP. The Geneva
minimalistic acoustic parameter set (GeMAPS) for voice research and affective computing. IEEE
Trans Affect Comput. 2015;7(2):190-202.
SECTION IV

ORIGINAL RESEARCH

Health care is one of the areas that could be

most dramatically reshaped by these new
technologies.

Distributed communication media could

become a significant enabler of consumer
health initiatives. In fact, they provide an
increasingly accessible communications
channel for a growing segment of the
population.

Moreover, in comparison to traditional

communication technologies, shared media
offers greater interactivity and better tailoring
of information to individual needs.

Wiederhold & Riva, 2004

 Annual Review of Cybertherapy and Telemedicine 2021 61

Exploratory Factor Analysis of the Virtual

Reality Stroop Task
Justin ASBEEa, b1 and Thomas D. PARSONSb, c, d
a
Department of Psychology, University of North Texas, Denton, TX, USA
b
iCenter for Affective Neurotechnologies (iCAN), University of North Texas, Denton,
TX, USA
c
Computational Neuropsychology & Simulation (CNS) Laboratory, University of North
Texas, Denton, TX, USA
d
College of Information, University of North Texas, Denton, TX, USA

Abstract. Executive functioning involves various controlled processing abilities

such as cognitive workload, attention, and inhibition. A common test of executive
functioning is the Stroop Color Word Interference Test. The dual process theory has
been used to explain results from the Stroop test and states that automatic processing
(e.g., word reading) is understood to be an overlearned behavior that requires little
effort or direct control from the participant, while controlled processing (e.g., color-
word interference) involves inhibition of prepotent (overlearned) responses. In the
current study, 85 participants from a university in the southwestern United States
(58% female; M age = 19.82, SD = 2.10) completed the Virtual Reality Stroop Task
(VRST). To examine the underlying constructs of the VRST, a factor analysis was
conducted. Both principal component analysis (PCA) and principal axis factoring
(PAF) were conducted. Further, both oblique and orthogonal rotation methods were
conducted. There were 4 factors with eigenvalues greater than one, which accounted
for 78.46% of the total variance, however based on scree plots, parallel analysis, and
the minimum average partial test, 3 factors were retained. Color naming and word
reading loaded onto an automatic processing factor. Simple interference and
complex interference were loaded onto a controlled processing factor. The VRST
may also discriminate between participants’ response accuracy (to Stroop stimuli)
relative to low and high stress environmental stimuli.

Keywords. Human Performance, Human-computer interactions, Virtual Reality,

Serious Games

1. Introduction

The Stroop Color Word Interference Test is a common test of executive functioning
[1]. However, novel technologies such as virtual reality (VR), virtual environments (VE),
and augmented reality (AR), which can be collectively categorized as extended reality
(XR), may lead to improvements in neuropsychological testing compared to traditional
paper-and-pencil tests. Previous work in our lab has found that the Virtual Reality Stroop
Task (VRST) has adequate convergent validity with several lower dimensional versions
of the Stroop Task such as a paper-and-pencil version Delis-Kaplan Executive Function
System (D-KEFS) and an automated computer delivery version from the Automated
Neuropsychological Assessment Metrics (ANAM). However, little research has been
conducted examining the factor structure of Stroop tests.
One theory which has been used to explain the Stroop effect is dual process theory.
Dual process theory suggests that automatic processing requires little effort for
overlearned behaviors, and controlled processing requires cognitive effort and involves
inhibition of these overlearned responses [2].

1
Corresponding Author: [email protected]
 Asbee et al. / Exploratory Factor Analysis of the Virtual Reality Stroop Task
62

While dual process theory is not the only theory researchers have used to explain the
Stroop effect, many agree that inhibiting responses requires mental effort, which many,
in turn, increase cognitive load, an ability associated with executive functioning [3]. The
VRST was created to utilize aspects of XR to improve upon the traditional Stroop task.
The VRST uses external threatening stimuli to attempt to increase cognitive and affective
load. Additionally, Stroop stimuli complexity (i.e., word reading vs color naming), and
interference complexity (i.e., static vs varying location of Stroop stimuli) are also
introduced [4]. These variables were included to manipulate the cognitive load and
mental effort required from participants. Indeed, previous work examining using the
VRST (high mobility multipurpose wheeled vehicle; HMMWV version) found that
optimal performance occurred when participants had a moderate level of arousal (i.e.,
cognitive and affective load).
However, while the VRST has been shown to have convergent validity with other
versions of the Stroop task, so far there is no research examining the underlying cognitive
constructure measured by the VRST. Further, little research in general has looked into
the factor structure of Stroop tests. One study which examined a computerized version
of the Stroop test found a two-factor solution [5]. The current study is designed to
examine the factor structure of the VRST. Researchers hypothesized that the VRST will
likely have a different factor structure than what was previously found due to additional
variables and improvements from incorporating XR technologies.

2. Methods

2.1. Participants

In the current study, 85 people from a university in the southern United States (58%
female; M age = 19.82, SD = 2.10) participated in the current study.

2.2. Virtual Environment

Virtual Reality Stroop Task (VRST): The high mobility multipurpose wheeled
vehicle (HMMWV) version of the VRST is a computerized 3D presentation of the Stroop
task. Participants viewed the VE using a head mounted display (HMD). Participants
responded to stimuli via key press on a keyboard. Within the VE, participants are placed
in a middle eastern type of environment and drive a simulated HMMWV on a desert road.
Participants also encounter two types of zones: safe zones and ambush zones. In safe
zones, participants experience few distractor stimuli such as gunfire, shouting, or
explosions. In the ambush zones, a greater number of the distractor stimuli are presented.
The Stroop task within the VRST includes color naming, word reading, simple
interference, and complex interferences conditions. The Stroop conditions are presented
in both the safe and ambush zones for a total of 8 conditions which were counterbalanced
across participants. Stroop stimuli were presented on the windshield of the HMMWV.
For the color naming condition, three colored X’s were presented. Participants responded
to the font color of the X’s. In the word reading conditions, participants read color words
and responded to the word presented. In the simple interference condition, color words
were presented in fonts that matched the written word or were different from the written
word. Participants were instructed to respond to the font rather than the written word.
Finally, in the complex interference condition, the task was the same as the simple
interference condition with the addition that Stroop stimuli could appear in various
locations of the windshield compared to previous conditions where Stroop stimuli were
presented in the center of the windshield. Participants were able to experience up to 50
items in each condition.

2.3. Analysis

To examine the underlying constructs of the VRST, an exploratory factor analysis

(EFA) was performed. Appropriateness of an EFA was conducted by examining the
correlation matrix [6], Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO), and
Bartlett's test of sphericity [7].
Asbee et al. / Exploratory Factor Analysis of the Virtual Reality Stroop Task 63

Feature extraction was performed using principal components analysis (PCA) and
principal axis factoring (PAF). Several methods were examined to determine the number
of factors to retain. While common, a cutoff suggested by Kaiser (1960) using
eigenvalues greater than 1 may lead to over-extraction [8]. Scree plots, parallel analysis,
and the minimum average partial (MAP) test were also examined.
Direct oblimin with delta set to 0 and varimax were both performed and compared.
In the current study, the pattern matrix is reported when performing an oblique rotation.

3. Results

Both KMO (0.78) and Bartlett's test of sphericity, χ2(120) = 1003.32, p < .001,
indicate that underlying factors likely exist for the VRST. There were 4 factors with
eigenvalues greater than 1 for the VRST; the scree plot indicated that 2 or 3 factors should
be extracted. Parallel analysis also indicated that 2 factors should be extracted using PCA
but identified more than 3 factors for extraction using PAF. However, factors with
eigenvalues less than 1 were not considered for extraction. Lastly, both the original and
revised MAP test indicated that 3 factors should be retained. Therefore, 3 factors were
extracted, accounting for 72.03 and 66.01 percent of variance from PCA and PAF,
respectively. Factor loadings for direct oblimin can be found in table 1.
Using PAF as a reference, factor one is likely measuring automatic processing speed.
Factor one was positively related to reaction times for color naming and word reading
conditions for both safe and ambush zones. It was also weakly related to response times
for simple interference in the safe and ambush zones, and response times for complex
interference in the ambush zone. Finally, this factor was negatively related to correct
responding for the safe zone color naming condition. As scores on this factor increase
overall, participants tend to take longer to respond and, in some conditions, provide fewer
correct responses.
The second factor is positively related to correct responding in the word reading
and simple interference conditions for both the safe and ambush zones. This factor may
be related to multi-tasking abilities or task switching. When performing the VRST,
participants also drive a virtual Humvee. Because this factor is poorly related to other
variables, it may indicate that participants who score high on this factor may be
performing task switching and when in more complex situations, this task switching
strategy may not be as effective. Therefore, participants may switch to other strategies
for performing the task or may simply use less effort as the task become increasingly
difficult.
The final factor may be related to performance under high cognitive load. This factor
was positively correlated with correct responding in the complex interference condition
in both the safe and ambush zone, correct responding in the simple interference condition
in the ambush zone, and negatively related to response times in the simple and complex
interference conditions for both the same and ambush zones. Again, this factor may
indicate when participants are trying to respond accurately when under high cognitive
load as it is related to slower response times, and increased correct responding when
participants are in conditions with increased cognitive demand.

Table 1. Factor Loadings VRST.

PCA PAF
Factor Factor
1 2 3 1 2 3
MRT color naming 0.91 0.04 0.01 0.92 0.04 0.04
MRT word reading 0.77 0.03 -0.02 0.68 0.05 -0.07
MRT simple interference 0.44 0.05 -0.56 0.45 0.01 -0.50
Safe Zone MRT complex interference 0.24 0.26 -0.80 0.22 0.22 -0.82
CR color naming -0.72 0.44 -0.10 -0.64 0.38 -0.08
CR word reading 0.05 0.73 0.04 0.07 0.58 0.05
CR simple interference -0.05 0.52 0.45 -0.09 0.50 0.35
CR complex interference 0.13 -0.05 0.96 0.12 0.02 0.91
MRT color naming 0.80 0.17 -0.22 0.81 0.16 -0.19
MRT word reading 0.91 -0.02 0.06 0.88 -0.01 0.06
 Asbee et al. / Exploratory Factor Analysis of the Virtual Reality Stroop Task
64

MRT simple interference 0.47 0.12 -0.61 0.48 0.08 -0.58

Ambush Zone MRT complex interference 0.48 0.22 -0.57 0.48 0.19 -0.55
CR color naming -0.18 0.36 0.45 -0.19 0.34 0.38
CR word reading -0.03 0.90 -0.07 -0.03 0.89 -0.14
CR simple interference 0.06 0.45 0.68 0.03 0.50 0.59
CR complex interference 0.02 0.23 0.71 -0.03 0.26 0.59
Note. Principal component analysis (PCA); PCA scores indicate scores from the component matrix; Principal
axis factoring (PAF); PAF scores indicate scores from pattern matrix; scores are presented from direct oblimin,
with delta set to 0; coefficients with magnitudes >.40 are bolded.

4. Discussion

The factor analysis of the VRST indicated that multiple cognitive constructs are
likely measured, as was suggested by previous work [4, 9, 10]. The total number of
correct responses for each section of the Stroop task and reaction times for each section
of the Stroop task were used in the analysis. The majority of the extraction rules indicated
that 3 factors should be extracted. The factors were related to automatic processing speed,
multi-tasking abilities or task switching, and performance under high cognitive load.
Results indicated that the VRST has well-defined cognitive constructs. Overall, similar
patterns of factor loadings were found when comparing PCA and PAF, but factor
loadings were generally smaller for PAF as would be expected [11]. However, there was
a difference when examining the factor loadings based on choice of rotational method.
Factor two accounted for more variability than factor three when using direct oblimin
but factor three accounted for more than factor two when using varimax.
The VRST was designed to examine endogenous attention, which is related to
participants actively directing their attention to specific stimuli or tasks, and exogenous
attention, which occurs when external stimuli influence attentional resources. In addition
to measuring automatic and controlled processing found in traditional Stroop tasks,
researchers have suggested that the VRST may be an accurate measure of the affective
impact of environmental stressors on a participant’s automatic and controlled processing,
executive functioning, simple attention, divided attentional abilities, and gross reading
speed [10].
Our lab has examined the validity of the VRST compared to other lower dimensional
Stroop tasks, finding that they were positively correlated [4, 9, 12]. Further, our lab has
found that the Stroop effect was observed even with differences between the various
Stroop tests [13, 14, 15]. Computerized assessments, including those utilizing VR, may
often assess cognitive domains not used by traditional paper-and-pencil measures [16,
17]. Simply using the computer interface may require more cognitive resources when
compared to traditional tests. However, VEs may have greater ecological validity as the
tests are more similar to real-world experiences, and the tests may be better predictors of
real-world outcomes [9, 18, 19]. A meta-analysis by Neguţ and colleagues (2015)
examined the validity of assessments utilizing VR compared to traditional paper-and-
pencil measures or computerized measures to VEs designed to measure the same
constructs. They found that on average, convergent validity scores were low but
acceptable, likely due to significant differences between traditional psychological
measures and measures utilizing VR [17].
In summary, the VRST seems to measure several cognitive constructs as was
previously theorized. The VRST had factors related to automaticity, multi-tasking
abilities or task switching, and performance under high cognitive load. It is likely that
the VRST likely taps into additional cognitive constructs and resources compared to
traditional measures. Future research may want to focus on a larger sample size than was
used in the current study to enhance the stability of factor structures.

References

[1] Alvarez JA, Emory E. 2006. Executive function and the frontal lobes: a meta-analytic review.
Neuropsychology review. 16(1):17-42.
[2] Barrett LF, Tugade MM, Engle RW. 2004. Individual differences in working memory capacity and dual-
process theories of the mind. Psychological bulletin. 130(4):553.
Asbee et al. / Exploratory Factor Analysis of the Virtual Reality Stroop Task 65

[3] McCabe DP, Roediger III HL, McDaniel MA, Balota DA, Hambrick DZ. 2010. The relationship between
working memory capacity and executive functioning: evidence for a common executive attention
construct. Neuropsychology. 24(2):222.
[4] Parsons TD, Courtney CG, Dawson ME. 2013. Virtual reality Stroop task for assessment of supervisory
attentional processing. Journal of clinical and experimental neuropsychology. 35(8):812-26.
[5] Rezaei M. 2019. Neuropsychological decomposing Stroop interference into different cognitive
monitoring: An exploratory factor analysis. Basic and clinical neuroscience. 10(5):475.
[6] Tabachnick BG, Fidell LS. 2013. Using Multivariate Statistics. New York (NY): Pearson.
[7] Howard MC. 2016. A review of exploratory factor analysis decisions and overview of current practices:
What we are doing and how can we improve?. International Journal of Human-Computer Interaction.
32(1):51-62.
[8] Henson RK, Roberts JK. 2006. Use of exploratory factor analysis in published research: Common errors
and some comment on improved practice. Educational and Psychological measurement. 66(3):393-416.
[9] Parsons TD, Courtney CG. 2011. Neurocognitive and psychophysiological interfaces for adaptive virtual
environments. InHuman-centered design of e-health technologies: Concepts, methods and applications
2011 (pp. 208-233). IGI Global.
[10] Wu D, Courtney CG, Lance BJ, Narayanan SS, Dawson ME, Oie KS, Parsons TD. 2010. Optimal arousal
identification and classification for affective computing using physiological signals: Virtual reality stroop
task. IEEE Transactions on Affective Computing. 1(2):109-18.
[11] Ngure JN, Kihoro JM, Waititu A. 2015. Principal component and principal axis factoring of factors
associated with high population in urban areas: a case study of Juja and Thika, Kenya. American Journal
of Theoretical and Applied Statistics. 4(4):258.
[12] Armstrong CM, Reger GM, Edwards J, Rizzo AA, Courtney CG, Parsons TD. 2013. Validity of the
Virtual Reality Stroop Task (VRST) in active duty military. Journal of Clinical and Experimental
Neuropsychology. 35(2):113-23.
[13] Parsons TD, Carlew AR. 2016. Bimodal virtual reality stroop for assessing distractor inhibition in autism
spectrum disorders. Journal of autism and developmental disorders. 46(4):1255-67.
[14] Parsons TD, Barnett MD. 2018. Virtual apartment stroop task: Comparison with computerized and
traditional stroop tasks. Journal of neuroscience methods. 309:35-40.
[15] Parsons TD, Barnett M. 2019. Virtual Apartment-Based Stroop for assessing distractor inhibition in
healthy aging. Applied Neuropsychology: Adult. 26(2):144-54.
[16] Parsons TD, Courtney CG. 2016. Interactions between threat and executive control in a virtual reality
stroop task. IEEE Transactions on Affective Computing. 9(1):66-75.
[17] Neguţ A, Matu SA, Sava FA, David D. 2015. Convergent validity of virtual reality neurocognitive
assessment: a meta-analytic approach. Transylvanian Journal of Psychology. 16(1).
[18] Parsons TD, Carlew AR, Magtoto J, Stonecipher K. 2017. The potential of function-led virtual
environments for ecologically valid measures of executive function in experimental and clinical
neuropsychology. Neuropsychological rehabilitation. 27(5):777-807.
[19] Zaki J, Ochsner K. 2009. The need for a cognitive neuroscience of naturalistic social cognition. Annals
of the New York Academy of Sciences. 1167:16.
 Annual Review of Cybertherapy and Telemedicine 2021 67

Cognitive stimulation using non-immersive

virtual reality tasks in children with
learning disabilities
Lucileide SANTOSa1 Jorge OLIVEIRAb, Pedro GAMITOb
a
CeiEd / Lusófona University
b
HEI-Lab / Lusófona University

Abstract. Executive functions comprise a set of higher-order mental processes that

are crucial for cognition, emotion and behavior regulation. This study aimed to
explore the benefits of a cognitive stimulation program using virtual reality (VR) for
improving executive functioning in children identified with learning disabilities.
The design of this study consisted of a pilot randomized controlled trial. This study
was approved by an ethics committee. Nineteen children were randomly distributed
by the experimental group consisting in non-immersive VR cognitive stimulation
with tasks and the control group without intervention. A battery to assess executive
functions in children was used for assessing the outcomes. The results suggest
improvements in most domains of executive functions from pre- to post-intervention.
A positive effect was also observed in the response to a questionnaire for parents
about behavioral aspects of their children. These improvements were found only in
the experimental group, which suggests a positive role of intervention in this
population. Despite being a pilot study, these results highlight the positive role non-
immersive VR on executive functions of cognitive interventions has.

Keywords. Cognitive Stimulation, Virtual Reality, Childhood

1. Introduction

Executive functions (EF) are a set of cognitive functions including basic and
complex cognitive functions such as goal-directed behavior, reasoning, problem solving,
and decision making [3]. Several studies emphasize the relationship between EF and
educational achievement, suggesting that deficits in executive functioning affect
information processing and emotional regulation, leading to school failure [4]. Therefore,
this study aimed to assess a cognitive intervention focused on executive functions. This
program consisted of a non-immersive virtual reality intervention directed at children in
primary education. The study design consisted of a pilot randomized controlled trial with
experimental and control groups.
The concept of executive functions was defined by the work of Luria [4] [7], who
hypothesized that the frontal lobes of the brain were responsible for controlling and
monitoring behavior. Further studies supported this relationship while establishing the
association between this region with other cognitive functions such as motor
programming, response inhibition, abstraction ability, problem solving, verbal regulation
of behavior, behavior modification according to environmental circumstances, and
integrity of personality and conscious behavior. Despite the importance of executive
functions for self-regulation of behavior, there is a lack of studies investigating the
efficacy of training programs directed at promoting executive functions in children.
Therefore, this study aimed to assess the usefulness and efficacy of a cognitive training
program for improving executive functions contextualized in a virtual reality
environment describing school tasks.

1
Corresponding Author: [email protected]
Santos et al. / Cognitive stimulation using non-immersive virtual reality tasks in children with learning
disabilities
68

Prior studies using virtual reality in children have suggested promising results in
different contexts – among children with hyperactivity and attention deficit [2] [10],
autism spectrum disorders [5] [11], cerebral palsy [1], and at the level of special
educational needs [6]

2. Methods

2.1. Sample

19 children aged between 6-10 years old were randomly divided into two groups:
experimental group with cognitive training using non-immersive VR (CT-VR), and
control group without CT-VR.

2.2. Measures

EF were assessed with Tartaruga da Ilha [Battery for the Evaluation of Executive
Functions in Children] - TI-BAFEC [9]. This test is presented as a book to children that
introduces stories about animals in 14 different tasks according to ‘cold’ (i.e., cognition
drive) and ‘hot’ (i.e., emotion driven EF), and a questionnaire for parents to assess
children’s behavior. The TI-BAFEC has shown good psychometric properties [5].

2.3. Procedure

This study was approved by the ethics committee of the host institution of this study.
The intervention was comprised of 8 sessions (two sessions/week) in 6h dosage of
intervention with cognitive tasks that involved cognitive abilities such as attention,
memory, language, and executive functions. Figure 1 depicts the tasks used in the
intervention program.
The program consisted of a combination of six different activities across eight
sessions. In the initial session, the participants received practical instructions regarding
the program (e.g., keyboard keys, how to interact with objects, main scene, and places)
and performed the letters soup activity which consisted of a word completion task of
finding the correct letter in four-letter words. The proposed domain for this task is verbal
ability and abstraction. In the second session, participants had to participate in two
activities. The first was the letter soup and the second was completed inside a virtual
sports gym where participants had to organize some equipment according to criteria, i.e.,
to arrange objects by shapes. The main purpose of this task was to involve categorization.
In the third session, the participants had to carry out two activities, the first being to
prepare a classroom for a chemistry class which consisted of observing the materials
placed on the virtual desk, and then going to the locker at the back of the room and
bringing the same materials to the virtual worktable (easiest difficulty, defined by the
number of objects). The other activity was the sports gym where participants had to
organize the objects by color. These tasks involved visual memory, cognitive flexibility,
and categorization. Regarding the fourth session, the participants had to prepare a
classroom of medium difficulty according to the instructions received by the virtual
teacher, while the second activity was the voltmeter where the objective was to associate
the movements of the voltmeter pointer to the corresponding colors, inhibiting a response
according to the instruction when the pointer marked the negative or positive voltage
(easiest difficulty level, defined by the pointer speed). In the fifth session, two activities
were also carried out. The first was the virtual sports gym, where the participants had to
attend to shape and color (e.g., blue object inside the blue box, or place the yellow books
inside the square yellow boxes) and the second activity used a voltmeter for medium
difficulty. Regarding the sixth session, it was carried out on the soccer field, similar to
the activity in the sports gymnasium, but this took place in an outdoor environment.
Participants had to organize the equipment by color. In the seventh session, the
participants carried out two activities, one conducted inside the sports gym where the
participants had to organize the equipment according to shape, and finally, the
participants had to perform the activity in the soccer field by color. In the eighth and last
session of the intervention, the participants had to take part in three activities: the sports
 Santos et al. / Cognitive stimulation using non-immersive virtual reality tasks in children with learning
disabilities 69

gym where they had to organize all the equipment that was out of the boxes, keeping
them inside their respective boxes respecting the colors and shapes, a second activity
using a voltmeter in the highest difficulty level, and the letter soup activity.

Figure 1. Examples of the tasks used for cognitive stimulation.

Note: Top-left (1) image describes the beginning environment at the school entrance; top-middle (2) the letter
soup for verbal ability and abstraction; top-right (3) the chemistry class for visual memory and cognitive
flexibility; bottom-left (4) the sports gym for categorization and cognitive flexibility; bottom-middle (5) the
voltmeter for inhibitory control; and at the bottom-right (6) the soccer field for categorization and cognitive
flexibility.

3. Results

The statistical analysis was conducted using non-parametric tests. To determine

which tests to use in the analysis of the results, tests of normality were carried out on the
distributions of the responses obtained in the QPPE (Questionnaire for Parents, Teachers
and Tutors) and in the TI-BAFEC (see Table 1). Given the small size of the sample, the
Shapiro-Wilks test was conducted to assess normality in these interval distributions. The
comparisons between assessments were compared with non-parametric tests for related
samples (Wilcoxon test) given by the Z values and p-value.

Table 1 depicts the descriptive statistics in mean and standard deviation for values
obtained by the two groups in the two assessment moments. The results presented the
descriptors Animals and Words, used in naming tasks, with the best results
corresponding to higher values (more frequently-named descriptors), suggesting that the
experimental group presented better, larger improvements compared to the control group.
In the remaining descriptors, the best results correspond to lower values, since they refer
to a smaller time or a smaller number of moves in the execution of the task. We can also
verify that the experimental group showed better improvements compared to the control
group. These data were analyzed using the Wilcoxon test for two related samples. The
results show statistically significant differences between pre- vs. post-intervention points
in the experimental group for most indices of the TI-BAFEC (p’s < .05), suggesting
improvements compared to controls, that did not show a significant change.

Table 1 - Values of the TI-BAFEC descriptors in the two evaluation moments

Initial assessment Final assessment

Control Group (n=10) M SD M SD Z
Animals 12.50 2.95 12.60 2.80 -.302
Words 13.70 4.69 13.40 4.79 -.832
Quick naming 33.50 8.84 34.70 6.70 -1.586
 Santos et al. / Cognitive stimulation using non-immersive virtual reality tasks in children with learning
disabilities
70

Animals in disguise_A 32.00 7.92 32.70 6.93 -.880

Animals in disguise_B 30.30 8.41 31.70 8.90 -2.235*
Colorless animals 44.00 12.76 44.80 12.80 -1.725
Animals wrong color_A 43.30 8.50 43.70 9.18 -.796
Animals wrong color_B 50.10 21.01 50.00 21.32 -.054
Game 79.00 8.79 80.40 6.54 -1.184
Experimental Group (n=9) M SD M SD Z
Animals 11.56 2.13 14.67 1.66 2.699**
Words 14.89 5.56 17.11 5.44 2.724**
Quick naming 38.00 12.75 34.44 12.89 -2.680**
Animals in disguise_A 45.89 9.98 40.67 8.90 -2.675**
Animals in disguise_B 45.56 10.21 41.67 9.53 -2.673**
Colorless animals 65.00 24.01 57.78 21.18 -2.670**
Animals wrong color_A 64.11 19.26 59.11 17.49 -2.558*
Animals wrong color_B 69.33 17.97 61.00 14.56 -2.666**
Game 90.33 15.69 83.22 8.63 -2.018*
Note: * p < .05; ** p < .01

The results obtained by the two groups, control and experimental, in the child's
behavior questionnaire were presented at the two moments, which, as already mentioned,
had an interval of 30 days between assessments (see Table 2). It should be noted that in
the eight descriptors presented in Table 2, the highest scores correspond to an
improvement in the child's behavior. Only two descriptors did not show differences at
post-assessment.

Table 2. Values of the Questionnaire for Parents, Teachers and Tutors (QPPE) descriptors in the two evaluation
moments

Initial assessment Final assessment

Control Group (n=10) M SD M SD Z
Can stay focused on activities 3.10 0.57 2.80 0.79 -1.342
Can you wait your turn 3.30 1.16 3.30 1.16 .000
Understands the point of view of -.577
3.10 0.57 3.00 0.67
others
Interrupts other people when
2.80 1.03 2.90 1.10 1.000
remembering something to say
Can modify behavior when
someone makes her see that she 3.50 0.53 3.30 0.48 -1.414
has made a mistake
Can find behavior when someone
makes her see that she was 3.50 0.71 3.40 0.70 -.577
wrong
Seems to choose behaviors that -1.414
3.20 1.32 3.00 1.25
end up harming him/her
Understands ironic sentences 3.10 0.57 3.10 0.74 .000
Experimental Group M SD M SD
Can stay focused on activities 2.00 1.32 3.22 0.97 2.598**
Can you wait your turn 2.33 1.00 3.33 0.50 2.460*
Understands the point of view of 2.000*
2.88 1.36 3.38 1.06
others
Interrupts other people when -2.762**
3.89 0.78 2.56 0.53
remembering something to say
 Santos et al. / Cognitive stimulation using non-immersive virtual reality tasks in children with learning
disabilities 71

Can modify behavior when

someone makes her see that she 2.38 0.74 3.63 0.74 2.640**
has made a mistake
Can find behavior when someone
makes her see that she was 2.44 0.73 3.56 0.73 2.887**
wrong
Seems to choose behaviors that
4.11 0.93 2.89 0.93 -2.810**
end up harming him/her
Understands ironic sentences 2.89 0.78 3.22 0.83 1.342
Note: * p < .05; ** p < .01

4. Conclusion

These results suggest that the use of VR techniques presents a good potential for
adaptation, constituting an efficient tool in terms of intervention in school difficulties.
Such intervention programs may also contribute to an improvement in behavior,
particularly in terms of social interactions, between the child and those closest to him.
Nevertheless, the small sample size and the brief intervention conducted in this study
may also undermine some of the results. Another limitation is related to the measure used
in our study that lacks a comprehensive validation for assessing executive functions in
children. Despite these limitations, the results from this pilot study point towards the
ability of VR cognitive stimulation to be used for improving executive functions with
children in systematic intervention protocol. It is important that future studies evaluate
these effects in larger samples, relying on follow-up assessments to understand the
stability of the effects.

References

[1] Aran, O. T., Şahin, S., Köse, B., Aǧce, Z. B., & Kayihan, H. (2020). Effectiveness of the virtual reality on
cognitive function of children with hemiplegic cerebral palsy: A single-blind randomized controlled trial.
International Journal of Rehabilitation Research, 12–19.
https://doi.org/10.1097/MRR.0000000000000378
[2] Areces, D., Dockrell, J., García, T., Gonzaález-Castro, P., & Rodríguez, C. (2018). Analysis of cognitive
and attentional profiles in children with and without ADHD using an innovative virtual reality tool. PLoS
ONE, 13(8). https://doi.org/10.1371/journal.pone.0201039
[3] Diamond, A. (2013). Executive functions. Annual Review of Psychology, 64, 135-68.
[4] Diamond, A., Barnett, W.S., Thomas, J. and Munro, S., (2007). Preschool program improves cognitive
control. Science (New York, NY), 318(5855), p.1387. doi:10.1126/science.1151148
[5] Dixon, D. R., Miyake, C. J., Nohelty, K., Novack, M. N., & Granpeesheh, D. (2020). Evaluation of an
Immersive Virtual Reality Safety Training Used to Teach Pedestrian Skills to Children with Autism
Spectrum Disorder. Behavior Analysis in Practice, 13(3), 631–640. https://doi.org/10.1007/s40617-019-
00401-1
[6] Lorusso, M. L., Travellini, S., Giorgetti, M., Negrini, P., Reni, G., & Biffi, E. (2020). Semi-immersive
virtual reality as a tool to improve cognitive and social abilities in preschool children. Applied Sciences
(Switzerland), 10(8). https://doi.org/10.3390/APP10082948
[7] Luria, A.R. (1973a). The frontal lobes and the regulation of behavior. Psychophysiology of the Frontal
Lobes, 3-26. https://doi.org/10.1016/b978-0-12-564340-5.50006-8
[8] Luria, A.R. (1973b). The working brain; an introduction to neuropsychology. Basic Books.
[9] Mesquita, M.I. (2011). Evaluation of executive functions in children: Psychometric, developmental, and
neuropsychological study [Evaluation of executive functions in children: Psychometric, developmental
and neuropsychological study. PhD thesis. Porto: Fernando Pessoa University.
[10] Shema-Shiratzky, S., Brozgol, M., Cornejo-Thumm, P., Geva-Dayan, K., Rotstein, M., & Leitner, Y.
(2019). Virtual reality training to enhance behavior and cognitive function among children with attention-
deficit/hyperactivity disorder: brief report. Developmental Neurorehabilitation, 22(6), 431–436.
https://doi.org/10.1080/17518423.2018.1476602
[11] Zhao, J.-Q., Zhang, X.-X., Wang, C.-H., & Yang, J. (2021). Effect of cognitive training based on virtual
reality on the children with autism spectrum disorder. Current Research in Behavioral Sciences, 2,
100013. https://doi.org/10.1016/j.crbeha.2020.100013
 Annual Review of Cybertherapy and Telemedicine 2021 73

Mind your time: The implications of

prolonged Instagram use and drive for
thinness in university students
Alexandra GHIŢĂa,1, Teuntje ELFRINKa, Adriana BÜLTERa, Svenja GABRIELa,
Marie GEISEa, Greta GREWEa and Gerben WESTERHOFa
a
Department of Health, Psychology and Technology, University of Twente, the
Netherlands

Abstract. Objective. The aim of the study was to explore the impact of prolonged
versus limited Instagram use regarding drive for thinness (DFT) in university
students. Method. 201 students, mainly from the University of Twente, participated
in this study. The students completed a survey with emphasis on socio-demographic
data, daily Instagram use, present or past mental health diagnoses, and Drive for
Thinness (DFT) scale, which assesses concerns about the body shape or weight, diet,
and fear of gaining weight. A cut-off score of 60 (daily) minutes was set for
Instagram use (N<60 = 119 respondents, N≥60 = 82 respondents). Results. There
was a significant group difference in DFT between students who engaged <60
minutes versus ≥60 minutes daily on Instagram. There was a positive correlation
between DFT and prolonged daily Instagram use (≥60 minutes), but not in the group
with limited Instagram use (<60 minutes). In addition, there were gender differences
in DFT and Instagram use. On average, females engaged in approximately twice the
amount of time on Instagram on a daily basis compared to males. Moreover, there
was a significant relationship between DFT and Instagram use in females, but not in
males. 18% of the respondents indicated a current or past mental health diagnosis,
however there were no differences in DFT or daily Instagram use between
respondents with or without a formal mental health condition. A simple regression
analysis indicated that daily time spent on Instagram predicts DFT in university
students. Conclusions. Drive for a thinner body is a major component in predicting
the development of formal eating disorders. This study shows the importance of
social media use in facilitating a strong desire to have a thinner body, particularly in
female students.

Keywords. Drive for Thinness, Social Media, Instagram use, University Students.

1. Introduction

Drive for thinness (DFT) is a multidimensional construct involving cognitive-

emotional and behavioral dimensions, and the core of it is weight control or weight loss
[1]. Perfectionism, food-related cognitive biases, stress, anxiety, and depressive
symptoms, as well as restrictive dieting and food avoidance may be the underlying
mechanisms of DFT [1,2]. DFT is regarded as one of the core risk factors in facilitating
the development and maintenance of formal eating disorders like anorexia nervosa and
bulimia nervosa [3,4]. This is concerning since there is a higher DFT prevalence in the
group of adolescent and young adults (AYA), particularly females [5]. DFT in the AYA
group is a significant component that may interfere with later adult mental wellbeing
[4,6].

1
Corresponding Author: [email protected].
Ghita et al. / Mind your time: The implications of prolonged Instagram use and drive for thinness in
university students
74

Social media is among the factors that facilitate DFT and body image dissatisfaction
implicitly, and some studies highlighted that it promotes unrealistic ideal body aesthetics
as a result of an excessive engagement with image/appearance-oriented social
networking sites (SNS) [7,8]. Therefore, the aim of the current study was to explore the
impact of prolonged versus limited SNS daily use (Instagram) on DFT in a sample of
university students.

2. Methods

2.1. Participants

201 students (Mage = 21 years, SD = 2.5) from several Dutch universities (mostly
the University of Twente) participated in this study. The inclusion criteria consisted of
being enrolled at a university and having a good understanding of English. Frequency
and descriptive data of the participants can be observed in Table 1 including gender,
nationalities (other nationalities included Lithuanian, Bulgarian, Finnish, Romanian,
Egyptian or Italian), education, and mental health status. 18% of the students informed
us about their past or current mental health condition (anxiety disorders, mood disorders,
trauma and stress-related disorders, and eating disorders). On average, students spent 54
minutes engaging on Instagram daily.

Table 1. Participant characteristics (N = 201)

Characteristics Total sample

N (%) / M (SD)
Age 21 (2.60)
Gender
Male 35 (17%)
Female 166 (73%)
Nationality
Dutch 17 (8%)
German 132 (62%)
Other 52 (30%)
Education
BSc 162 (76%)
Hogeschool* 23 (16.5%)
MSc 11 (5.2%)
PhD 5 (2.3%)
Mental health condition
Yes 37 (18%)
No 164 (82%)
*BSc, bachelor; Hogeschool, university of applied sciences; MSc, master;
PhD, doctoral program.

2.2. Instruments

Socio-demographic data collected for this study were age, gender, education, daily
Instagram use (in minutes), and current or previous (formal) mental health diagnoses
(established by a clinician).

Drive for thinness (DFT) is a subscale of the Eating Disorder Inventory-2 (EDI-2) [9],
which aims to assess concerns about body shape or weight, diet, and fear of gaining
weight. The subscale measures and distinguishes problematic eating symptoms (anorexia
nervosa) from nonclinical populations. DFT consists of seven items with a six-point
Likert scale response (ranging from “never” to “always”). Within the EDI-2, the DFT
subscale showed an internal reliability of 0.89 within a nonpatient population [10]. The
DFT subscale had a high level of internal consistency as determined by a Cronbach's
alpha of 0.82 in this study.
Ghita et al. / Mind your time: The implications of prolonged Instagram use and drive for thinness in university
students 75

2.3. Procedure

This study is part of a larger project and was approved by the Ethics Committee of
the Faculty of Behavioral, Management and Social Sciences at the University of Twente
in the Netherlands. All students participated in this study voluntarily and with their
written informed consent. The participants filled out a survey using Qualtrics software,
including their socio-demographic data and the DFT subscale. The Qualtrics link was
distributed on social media platforms and the SONA system of the university (students
received 0.25 credit points for completing the survey).

2.4. Statistical analysis

Data normality was examined using the Shapiro-Wilk test (p> 0.05), and parametric
tests were utilized. Descriptive and frequency analyses were carried out to examine the
sample of the study. To further analyze the data, a cut-off score of 60 (daily) minutes
was set for Instagram use (N<60 = 119 respondents, N≥60 = 82 respondents). T-tests for
independent samples were run to assess gender differences regarding DFT and Instagram
use, differences between students who engage in limited versus prolonged time on
Instagram regarding DFT, and differences between participants with or without a mental
health condition in terms of their DFT and daily Instagram use. Pearson correlations were
used to test the relationship between prolonged Instagram use (≥60 minutes) and DFT,
which were analyzed separately for male and female students. In addition, a linear
regression analysis was conducted to explore the degree to which time engaging on
Instagram (minutes/day) can predict DFT (total score) in university students.

3. Results

The sample data indicated there was a significant group difference in DFT [M(total
DFT) = 25.21, SD = 8.01] between students who engaged <60 minutes (MDFT = 23.42,
SD = 8.12) versus ≥60 minutes (MDFT = 27, SD = 7.66) daily on Instagram [t(199) =
2.985, p < 0.001, d = 0.43, M<60 = 23 minutes/day vs. M≥60 = 100 minutes/day]. There
was a positive correlation between DFT and prolonged daily Instagram use (≥60
minutes) (r = 0.349, p < 0.001), but no relationship was found between DFT and limited
Instagram use (<60 minutes) (p > 0.05).
Regarding gender differences, the results depicted significant differences in
DFT [t(199) = 3.424, p < 0.001, d = 0.63) and Instagram use [t(199) = 2.931, p = 0.004,
d = 0.54). On average, female students engage in approximately twice the amount of time
on Instagram on a daily basis compared to males (M male = 31 minutes vs. Mfemale = 59
minutes). A positive correlation between DFT and daily time using Instagram was found
in females (r = 0.313, p < 0.001) but not in males (p > 0.05).
In terms of mental health conditions, 18% of the respondents (N = 37) indicated
a current or past mental health diagnosis, however there were no differences in DFT or
daily Instagram use between respondents with or without a formal mental health
diagnosis (p > 0.05).
Finally, a linear regression analysis established that daily time spent on
Instagram could statistically significantly predict DFT in university students: F(1, 199)
= 21.03, p < 0.001). Average daily Instagram use accounted for 9.6% of the variation in
DFT with an adjusted R2 = 9.1%. The regression equation was: predicted DFT = 28.05
+ 0.048*(time engaging in Instagram use in minutes/day).

Table 1. Regression analysis summary for time (engaging on Instagram daily) predicting drive for
thinness in university students.
Variable B 95% CI β t p

(Constant) 20.05 [26.46 29.54] 35.82 0.00

Time 0.04 [-0.06 0.02] 0.30 4.58 0.00
R2adjusted = 0.09; CI = confidence interval for B.
Ghita et al. / Mind your time: The implications of prolonged Instagram use and drive for thinness in
university students
76

4. Discussion

This study depicted and confirmed previous literature regarding the impact of SNS
daily use on young individuals’ DFT status. The data showed that the more time engaging
on SNSs like Instagram, the greater the DFT scores were in the sample. Previous research
highlighted that appearance-related comparisons to an “ideal body” depicted on SNS
like Instagram may have an important role in determining DFT especially in young
university students [11]. In addition, a mixed-methods systematic review emphasized
that SNS exposure to unrealistic (celebrity) bodies enables implicit physical comparisons
associated with a greater body image dissatisfaction, a greater desire to have a thin(er)
body, and problematic eating behaviors including dieting [12].
Despite the significant difference in DFT between participants spending limited
versus prolonged time on Instagram, the respondents with limited Instagram use (<60
minutes) had “higher than normal” scores on the DFT subscale. This is concerning since
the sample in this study represented a nonclinical population. Therefore, the high scores
on the DFT subscale may posit further mental health-related consideration in the
community of university students. Subsequently, a significant link was found between
DFT and prolonged Instagram use, which indicates that the more time students engage
on Instagram, the higher their desire to have a thin body may be. These results are in line
with previous research confirming the concerning impact of SNS on students’ drive for
a thinner body and concerns with body shape, as well as fear of gaining weight [7,13].
In terms of gender differences, our data are in line with previous studies showing that
young females are more affected by exposure to image-based SNSs like Instagram. One
possible explanation may be females’ greater amount of time spent engaging on
Instagram daily compared to males, as seen in the current study. An earlier study
indicated different DFT-related factors for women and men. For young females, media
influence (e.g. female models) and social pressure were significant mechanisms that
increased DFT, although media influence (e.g. male models) and media internalization
were found to be significant factors in DFT in men [14]. Nevertheless, the results of our
study should be viewed in light of its most important limitation: gender imbalance.
Lastly, engaging daily on social media is a predictor of DFT in university
students. A previous study determined that media influence was associated with
increased DFT, muscularity, and body image concerns in young females over the course
of one year [15]. A solid body of literature indicated the negative impact of engaging
with social media platforms on individuals’ body image-related concerns and
problematic eating behaviors [12]. This posits further attention regarding the
psychosocial support system since body concerns at a young age may facilitate risky
body-related behavioral patterns, as seen in previous research [15].

References

[1] Williamson, DA, White, MA, York-Crowe, E, Stewart, TM. (2004). Cognitive-behavioral theories of
eating disorders. Behav Modif. 2004, 28(6): 711-738.
[2] Papalini S, Beckers T, Claes L, Vervliet B. The drive for thinness: Towards a mechanistic understanding
of avoidance behaviors in a non-clinical population. Behav Res Ther. 2021, 142, 103868.
[3] Peñas-Lledó E, Bulik CM, Lichtenstein P, Larsson H, Baker JH. Risk for self-reported anorexia or bulimia
nervosa based on drive for thinness and negative affect clusters/dimensions during adolescence: A three‐
year prospective study of the TChAD cohort. Int J Eat Disord. 2015, 48(6), 692-699.
[4] Krug I, Binh Dang A, Granero R, Agüera Z, Sánchez I, Riesco N, Jimenez-Murcia S, Menchón J,
Fernandez-Aranda F. Drive for thinness provides an alternative, more meaningful, severity indicator than
the DSM‐5 severity indices for eating disorders. Eur Eat Disord Rev. 2021, 29(3), 482-498.
[5] Li M, Chang EC. Predicting drive for thinness and bulimic symptoms in female adults: might an act-
frequency approach to the measurement of eating attitudes be useful?. Curr Psychol. 2021, 1-9.
[6] Bodell LP, Brown TA, Keel PK. Weight suppression predicts bulimic symptoms at 20-year follow-up: The
mediating role of drive for thinness. J Abnorm Psychol. 2017, 126(1), 32.
[7] Wiederhold BK. Instagram: Becoming a Worldwide Problem?. Cyberpsychol Behav Soc Netw. 2019, 22,
567-568.
[8] Moreno MA, Standiford M, Cody P. Social media and adolescent health. Curr Pediatr Rep. 2018, 6(2),
132-138.
[9] Garner DM, Olmstead MP, Polivy J. Development and validation of a multidimensional eating disorder
inventory for anorexia nervosa and bulimia. Int J Eat Disord. 1983, 2(2), 15-34.
[10] Nevonen L, Clinton D, Norring C. Validating the EDI-2 in three Swedish female samples: eating
disorders patients, psychiatric outpatients and normal controls. Nord. J. Psychiatry. 2006, 60(1), 44-50.
Ghita et al. / Mind your time: The implications of prolonged Instagram use and drive for thinness in university
students 77

[11] Hendrickse J, Arpan LM, Clayton RB, Ridgway, JL. Instagram and college women's body image:
Investigating the roles of appearance-related comparisons and intrasexual competition. Comput Hum
Behav. 2017, 74, 92-100.
[12] Rounsefell K, Gibson S, McLean S, Blair M, Molenaar A, Brennan L, Truby H, McCaffrey TA. Social
media, body image and food choices in healthy young adults: A mixed methods syst12ematic review.
Nutr Diet. 2020, 77(1):19-40.
[13] Fardouly J, Vartanian LR. Social media and body image concerns: Current research and future directions.
Curr Opin Psychol. 2016, 9, 1-5.
[14] Fernandez S, Pritchard M. Relationships between self-esteem, media influence and drive for thinness.
Eat Behav. 2012, 13(4), 321-325.
[15] Girard M., Rodgers RF, Chabrol H. Prospective predictors of body dissatisfaction, drive for thinness, and
muscularity concerns among young women in France: A sociocultural model. Body Image. 2018, 26,
103-110.
 Annual Review of Cybertherapy and Telemedicine 2021 79

Influence of Photorealism and Non-

Photorealism on Connection in Social VR
Nienke BIERHUIZEN a, c1, Wendy POWELL b, Tina MIOCH c, Omar NIAMUT c and
Hans STOKKING c
a
Department of Data Management and Biometrics, University of Twente, The
Netherlands
b
Department of Cognitive Science and Artificial Intelligence, Tilburg University, The
Netherlands
c
TNO: Netherlands Organisation for Applied Scientific Research, The Netherlands

Abstract. Traveling for business meetings is not only costly but also has a negative
influence on the environment. Many video conferencing platforms have tried to
reduce the need to travel, but people still find it relevant to meet face-to-face.
Remote meetings via virtual reality (VR) allow users to still have the feeling of being
together in the same space. In VR, avatars are used as digital user representations.
This study investigated whether photorealistic avatars influence the connection
users feel with each other during a VR remote meeting, and whether congruence
between environment and avatar realism influences this connection. A 2x2 within-
subject experiment was conducted whereby twelve participants had remote meetings
in VR with photorealistic and non-photorealistic avatars and environments. Results
indicate that when both participants are represented by live video footage of
themselves (photorealistic), they feel more connected with each other than when
they are represented by a non-photorealistic avatar. Congruence between the avatar
and environment did not seem to influence connection. These results may help to
improve the value of future remote business meetings.

Keywords. Virtual Reality, Social VR, Connection, Photorealism, Avatars

1. Introduction

During the Covid-19 crisis, synchronous communication platforms have been more
important than ever. Video conferencing is often preferred over an audio-only voice call
as it facilitates the addition of important non-verbal visual information such as gestures
and eye-gaze [12]. However, video conferencing still has significant limitations with
respect to communication elements such as joint interactions, gaze direction, and sound
localization [3].
Virtual Reality (VR) offers the potential to address some of these limitations as it
can give the illusion of being co-located. Recent research suggests that VR offers a
greater sense of connection with another person when compared to Skype [10]. VR may
also more closely approximate face-to-face meetings by immersing the user in the virtual
conversational space [10].
Personalisation of avatars has been shown to increase presence and body ownership
[16]. However, higher levels of realism can result in a lower sense of co-presence [7,13],
which may be attributable to the uncanny valley effect [14,11]. A potential solution to
this is proposed by the TogetherVR platform. Instead of using hyper-realistic graphical
avatars, TogetherVR uses depth cameras to transfer live user images into a virtual
environment (VE) [2,5].
An additional important factor is the level of realism of the VE, with users
experiencing better social communication with higher perceived realism [17]. However,
incongruence in realism between the environment and the objects within it can reduce
the feeling of presence [15].

1
Corresponding Author:[email protected]
Bierhuizen et al. / Influence of Photorealism and Non-Photorealism on Connection in Social VR
80

We hypothesise that users will feel a deeper social connection when participants
are represented by a photorealistic avatar and when there is congruence in the realism of
the avatar and the VE. Additionally, we explore if congruence in realism affects spatial
presence and if this could explain why people feel more connected in certain
environments. Spatial presence in this context is defined as the extent to which
oneself perceives to be involved in and interacting within the VE [10].

2. Methodology

2.1. Experimental Setup

A 2x2 within-subjects design study was conducted to test the hypotheses. The factors
were (i) avatar type and (ii) VE type. The four conditions are presented in Table 1.
Table 1. The four meeting conditions for the study.
Photorealistic Avatar Non-realistic Avatar
Photorealistic VE A C
Non-realistic VE B D

Participants were seated on a chair 1.8m in front of a Kinect Azure depth camera.
The VR head mounted display (HMD) was the HP Reverb. Conditions A+B were
executed using TogetherVR, which alpha-blends video recorded users into the VE [5].
Mozilla Hubs was used for conditions C+D. In these conditions, the users’ head and
hand movements were mapped to the avatar using the HMD and controllers. In addition,
the mouth movement of the avatars corresponded with the microphone’s input. Each
Mozilla Hub avatar was chosen by the participant based on personal characteristics out
of 25 avatars. The VEs were created in Unity and modified using either photographic
(realistic) or digitally generated (not realistic) materials. Dimensions of the room and
furniture were the same in both VEs. Users received spatial audio from both the platforms
directly via the HP Reverb headsets. Figure 1 shows the models of the two VEs and the
participant view of the two avatar types.

Figure 1. From left to right: photorealistic VE, non-photorealistic VE, photorealistic avatar and non-
photorealistic avatar.

2.2. Procedure

Twelve participants each attended two sessions consisting of 2x5-minute

‘business’ meetings (A+B or C+D). They were free to choose any topic of conversation
except discussion of the experiment. The conversational partners were changed between
sessions. During all conditions, participants had to hold the controllers (which were only
necessary for movement tracking during conditions C+D).
In both VE conditions, the researcher verbally guided the participants to their
seats and then left the VE.

2.3. Questionnaires

Before the experiment, participants were asked to indicate their age range, gender,
VR experience level, and acquaintance level of the conversation partner. ‘Extraversion’
and ‘openness to experience’ personality questions from the 60-item HEXACO PI-R [1]
were also asked. It was expected that people with high scores on extraversion would be
Bierhuizen et al. / Influence of Photorealism and Non-Photorealism on Connection in Social VR 81

more comfortable with the conversation task and people with a high score on openness
would be more comfortable in the virtual environment in general.
After each condition, participants completed the Quality of Communication
(QoC) questionnaire [4] which includes the four dimensions (i) face-to-face, (ii)
involvement, (iii), co-presence, and (iv) partner-evaluation [4]. The single question
“How well did you feel a connection with your conversation partner” was an additional
measure for connection. Spatial presence was measured using ‘presence’ components of
Li’s user experience questionnaire [10]. All connection and spatial presence measures
were scored on a 5-point Likert scale. Furthermore, participants had to answer six
questions from the simulator sickness questionnaire [8] which were used to judge
whether they were able to continue the experiment and to explore whether possible
sickness had influenced the results. Lastly, they could add any comments about the
experience of each condition in an open question section.

2.4. Participants

The experiment included 12 participants. Because of the design of the experiment,

this led to 48 individual evaluations of 24 experiments. Since this experiment was
executed during the Covid-19 lockdown, participants were selected based on whether
they were able to have the correct setup at home. Ten of the participants were male and
the tested population’s age varied from 25 to 64. All participants had prior experience
with VR (occasionally or regularly), but it differed per couple how well they already
knew each other.

3. Results

Table 2. Mean scores (with standard deviations) for the different groups. Results in bold are significantly
higher than the comparator.
Avatar Environment Congruence
Measurement Non-real Photo-real Non-real Photo-real Non-real Photo-real
QoC 3.79 (0.55) 4.15 (0.38) 3.91 (0.51) 4.03 (0.50) 4.00 (0.44) 3.94 (0.56)
Connection 3.71 (0.75) 4.04 (0.55) 3.75 (0.68) 4.00 (0.66) 3.96 (0.55) 3.79 (0.78)
Spatial Presence 3.02 (0.99) 3.63 (0.72) 3.10 (0.94) 3.54 (0.83) 3.26 (0.79) 3.38 (1.02)

A paired t-test indicated a significant difference in connectedness (QoC) between the

avatar types, with higher levels of connection reported with the more realistic avatar
(t(23)=-2.55, p<0.05). The differences were most evident in the subscales face-to-face
(t(23)= 2.12, p<0.05) and co-presence (W=12.5, p<0.05). The connection question
showed a similar difference of 0.33 with a Wilcoxon signed-rank test between the two
avatar groups (W=19.5, p<0.05). Pearson’s r showed a moderate positive correlation
between QoC and the connection question (t(46)=6.12, p<0.05) with r = 0.67. In addition,
another paired t-test indicated that spatial presence was significantly higher in the
photorealistic environment than in the non-photorealistic one (t(23)=-2.65, p<0.05). See
Table 2 for an overview of the mean scores on the different measurements.
Paired t-tests did not find a significant difference between congruent and
incongruent conditions for both QoC (t(23)=0.76, p=0.45) and spatial presence (t(23)=-
0.63, p=0.53). A Sobel mediation test indicated that spatial presence was not a mediator
variable between congruence and QoC (z=0.45, p=0.66).
Figure 2 presents the results on QoC for each condition in a boxplot. The
boxplot shows a trend towards the importance of photorealism in user representation and
environment. Differences between the conditions are not significant (F(3,44)=2.53,
p=0.07) according to a one-way analysis of variance (ANOVA).
Pearson’s r shows that the level of extraversion (t(46)=1.76, p=0.08) and
openness (t(46)=-1.38 p=0.17) measured by the HEXACO PI-R [1] did not influence
QoC. Nor was any previous experience in VR (occasionally or regularly) found to have
any influence (t(43)=-0.13, p=0.90) with a t-test. A Wilcoxon signed-rank test suggested
that order effect of the conditions did not influence scores on QoC (W=125.5, p=0.72).
No participant reported simulator sickness during the experiment.
Bierhuizen et al. / Influence of Photorealism and Non-Photorealism on Connection in Social VR
82

Figure 2. Connection measured by QoC [4] per condition.

4. Discussion

The results indicate that photorealistic avatars might increase the feeling of
connectedness between conversational partners. However, the difference was less than
0.5 on a 5- point Likert scale. The finding that co-presence was higher for the realistic
avatar appears to contradict the findings of Jo et al. [7]. However, in that study the avatar
was a digital creation rather than photorealistic, so there may have been an uncanny
valley effect in Jo’s study.
The experimental design meant that the two different avatar types were
presented via two platforms: Mozilla Hubs and TogetherVR. It has to be studied whether
the found effect is also present using different types of avatars. The different platforms
may have also had an effect on the results. Mozilla Hubs had a delay of approximately
550ms, whereas TogetherVR’s delay was only 400ms. Although these differences in
delay have been previously reported as having a differential effect on user satisfaction
with communication [6], this effect may have been reduced since the experimental task
was a free conversation [9].
The lack of significance between congruent and non-congruent conditions did
not support our hypothesis. Results suggest that both realism of the avatar and
environment might be more important with respect to connection than the congruence
between the two (Figure 2). This lack of significance may be due to the mediating effect
of the different platforms, but a larger sample size would be required in order to explore
this further. The same trend has been observed for spatial presence, not supporting our
hypothesis. One limitation here might be that the differences between the two
environments was too large. Comments from the participants in the open question section
described that they experienced the non-photorealistic environment as a bare room that
felt less inviting (Figure 1).
Future studies are recommended to study the effect of realism using the same
platform to control for any underlying variables related to the differences of the two
platforms. Also, qualitative studies using VR conferencing for a longer period of time
for business meetings are highly recommended.
With the aim of reducing travel, more studies to understand how remote
meetings could become as valuable as face-to-face meetings should follow.

Acknowledgements

This project was partially funded by TNO’s Early Research Project ‘Social eXtended
Reality’.

References

[1] Ashton MC, Lee K. The HEXACO-60: A short measure of the major dimensions of personality. J Pers
Assess. 2009;91(4):340–5.
Bierhuizen et al. / Influence of Photorealism and Non-Photorealism on Connection in Social VR 83

[2] De Simone F, Li J, Debarba HG, Ali A El, Gunkel SNB, Cesar P. Watching videos together in social virtual
reality: An experimental study on user’s QoE. In: 26th IEEE Conference on Virtual Reality and 3D User
Interfaces, VR 2019 - Proceedings. IEEE; 2019. p. 890–1.
[3] Eisert P. Immersive 3D video conferencing: challenges, concepts, and implementations. In: Visual
Communications and Image Processing 2003. International Society for Optics and Photonics; 2003. p.
69–79.
[4] Garau M, Slater M, Bee S, Sasse MA. The impact of eye gaze on communication using humanoid avatars.
In: Conference on Human Factors in Computing Systems - Proceedings. 2001. p. 309–16.
[5] Gunkel SNB, Stokking HM, Prins MJ, Van Der Stap N, Ter Haar FB, Niamut OA. Virtual reality
conferencing: Multi-user immersive VR experiences on the web. In: Proceedings of the 9th ACM
Multimedia Systems Conference, MMSys 2018. 2018. p. 498–501.
[6] ITU-T I. Recommendation G. 114. One-w Transm time. 2003;
[7] Jo D, Kim K, Kim GJ. Effects of Avatar and Background Types on Users’ Co-presence and Trust for Mixed
Reality-Based Teleconference Systems. In: Casa 2017. 2017. p. 27–36.
[8] Kennedy RS, Lane NE, Berbaum KS, Lilienthal MG. Simulator Sickness Questionnaire: An Enhanced
Method for Quantifying Simulator Sickness. Int J Aviat Psychol. 1993;3(3):203–20.
[9] Kitawaki N, Itoh K. Pure delay effects on speech quality in telecommunications. IEEE J Sel Areas Commun.
1991;9(4):586–93.
[10] Li J, Kong Y, Roggla T, De Simone F, Ananthanarayan S, De Ridder H, et al. Measuring and
understanding photo sharing experiences in social virtual reality. In: Conference on Human Factors in
Computing Systems - Proceedings. 2019. p. 1–14.
[11] Mori M, MacDorman KF, Kageki N. The uncanny valley. IEEE Robot Autom Mag. 2012;19(2):98–100.
[12] Muhlbach L, Bocker M, Prussog A. Telepresence in videocommunications: A study on stereoscopy and
individual eye contact. Hum Factors. 1995;37(2):290–305.
[13] Nowak KL, Biocca F. The Effect of the Agency and Anthropomorphism on users’ Sense of Telepresence,
Copresence, and Social Presence in Virtual Environments. Presence Teleoperators Virtual Environ.
2003;12(5):481–94.
[14] Shin M, Kim SJ, Biocca F. The uncanny valley: No need for any further judgments when an avatar looks
eerie. Comput Human Behav. 2019;94:100–9.
[15] Steptoe W, Julier S, Steed A. Presence and discernability in conventional and non-photorealistic
immersive augmented reality. In: ISMAR 2014 - IEEE International Symposium on Mixed and
Augmented Reality - Science and Technology 2014, Proceedings. IEEE; 2014. p. 213–8.
[16] Waltemate T, Gall D, Roth D, Botsch M, Latoschik ME. The impact of avatar personalization and
immersion on virtual body ownership, presence, and emotional response. IEEE Trans Vis Comput Graph.
2018;24(4):1643–52.
[17] Wijnmaalen M. The influence of real or virtual environments on user experience and communication
efficiency in social XR. (TNO internal)
 Annual Review of Cybertherapy and Telemedicine 2021 85

Idealization on Dating Apps: Seeing Fewer

Photos of the Potential Partner Leads to
Expectancy Violation and Lower Attraction
Simona SCIARAa,1, Clelia MALIGHETTIa, Giorgia MARTINIb, Giuseppe RIVAa, c,
Camillo REGALIAa
a
Department of Psychology, Catholic University of the Sacred Heart, Milan, Italy.
b
Faculty of Psychology, Vita-Salute San Raffaele University, Milan, Italy.
c
Applied Technology for Neuro-Psychology Lab, Istituto Auxologico Italiano, Italy.

Abstract. Online dating apps facilitate the initiation of romantic relationships by

helping users connect with new partners and meet them in subsequent face-to-face
appointments. However, switching from online to face-to-face dating can induce
expectancy violation and diminish attraction. Drawing on expectancy violation
theory, we hypothesized that seeing just a few photos of the potential partner on
their dating app profile can lead to these negative effects. Users who cannot rely on
many photos for forming their impression are expected to idealize the person and
show, in the moving from online to offline dating, lower levels of attraction, lower
pleasantness of the person’s characteristics, and worse expectancies about their
personality. To test this hypothesis, 57 single young adults were randomly assigned
to one of two experimental conditions: half of them viewed a dating app profile with
18 photos of the potential partner; the other half viewed the same profile but with
just 4 photos. Participants then filled out a questionnaire assessing their impressions
(i.e., attraction, pleasantness, and expected personality). Later, participants watched
a video interview of the person and completed a new questionnaire assessing their
updated impressions. Results supported our hypothesis. While participants who had
seen more photos maintained their impression as positive and stable, participants
who had seen fewer photos showed, after the video, lower physical attraction, lower
pleasantness of the person’s characteristics (e.g., gestures), and worse expected
personality traits. These results have important implications for the study of
romantic attraction and online behaviors.

Keywords. Idealization, Expectancy Violation, Attraction, Dating apps, Social

Media

1. Introduction

Online dating applications (apps) are social media platforms that facilitate the
initiation of a new romantic relationship. Tinder, for instance, helps users connect with
new potential partners and possibly meet them in subsequent face-to-face appointments
[1]. Although moving from online to offline meetings can enhance some aspects of the
social interaction [2], previous research found that switching from online to face-to-face
dating—the so-called “modality switching”—frequently induces expectancy violation
and, in turn, a reduced romantic attraction [2, 3]. In this sense, the type and the timing of
the online interaction can represent relevant factors. Impressions formed through an
extended period of online communication fail to match the physical reality experienced
during in-person meetings and lead to expectancy violation [2].
Previous scholars tried to identify the causes of this mismatch and the consequent
expectancy violation [4]. As computer-mediated communication filters out many social
and affective cues associated with human interaction [5], a certain lack of information
regarding the potential partner may be responsible for expectancy violation on dating
apps. Some characteristics of the potential partner that normally influence how people

1
Corresponding Author: [email protected].
Sciara et al. / Idealization on Dating Apps: Seeing Fewer Photos of the Potential Partner Leads to
Expectancy Violation and Lower Attraction
86

form impressions in face-to-face dating [6, 7] such as linguistic indicators (e.g.,

voice, vocabulary, tone, and accent) and nonverbal cues (e.g., gestures, posture,
proximity), are indeed absent or strongly attenuated in online dating. The paucity of these
features leads the individual to imagine the unknown features, thus running into a
potential violation of expectancy [4]. When information is missing, people still form their
impression by applying cognitive schemes [8, 9]. They can rely on “typical types” by
completing what they do not know about the potential partner with typical features for
that social category (e.g., stereotypical features), or “ideal types” by completing missing
information with an idealized version of the unknown features [10, 11]. In intimate
relationships, especially if the person already evaluates the partner as desirable and
attractive, idealization of missing information is the most likely to occur [12].
The consequences of idealization on dating apps should be consistent with
expectancy violation theory (EVT) [13]. According to EVT, while positive violations
increase the attraction of the person who violated the expectancies, negative violations
decrease the attraction of the violator. Then, since idealization is representing something
as perfect or better than in reality, idealizing potential partners on dating apps because of
a lack of information should produce negative violations that, in turn, will diminish
romantic attraction when the dating switches from the online to the offline modality.

1.1. The Present Research

Based on EVT [13] and previous research on computer-mediated communication

and online dating [2, 3, 4, 5], seeing just a few photos of a potential partner on their
dating app profile should lead to idealization and, in the moving from online to face-to-
face dating, expectancy violation and lower attraction. To test this hypothesis, we
conducted an experiment in which participants viewed different versions of a dating app
profile before seeing a video of the target person. We expected participants who viewed
fewer photos to idealize the unknown features of the target and thus experience, after the
video, a negative expectancy violation with (a) lower levels of attraction, (b) lower
pleasantness of the person’s characteristics, and (c) worse expectancies about their
personality.

2. Methods

2.1. Participants, Design, and Procedure

Fifty-seven young adults (98.2% females; Mage = 24.05, SDage = 3.50)

volunteered in a 2 x 2 repeated-measure experiment. All participants were single and
attracted to men (inclusion criteria). The study had 80% power to detect an effect size of
at least f(U) = .38 in within-between interactions (α = .05; non-centrality parameter λ =
8.13; G*Power 3.1).
After they gave their informed consent, participants completed an online survey
that consisted of the following parts: some demographic questions, a section that showed
the dating app profile of a target person (a young adult man) entailing the manipulation
of the number of photos, the first assessment of the dependent variables (i.e., attraction,
pleasantness, and expected personality), a video interview of the target person (1.5 mins)
and, finally, the second assessment of the dependent variables. At the end of the survey,
participants were fully debriefed and thanked for their participation.

2.2. Materials

To manipulate the number of photos on the dating app profile, participants were
randomly assigned to one of two experimental conditions. Half of them viewed a profile
with 18 photos of the target person; the other half viewed the same profile but with just
4 photos including 1 headshot and 3 photos randomly extracted from the set.
After the first measurement of the dependent variables, participants watched a
video of the potential partner that was intended to simulate the modality switch from
online to offline dating [2]. For this reason, it purposely revealed those partner’s pieces
Sciara et al. / Idealization on Dating Apps: Seeing Fewer Photos of the Potential Partner Leads to
Expectancy Violation and Lower Attraction 87

of information that can be usually discovered during the first date such as the person’s
gestures, posture, voice, and attitudes in social exchanges (e.g., proximity) 1.
The dependent variables were assessed twice, both before and after the video. They
included physical attraction (6 items; α = .83), the perceived pleasantness of three
objective characteristics 2 of the target person (i.e., gestures/posture, voice, social
attitudes), and the expectancy that the target possesses some personality traits including
positive (e.g., sociability, warmth; 7 items, α = .83) and negative traits (i.e., jealousy,
aloofness; 2 items, r = .30). Answers’ scales ranged from 0 (not at all) to 10 (very much)
for attraction, and from 1 (not at all) to 5 (very much) for all other measures.

3. Results

Results supported our hypothesis (see Table 1). As expected, we documented a

between-within interaction effect of the number of photos and the modality switch on
physical attraction, F(1, 54) = 5.16, p = .027, perceived pleasantness of the person’s
gestures/posture (marginally significant), F(1, 53) = 3.05, p = .086, and expectancy about
personality, both regarding positive traits, F(1, 53) = 7.03, p = .011, and negative traits,
F(1, 53) = 6.70, p = .012. Specifically, participants reported comparable impressions
before the video interview, with no difference between the two manipulated conditions.
Instead, after the video, participants who formed their first impressions based on fewer
photos updated them with a worse evaluation of the target while other participants
confirmed their impressions (Figure 1).
Against predictions, we did not find an interaction effect of the number of photos
and the modality switch on the pleasantness of the target’s voice, F(1, 53) = .77, p = .383.
Also, perceived pleasantness of the target’s social attitudes was unexpectedly affected
by the number of photos already before the switch, with participants in the ‘fewer photos’
condition reporting lower levels of perceived pleasantness than their counterparts, F(1,
53) = 10.81, p = .002.

Figure 1. Participants’ levels of attraction towards the potential partner before and after the video
depending on how many photos they previously viewed on his dating app profile. Error bars represent SEs.

In sum, while participants who had seen more photos maintained their impression
as positive and stable over time, participants who had seen fewer photos on the dating
app profile showed, after the video, a worse impression of the target person. They
reported lower physical attraction, lower perceived pleasantness of his gestures and
posture, and worse expectancies about his personality (e.g., less sociability, more
jealousy).

1
To facilitate the identification with a real face-to-face date, the video showed the potential partner from
the perspective of a young woman that was framed from behind while interacting with him spontaneously.
2
In the first measurement of perceived pleasantness of the target person’s objective features, we asked
participants to first imagine some specific characteristics and then evaluate their expectations (e.g., expected
voice). In the second measurement (i.e., after the video), we asked them to evaluate the pleasantness of the
same characteristics but, this time, referring to what they actually observed in the video (e.g., actual voice).
Sciara et al. / Idealization on Dating Apps: Seeing Fewer Photos of the Potential Partner Leads to
Expectancy Violation and Lower Attraction
88

Table 1. Participants’ mean levels of physical attraction, perceived pleasantness of the target person’s features,
and expectancy about his personality (positive and negative traits). Standard deviations are in parentheses.
Fewer Photos n = 27 More Photos n = 30
Attraction Pre-Video 5.16 (2.10) 4.73 (2.07)
Post-Video 4.24 (2.30) 4.83 (2.13)
Pleasantness of Gestures/Posture Pre-Video 3.48 (.80) 3.47 (.57)
Post-Video 3.00 (.85) 3.41 (.68)
Pleasantness of Voice Pre-Video 3.44 (.58) 3.20 (.81)
Post-Video 3.54 (.81) 3.62 (.86)
Pleasantness of Social Attitudes Pre-Video 3.41 (.84) 3.80 (.55)
Post-Video 3.27 (1.04) 3.86 (.58)
Expected Personality (Positive) Pre-Video 3.70 (.51) 3.67 (.37)
Post-Video 3.46 (.61) 3.73 (.44)
Expected Personality (Negative) Pre-Video 2.39 (.58) 2.25 (.47)
Post-Video 2.54 (.56) 2.12 (.56)

4. Discussion

Drawing on EVT [13] and previous research on online dating [2, 3], we predicted
and found that seeing just a few photos of the potential partner on their dating app profile
can lead users to idealize what they do not know about the potential partner, thus risking
experiencing, in moving from online to offline dating, a negative expectancy violation.
Future research is encouraged to replicate and confirm the current findings, possibly
extending them to other populations (e.g., people attracted by women and/or both
genders) or testing the specific role of idealization in mediating the relationship between
a lack of information and expectancy violation.
Our results complement previous work on idealization and disillusion in intimate
relationships [12]. Since any lack of information may lead to idealization, expectancy
violations as a result of a mismatch between idealized and actual features should occur
in various social circumstances including traditional dating, stable relationships, and
simple friendships—e.g., in the switching from dating to living together.
More broadly, our results have implications for the study of online behaviors. They
add to previous research on computer-mediated communication and the effects of a lack
of social cues on impression formation [4]. Also, since idealization may affect any
impression formation, the proposed process should be relevant for any professionals and
scholars interested in the impact of social media use, including those in the organizational
field (e.g., for studying/predicting expertise recognition in online teamwork) [14].

References

[1] LeFebvre LE. Swiping me off my feet: Explicating relationship initiation on Tinder. J Soc Pers Relat.
2008; 35(9):1205-1229. https://doi.org/10.1177/0265407517706419
[2] Ramirez A Jr, Wang Z. When Online Meets Offline: An Expectancy Violations Theory Perspective on
Modality Switching. J Commun. 2008; 58:20-39. https://doi.org/10.1111/j.1460-2466.2007.00372.x
[3] Antheunis ML, Schouten AP, Walther JB. The hyperpersonal effect in online dating: Effects of text-based
CMC vs. videoconferencing before meeting face-to-face. Media Psychol. 2020; 23(6):820-839.
https://doi.org/10.1080/15213269.2019.1648217
[4] Jacobson D. Impression formation in cyberspace: Online expectations and offline experiences in text-
based virtual communities. J Comput.-Mediat Comm. 1999; 5(1):JCMC511.
https://doi.org/10.1111/j.1083-6101.1999.tb00333.x
[5] Lea M, Spears R. Paralanguage and social perception in computer‐mediated communication. J Organ
Comput. 1992; 2:321-341. https://doi.org/10.1080/10919399209540190
[6] Ekman P, Keltner D. Universal facial expressions of emotion: An old controversy and new findings. In:
U Segerstrale and P Molnar (Eds.), Nonverbal Communication: Where Nature Meets Culture. NJ:
Lawrence Erlbaum Associates; 1996. p. 27–46.
[7] Goffman E. The presentation of self in everyday life. Garden City, New York: Doubleday Anchor Books;
1959.
Sciara et al. / Idealization on Dating Apps: Seeing Fewer Photos of the Potential Partner Leads to
Expectancy Violation and Lower Attraction 89

[8] Stangor C, Schaller M. Stereotypes as individual and collective representations. In: CN Macrae, C
Stangor, M Hewstone (Eds.), Stereotypes and stereotyping. New York: The Guilford Press; 1996. p. 3-
37.
[9] Zebrowitz LA. Physical appearance as a basis of stereotyping. In: CN Macrae, C Stangor, M Hewstone
(Eds.), Stereotypes and Stereotyping. New York: The Guilford Press; 1996. p. 79-120.
[10] Lakoff G. Women, fire, and dangerous things: What categories reveal about the mind. Chicago:
University of Chicago Press; 1987.
[11] Lakoff G. Cognitive models and prototype theory. In: U Neisser (Ed.), Concepts and conceptual
development: Ecological and intellectual factors in categorization. Cambridge: Cambridge University
Press; 1987. p. 63-100.
[12] Niehuis S, Lee K‐H, Reifman A, Swenson A, Hunsaker S. Idealization and Disillusionment in Intimate
Relationships: A Review of Theory, Method, and Research. J Fam Theory Rev. 2011; 3:273-302.
https://doi.org/10.1111/j.1756-2589.2011.00100.x
[13] Burgoon JK. Interpersonal expectations, expectancy violations, and emotional communication. J Lang
Soc Psychol. 1993; 12:13-21. https://doi.org/10.1177/0261927X93121003
[14] Bazarova NN, Yuan, YC. Expertise Recognition and Influence in Intercultural Groups: Differences
Between Face-to-Face and Computer-Mediated Communication. J Comput.-Mediat Comm. 2013;
18(4):437-453. https://doi.org/10.1111/jcc4.12018
 SECTION V

CLINICAL OBSERVATIONS

Cybertherapy is a field that is growing rapidly

due to today’s technology and information boom.

Virtual reality and advanced technologies

have been used successfully in a variety of
healthcare situations including treatment of
anxiety disorders and phobias, treatment of eating
and body dysmorphic disorders,
neuropsychological assessment and rehabilitation,
and for distraction during painful or unpleasant
medical procedures.

The novel applications of these technologies

yield many advantages over traditional treatment
modalities, and the disadvantages that
accompanied the first trials of virtual reality are
quickly being addressed and eliminated.

Virtual reality peripherals such as data gloves,

physiological monitoring, and Inter-net worlds are
swiftly demonstrating their usefulness in
cybertherapy applications.

Wiederhold & Wiederhold, 2004

 Annual Review of Cybertherapy and Telemedicine 2020 93

Going beyond body exposure therapy.

Presenting an innovative Virtual Reality and Eye-
Tracking body-related attentional bias task.
Bruno PORRAS-GARCIA1a, Alana SINGH a, Helena MIQUEL a, Gemma TAÑA-
VELASCO a, Natalia BRISEÑO-OLORIZ a, Jesus FLETA-DÍAZ a, Emma IGLESIAS a,
Marta FERRER-GARCIA a and José GUTIÉRREZ-MALDONADO a
a
Department of Clinical Psychology and Psychobiology, University of Barcelona,
Passeig de la Vall d’Hebron, 171, 08035 Barcelona, Spain.

Abstract. The present study provides preliminary findings of an innovative body-related

attentional bias modification task (ABMT) using Virtual Reality (VR) and Eye-tracking
(ET) technologies. Analyses were carried out on a sample composed of college women
(n = 35) who were divided into groups of those with high body dissatisfaction (n = 16)
and those with low body dissatisfaction (n = 19). All participants were exposed to an
immersive virtual environment in which they were embodied in a virtual body that
resembled the measurements of their real body. Subjects performed the body-related
attentional bias modification task for 20-minutes. Eating Disorder (ED) symptomatology
disturbances and body-related attentional bias (AB) were measured before and after a
single session of the VR-ABMT. Results showed a significant (p<.05) reduction of fear
of gaining weight after the intervention among women with high body dissatisfaction.
Our results indicated promising evidence in favor of using this ABMT, particularly
among women with high body dissatisfaction. In addition, the current research provides
a new application of VR and ET technologies that might open a wide range of possibilities
for designing and developing new body-related interventions among patients with EDs
and women with body image disturbances.

Keywords. Attentional Bias Modification Training, Virtual Reality, Eye-Tracking, Body

Image Disturbances, Fear of Gaining Weight.

1. Introduction

In a phenomenon known as attentional bias (AB), described as the propensity to pay

more attention to certain types of stimuli or information (e.g., disorder-relevant information)
over other sorts of information [1], adult and young patients with eating disorders (EDs) and
women with high body dissatisfaction show a tendency to focus more on self-reported
unattractive body parts than other body parts [2-3]. Dysfunctional body-related AB may be
responsible for decreasing the effectiveness of body exposure-based treatments used in
patients with EDs. For example, some patients may tend to avoid or on the contrary, overlook
at those self-reported unattractive body areas, interfering with the exposure-based task.

For this reason, it is necessary to develop new treatment techniques by adding specific
components that aim to reduce the body-related AB. One way to target body-related fears is
through exposure therapy. Previous studies have shown promising results in this field
through the use of mirror exposure therapy [4] or Virtual Reality [5].These techniques usually

1
Corresponding Author: [email protected]
Porras-Garcia et al. / Going beyond body exposure therapy. Presenting an innovative Virtual Reality and Eye-
Tracking body-related attentional bias task.
94

involve the patients systematically observing their body or specific body parts for a certain
amount of time [4]. The current project aims to go further and includes AB modification
techniques within the body exposure therapy as a method to reduce body-related AB, body
dissatisfaction, fear of gaining weight, and body anxiety among patients with EDs (for further
details, see clinicaltrials.gov, NCT04786951). The main aim of this study is to develop an
innovative body-related attentional bias modification task using VR and ET technologies.
Particularly, this study provides preliminary findings of a single session of ABMT among
healthy women with high and low body dissatisfaction (BD) levels.

2. Method

2.1. Sample

Thirty-five women at the University of Barcelona participated in the study and were
recruited through campus flyers and advertisements in social network groups. The exclusion
criteria were a self-reported diagnosis of a current ED, a Body Mass Index (BMI) of less than
17 or more than 30, or a self-reported current severe mental disorder diagnosis. Each
participant was given an identification code to guarantee the confidentiality of the data. This
study was approved by the ethics committee of the University of Barcelona.

2.2. Measures

The following measures were utilized before and after the task (pre-post evaluation):
- Full body illusion (FBI), body anxiety, and fear of gaining weight (FGW) were assessed
by means of visual analog scales (VAS) estimating the intensity of the illusion and the
FGW and anxiety related to the whole body from 0 to 100.
- The Physical Appearance State and Trait Anxiety Scale [PASTAS; 6] was used as a self-
reported questionnaire that assesses body anxiety. The PASTAS is comprised of two
self-report scales measuring weight-related and non-weight-related anxiety, but only the
weight scale (W) with 8 items was used in the current study.
- The 10-item body dissatisfaction scale of the Eating Disorder Inventory (EDI-3 BD;
Garner, 2004) was used to assess body dissatisfaction with the whole body and specific
body parts. Particularly, the Spanish version of the EDI-3 was used in this study.

AB measures: In accordance with the Weight Scale items of the PASTAS, the same areas
of interest (AOIs) were individually drawn onto a 2D frontal view picture of a female avatar
and were labeled as weight-related body parts (W-AOIs), i.e., thighs, buttocks, hips, stomach,
legs and waist.
The participant’s visual fixation was estimated by the following variables:
- Number of fixations on W-AOIs (NF): number of available fixations on the specified
area of interest group (i.e., weight-related AOIs).
- Complete fixation time on W-AOIs (CFT): sum of the fixation duration at the specified
area of interest group (i.e., weight-related AOIs) in milliseconds.

2.3. Instruments

Hardware: Participants were exposed to an immersive virtual environment using a VR head-

mounted display (HMD) (HTCVIVE Pro Eye) with a precise ET device included (Tobii ET).
In addition to the two controllers this HMD usually provides, three additional body trackers
were used to achieve full body motion tracking.
Software: The female avatar was designed using the software Blender v. 2.78. A young
female avatar wearing a basic white t-shirt with blue jeans and black trainers was created.
The avatar also wore a swim cap to avoid any influence of hairstyle. The Unity 3D 5.6.1
(Unity Technologies) software was used to design the VR room, develop the programming
Porras-Garcia et al. / Going beyond body exposure therapy. Presenting an innovative Virtual Reality and Eye-
Tracking body-related attentional bias task. 95

code, and incorporate the virtual avatars within. The virtual environment consisted of a
unique room without any furniture except for a large mirror on the wall placed 1.5 m in front
of the patient. Participants could see their whole body reflected in the mirror, even when they
were moving.

2.4. Procedure

The virtual avatar was generated by taking a frontal and lateral photo of the participant.
To match the silhouette of the avatar to the actual silhouette of the participant, different parts
of the pictures were adjusted. Simultaneously, the other researchers administered the pre-
assessment questionnaires and answered the participant’s questions. Next, the full body
illusion (FBI) was induced over the virtual body (i.e., to perceive and regard a virtual body
as their own real body) using two procedures: visuo-motor and visuo-tactile stimulation. Both
procedures lasted three minutes. Once the FBI was induced, the participant’s gaze was
tracked while they were asked to observe their virtual body in the mirror for 30 seconds to
assess body-related AB. During this process, and as a cover story, participants were told to
stand still and avoid abrupt head movements while the virtual avatar position was being
recalibrated.
The ABMT was based on an adaptation of the AB induction procedure proposed by
Smeets et al. [7]. The training was developed by selecting a series of geometric figures (e.g.,
square, rectangle, and circle) that roughly matched specific parts of the participant's body.
Each of these figures had different colors and sizes. Participants were instructed to detect and
identify the figures that appeared on different parts of the avatar's body. Specifically,
participants were asked to focus their attention on that body part for 4 seconds while it was
progressively illuminated. Afterwards, the figure appeared on another part of the body. In
45% of the trials, the geometric figures appeared on weight-related body parts, and in another
45% of the trials, the figures appeared on non-weight-related body parts. In the remaining
trials (10%), the figures appeared on three neutral objects located next to the avatar.

2.5. Statistical analyses

The analysis software Ogama (Open Gaze Mouse Analyzer) was used to transform the
eye-tracking raw information into suitable quantitative data. An additional data
transformation was conducted by calculating the difference for each attentional variable
between weight-related and non-weight-related AOIs. For further details, see Porras-Garcia
et al. [5]. The outcome of the intervention was analyzed by the statistical software IBM SPSS
Statistics v.25. The participants were divided into high vs. low BD levels using the median
score of the EDI-BD as a cut-off point (Me BD = 8). Finally, to investigate whether there
was an attentional bias modification a mixed between (BD_levels)-within
(Assessment_Time) analyses of variance (ANOVA) was conducted. All the assumptions
were partially met; there was homogeneity of variances and sphericity, but some data were
not normally distributed in some variables.

3. Results

The analyses did not show statistically significant group*time interactions in any of the
measures assessed. However, the analysis further revealed main effects of time in fear of
gaining weight (F (1,31) = 4.553, p = .041, partial η2 =.128). As can be revised in Figure 1.h.,
all women, regardless of their body dissatisfaction levels, showed a tendency to reduce the
fear of gaining weight levels after the intervention. When women with high and low body
dissatisfaction were considered separately, the reduction in fear of gaining weight between
the pre-post assessment was only significant among women with high BD (F (1,23) = 5756,
p = .022, partial η2 =.158).
Porras-Garcia et al. / Going beyond body exposure therapy. Presenting an innovative Virtual Reality and Eye-
Tracking body-related attentional bias task.
96

a b

c d
a

f g h
a f f

Figure 1. Means of the ED and AB measures between women with high and low body dissatisfaction, before
and after the ABMT. Error bars represent standard errors.

4. Conclusions

Our results, although still preliminary, suggest that this procedure can be useful to
reduce the FGW reported by healthy women, particularly those with higher body
dissatisfaction. These results are noteworthy, since FGW is usually considered one of the
more difficult fears to reduce in ED treatments due to the impossibility of directly confronting
it through in vivo exposure therapies as with other sorts of fears [8]. However, these findings
should be carefully considered since other ED or AB measures were not significantly reduced
after the ABMT (e.g., body dissatisfaction). These results are partially in line with those
reported by Smeets et al. [7] in which exposure to all body areas in a group of healthy women
did not lead to a reduction of ED symptomatology. On the other hand, those women who
only attended to their self-reported, most attractive body parts showed higher body
satisfaction levels after the task [7]. Therefore, more studies are required to assess whether
the sort of ABMT procedure toward the body or the number of sessions (e.g., a long-term
intervention) might further improve the effectiveness of this intervention, particularly on
those individuals with high body image disturbances or patients with EDs.
Future assessment and treatment of body image and EDs might benefit from the rapid
technological advancement in VR and ET technologies and the countless possibilities that
both technologies might provide to this field. The current study presented a pioneering
Porras-Garcia et al. / Going beyond body exposure therapy. Presenting an innovative Virtual Reality and Eye-
Tracking body-related attentional bias task. 97

ABMT procedure taking advantage of both VR and ET technologies. The combination of

both technologies might open a wide range of possibilities for designing and developing new
body-related interventions that gradually retrain automatic body-related attentional processes
in patients with EDs.

Acknowledgements: This study was supported by the Spanish Ministry of Science and
Innovation (Project PID2019-108657RB-I00, and by AGAUR, Generalitat de Catalunya,
2017SGR1693.

References

[1] D. A. Williamson, M. A. White, E. York-Crowe, T. M. Stewart, Cognitive-behavioral theories of eating

disorders, Behavior Modification 28 (2004), 711-738.
[2] Jansen, C. Nederkoorn, S. Mulkens, Selective visual attention for ugly and beautiful body parts in eating
disorders, Behaviour research and therapy 43 (2005), 183-196.
[3] Bauer, S. Schneider, M. Waldorf, K. Braks, T. J. Huber, D. Adolph, S. Vocks, Selective visual attention towards
oneself and associated state body satisfaction: an eye-tracking study in adolescents with different types of
eating disorders, Journal of Abnormal Child Psychology 45 (2017), 1-15.
[4] S. S. Delinsky and G. T. Wilson, Mirror exposure for the treatment of body image disturbance, International
Journal of Eating Disorders 39 (2006).
[5] Porras-Garcia, M. Ferrer-Garcia, E. Serrano-Troncoso, M. Carulla-Roig, P. Soto-Usera, H. Miquel-Nabau, ... J.
Gutiérrez-Maldonado, AN-VR-BE. A Randomized Controlled Trial for Reducing Fear of Gaining Weight and
Other Eating Disorder Symptoms in Anorexia Nervosa through Virtual Reality-Based Body Exposure. Journal
of clinical medicine 10(2021), 682.
[6] L. Reed, J. K. Thompson, M. T. Brannick, W. P. Sacco, Development and validation of the Physical Appearance
State and Trait Anxiety Scale (PASTAS), Journal of Anxiety Disorders 5 (1991), 323–332.
[7] Smeets, A. Jansen, A. Roefs, Bias for the (un) attractive self: On the role of attention in causing body (dis)
satisfaction, Health Psychology, 30 (2011), 360
[8] R. F. Rodgers, R. DuBois, M. R. Frumkin, D. J. Robinaugh, A network approach to eating disorder
symptomatology: Do desire for thinness and fear of gaining weight play unique roles in the network?, Body
Image, 27 (2018),1–9.
 Annual Review of Cybertherapy and Telemedicine 2021 99

The way we look at our own body really

matters! Body-related attentional bias as
a predictor of worse clinical outcomes
after a virtual reality body exposure
therapy
Marta FERRER-GARCIAa1, Bruno PORRAS-GARCIAa, Helena MIQUELa, Eduardo
SERRANO-TRONCOSOb, Marta CARULLA-ROIGb and
José GUTIÉRREZ-MALDONADOa
a
Department of Clinical Psychology and Psychobiology, Universitat de
Barcelona.
b
Department of Child and Adolescent Psychiatry and Psychology, Hospital Sant
Joan de Déu of Barcelona.

Abstract: Body-related attentional bias (AB) experienced by anorexia nervosa (AN)

patients has been associated with body image disturbances and other eating disorders
(ED)-related symptoms. The aim of this study was to assess whether the body-related AB
reported by AN patients before a virtual reality (VR)-based body exposure therapy
predicted worse clinical outcomes after treatment. Thirteen AN outpatients participated
in the study. AB was recorded using an eye-tracker incorporated in a VR-Head Mounted
Display. Results showed that AN patients attended to their weight-related body parts for
longer and more frequently than to their non-weight-related body parts. Statistically
significant (p<.05) negative and positive correlations between pre-intervention body-
related AB measures and the difference between pre- and post-assessment fear of gaining
weight, body dissatisfaction, and body appreciation measures were also found. Showing
higher body-related AB before the intervention marginally predicted a lower reduction of
fear of gaining weight (p =.08 and p=.07) and body dissatisfaction (p = .05 and p=.06) at
post-treatment, and significantly predicted a lower increase of body appreciation scores
after the intervention (p<.001). Results suggest that body-related AB may reduce the
efficacy of VR-based body exposure therapy in patients with AN.

Keywords: Anorexia Nervosa, Body-Related Attentional Bias, Virtual Reality, Eye-

Tracking, Body Exposure Therapy, Treatment Outcomes

1. Introduction

Body-related attentional bias (AB), understood as the tendency to selectively attend to

body appearance-related cues in preference to other information, is a phenomenon observed
in Anorexia Nervosa (AN) patients [1]. Previous research found that body-related AB (i.e.

AB toward self-reported unattractive body parts) is strongly related to body image

disturbances [2, 3]. However, more research is needed to explore the relationship between
body-related AB and other core symptoms of AN such as fear of gaining weight.

1
Corresponding Author: [email protected].
.
Ferrer-Garcia et al. / The way we look at our own body, really matters! Body-related attentional bias as a
predictor of worse clinical outcomes after a virtual reality body exposure therapy
100

Based on the preliminary findings of a randomized clinical trial with AN patients [4],
the objective of this study is to assess whether the body-related AB reported by AN patients
before a virtual reality (VR)-based body exposure therapy predicts worse clinical outcomes
after treatment. Specifically, patients showing higher body-related AB before the
intervention were expected to report poorer outcomes (i.e., lower reduction of fear of gaining
weight and body dissatisfaction, and lower increase of body appreciation) after the VR-based
body exposure than patients showing lower body-related AB at pre-treatment.

2. Method

2.1. Sample

Thirteen AN outpatients (11 women and 2 men) receiving day-ward treatment at the
Eating Disorders Units of the Hospital de Sant Joan de Déu and the Hospital de Bellvitge
(Barcelona, Spain) participated in the study. The inclusion criteria were being 13 years or
older and a body mass index over 19. The exclusion criteria were serious mental disorders
with psychotic or manic symptoms (e.g., schizophrenia or bipolar disorders), sensory
complications that precluded exposure (e.g., visual, tactile, or auditory deficits), epilepsy,
clinical cardiac arrhythmia, and pregnancy.

2.2. Measures

Fear of gaining weight (FGW), body appreciation (BA), and body dissatisfaction (BD)
were assessed before and after the intervention. Self-reported FGW levels were assessed
using a visual analog scale (VAS) ranging from 0 (not at all) to 100 (completely). Body
image-related measures were assessed using the Body Appreciation Scale (BAS) [5], and the
Body Dissatisfaction scale (BD) of the Eating Disorder Inventory (EDI-3) [6].
AB measures included the number of fixations and the complete fixation time on
weight-related body parts (Weight-related areas of interest, W-AOIs) and on non-weight-
related body parts (Non-weight-related areas of interest, NW-AOIs). In accordance with the
Physical Appearance State and Trait Anxiety Scale (PASTAS) [7], weight-related areas
included thighs, buttocks, hips, stomach, legs, and waist, and non-weight related areas
included the remaining body parts (i.e., head, shoulders, arms, décolletage, neck, and chest).
The number of fixations and the complete fixation time are considered reliable measures for
the assessment of body-related AB and have been widely applied in previous studies using
eye tracker (ET) technology [8].

2.3. Technical Features

Patients were exposed to an immersive virtual environment using a VR head-mounted

display (HMD-HTC-VIVE). The virtual scenario was designed with the Unity 3D 5.6.1
software and consisted of a small room with a large mirror on the wall placed 1.5 m in front
of the patient’s avatar, so that patients could see their whole virtual body reflected in it. The
avatars (male and female versions) were designed using the software Blender v.2.78. They
wore standard clothes (a white t-shirt, blue jeans, and black trainers) and their hair was
covered by a gray hat to reduce any influence of hairstyle. The avatars also wore an HMD as
the patients did, and their height and size could be adjusted based on the patient’s measures.
In addition to the two controllers of HTC-VIVE, three additional body trackers were used to
achieve full body motion tracking of the avatar. The VR HMD-FOVE-Eye Tracking was
used to detect and register the participant’s eye movements while looking at the avatar in the
mirror.
Ferrer-Garcia et al. / The way we look at our own body, really matters! Body-related attentional bias as a
predictor of worse clinical outcomes after a virtual reality body exposure therapy
101

2.4. Procedure

The study was approved by the ethics committees of the University of Barcelona
(Institutional Review Board IRB00003099) and the hospitals that participated in the study.
During the first session (pre-assessment), and once the patient signed the informed consent,
an avatar (i.e., virtual body) with the same measures of the patient was created and FWG,
BAS, and EDI-3-BD questionnaires were administered. Then, the patient was exposed to the
virtual environment and full body illusion (FBI) was induced over the virtual body using
visuo-motor and visuo-tactile stimulations. Once the FBI was induced, the patient’s gaze was
tracked while they were asked to observe their virtual body in the mirror for 30 seconds to
assess body-related AB. A more detailed description of the avatar development, the visuo-
motor and visuo-tactile stimulations, and the eye tracker assessment task procedures are
provided elsewhere [9, 10].
Treatment consisted of standard cognitive-behavioral therapy plus five sessions of VR-
based body exposure therapy. During exposure sessions, patients were exposed (in the first-
person perspective and on a mirror) to an avatar simulating their own body. In the first
exposure session, the virtual body had the real-size silhouette and body mass index (BMI) of
the participant. Throughout subsequent sessions (maximum 60 minutes, once a week), the
virtual body progressively increased its size until showing a healthy weight. Once
participants finished the fifth (and last) exposure session, the assessment questionnaires were
administrated again.

2.5. Statistical analyses

The OGAMA (Open Gaze and Mouse Analyzer) software was used to transform the
raw eye-tracking data into suitable quantitative data. In addition, the difference between
weight-related and non-weight-related AOIs was calculated so that a positive outcome meant
that the patient had been looking more at the weight-related body parts than at the non-
weight-related body parts, and a negative outcome meant the opposite. On the other hand, a
difference close to zero indicated that the patient had attended to both the weight-related and
the non-weight-related body parts (i.e., there is no attentional bias).
Pearson correlation and linear regression analyses were conducted to assess the
association between the attentional bias showed by the patients before the treatment and the
outcomes of the intervention (differences between pre- and post-treatment scores in FGW,
BD scale of the EDI-3, and BAS). Assumptions were partially met, as some variables were
not normally distributed [11]. Analyses were conducted with the software IBM SPSS
Statistics v.25.

3. Results

Prior to the treatment, the mean of complete fixation time of patients was 5,197 ms
(SD=9,368.79) and the mean number of fixations was 18,77 (SD=15.87), indicating that
participants showed an attentional bias to weight-related body parts (i.e. when looking at
their avatar, they attended to their W-AOIs for longer and more frequently than to their NW-
AOIs).
Furthermore, Pearson correlation analyses showed statistically significant (p<.05)
negative and positive correlations between pre-intervention body-related AB measures and
the difference between pre- and post-assessment fear of gaining weight, body dissatisfaction
and body appreciation (Table 1).
Ferrer-Garcia et al. / The way we look at our own body, really matters! Body-related attentional bias as a
predictor of worse clinical outcomes after a virtual reality body exposure therapy
102

Table 1. Pearson correlations between attentional bias measures at pre-treatment and the difference between
scores of fear of gaining weight, body dissatisfaction, and body appreciation before and after treatment.

Measures FGW (pre- EDI-3-BD (pre- BAS (pre-

post post treatment) post
treatment) r (p) treatment)
r (p) r (p)

Number of -.541* (.043) -.575* (.025) .777**

fixations at pre- (.001)
treatment

Complete fixation -.562* (.036) -.552* (.031) .720**(.004)

time at pre-
treatment
Note: FGW (Fear of Gaining Weight), EDI-3-BD (Body dissatisfaction scale of the
Eating Disorder Inventory 3, BAS (Body Appreciation Scale).
* = statistically significant at p<.05 level
** = statistically significant at p<.01 level

Finally, linear regression analyses (Table 2) showed that having higher body related AB
levels before the intervention marginally predicted a lower reduction of fear of gaining
weight (p=.086 and p=.072) and body dissatisfaction (p=.050 and p =.063) after the
intervention. In addition, having higher body related AB levels before the intervention also
significantly predicted a lower increase of body appreciation scores after the intervention
(p<.001).

Table 2. Summary of linear regression analyses for attentional bias measures (number of fixations and complete
time of fixation) predicting VR-based body exposure therapy outcomes.

Predictors Depende Beta t p R2 R2 F p

nt adj.
variables

Number of FGW -.541 -.1.92 .086 .292 .214 3.721 .086

fixations 9
at pre- EDI-3- -.575 - .050 .331 .264 4.945 .050
treatment BD 2.224
BAS .777 3.898 .003 .603 .563 15.19 .003*
4 *

Complete FGW -.562 - .072 .316 .240 4.151 .072

time of 2.037
fixation at EDI-3- -.552 - .063 .305 .235 4.384 .063
pre- BD 2.094
treatment BAS .720 3.278 .008 .518 .470 10,74 .008*
4 *
Ferrer-Garcia et al. / The way we look at our own body, really matters! Body-related attentional bias as a
predictor of worse clinical outcomes after a virtual reality body exposure therapy
103

Note: FGW (Fear of Gaining Weight), EDI-3-BD (Body dissatisfaction scale of the Eating Disorder
Inventory 3 (EDI-3 BD), BAS (Body Appreciation Scale).
* = statistically significant at p<.05 level
** = statistically significant at p<.01 level

4. Conclusion

As expected and consistent with previous research [2, 3], higher levels of body-related
AB at pre-treatment were strongly associated with poorer outcomes (i.e., lower reduction of
fear of gaining weight and body dissatisfaction, and lower increase of body appreciation)
after the intervention. Consequently, despite a promising reduction in eating disorder
symptomatology after the VR-based body exposure therapy [4], our results suggest that
body-related AB may have reduced the efficacy of the intervention in some ED measures.
The combination of VR and eye-tracking technology could make it possible to control, and
even reduce, body-related AB, and thus represents a useful way to improve body exposure
therapies in AN.

Acknowledgments. This study was supported by the Spanish Ministry of Science and
Innovation (Ministerio de Ciencia e Innovación, Spain/Project PID2019-108657RB-I00:
Modification of attentional bias, with virtual reality, for improving anorexia nervosa
treatment and by AGAUR, Generalitat de Catalunya, 2017SGR1693.

References

[1] Williamson DA, White MA, York-Crowe E, Stewart TM. Cognitive-behavioral theories of eating disorders.
Behav. Modif. 2004; 28: 711-738.
[2] Kerr-Gaffney J, Harrison A, Tchanturia K. Cognitive and affective empathy in eating disorders: A systematic
review and meta-analysis. Front Psychiatry. 2019; 10.
[3] Rodgers RF, DuBois RH. Cognitive biases to appearance-related stimuli in body dissatisfaction: A systematic
review. Clin Psychol Rev. 2016; 46: 1-11.
[4] Porras-Garcia B, Ferrer-Garcia M, Serrano-Troncoso E, Carulla-Roig M, Soto-Usera P, Miquel-Nabau H, ...,
Gutiérrez-Maldonado J. AN-VR-BE. A Randomized Controlled Trial for Reducing Fear of Gaining Weight
and Other Eating Disorder Symptoms in Anorexia Nervosa through Virtual Reality-Based Body Exposure. J
Clin Med. 2021; 10(4): 682.
[5] Avalos L, Tylka TL, Wood-Barcalow N. The Body Appreciation Scale: Development and psychometric
evaluation. Body Image. 2005; 2: 285–297.
[6] Garner D. Eating Disorder Inventory-3: Professional Manual. Psychological Assessment Resources: Lutz, FL,
USA, 2004.
[7] Reed DL, Thompson JK, Brannick MT, Sacco WP. Development and validation of the Physical Appearance
State and Trait Anxiety Scale (PASTAS). J. Anxiety Disord. 1991; 5: 323–332.
[8] Kerr-Gaffney J, Harrison A, Tchanturia K. Eye-tracking research in eating disorders: A systematic review.
Int. J. Eat. Disord. 2019; 52: 3–27.
[9] Porras-Garcia B, Ferrer-Garcia M, Serrano-Troncoso E, Carulla-Roig M, Soto-Usera P, Miquel-Nabau H, …
Gutiérrez-Maldonado, J. Validity of Virtual Reality Body Exposure to Elicit Fear of Gaining Weight, Body
Anxiety and Body-Related Attentional Bias in Patients with Anorexia Nervosa. J. Clin. Med. 2020, 9: 3210.
[10] Porras-Garcia B, Serrano-Troncoso E, Carulla-Roig M, Soto-Usera P, Ferrer-Garcia M, Figueras-Puigderrajols
N, … Gutiérrez-Maldonado J. Virtual Reality Body Exposure Therapy for Anorexia Nervosa. A Case Report
with Follow-Up Results. Front Psychol. 2020; 11: 956.
[11] Schmidt AF, Finan C. Linear regression and the normality assumption. J. Clin. Epidemiol. 2018; 98:146-151.
 Annual Review of Cybertherapy and Telemedicine 2021 105

A Virtual Reality tool using embodiment and

body swapping techniques for the treatment
of obesity: A pilot usability study
Dimitra ANASTASIADOUa,1, Bernhard SPANLANGb, Mel SLATERb,c,d, Julia
VAZQUEZ-DE SEBASTIANa, Josep Antoni RAMOS-QUIROGAe,f, g, Gemma
PARRAMON PUIGe,f, g, Andreea CIUDINh, Marta COMASh, Pilar LUSILLA-PALACIOS
e,f,g

a
Psychiatry, Mental Health and Addictions research group, Vall d´Hebron Research
Institute
b
Virtual Bodyworks S.L.
c
Universitat de Barcelona
d
Department of Clinical Psychology and Psychobiology, Institut de Neurociències de la
Universitat de Barcelona
e
Psychiatry Department; Vall d´Hebron University Hospital
f
Department of Psychiatry and Legal Medicine, Universitat Autònoma de Barcelona
g
Biomedical Network Research Centre on Mental Health (CIBERSAM)
h
Endocrinology and Nutrition Department; Vall d´Hebron University Hospital

Abstract. The objective of the present study, which is framed within the European
Union’s H2020 project titled SOCRATES, is to examine the usability of a Virtual Reality
(VR) embodiment tool for treating obesity. Six healthy adult participants with a desire to
make lifestyle changes in terms of eating healthier and doing more physical activity were
recruited and were randomly assigned to the experimental group (EG) or the control
group (CG). Participants from the EG engaged in a self-conversation aiming at enhancing
their self-awareness and, through embodied perspective taking (body swapping), they
were embodied alternately in their own virtual representation and in a counsellor’s virtual
body. Participants from the CG, embodied in their own virtual bodies, participated in a
“scripted dialogue” with a counsellor of their choice who asked them about their
perceived barriers for engagement with a healthier lifestyle and gave them practical
recommendations about how to make lifestyle changes. A mixed-methods design was
used, involving a semi-structured interview examining the level of users’ satisfaction with
the 2 virtual experiences and their uncovered needs, as well as self-report questionnaires
including those addressing readiness to change habits, body ownership during the VR
experiences, and system usability. The pilot usability study was conducted in July 2021.
Then from September 2021 onwards, once modifications to the prototype are carried out
based on the usability testing and the final VR tool is ready for use in a clinical setting, a
Randomised Controlled Trial will be conducted with 96 participants with obesity to
assess its efficacy compared to usual care.

Keywords. Virtual Reality, Embodiment, Motivational Interviewing, Obesity, Usability.

1. Introduction

The rising prevalence of overweight and obesity in several countries has been described
as a global pandemic. Guidelines support psychological and behavioural weight management

1
Corresponding Author: [email protected].
Anastasiadou et al. / A Virtual Reality tool using embodiment and body swapping techniques for the treatment of
obesity: A pilot usability study
106

interventions including Cognitive Behavioural Therapy (CBT) as a gold standard of

psychological treatment [1].
However, treatments often fail in the long-term due to lack of motivation to maintain
changes among patients. In fact, most patients experience significant weight regain within a
year of completing treatment [2]. A recent study shows that people with obesity along with
a personal motivation to lose weight were more likely to report successful weight loss
compared to unmotivated people [3].
Motivational Interviewing (MI) is a person-centered approach aiming to help patients
strengthen their motivation and commit to change, in line with their own desired outcomes
and values [4]. The integration of MI with weight loss interventions has been associated with
improvements in health-related outcomes, especially in the long-term [5].
In addition, the incorporation of new technologies among the clinical treatments for
obesity has increased steadily in the past few years and has taken on a special relevance
during the Covid-19 crisis [6]. In particular, the use of Virtual Reality (VR) with embodiment
techniques following CBT principles has proved its preliminary effectiveness in the
treatment of eating disorders (ED) and obesity [7]. However, most of these studies essentially
deliver the same CBT in a VR environment, and do not exploit VR benefits to promote
treatment adherence and real-life behavioural changes – key aspects for having enduring
effects.
The SOCRATES project1, within which the present study is framed, aims to use VR
from a user-centered perspective that focuses on the specific needs of people with obesity.
For this goal, the project will adapt an existing platform known as ConVRself and will use
embodiment and body swapping techniques to promote a self-conversation based on MI and
CBT principles. By doing so, patients will try to solve some of the root causes of their
condition including lack of self-awareness, self-stigmatization, and lack of self-
determination. The tool has proved its preliminary effectiveness in using the self-
conversation with university students [8].
According to the objectives of the SOCRATES project, the clinical validation study of
the ConVRself tool will be performed in 2 phases: In Phase I, which is included in this
contribution, we will investigate the usability of the tool specifically adapted to the needs of
people who desire to make lifestyle changes in terms of eating healthier and being more
physically active. Specifically, through this pilot usability study, we will identify and gain
information about the specific uncovered needs of people who want to make lifestyle changes
and the product details. In Phase II, once modifications to the prototype are completed based
on the usability testing and the final VR tool is ready to be tested in a clinical setting, we will
conduct a Randomised Controlled Trial (RCT) to assess the efficacy of ConVRself as a
therapeutic tool that promotes behavioural change among people living with obesity (Phase
II). Our hypotheses for Phase I are that the participants from the Experimental Group (EG)
would report greater levels of perceived acceptability and usability of the platform compared
to the control group (CG), while body ownership, agency, and self-recognition would not
differ among groups.

2. Methods

2.1. Sample

Six healthy adult volunteers who expressed a desire to make lifestyle changes in terms of
eating healthier and/or being more physically active were recruited. The initial goal to recruit
8 patients with obesity from the Vall d´Hebron University Hospital could not be met due to
the Covid-19 pandemic restrictions. In particular, face-to-face visits of these patients in the
hospital were restricted and carried out mostly virtually. As a result, we were forced to carry

1
The SOCRATES project (Self Conversation in Virtual Reality Embodiment to Enhance Healthier Lifestyles Among
Obese People) (Grant Agreement No 951930) is a project supported by the European Union’s Horizon 2020 research
and innovation programme.
Anastasiadou et al. / A Virtual Reality tool using embodiment and body swapping techniques for the treatment of
obesity: A pilot usability study 107

out the study with healthy volunteers who were visited at their homes, taking the appropriate
precautions.

2.2. Design and Procedure

A mixed-methods design was used, involving semi-structured interviews and self-report

questionnaires, with two parallel groups. Ethical approval was obtained from the Ethics
Committee of the Vall d´Hebron University Hospital [PR(AG)224/2021].
All participants signed the informed consent to participate in the study. Then, a
photography procedure was carried out in which 2 photographs of their front and side were
taken for the construction of their avatars. Once the avatars were ready, a brief clinical
interview was conducted by a clinical researcher from Vall d’Hebron Institute of Research
(VHIR) to collect sociodemographic and clinical information from participants, confirm their
eligibility to the study, and assess their readiness to change before the experiment. Then, all
participants were randomly assigned to the EG and the CG. Before the beginning of the VR
session, a system calibration and a familiarization stage took place. Then, participants had
the opportunity to choose among different gender-matched avatar options for their virtual
counsellor. Once the experimental session started, participants from the EG were exposed to
a 20-minute VR session. In this session, participants engaged in a self-conversation through
embodied perspective taking (body swapping), according to which they were embodied
alternately in their own virtual representation and in their counsellor’s virtual body.
Participants started describing to their counsellor the change that they would like to
implement in their lifestyle in a simple statement, and were then body swapped to the
counsellor’s body and responded to that statement by giving positive feedback or asking
further questions regarding perceived barriers and facilitators for the change. During the
whole virtual session, the clinical researcher accompanied the participants and gave them
specific tips about how to formulate questions and give feedback to increase motivation,
inspired by the MI principles. Participants from CG were embodied in their own virtual
representation and listened to a “scripted” counsellor of their choice that asked about the
perceived barriers for engagement with a healthier lifestyle and gave practical
recommendations about how to achieve a healthier lifestyle. No body swapping took place
for this group. Finally, right after the VR sessions, participants from both groups provided
open-ended feedback regarding their satisfaction with the virtual experiences, specific
uncovered needs, and responded to two self-report questionnaires.

2.3. Measurements

Sociodemographic variables - Baseline: Participants’ age, gender, level of education, civil

and employment status, and Body Mass Index (BMI) was collected.
Readiness to change ruler [9] - Baseline: The Readiness Ruler uses a Visual Analogue Scale
ranging from 1 to 10 to assess participants’ “preparation” to change.
Suitability Evaluation Questionnaire (SEQ) - Post-Experiment: The SEQ is a 13-item
questionnaire designed to measure satisfaction, acceptance, and security of use in VR
systems.
Body Ownership Questionnaire - Post-Experiment: At the end of the VR sessions,
participants gave a subjective rating of their illusion of body ownership through a 7-point
Likert scale, where -3 means “not at all” and 3 means “very much”. The questions were taken
from a previous study evaluating ConVRself [8].
Users’ experience - Post-Experiment: A brief debriefing session was carried out after the VR
sessions to assess participants’ satisfaction with the VR experience, ease of use, and
acceptability of the ConVRself tool.

2.4. Data Analysis

Thematic content analysis was used to analyse the qualitative data from the interviews
while descriptive data were analysed using the Excel program.
Anastasiadou et al. / A Virtual Reality tool using embodiment and body swapping techniques for the treatment of
obesity: A pilot usability study
108

3. Results

3.1. Participants’ Characteristics

Three male and three female participants took part in the study. The mean age of the
sample was 51.2 years (SD=13.26) with 5 out of 6 participants having completed university
studies and being employed. Regarding their civil status, 50% were single and 50% married.
The mean BMI was 24.06 (SD=2.53) with only 1 participant presenting as overweight. The
most frequently reported lifestyle change was to do more physical activity and exercise
(100%) while 66% of participants expressed the desire to lose weight and 33.3% to eat
healthier. Regarding participants’ readiness to lose weight, the mean score was 7.3/10 and
their readiness to exercise more was 7.8/10.

3.2. Participants’ Experiences

Table 1 summarizes qualitative information from the post-experiment interview for the
EG. Participants from the CG, although they enjoyed the experience and found the tool easy
to use, they missed a more interactive conversation and personalised experience. The mean
of the total SEQ for the 6 participants was 37.8 (SD=5.34; Range=33-46), with participants
from the EG showing more satisfaction and acceptance of the tool (M=41.7; SD=5.13)
compared to CG (M=34.0; SD=1). Regarding particular items, participants enjoyed the
system (Q1, M=4; SD=0.89), they did not feel confused or disoriented (Q8, M=1.8; SD=1.33),
and they thought that ConVRself would be helpful for their particular lifestyle change (Q11,
M=3.67; SD=0.82). The median of all participants for “Body Ownership” of the self-avatar
was 1 (Range=-3 to 2), for “Agency” was 2 (Range=-1 to 3), and for “Self-recognition” was
1 (Range=-1 to 2).

Table 1. Perceived advantages and disadvantages of the VR experience by participants from the EG.
ADVANTAGES
Domain Key themes Examples of users’ statements
Characteristics of Perceived ease - “At the beginning I needed some instructions and guidance
the VR platform of use from the therapist but then it was easy for me to use it”
Perceived - “It confronts you with your goals and helps you to better
usefulness organize your ideas”
- “It helped me to reflect and feel more aware of the steps that
I must follow to achieve my goals in a more concrete and
tangible way”; “I have now set a day to take my first step”
- “I would recommend the platform to a friend who wants to
make any lifestyle change”
- “I have been able to immerse myself into the virtual
experience, gain confidence and not feel ashamed when
explaining my problem to someone I did not know”
- “ConVRself can be a time- and cost-saving platform to be
used as a complementary tool to face-to-face treatment”
- “Sometimes it is better to train people in this way, rather than
to take them to the real therapeutic office”
- “The whole experience was interactive and real, I felt that I
was literally in that place”
Design – - “I found the virtual environment attractive and cosy, and at
platform the same time, being empty, it was relaxing and promoted my
concentration”
- “I liked the personalised avatars”
- “I found the counsellor really kind and empathetic"
DISADVANTAGES
Domain Key themes Examples of patients’ statements
Characteristics of Negative - “I would improve the sensitivity of the controllers as it was
the VR platform perception of difficult to use them properly and to press the buttons that I
ease of use was asked to in order to continue with the experience”
Anastasiadou et al. / A Virtual Reality tool using embodiment and body swapping techniques for the treatment of
obesity: A pilot usability study 109

Negative - “During the self-conversation, I felt disconnected, and a bit

perception of stressed with the change of roles (being patient or counsellor).
usefulness I did not know if I had to talk being myself or the counsellor”
- “To use it properly, I consider it essential to receive previous
training on the correct use of the platform and also on how to
correctly formulate questions for the self-conversation”
Problems with - “The self-avatar did not look like me, it was fatter than me,
the design of and this was annoying during the experience”
the platform - “I would prefer a more realistic virtual environment”
Lack of - “It would have been better to personalize the background of
personalization the virtual environment. For instance, I would have liked it if I
could choose among different backgrounds, such as the
therapist’s office or a beach. This could increase my
introspection capacity”
Individual factors Age, gender, - “I think that elderly people will have more problems in using
and education the platform”
Characteristics of Negative - “The HMD was heavy and uncomfortable to wear after a bit”
Oculus Quest 2 perception of
ease of use

4. Conclusion

All participants found the VR experience to be novel, interesting, and enjoyable. In

addition, moderate usability and high satisfaction of participants with the platform was found.
Some difficulties in using the VR tool were observed, but it is worth highlighting that most
of these issues are considered to be easily overcome with further training of the users on MI
and the correct use of the platform. In addition, a supporting strain to better support the HMD
device will be used. Regarding platform improvements, we are willing to improve avatar
voices and offer users more options to personalize their virtual environments.
No privacy and security issues were reported, nor were difficulties related to the gender
or education of the participants. Regarding the age factor for older participants, learning to
use these technologies could be harder. Little incidence of simulator sickness was reported,
and no further emotional and mental issues occurred. ConVRself could be time and cost-
saving if all the issues reported are overcome.
The present pilot usability study has provided an important advancement in the
development of the SOCRATES project and in particular, the preparation of Phase II. It has
showed that the ConVRself tool is well-accepted by participants and is now ready to be tested
in a clinical setting with real patients living with obesity. In turn, through the Phase II
study, we expect to demonstrate the effectiveness of the embodiment body-swapping
paradigm to treat obesity while also integrating MI and CBT techniques in a VR context.

References

[1] Wharton S, Lau DCW, Vallis M, Sharma AM, Biertho L, Campbell-Scherer D, et al. Obesity in adults:
a clinical practice guideline. Can Med Assoc J [Internet]. 2020 Aug 4;192(31):E875–91. Available from:
http://www.cmaj.ca/lookup/doi/10.1503/cmaj.191707
[2] Wadden TA, Butryn ML, Wilson C. Lifestyle Modification for the Management of Obesity.
Gastroenterology [Internet]. 2007 May 1 [cited 2021 Mar 15];132(6):2226–38. Available from:
https://linkinghub.elsevier.com/retrieve/pii/S0016508507005781
[3] Dhurandhar N V., Kyle T, Stevenin B, Tomaszewski K. Predictors of weight loss outcomes in obesity
care: results of the national ACTION study. BMC Public Health [Internet]. 2019 Dec 30;19(1):1422.
Available from: https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-019-7669-1
[4] Miller W, Rollnick S. La entrevista motivacional: ayudar a las personas a cambiar. Barcelona: Paidós;
2015.
[5] Barrett S, Begg S, O’Halloran P, Kingsley M. Integrated motivational interviewing and cognitive
behaviour therapy for lifestyle mediators of overweight and obesity in community-dwelling adults: a
systematic review and meta-analyses. BMC Public Health [Internet]. 2018 Dec 5;18(1):1160. Available
from: https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-018-6062-9
[6] De Amicis R, Cancello R, Capodaglio P, Gobbi M, Brunani A, Gilardini L, et al. Patients with Severe
Anastasiadou et al. / A Virtual Reality tool using embodiment and body swapping techniques for the treatment of
obesity: A pilot usability study
110

Obesity during the COVID-19 Pandemic: How to Maintain an Adequate Multidisciplinary Nutritional
Rehabilitation Program? Obes Facts [Internet]. 2021;14(2):205–13. Available from:
https://www.karger.com/Article/FullText/513283
[7] Freeman D, Reeve S, Robinson A, Ehlers A, Clark D, Spanlang B, et al. Virtual reality in the assessment,
understanding, and treatment of mental health disorders. Psychol Med [Internet]. 2017 Oct
22;47(14):2393–400. Available from:
https://www.cambridge.org/core/product/identifier/S003329171700040X/type/journal_article
[8] Slater M, Neyret S, Johnston T, Iruretagoyena G, Crespo MÁ de la C, Alabèrnia-Segura M, et al. An
experimental study of a virtual reality counselling paradigm using embodied self-dialogue. Sci Rep
[Internet]. 2019 Dec 29;9(1):10903. Available from: http://www.nature.com/articles/s41598-019-46877-
3
[9] Miller W, Rollnick S. Motivational interviewing: Preparing people for change. 2nd editio. New York:
The Guilford Press; 2002.
 Annual Review of Cybertherapy and Telemedicine 2021 111

Aphasia360°: A virtual reality intervention

for anomia rehabilitation in post-stroke
patients
Claudia REPETTOa,1, Alice CANCERa, Claudia RODELLAa, Marta CAMPAGNAb,
Alessandra MAIETTIb
a
Dept. of Psychology, Università Cattolica del Sacro Cuore, Milan, Italy
b
Dipartimento di Riabilitazione e Recupero Funzionale, Fondazione Poliambulanza,
Brescia

Abstract. Aphasia is an acquired deficit following acute damage to the central nervous
system that involves the difficulty or impossibility of understanding and formulating
language. A typical disorder of non-fluent forms of aphasia is anomia. Anomia refers to
the difficulty in finding words, in particular when trying to name objects and actions.
According to the Embodied Cognition approach (EC), language is tightly connected to
the motor system. In this view, language rehabilitation programs should stimulate
language through the activation of the motor system. In this approach, since anomic
deficits are often due to a weak link between the meaning of the word and its lemma, the
Hebbs’ principles of coincident and correlated learning can be exploited, i.e., by
intensifying the synchronous activation of lexicon and semantics and connecting them
with the motor counterpart. In this study, we present an innovative training, based on the
EC framework, in which we will make use of new technologies for anomia rehabilitation
in post-stroke patients. Specifically, we will use immersive 360° videos representing
everyday actions displayed from the first-person point of view, experienced through a
head-mounted display. The training will be administered 3 times a week for 4 weeks. The
control group will watch standard videos representing the same actions recorded from the
third-person perspective. Naming abilities will be tested before and after the training
together with other cognitive and psychological measures. We expect that the group who
will undergo the 360° video-based training will show greater improvement of
performance compared to the control group.

Keywords. Immersive 360° Videos; Aphasia; Anomia; Embodied Cognition,

Rehabilitation, Virtual Reality

1. Introduction

Language is an essential function for human beings because it is the one that allows
verbal communication and includes skills such as the expression and understanding of
phonological, semantic, and syntactic aspects. Aphasia is an acquired deficit following acute
damage to the central nervous system that involves the difficulty or impossibility of
understanding and formulating language [1]. The loss of one or more language skills has a
serious impact on the quality of life of patients [2] as it causes difficulties in many areas of
life, i.e., on a personal, social, and socio-economic level [3,4].

1
Corresponding Author: [email protected].
Repetto et al. / A virtual reality intervention for anomia rehabilitation in post-stroke patients
112

For this reason, the rehabilitation of aphasic disorders is considered of primary

importance starting from the subacute phase of the index event and sometimes, compatibly
with the resources available, accompanies the patient even in the chronic phase in an attempt
to continue improving their communicative skills.
A typical disorder in non-fluent forms of aphasia within the chronic phase is anomia.
Anomia refers to the difficulty in finding the lexicon, which can manifest itself in the
difficulty in naming objects and actions [5], and may be due both to a deficit in access to the
lexicon and to an impairment of the semantic warehouse. If the deficit is at the semantic level,
an effective rehabilitation method seems to be the Semantic Feature Analysis [6] which is
based on semantic suggestions such as belonging to categories, use of objects, synonyms, etc.
If, on the other hand, the deficit is at the level of the lexicon, the treatment is aimed at
restoring the phonological/orthographic representation of the word (Phonological
Components Analysis; [7]) through the administration of suggestions such as the initial
phoneme, the rhyme, etc. In many cases, rehabilitation therapists combine the two strategies
into integrated training.
These approaches are based on a vision of the mind and the cognitive system as a
computer that manipulates abstract symbols following predefined rules. However, for some
decades, the study of the mind and cognitive processes has led researchers to propose a new
theoretical paradigm called Embodied Cognition, according to which cognitive processes,
including language, are based on multimodal representations whose inputs derive from
activation of sensory and motor systems [8]. In other words, our mental operations, although
considered to be of "high level", such as reasoning, memory, and language itself, depend on
our body to the point that the same neural structures involved in sensory, perceptual, and
motor processes are also active when we elaborate concepts, make inferences, recall
memories, and use language. Over the years, this theoretical perspective has received robust
confirmation from experimental data, especially in regards to the interaction between the
motor system and the linguistic system. It is therefore inevitable to also think about the
repercussions that this research could have on enhancement strategies for the recovery of
language skills. Pulvermuller & Berthier [9], in a reference article in this area, suggest
important reflections that arise from the lessons learned from imaging studies that confirm
the involvement of the motor system in linguistic processes. If language and action are
connected, then the approach that structures the rehabilitation treatment could be overcome
only by acting on language skills in favor of a "neuroscience-based" strategy in which
language is acted upon by stimulating the motor system. Moreover, since anomic deficits are
often due to a weak link between the meaning of the word and its lemma, the Hebbs’
principles of coincident and correlated learning can be exploited, i.e., by intensifying the
synchronous activation of both systems (lexicon and semantics) and connecting them with
the motor counterpart. Previous studies on aphasic patients have shown how treatment
through the use of observation-execution of actions is beneficial for the recovery of naming
skills [10,11]. With this study, we want to take a further step in the rehabilitation of anomia
by exploiting the connection between the motor and linguistic systems on the one hand and
the potential offered by new technologies on the other.

2. Method

2.1 Subjects
Subjects are aphasic patients aged between 18 and 80 years with naming disorder in the
post-acute phase (index event occurring at least 6 months earlier) and who have already
received rehabilitation treatment in the acute phase. The naming disorder will be evaluated
with the oral naming subscales of nouns and verbs of the ENPA battery (scores below the
cut-off: 8.2 and 6.1 respectively). Patients will be selected from among those requesting
access to outpatient aphasia rehabilitation treatment.
Patients with the following conditions will be excluded from the study: concomitant or
pre-existing (with respect to the index event) neurological and psychiatric deficits, epilepsy,
balance disorders, neglect, and impaired reading and writing.
Repetto et al. / A virtual reality intervention for anomia rehabilitation in post-stroke patients 113

The analysis of the calculation of the sample size indicated the number of 40 patients as
necessary to have a statistical power of the ANOVA with mixed design (2x2) equal to 0.96,
considering a Cohen's average effect size d = 0.6 and setting an alpha limit = 0.05.

2.2 Material

We built 80 short sentences including a verb with a first singular person and an object.
The sentences describe everyday actions such as “I water the plants”. For each sentence, a
360° video has been recorded displaying the action in the first-person perspective. For this
purpose, the camera was placed in the middle of the forehead of the actor performing the
action. As a result, during the playback, the user has the impression of being the agent of that
action. In addition, the same action has also been recorded from the third-person perspective
in a standard non-spherical video. In watching this video, the user sees someone else
performing the action. Both videos have been enriched with the written sentence and an audio
file of the spoken sentence.

2.3 Assessment
The baseline assessment will include the following (T0):
• Aphasia examination using the ENPA battery [12]. It consists of a battery of
tests that evaluate different areas of language: repetition, reading, writing,
naming, comprehension, numbers, and calculation. The obtained profile
evaluates the presence, severity, and main characteristics of aphasia in the
subject examined. This evaluation serves as a screening to identify eligible
patients and as an initial assessment of the anomic deficit.
• A health-related quality of life measure (EQ-5D-3L [13]). It consists of 2 parts:
the EQ-5D descriptive system and the EQ visual analog scale (EQ VAS). The
description system includes the following five dimensions: mobility, personal
care, habitual activities, pain/discomfort, and anxiety/depression. Each
dimension has 3 levels: no problems, some problems, and extreme problems.
The VAS EQ records the patient's self-estimated health on a vertical visual
analog scale where extremes are labeled "best health imaginable" and "worst
health imaginable".
• Functional Outcome Questionnaire - Aphasia (FOQ-A, [14]), which measures
the perception that family members have of the patient's communication skills
in everyday contexts. The questionnaire consists of 32 items that evaluate
different dimensions of communication skills: ability to communicate basic
needs, ability to make routine requests, ability to communicate new
information, attention, and other communication skills.
• Before starting the training, patients will undergo an initial test which will be
used to select the set of personalized stimuli based on their real difficulties. In
order to select the linguistic stimuli that will be used during the specific
training of each patient, all the stimuli prepared on standard video will be
presented 3 times on consecutive days and their naming will be requested.
Stimuli that is named at least once will be excluded from the treatment set.
From the remaining stimuli, a maximum of 50 will be selected for training.
The post treatment assessment (T1) will include the same measures. The treatment
protocol is illustrated in Figure 1.
Repetto et al. / A virtual reality intervention for anomia rehabilitation in post-stroke patients
114

Figure 1: The treatment protocol.

2.4 Treatment
The treatment will be done 3 times a week for 4 weeks. Patients will be randomly
assigned to one of the two treatment conditions: the Experimental Group (EG) will undergo
the 360° video-based treatment, whereas the Active Control group (AC) will undergo the
standard video-based treatment. For both groups, the videos will be displayed through the
Oculus Go to make the two conditions more uniform for all the contextual variables that are
not specifically manipulated at the experimental level. The patient's task is to listen to the
sentence pronounced while observing the scene, and to read the written sentence appearing
at the end of the video.
In each session, all the stimuli to be rehabilitated will be presented in random order.

3. Expected Results

As a primary endpoint, we expect greater improvement in naming abilities in patients

trained with 360° video-based treatment compared to those undergoing standard video
training. As a secondary endpoint, we expect greater improvement in the quality of life in
relation to the state of health and communication skills reported by the caregiver in patients
undergoing training with 360° videos compared to those undergoing training with standard
videos.

References

[1] Rosenbek JC, Lapointe LL, Wertz RT. Aphasia: A clinical approach. In: Aphasia:
A clinical approach. Pro Ed; 1989.
[2] Hilari K. The impact of stroke: Are people with aphasia different to those without?
Disability and Rehabilitation. 2011;33(3):211–8.
[3] Hofgren C, Björkdahl A, Esbjörnsson E, Stibrant-Sunnerhagen K. Recovery after
stroke: Cognition, ADL function and return to work. Acta Neurologica
Scandinavica. 2007;115(2):73–80.
[4] Parr S. Living with severe aphasia: Tracking social exclusion. Aphasiology.
2007;21(1):98–123.
Repetto et al. / A virtual reality intervention for anomia rehabilitation in post-stroke patients 115

[5] Goodglass H, Kaplan E, Barresi B. BDAE-3: Boston Diagnostic Aphasia

Examination–Third Edition. 3rd ed. Philadelphia:PA, PA: Lippincott Williams &
Wilkins.; 2001.
[6] Boyle M, Coelho CA. Application of Semantic Feature Analysis as a Treatment for
Aphasic Dysnomia. 1995 Nov;4(4):94–8.
[7] Leonard C, Rochon E, Laird L. Treating naming impairments in aphasia: Findings
from a phonological components analysis treatment.
http://dx.doi.org/101080/02687030701831474. 2008 Sep;22(9):923–47.
[8] Barsalou LW. Grounded cognition. Annual Review of Psychology. 2007/08/21.
2008;59:617–45.
[9] Pulvermuller F, Berthier ML, Pulvermüller F, Berthier ML. Aphasia therapy on a
neuroscience basis. Aphasiology. 2008/10/17. 2008;22(6):563–99.
[10] Marangolo P, Cipollari S, Fiori V, Razzano C, Caltagirone C. Walking but not
barking improves verb recovery: Implications for action observation treatment in
aphasia rehabilitation. PLoS ONE. 2012;7(6):1–8.
[11] Marangolo P, Bonifazi S, Tomaiuolo F, Craighero L, Coccia M, Altoè G, et al.
Improving language without words: First evidence from aphasia.
Neuropsychologia. 2010;48(13):3824–33.
[12] Miceli G, Laudanna A, Burani C, Capasso R. Batteria per l’analisi dei deficit afasici.
Cepsag, Roma. 1994;
[13] Janssen B, Szende A. Population norms for the EQ-5D. In: Self-Reported
Population Health: An International Perspective Based on EQ-5D. Springer
Netherlands; 2014. p. 19–30.
[14] Spaccavento S, Cafforio E, Cellamare F, Colucci A, Di Palma A, Falcone R, et al.
Italian adaptation of the functional outcome questionnaire–aphasia: initial
psychometric evaluation. Disability and Rehabilitation. 2018 Nov 20;40(24):2925–
30.
 Annual Review of Cybertherapy and Telemedicine 2021 117

Rescripting emotional eating with virtual

reality: a case study
Clelia MALIGHETTIa,1, Ciara SCHNITZERb, Georgia POTTERb, Katherine NAMETHc,
Theresa BROWNb, Emily VOGELb, Giuseppe RIVAa,d, Cristin RUNFOLAc, Debra
SAFERc

a
Department of Psychology, Catholic University of the Sacred Heart, Milan, Italy.
b
PGSP-Stanford PsyD Consortium, Palo Alto, CA, United States.
c
Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA,
United States.
d
Applied Technology for Neuro-Psychology Lab, Istituto Auxologico Italiano, Italy.

Abstract. Emotional eating (EE), or eating in response to negative emotions, is a

manifestation of difficulties with emotion regulation (ER) for individuals with eating
disorders. Although existing ER-focused interventions for eating disorders show promise
in reducing symptoms, a significant subset of patients (~50%) experience suboptimal
outcomes, with residual symptoms including EE. Innovative approaches targeting ER,
and its manifestation as EE, are needed. We developed a novel ER-focused virtual reality
(VR) immersive intervention aimed at reducing EE among individuals with a history of
a diagnosed binge-type eating disorder, designed to be delivered over seven weekly
sessions. Here, we present a case study of a patient with binge-eating disorder who
received this intervention in a “real-world” clinical setting. The immersive experiences
were delivered by a doctoral-level psychology student using a stand-alone head-mounted
display (HMD). They included rescripting techniques aimed to increase awareness and
identification of emotional states, resilience, and self-control over eating. Pre- and post-
treatment assessments were administered to evaluate EE (Dutch Eating Behavior
Questionnaire, DEBQ), eating disorder symptoms (Binge Eating Scale, BES), and
emotion dysregulation (Difficulties in Emotion Regulation Scale, DERS). Results show
a decreasing trend in emotional eating, emotion dysregulation, and binge episodes at the
end of the treatment compared to baseline. This single case suggests value in further
evaluation of this novel emotion regulation approach.

Keywords. Emotional Eating, Virtual Reality, Eating Disorders, Emotion Regulation,

Binge Eating

1. Introduction

Emotion regulation (ER), a multifaceted process, is defined as the initiation,

maintenance, and modification of the occurrence, intensity, and duration of feeling states [1].
ER has emerged as a significant transdiagnostic construct across eating disorders and a
relevant intervention target that has proven to be largely modifiable and thus therapeutically
addressable. One of the key difficulties with ER in eating disorders is emotional eating (EE),
commonly defined as eating in response to negative emotions [2]. Although existing ER-
focused interventions for eating disorders show promise in reducing symptoms, a significant
subset of patients (~50%) experience suboptimal outcomes, with residual symptoms
including EE. Thus, innovative approaches targeting ER and its manifestation as EE are
needed.

1
Corresponding Author, [email protected]
Malighetti et al. / Rescripting emotional eating with virtual reality: a case study
118

Virtual reality (VR)-enhanced treatments for eating disorders show superior efficacy,
faster results, and better maintenance compared to non-VR delivery [3-6]. Putative
mechanisms of action involve heightened presence and greater emotional arousal which in
turn facilitate reappraisal of stimuli and modification of emotional responses that motivate
and maintain new actions. Another strategy to improve ER-focused interventions for
eating disorders involves the use of immersive experiences with techniques rooted in
cognitive science and basic research such as imagery rescripting, metaphorical journeys, and
mindfulness. Imagery rescripting techniques effectively change emotional core beliefs [7]
and the associated somatic markers [8]. Somatic markers are physical sensations that are
associated with emotions. The somatic marker hypothesis proposes that emotional processes
guide behavior and strongly influence decision-making [8]. Re-scripting the somatic marker
helps the patient have a new emotional experience that involves the body and thus could
influence the patient's behavior and decisions (e.g., EE episodes). Immersive metaphorical
journeys, which involve overcoming personal obstacles, provide opportunities to acquire new
ER strategies or rediscover existing skills. Mindfulness-focused eating disorder treatments
significantly reduce EE and binge eating [9] with putative mechanisms including greater
mindful awareness of internal experiences. The present novel ER-focused VR immersive
intervention was developed to integrate these strategies to address EE in an immersive
environment. To date, almost all studies of immersive technologies with eating disorders
were conducted under highly controlled research settings. Here we present a preliminary case
report of a patient with EE and binge-eating disorder who received a novel intervention aimed
at increasing awareness, identification of emotional states, and self-control over eating. The
intervention was delivered in a “real-world” clinical setting.

2. Methods

2.1. Participant

Marina (a pseudonym) is a 54-year-old woman with EE and DSM-5 binge-eating

disorder who was referred from the Adult Eating Disorders Program within the Stanford
University School of Medicine. Marina had tried multiple treatments to address her
symptoms (1 year of prior cognitive behavior therapy and 6 months of dialectal behavior
therapy) yet complained of ongoing EE and binge eating. The participant was informed about
the purpose of the research and consented to participate as a pilot patient of a larger and novel
ER-focused immersive intervention study.

2.2. Measures

An assessment battery was administered pre- and post-the 7-session ER-focused VR

immersive intervention. Specifically, the Dutch Eating Behavior Questionnaire emotional
eating sub scale (DEBQ-EE) [10] was used to evaluate emotional eating, the Binge Eating
Scale (BES) [11] to evaluate binge eating symptom severity, the Difficulties in Emotion
Regulation Scale (DERS) [12] to evaluate emotion dysregulation, and the Eating Disorder
Recovery Self-Efficacy Questionnaire (EDRSQ) [13] to assess self-efficacy to cope with
eating disorder symptoms. In addition, the Beck Depression Inventory Scale (BDI-II) [14]
was used to assess the presence of depressive symptoms, the Rosenberg questionnaire [15]
to evaluate self-esteem, and the State-Trait anxiety inventory (STAI Y1, STAI Y2) [16] to
assess the presence of state and trait anxiety.

2.3. Treatment

The ER-focused VR immersive intervention includes an initial assessment to allow

greater personalization of the treatment (one session) followed by an emotion regulation
component (two sessions) and an emotional re-scripting component (four sessions) for a total
of seven 50–60-minute sessions. The emotion regulation component aims to increase the
Malighetti et al. / Rescripting emotional eating with virtual reality: a case study 119

patient’s ability to recognize emotional states. Both sessions begin with a focus on attention
exercise in which the therapist reads a script to help the patient focus on the present moment
and bodily sensations. This exercise is followed by the immersive ER VR experience in
which the patient is guided to implement mindfulness-based strategies to explore landscapes
while identifying internal states. The 3-D immersive videos were administered through a
stand-alone head mounted display, Oculus Go. Afterwards, the therapist helps the patient
identify a safe place by recalling the immersive experience and identifying a moment where
they experienced pleasant emotions. Finally, the anchoring phase consists of the therapist
helping the patient to link the positive emotion experienced in the virtual environment to a
real-life experience. Once the patient retrieves the real-life experience, the therapist
encourages them to generate vivid details about the real-life experience while making a hand
gesture (closing their thumb between their four fingers) in order to support the patient's ability
to re-experience this moment outside of session when needed. The emotional re-scripting
component aims to increase the patient’s confidence in their ability to experience and
skillfully manage challenging emotional states. Each session began with an initial focus on
attention exercise, followed by an immersive metaphorical journey that was designed toward
healing, with the patient facing obstacles and acquiring or rediscovering skills to achieve a
certain outcome. The patient is asked to hold a real object (e.g., key), a sensorimotor
reinforcement aimed to increase involvement of the body in the immersive experience. After
the video is viewed, the patient is asked to identify their emotions and localize them in their
body. If the patient predominantly experiences negative emotions, the therapist leads the
patient through a desensitization experience to help them reduce the negative experience
through the use of awareness, acceptance, and physicalizing exercises. If the predominant
emotion after the immersive experience is positive, the therapist first amplifies the emotion
and then anchors it to a real-life experience of the patient. At the conclusion of each session,
the patient is given an mp3 audio version of the session and asked to listen to it while holding
the real-life object associated with the scene/emotion.

3. Results

Results showed a decreasing trend in emotional eating (DEBQ-EE score), emotion

dysregulation (DERS score), depression (BDI score), and anxiety (STAI Y1-Y2) from
baseline to end of treatment. Furthermore, the patient showed an increasing trend in self-
esteem (Rosenberg) and confidence to eat and perform eating-related activities (EDRSQ
score) (see Table 1). The objective changes observed matched the patient's subjective
experience of her emotional and cognitive shifts. For example, the first two sessions focused
on emotion regulation helped her regain the ability "to breathe". Anxiety, as she called it,
prevented her from letting her breath reach her abdomen; it stayed in her throat. For some
time, Marina had been experiencing panic attacks that led her to faint. The feeling of suddenly
losing consciousness, with dizziness and lightheadedness, led her to turn to food as a possible
solution to this dizziness, which led her to binge eat. The mindfulness exercises and the
immersive experiences guided Marina in learning how to relax and become aware of the
sensations in her body. She learned to welcome these sensations rather than fear them, and
to relax the parts of her body that were holding tension and anxiety. Anchoring the positive
emotion in the present and safe place allowed her to recover an old sense of freedom and
independence that she had long buried in her memory. In the first two rescripting sessions,
Marina experienced mainly negative emotions, including a longstanding fear associated with
facing obstacles and of being insufficiently able to overcome them. Concretizing the fear by
giving it a color, a shape, and a name, allowed her to see it recede from herself and to remove
it from the bodily sensation of immobility she had been experiencing for some time. Through
these experiences, she was able to give a new meaning and a role to her fears, rewriting them.
The two positive emotions that emerged in subsequent sessions were a strong sense of
freedom and awe. Through the amplification of these two sensations, Marina reported feeling
Malighetti et al. / Rescripting emotional eating with virtual reality: a case study
120

safe, calm, and confident in herself. Therefore, at each stage of the journey, Marina reinforced
emotional stability, resulting in better management of her emotions.

Table 1: Pre-post treatment scores

Measure Pre-treatment Post-treatment Cut-offs
score score

DEBQ-EE 5 1.07 < 3.25= absence of clinical EE

DERS 145 57 > 80= clinical range

BDI 32 5 <10 = absent; 10-19 = mild; 20-29 =

moderate; >30 = severe

BES 39 1 <17 = absent; >17 = possible; >27 =

binge eating disorders diagnosis

STAI Y1 74 28 0-40 = absent; 40-50 = mild;

50-60 = moderate; > 60 =
severe

STAI Y2 76 36 0-40 = absent; 40-50 = mild; 50-60 =

moderate; > 60 = severe

EDRSQ- Norm Eat 3.2 4.8 No published cut off scores available

EDRSQ- Body I 1.3 2.8 No published cut off scores available

Rosenberg 18 36 <15= low self-esteem

BDI=Beck Depression Inventory; BES=the Binge Eating Scale; DERS= the Difficulties in Emotion Regulation
Scale; DEBQ-EE=the emotional eating subscale of Dutch Eating Behavior Questionnaire; STAI Y1, STAI-Y2=the
State-Trait anxiety inventory; EDRSQ=the Eating Disorder Recovery Self-Efficacy Questionnaire

3. Conclusion

This case report offers a preliminary account of a novel ER-focused VR immersive

intervention in a patient with BED who, despite prior treatment, continued to experience
emotional eating as well as residual binge eating. Our results showed that the patient had a
positive response to the intervention, as indicated by shifts in pre/post scores on the
questionnaires. In addition, the patient reported increased self-awareness of her emotional
states, increased self-confidence, decreased urges to emotionally eat, and cessation of binge
eating by the end of treatment. These positive findings support further research of this novel
intervention in larger patient samples that include comparison controls and long term follow
up.

References

[1] Gross JJ. The emerging field of emotion regulation: an integrative review. Rev Gen Psychol. 1998; 2: 271-
299.
[2] Frayn M, Livshits S, Knäuper B. Emotional eating and weight regulation: a qualitative study of compensatory
behaviors and concerns. J Eat Disord.2018; 6: 1-10
[3] Brown T, et al. Bringing virtual reality from clinical trials to clinical practice for the treatment of eating
disorders: an example using virtual reality cue exposure therapy. J Med Internet Res.2020; 22: e16386.
[4] Riva G, Malighetti C, Serino S. Virtual Reality in the Treatment of Eating Disorders. Clin Psychol Psychoter.
2021; 28(3): 477-488
Malighetti et al. / Rescripting emotional eating with virtual reality: a case study 121

[5] Malighetti C, et al. Manipulating body size distortions and negative body related memories in patients with
anorexia nervosa: a virtual reality-based pilot study. Annu Rev CyberTherapy Telemed. 2020; 18: 177-181
[6] Chirico A, Malighetti C, et al. Towards an advancement of multisensory integration deficits in anorexia
nervosa: Exploring temporal discrimination processing of visuo-auditory stimuli. Annu Rev CyberTherapy
Telemed. 2019; 17: 53-58
[7] Riva G. The key to unlocking the virtual body: virtual reality in the treatment of obesity and eating disorders.
J Diabetes Sci Technol. 2011; 5: 283-29.
[8] Damasio AR. The somatic marker hypothesis and the possible functions of the prefrontal cortex. Philos T R
Soc Lond. B. 1996; 351:1413-1420.
4. Lattimore P. Mindfulness-based emotional eating awareness training: taking the emotional out of eating. Eat
Weight Disord. 2020; 25: 649-657.
[9] Van Strien T, et al. The Dutch Eating Behavior Questionnaire (DEBQ) for assessment of restrained, emotional,
and external eating behavior. Int J Eat Disord. 1986; 5:295-315.
[10] Timmerman GM. Binge eating scale: further assessment of validity and reliability 1. J Appl Biobehav
Res. 1999;4: 1-12.
[11] Dan-Glauser ES. The difficulties in emotion regulation scale (DERS). Swiss J Psychol. 2012.
[12] Pinto AM, et al. The Eating Disorder Recovery Self-Efficacy Questionnaire (EDRSQ): change with treatment
and prediction of outcome. Eat Behav. 2008; 9:143-153.
[13] Steer RA, et al. Mean Beck Depression Inventory–II scores by severity of major depressive episode. Psychol
Rep. 2001; 88:1075-1076.
[14] Goldsmith RE. Dimensionality of the Rosenberg self-esteem scale. Soc Behav Pers. 1986;1: 253.
[15] Spielberger CD, et al. The factor structure of the State-Trait Anxiety Inventory. In: Sarason IG, Spielberger
CD, editors. Stress and Anxiety. Hemisphere: Washington, DC, USA, 1980. p.23
SECTION VI

WORK IN PROGRESS

It is important to emphasize the importance of

developing technological strategies (such as
artificial intelligence or augmented reality) that
can provide either new enhanced experiences or
technological systems also nurtured by artificial
intelligence techniques developed by humans.

These new mixed ICT tools might evolve into

experts in “helping others,” with the objective of
making our net-shared experience increasingly
more competitive, creative, and capable in the task
of helping others. Of course, this has significant
ethical implications, which will also need to be
explored at greater depth.

Botella, Riva, Gaggioli, Wiederhold,

Alcaniz, and Banos, 2012
 Annual Review of Cybertherapy and Telemedicine 2021 125

The immersive 3D objects’ library for

applying non-invasive brain stimulation in
research on the motor control and the
mirror neurons system: a call for
collaboration1
Andrey V. VLASOVa,b,c,†, Fanir KILMUKHAMETOVd, and Matteo FEURRAb
a
Izmerov Research Institute of Occupational health;
b
Center for Cognition and Decision Making,
Institute for Cognitive Neuroscience, HSE University;
c
Moscow Center for Diagnostics and Telemedicine;
d
DeMontroyal.

Abstract. We have developed and tested a library of 3D objects for the study of
motor control in immersive reality (https://p3d.in/TK7k3). The use of this stimulus
material (e.g., as on the Figure 4a) opens up new opportunities for evaluating
physiological parameters when using the method of neurotherapy in virtual reality.
In combination with the Non-Invasive Brain Stimulation (NIBS) methods such as
Transcranial Magnetic Stimulation (TMS), we propose to explore the effects of
functional activity of the mirror neuron system on a large scale for further
approbation of advanced and promising neurorehabilitation protocols.

Keywords. 3D, Mirror Neurons System (MNS), Non-Invasive Brain Stimulation

(NIBS), Immersive Reality for Neurorehabilitation

1. Introduction

The study of the motor control in immersive reality [environments] allows us to

identify physiological differences between traditional approaches of neurotherapy (for
example, mirror therapy after a stroke) and "non-traditional" using virtual reality (VR)
coupled with TMS for the functional effect(s) of the mirror neuron system (MNS). Also,
the activation of the human MNS can be a good objective criterion for more widely
practical application in rehabilitation programs.

2. The research of the MNS system with NIBS

The classic approach to the investigation of MNS involves single pulse TMS
application over the motor cortex in order to elicit motor evoked potentials (MEPs) from
the contralateral (to the site of stimulation) arm muscles in human volunteers. This
requires the subject to observe an experimenter grasping objects, so called "transitive
hand actions", or performing meaningless arm gestures – "intransitive arm movements".

1
Acknowledgement: the researchers used facilities of the HSE unique equipment (Reg. №354937)
supported by the RF Government grant (ag. №075-15-2021-673), partially it was supported (granted) by
DeMontroyal.
†
Corresponding Authorç [email protected].
Vlasov et al. / The immersive 3D objects’ library for applying non-invasive brain stimulation in research on
the motor control and the mirror neurons system: a call for collaboration
126
Evidence (of Fadiga et al., 1995 [1]) has shown that the observation of both transitive
and intransitive actions determined an increase of the recorded MEPs [2]. Therefore, we
have developed 3D hands in open-source 3DCG software toolset Blender version 2.92.01
(Blender Foundation, Netherlands) which were sketched from a scratch, modeled
and mashed, and integrated in a simple environment (a light table with a white ball and
grey-colored interior). The grasping movement is animated to the hand as well.
Additionally, we have created 2 to 4-second video clips involving the hands’ movement
(you can see 2D shots from clips on Figures 1a, 1b, 1c, 2a, 2b, and 2c). A hand-made
hand is an (virtual) asset. It may be exported and used in the process of developing an
experimental design on a virtual platform (e.g., on cross-platform game engine Unity).

2.1. Action observation and the MNS

The mirror neurons are not reacting to the mere observation of a movement (purely
motoric), but their activity is also related to the intention of the action execution. This
account identifies a precise role for the MNS in our ability to infer intentions from other
people’s actions [3]. Here, we have developed the stimuli for NIBS testing in a VR-
contained hand grasping a ball (the movement from the front view point, 0°: Figures 1a-
1c, and the movement from the opposite view point, -180°: Figures 3a-3c). So, we had
prepared and adopted those stimuli for testing subjects in the project studying the effects
of MNS. Interestingly, by animating the 3D hand and customizing movements as a
constructor (see library: [4]), the researcher might play with them or create a design of
his/her study in an easy way. Importantly, any design will be standardized and circulated
(lab2lab) ecologically.

2.2. Study of the empathy for pain, racial bias, and the MNS

Evidence showed a reduction in amplitude of MEPs which is specific to the muscle

of a hand being pricked during observation [5]. The empathic behavior may be based on
‘mirror-matching’ simulation of others’ states. Its inference about the sensory qualities
of others’ pain and their automatic embodiment in the observer’s motor system may be
crucial for the social learning of reactions to pain [5]. For testing the empathy for pain,
we have developed (other) models of damaged hands (Figure 4a).
Additionally, it has been shown that observation of empathy for pain-related stimuli
from black and white humans elicits an implicit racial bias [6]. The research found that
observing pain in the same racial group induced a reduction in corticospinal excitability
that is specific to the muscle (specific location of the observed hand) which participants
observed being affected. For testing the racial bias of the MNS, we have created an
additional model of a hand with black skin (Figure 4b) wherein we switch the model of
the hand and use it for the same grasping movements (presented in Figures 1a-1c)
according to different race groups (i.e., African-American and Caucasian).

Figure 1a. Grasping a ball (1st frame of movement). Figure 1b. Grasping a ball (2nd). Figure 1c. Grasping a
ball (3rd). CC BY-NC-ND [4]

Fig. 2a. Backing a ball (1st frame of movement). Figure 2b. Backing a ball (2nd). Figure 2c. Backing a ball
(3rd). CC BY-NC-ND [4]

1
https://www.blender.org/
Vlasov et al. / The immersive 3D objects’ library for applying non-invasive brain stimulation in research on
the motor control and the mirror neurons system: a call for collaboration 127

Figure 3a. Grasping a ball: opposite view (1st frame of movement). Figure 3b. Grasping a ball: opposite
view (2nd). Figure 3c. Grasping a ball: opposite view (3rd). CC BY-NC-ND [4]

Figure 4a. Damaged hands. CC BY-NC-ND [4]

Figure 4b. Hands with white skin and hands with black skin. CC BY-NC-ND [4]

3. The advanced research with NIBS in immersive reality: a call for collaboration

We have developed a set of 3D objects and examples of standard movements (e.g.,

Figures 1a-1c; for deep exploring go to: [4]) for its application in neuroscience research.
It is by applying this library in collaborative work that an immersive and unified stimuli
approach could be tested and implemented widely. 3D virtual stimuli can be useful in
advanced study of neurotherapy and personalized remote neurorehabilitation. We
propose to scientific groups with a specialty in (human) motor control to discuss and
share ideas about mirror therapy and VR in order to create novel neurophysiological
projects. We expect that the mirror neurons’ effect measured using TMS-induced MEPs
can be a predictor of the effectiveness of the therapy. For example, by using a training
MNS’ protocol in assisted VR, measured MEP index will allow for determining and
specifying an optimal approach to the use of therapy, including in VR.

References

[1] Fadiga L, Fogassi L, Pavesi G, Rizzolatti G. Motor facilitation during action observation: a magnetic
stimulation study. J. Neurophysiol. 1995; 73(6): 2608-11. doi: 10.1152/jn.1995.73.6.2608.
[2] Feurra M, Blagovechtchenski E, Nikulin VV, Nazarova M, Lebedeva A, Pozdeeva D, Yurevich M, Rossi
S. State-Dependent Effects of Transcranial Oscillatory Currents on the Motor System during Action
Observation. Sci. Rep. 2019; 9: 12858. doi: 10.1038/s41598-019-49166-1
[3] Iacoboni M, Molnar-Szakacs I, Gallese V, Buccino G, Mazziotta JC, Rizzolatti G. Grasping Intentions
with Mirror Neurons. PLoS Biol. 2005; 3(3), e79. doi: 10.1371/journal.pbio.0030079
[4] CYPSY 3D objects’ library: https://github.com/toandreyhse/cypsy/ Error! Hyperlink reference not
valid.
[5] Avenanti, A, Bueti D, Galati G, Aglioti SM. Transcranial magnetic stimulation highlights the
sensorimotor side of empathy for pain. Nat. Neurosci., 2005; 8: 955–960. doi: 10.1038/nn1481
[6] Avenanti A, Sirigu A, Aglioti SM. Racial bias reduces empathic sensorimotor resonance with other-race
pain. Curr Biol. 2010; 20(11): 1018-22. doi: 10.1016/j.cub.2010.03.071
 Annual Review of Cybertherapy and Telemedicine 2021 129

Meta Cognition on the Internet: Expected

Meta-Accuracy for Human and AI Virtual
Assistants’ First Impressions about Us1,2
Elena TSANKOVAa,3 and Ergyul TAIR a
a
Institute for Population and Human Studies, Bulgarian Academy of Sciences

Abstract. On the Internet, we encounter different types of interaction partners –

some human, some AI. How we believe they perceive us upon our very first
encounter sets the ground and may determine the outcome of the subsequently
unfolding interaction. In a between-subject online experimental design, we studied
the expected accuracy with which imagined human and AI Internet virtual assistants
form first impressions. We measured expected meta-accuracy as the absolute
difference between participants’ self- and meta- (perceived interaction partner’s
viewpoint) perception along the two central dimensions of social cognition (warmth
and competence). We anticipated that 1) expected meta-accuracy would be higher
for human than for AI virtual assistants and 2) expected meta-accuracy would
increase with increasing perceived human-mind-like abilities of the virtual assistant.
We found proof for the second line of predictions and uncovered evidence for
primacy of the warmth dimension in Internet meta cognition. Our results are in line
with and extend previous knowledge about mind perception to the field of Internet
meta cognition. Our work is pioneering in the study of meta cognition of first
impressions on the Internet and promises an interesting route for uncovering
previously unexplored aspects of online communication. With the ultimate goal of
applying theoretical knowledge to the practical facilitation of Internet
communication, we present ideas for further exploration of the area of online meta
cognition. Our suggestions include different operationalizations of the primary
phenomenon of interest, varying embodiment levels, and strengthening
experimental manipulation to probe for the Uncanny Valley phenomenon in Internet
meta cognition.

Keywords. Meta Cognition, Meta-Accuracy, First Impressions, Internet, Virtual

Assistants

1. Introduction

How we believe others perceive us on the Internet sets the ground for our subsequent
interactions with them. Here we studied the expected accuracy with which people
understand the first impressions that human and AI virtual assistants (VAs) may form
about them (meta-accuracy). Given the central role of mind attribution to and felt
presence of VAs for their perception on the social and cognitive level, e.g., [1,2,3], we
predicted: 1) higher expected meta-accuracy in human than in AI VA encounters, and 2)
positive correlations between expected meta-accuracy, perceived human-mind-like
abilities of the VAs, expected meta-accuracy, felt VA presence, and previous VA
experience.To match the social and cognitive levels of VA perception, we
operationalized first impressions in terms of the universal dimensions of social
cognition—warmth and competence [4].

1
This research was supported by grant КП-06-ДБ-3 from the National Scientific Program "Petar Beron.
Science and Innovation with Europe" of The Bulgarian National Science Fund
2
Pre-registration protocol and extensive methodological detail: osf.io/rgxbq
3
Corresponding Author [email protected]
2. Method

In a between-subject design, 211 (106 female) participants aged 18–36 years (M =

24.39, SD = 4.95) participated in a simulated first-impression-type of online encounter
with a hypothetical human or AI VA. Participants imagined landing on a website that
relies on an unembodied VA for enhancing user experience. They read a welcome
message from the assistant and responded to the assistant’s request to introduce their
personality briefly in writing (so that the VA could better help them). Once participants
submitted their description, a brief pre-programmed message from the VA thanked them,
encouraged them to continue browsing the website, and to ask questions if they needed
assistance. This marked the end of the simulated encounter. Adhering to definitions of
warmth and competence [2], participants then rated on 7-point Likert-type scales (1 =
lowest extreme, 7 = highest extreme) as to how warm and how competent they believed
they had described themselves (self-perception). Next, on the same type of scales,
participants took the perspective of the VA and indicated as to how warm and how
competent they believed the VA had seen them (meta-perception). Finally, also on
similar scales, participants rated the VA’s ability to think and feel, the strength with
which they felt the VA’s presence, and their past experience with the type of VA they
had just encountered.

3. Results

We computed a Meta-Accuracy Score for each dimension (warmth and competence)

as the absolute value of the difference between participants’ self- and meta-perception.
A larger Meta-Accuracy Score signifies a larger difference between self- and meta-
perception and thus indicates lower expected meta-accuracy.
Contrary to our predictions, mean Meta-Accuracy Score for the human VA
condition was not higher than that for the AI VA condition for either dimension (one-
tailed independent samples Student’s t-tests, both ps > .1). However, participants in the
human VA condition believed they had described themselves as warmer than those in
the AI assistant condition (one-tailed independent samples Student’s t-test, t(209) =
−1.83, p = .035, Cohen’s d = 0.25; Mhuman = 4.96, SDhuman = 1.36; MAI = 4.62, SDAI
= 1.35). In addition, participants perceived the human VA as more capable of thinking
(one-tailed independent samples Student’s t-test, t(209) = −2.64, p = .004, Cohen’s d =
0.36; Mhuman = 4.52, SDhuman = 1.81; MAI = 3.87, SDAI = 1.74) and feeling (one-tailed
independent samples Welch’s t-test, t(187.84) = −3.64, p < .001, Cohen’s d = 0.50;
Mhuman = 3.19, SDhuman = 2.03; MAI = 2.30, SDAI = 1.51) than the AI VA. There was also
a tendency for the human assistant to be perceived as more present than the AI one (one-
tailed independent samples Welch’s t-test, t(208.09) = −1.59, p = .057, Cohen’s d = 0.22;
Mhuman = 3.78, SDhuman = 1.39; MAI = 3.45, SDAI = 1.56).
As predicted, meta-accuracy increased with increasing perceived human-mind-like
abilities and felt presence of the VA (Table 1). Experience with the respective VA type
was not significantly correlated with meta-accuracy (all ps > .1).

Table 1. Pearson correlation coefficients for the associations between expected meta-accuracy and perceived
virtual assistant human-mind-like abilities, felt VA presence, and past VA experience.
VA VA VA VA
Thinking Feeling Experience Felt
Ability Ability Presence

r = −.12* r = −.21** r = − .06 r = − .15*

Meta-Accuracy Warmth rhuman = −.15 rhuman = −.24** rhuman = − .06 rhuman = − .11
rAI = −.07 rAI = − .15 rAI = − .05 rAI = − .18*
r = −.12* r = − .14* r = − .01 r = − .12*
Meta-Accuracy Competence rhuman = −.08 rhuman = − .18* rhuman = − .12 rhuman = − .09
rAI = −.14 rAI = − .08 rAI = − .12 rAI = − .14
Note. All tests one-tailed, for negative correlation (higher Meta-Accuracy Score indicates larger difference
between self- and meta-perception and thus lower meta-accuracy)
* p < .05, ** p < .01, *** p < .001
Tsankova et al. / Meta Cognition on the Internet: Expected Meta-Accuracy for Human and AI Virtual
Assistants’ First Impressions about Us 131

4. Discussion

We present evidence that human-mind-like abilities of VAs are associated with

expected meta-accuracy of first impressions in Internet encounters. Our results are in line
with previous findings [1,2] and extend them into the field of Internet meta cognition.
Our observations also suggest primacy of warmth over competence [4] in meta cognition
online. We propose that further research explores broader operationalizations of meta-
accuracy [5] and includes varying degrees of embodiment [3] and human-likeness to
probe for the presence of the Uncanny Valley phenomenon in Internet meta-accuracy [1].
Our work undertakes the study of a largely unexplored but highly relevant in the
digital age area of human cognition – meta-accuracy of first impressions on the Internet.
We outline an initial idea of how the known principles of human-AI interaction may
guide meta cognition on the Internet.

References

[1] Stein J-P, Ohler P. Venturing into the uncanny valley of mind—The influence of mind attribution on the
acceptance of human-like characters in a virtual reality setting. Cognition. 2017 Mar;160:43-50.
doi:10.1016/j.cognition.2016.12.010
[2] Shank DB, Graves C, Gott A, Gamez P, Rodriguez S. Feeling our way to machine minds: people’s
emotions when perceiving mind in artificial intelligence. Comput. Hum. 2019 Sep;98:256-266.
doi:10.1016/j.chb.2019.04.001
[3] Kim K, Boelling L, Haesler S, Bailenson J, Bruder G, Welch GF. Does a digital assistant need a body? The
influence of visual embodiment and social behavior on the perception of intelligent virtual agents in AR.
In: Proceedings of the 2018 IEEE International Symposium on Mixed and Augmented Reality (ISMAR);
2018 Oct 16-20; IEEE, 2018, p.105-114. doi:10.1109/ISMAR.2018.00039
[4] Fiske ST, Cuddy AJC, Glick P. Universal dimensions of social cognition: warmth and competence. Trends
Cogn. Sci. 2007 Feb;11(2):77-83. doi:10.1016/j.tics.2006.11.005
[5] Carlson EN. Do psychologically adjusted individuals know what other people really think about them? The
link between psychological adjustment and meta-accuracy. Soc Psychol Personal Sci. 2016 Apr;7(7):717-
725. doi:10.1177/1948550616646424
 Annual Review of Cybertherapy and Telemedicine 2021 133

Virtual reality for relaxation in a pediatric

hospital setting: an interventional study
with a mixed-methods design
Sylvie BERNAERTSa,1, Bert BONROY b, Jo DAEMS b, Romy SELS b, Dieter
STRUYFc, Wessel VAN DE VEERDONK d
a
Expertise Unit Psychology, Technology & Society, Thomas More University of
Applied Sciences, Antwerp, Belgium
b
Mobilab & Care, Thomas More University of Applied Sciences Kempen, Geel,
Belgium
c
Creative and Innovative Business, Thomas More University of Applied Sciences,
Mechelen, Belgium
d
Access to care, Thomas More University of Applied Sciences, Mechelen, Belgium

Abstract. Accumulating evidence supports virtual reality (VR) as a feasible and

effective method to alleviate anxiety and pain in pediatric patients during specific
medical procedures. However, adoption of VR in clinical practice is limited. To
address implementation barriers, this intervention study with a mixed-methods
design focuses on the feasibility, acceptability, tolerability, and preliminary
effectiveness of Relaxation-VR, a VR application aimed to provide relaxation as it
is used for anxiety, stress, and pain reduction for children in hospital. Primary
outcomes include intervention completion, technical issues, the pediatric Simulator
Sickness Questionnaire (tolerability), and visual analogue scales (VAS) addressing
ease of use, likeability (feasibility), and future use (acceptability). Secondary
outcomes include pre-to-post-changes in the Self-Assessment Manikin, VAS, and
Faces Pain Rating Scale-Revised to measure happiness and stress, anxiety, and pain,
respectively. We present preliminary data of 51 participants of this ongoing study.
A minority of participants (10/51) quit the intervention prematurely for reasons
including discomfort, disliking the application, technical issues, and willingness to
see the medical procedure being performed. Only 5 out of 51 participants reported
technical issues including start-up issues and low battery levels. Ease of use,
likeability, and future use of the intervention were favorably scored. No adverse
events and minimal VR sickness symptoms were reported. Compared to baseline,
participants reported less anxiety, less pain, less tension, and more happiness while
using Relaxation-VR. These preliminary findings indicate that Relaxation-VR is
acceptable, feasible, and tolerable, and can reduce anxiety, tension and pain, and
increase happiness in pediatric patients with various medical conditions.

Keywords. Virtual Reality, Pediatrics, Relaxation, Anxiety, Stress, Implementation

1. Introduction

Accumulating evidence supports virtual reality (VR) as a feasible and effective

method to alleviate anxiety and pain for pediatric patients during specific medical
procedures (1,2). However, adoption of VR in clinical practice is limited due to multiple
implementation barriers. To improve translation from research to practice and address
implementation barriers, the current study focuses on the feasibility, acceptability,
tolerability, and preliminary effectiveness of Relaxation-VR, a VR application
(prototype) aimed to provide relaxation, used in this study to reduce anxiety, stress, and
pain for children in hospital as assessed by pediatric patients, their parents, and clinical
staff.

1
Corresponding author: [email protected]
Bernaerts et al. / Virtual reality for relaxation in a pediatric hospital setting: an interventional study with a
mixed-methods design
134

2. Methods

The study sample consists of 55 pediatric in- and outpatients aged 4 to 16 years,
recruited in two Belgian hospitals. Data of 51 participants has been included in analyses.
Relaxation-VR was used with different types of patients with various reasons for hospital
admittance (e.g., examination and/or treatment of gastrointestinal complaints,
appendicitis, eating disorder, surgery). 19 participants used Relaxation-VR during a
medical procedure, and 30 participants used Relaxation-VR as a means to relax during a
longer hospital stay (missing N=1).
Relaxation-VR is a VR application (prototype) aimed to relax its users. In this study,
we assess whether this application can be used to reduce anxiety, stress, and pain by
distracting the patient in a relaxing and interactive environment. Relaxation-VR consists
of three levels (breathing exercises, meditation exercises, interactive games) and is
administered via a commercially available VR headset (Oculus Go). To assess the
feasibility, acceptability, tolerability, and preliminary effectiveness of Relaxation-VR as
an intervention for a variety of pediatric patients, we conducted an interventional study
with a mixed-methods design. Primary outcomes include intervention completion,
technical issues, visual analogue scales (VAS) addressing ease of use and likeability
(feasibility), future use (acceptability), and the pediatric Simulator Sickness
Questionnaire and adverse event reporting (tolerability). These outcomes were measured
after completion of the VR intervention. Secondary outcomes include the Self-
Assessment Manikin to measure happiness and stress, VAS to measure anxiety, and the
Faces Pain Rating Scale-Revised to measure pain. These measures were administered at
baseline and after completion of the VR intervention to assess pre-to-post-changes. Data
collection is completed, but data analyses are ongoing.

3. Results

Concerning feasibility, a minority of participants (10/51) quit the intervention

prematurely for reasons including discomfort, disliking the application, technical issues,
and willingness to see the medical procedure being performed. Only 5 out of 51
participants reported technical issues including start-up issues and low battery levels.
Ease of use (M=15.31, SD=22.22) and likeability (M=13.92, SD=21.12) of the
intervention were favorably scored (0= easy to use/fun, 100= difficult to use/not fun at
all). Future use of the intervention (acceptability) was also favorably scored (M=16.10,
SD=26.19) (0= I want to use it again, 100= I do not want to use it again at all). Regarding
tolerability, no adverse events and minimal VR sickness symptoms were reported.
Compared to baseline, participants reported less anxiety, less pain, less tension (stress),
and more happiness while using Relaxation-VR (Table 1).
Table 1. Mean, standard deviation, t-value and p-value of the secondary outcome measures.
Baseline Post
Outcome measure N M (SD) M (SD) t-value
Anxiety (VAS) 50 32.82 (28.09) 14.34 (18.48) 5.53*
Pain (FPRS-R) 51 2.65 (2.37) 1.55 (1.69) 3.80*
Tension (SAM) 50 4.86 (1.90) 2.92 (2.33) 7.16*
Happiness (SAM) 51 5.94 (2.03) 7.29 (2.07) -4.99*
Note. FPRS-R = Faces Pain Rating Scale – Revised, VAS = Visual Analogue Scale, SAM = Self-Assessment
Manikin; *p<.001

4. Conclusion

These preliminary findings indicate that Relaxation-VR is acceptable, feasible, and

tolerable for a variety of pediatric patients and that the use of Relaxation-VR can reduce
anxiety, pain, and tension (stress), and increase happiness in pediatric patients with
various medical conditions.
Bernaerts et al. / Virtual reality for relaxation in a pediatric hospital setting: an interventional study with a
mixed-methods design 135

References

[1] Eijlers R, Utens EMWJ, Staals LM, de Nijs PFA, Berghmans JM, Wijnen RMH, et al. Systematic
Review and Meta-analysis of Virtual Reality in Pediatrics: Effects on Pain and Anxiety. Anesth
Analg. 2019;129(5):1344–53.

[2] Malloy KM, Milling LS. The effectiveness of virtual reality distraction for pain reduction: A
systematic review. Clinical Psychology Review. 2010;30(8): 1011–8.
SUBJECT INDEX Regenerative VR; 3
Relaxation; 133
Remote Meeting; 79
Social Media; 3; 73; 85
Social Stress; 41
SPANLANG; 105
3D objects; 125 STOKKING; 79
Affordances; 21 Stress; 133
AI Internet; 129 STRUYF; 133
Anorexia Nervosa; 99 Usability; 47
Aphasia; 111 Video Game; 35
Attentional Bias; 93; 99 Virtual Assistant; 129
Autism Spectrum Disorder; 47; 53 Virtual Reality; 3; 11; 41; 47; 61; 133
Avatar; 79 3
Body Dissatisfaction; 93; 99
Body Exposure Therapy; 99
Body Image; 73; 93; 99
Body Ownership; 11
Body Swapping; 105
Care Technology; 15
Cave Assisted Virtual Environment; 47
Children; 47; 67
Cognitive Stimulation; 67
COVID-19 pandemic; 29; 79
Cyberball-Game; 41
Dark Triad; 35
Dementia; 15
Drive for Thinness; 73
Eating Disorders; 93; 117
Ecological Validity; 41
Embodied Cognition; 111
Embodiment; 105
eMental Health; 29
Emotion Regulation; 117
Emotional Eating; 117
Empathic Interaction; 29
Empathy; 29
Ethics; 15
Executive Functioning; 61; 67
Eye-Tracking; 93; 99
Gaming Disorder; 35
Immersive 360° Videos; 111
Instagram; 73
Internet of Things; 3
Language; 111
Machine Learning; 53
Neurorehabilitation; 125
Non-Invasive Brain Stimulation; 125
Obesity; 105
Online Behavior; 85
Online Dating App; 85
Online Psychological Treatment; 29
Ostracism; 41
Photorealism; 79
Presence; 11
Randomized Controlled Trial; 67, 105
AUTHOR INDEX MARÍN MORALES; 53
MARTINI; 85
MEDAKOVSKAYA; 35
MINISSI; 47; 53
MIOCH; 79
ALCAÑIZ; 47; 53 MIQUEL; 93; 99
ANASTASIADOU; 105 MULVANEY; 41
ASBEE; 61 MUNSCH; 11
BARTOLOTTA; 3 NAMETH; 117
BERNAERTS; 133 NIAMUT; 79
BIERHUIZEN; 79 OLIVEIRA; 67
BOGACHEVA; 35 PARRAMON PUIG; 105
BONROY; 133 PARSONS; 61
BRANKAERT; 15 PORRAS-GARCIA; 93; 99
BRISEÑO-OLORIZ; 93 POTTER; 117
BROWN; 117 POWELL; 79
BÜLTER; 73 RAMOS-QUIROGA; 105
CAMPAGNA; 111 REGALIA; 85
CANCER; 111 REPETTO; 111
CARULLA-ROIG; 99 RIVA; 3; 85; 117
CHICCHI GIGLIOLI; 47; 53 RODELLA; 111
CHRISTOPHERS; 41 ROONEY; 41
CIUDIN; 105 RUBO; 11
COMAS; 105 RUNFOLA; 117
COVID-19 pandemic; 29; 79 SAFER; 117
DAEMS; 133 SAHRAOUI; 11
DE KORT; 29 SAJNO; 3
DI LERNIA; 3 SANSONI; 3
ELFRINK; 73 SANTOS; 67
EPISHIN; 35 SCHNITZER; 117
FEIJT; 29 SCIARA; 85
FERRER-GARCIA; 93; 99 SELS; 133
FEURRA; 125 SERINO; 3
FLETA-DÍAZ; 93 SERRANO-TRONCOSO; 99
GABRIEL; 73 SINGH; 93
GAGGIOLI; 3 SIRERA; 47; 53
GAMITO; 67 SLATER; 105
GEISE; 73 SPANLANG; 105
GHIŢĂ; 73 STOKKING; 79
GÓMEZ-ZARAGOZÁ; 53 STRUYF; 133
GREWE; 73 TAIR; 129
GUTIÉRREZ-MALDONADO; 93, 99 TAÑA-VELASCO; 93
IGLESIAS; 93 TSANKOVA; 129
IJSSELSTEIJNa; 15 TUMMERS-HEEMELS; 15
KILMUKHAMETOV; 125 VAN DE VEERDONK; 133
LEADER; 21 VAZQUEZ-DE SEBASTIAN; 105
LUSILLA-PALACIOS; 105 VLASOV; 125
MADDALON; 53 VOGEL; 117
MAIETTI; 111 WESTERHOF; 73
MALIGHETTI; 85; 117 WESTERINK; 29
MANTOVANI; 47 WIEDERHOLD; 3
.