Dynamic Interactive Social Cognition training in Virtual Reality (DiSCoVR) for people with a psychotic disorder

Understanding the social world in which we live, requires the ability to
interpret emotions, intentions and behavior of other people. Deficits in social
cognition, such as facial affect recognition (FAR), prosody and Theory of Mind
(ToM), the ability to attribute mental states to oneself and others and to
understand that others have beliefs, desires, and intentions that are different
from one's own, are common in people with psychotic disorders (Savla, Vella,
Armstrong, Penn, & Twamley, 2013). People with a psychotic disorder fail to use
important communicative cues and do not properly consider social contextual
information. This limits social functioning in work and relationships (Irani,
Seligman, Kamath, Kohler, & Gur, 2012); it causes difficulties in finding and
keeping a job, and establishing and maintaining relationships, as a result of
which they can become socially isolated.
SCT is aimed at improvement of social cognition through drill and
practice or strategy training. SCTs generally consist of different modules that
focus on a specific aspect of social cognition (e.g., learning about emotions,
understanding situations). Most SCTs feature a combination of repeated practice
(restorative approach) and learning strategies to cope with deficits
(compensatory approach) (Paquin, Wilson, Cellard, Lecomte, & Potvin, 2014).
Social cognitive abilities are taught in a hierarchical manner, starting with
fundamental skills (e.g., facial affect recognition) that are built upon with
increasingly complex skills. Meta-analyses (Kurtz, Gagen, Rocha, Machado, &
Penn, 2016; Kurtz & Richardson, 2012) have found that current SCT interventions
for people with psychotic disorders have a moderate to large effect on
performance on social cognition tasks.
The effects of SCT on daily life are most substantial when the
intervention is combined with psychosocial rehabilitation (Kurtz & Richardson,
2012; Medalia & Saperstein, 2013). The stimuli that are typically used in
training are distinct, however, from the dynamic, complex, interactive nature
of real social interactions. Both FAR and ToM training are typically provided
utilizing static pictures, text, comics or video clips. Thus, there is no
opportunity for the participant to interact with the stimuli. Additionally, the
tasks are isolated from other social and cognitive processes that occur
simultaneously in real life situations, such as social context, conversations,
body movements, background stimuli, limited time for responding before a
situation changes, etc. Providing SCT directly in patients* daily lives might
be more effective (i.e., having them practice with a therapist in real-life
social interactions), but this is not practically feasible (Peyroux & Franck,
2014); it would require therapists to accompany patients in a variety of
real-life social situations. Given the number of sessions typically needed,
this is too time- and resource-consuming, and it is highly inefficient to wait
for specific practice situations to occur. In addition, the presence of a
therapist changes the situation, which further limits real-life training.
A promising solution may be to provide SCT using Virtual Reality (VR).
The defining characteristic of VR is the experience of a sense of presence in
an interactive three-dimensional world. A computer generates an image, which is
presented to the user through a screen or a head-mounted display. The virtual
world changes according to the participant*s movements and actions, because a
tracker feeds back the user*s position and orientation to the computer to
update the image. The result is an interactive experience that elicits
psychological and physiological responses remarkably similar to those in the
real world (Veling, Moritz, & van der Gaag, 2014). This considerable ecological
validity offers a unique possibility to explore interactions between
individuals and complex everyday environments in an experimentally controlled
way. It allows practicing social cognitive skills in interaction with (virtual)
others.
VR may solve current problems in SCT, because it is sufficiently
realistic and interactive to induce genuine psychological reactions (Veling,
Moritz, et al., 2014), yet is controllable, which facilitates structured
training of social cognitive skills (such as FAR and ToM) in dynamic daily life
situations. In VR, specific situations or scenarios can be customized, repeated
and varied, and it allows direct feedback. VR has been shown to be feasible and
effective for treatment of mental disorders (e.g., anxiety; Opri* et al.,
2012), and it is a safe and valid way to study psychotic disorders (Veling,
Moritz, et al., 2014).(Veling, Moritz, et al., 2014). Pilot studies showed
promising treatment effects, including improvement of social skills in
psychotic disorders (Park et al., 2011; Rus-Calafell, Gutiérrez-Maldonado, &
Ribas-Sabaté, 2014). These studies did not specifically target social
cognition, however. A study (n=8) by Kandalaft, Didehbani, Krawczyk, Allen and
Chapman (2013) found significant improvements on several social cognitive
domains and social functioning after a VR SCT for young adults with autism
spectrum disorders. The extent to which a specialized social cognition training
is effective for people with a psychotic disorder remains unknown, however. One
case study (n=2) is available (Peyroux & Franck, 2016), in which two
participants received a VR SCT. Substantial improvements in a variety of social
cognitive and functional domains were observed. However, since this study
included only two participants, it serves mostly as an indication rather than
confirmation of the efficacy of VR SCT.
In recent years, our collaborating research group has developed and
tested VR social environments for research of psychotic disorders (Brinkman et
al., 2011; Veling, Brinkman, Dorrestijn, & van der Gaag, 2014). These
environments (café, bus, supermarket and street) were used to study social
behavior in VR (Counotte et al., 2017; Geraets et al., 2017; Veling,
Pot-Kolder, Counotte, van Os, & van der Gaag, 2016). Additionally, the virtual
environments were successfully utilized to provide a VR treatment targeting
paranoia and social anxiety in people with a psychotic disorder (Pot-Kolder,
Veling, Geraets, & van der Gaag, 2016). The café, supermarket and street worlds
served as a basis for developing the VR SCT environments.
In the previous year, we have conducted a pilot study on the
acceptability and feasibility of this VR SCT, called DiSCoVR (Dynamic
Interactive Social Cognition training in Virtual Reality, registered under:
NL55477.042.16). Preliminary analysis on the first twelve participants shows
that they enjoyed the intervention, thought it was moderately useful for daily
social contact, enjoyed the combination of VR and a therapist, and judged the
difficulty level of the intervention to be adequate. At the time of writing,
two of the twenty-one currently included participants have dropped out.
Preliminary testing suggests that DiSCoVR significantly improves emotion
perception and self-esteem. Our next step is to test DiSCoVR in a larger study,
to be able to draw conclusions on its efficacy on daily life social functioning
and social cognition.

The primary objective of the study is to determine whether DiSCoVR improves
social cognition. We are interested in the effect of DiSCoVR on the following
social cognitive domains:
- Emotion Perception.
- Social Perception.
- Theory of Mind.

The second objective of DiSCoVR is to investigate the impact of DiSCoVR on
social functioning. Additionally, the effect of DiSCoVR on secondary domains of
interest is investigated, including self-esteem, depression and social anxiety.

The study will be a multicenter Randomized Controlled Trial with two groups: a
group receiving the VR SCT, DiSCoVR, and an active control group, virtual
reality relaxation therapy (VRelax), matching the intervention as closely as
possible in terms of visits, content and attention received. Participants will
be randomly allocated to one of these groups. Both the intervention and the
active control conditions consist of sixteen 45 to 60-minute sessions, provided
over the course of eight weeks. Social cognition and social functioning will be
assessed at three points in time: at baseline, post-treatment, and at three
months after completing either DiSCoVR or VRelax. These assessments will be
performed by research assistants who are blind to the participant's study
condition. The DiSCoVR group will receive one additional assessment, between
sessions 1 and 2. In this assessment, a VR task and two questionnaires will be
completed.

The SCT consists of sixteen sessions, which last 45-60 minutes each. During
these sessions, participants in the VR SCT group navigate virtual environments
developed to train social cognitive skills. The VR SCT is provided by a
therapist, that is, a psychologist or other mental healthcare professional who
has been trained to apply the treatment protocol. This therapist has the
following tasks:

1. Operating the VR system and assisting the participant in the use of the VR
technology (as explained above).
2. To tailor the training (e.g., difficulty level) to the abilities and needs
of the participant.
3. To formulate strategies with the participant which they can use in the
exercises, and to evaluate the performance of the participant during the
exercises and tweak their strategies accordingly.
4. To observe the behavior of the participant in the virtual environment and
provide feedback on the utility and functionality thereof (e.g., gaze,
interactive behaviors). For example, if a participant avoids eye contact with
avatars, a therapist may comment on this and encourage participants to explore
facial features to improve affect recognition.
5. To control the dialog function in the latter part of the training (explained
below).
6. Reporting (e.g., session content, duration, protocol deviations).

DiSCoVR consists of three modules, targeting different domains of social
cognition: emotion perception; social perception & ToM; and social interaction
training. At the end of each session, participants are given an (optional)
homework assignment, which are intended to promote the use of the
intervention*s techniques in daily life.

Module 1: Emotion Perception (sessions 1-5)
Participants walk around the virtual environment and encounter virtual
characters (avatars) who show dynamic facial emotions. Participants will be
trained to recognize these emotions by using strategy coaching (i.e., helping
the participant to choose the most appropriate strategy to complete a task),
practice, and attentional direction to salient features (i.e., the face and
mouth, which provide important affective cues). Participants are encouraged to
explore the avatars* facial features, and identify the emotion that they
portray out of six basic emotions (happiness, surprise, fear, disgust, anger
and sadness). Participants choose the correct emotion by selecting it with
their joystick in a multiple-choice menu that is shown in their field of
vision. Homework exercises during this module include recognizing emotions in
the home environment, so that participants learn to employ FAR strategies in
their daily lives.


Module 2: Social Perception & ToM (sessions 6-9)
In this part of the intervention, participants witness conversations between
avatars. The goal of this module is to place emotions into a narrative. By
introducing context, participants are taught to consider situational
information in order to judge the thoughts, emotions and behaviors of others.
The social scenarios generally include multiple versions and/or endings, in
which many factors are equal but one crucial factor differs: e.g., an avatar is
having a good day or a bad day, which is reflected in their reactions. At
different points during the scenario, the participant is prompted to assess the
emotions and/or thoughts of the avatar. If the participant gives a wrong
answer, the avatars will more explicitly state their thoughts, intentions
and/or emotions, after which participants are prompted to try again.
Participants analyze social situations by assessing (the relations between)
thoughts, emotions and behavior. Using this *GGG-model* (gedachten, gevoelens,
gedrag), participants learn to understand why others act in a certain way and
how mental states are influenced by situations and other people. At this stage
in the intervention, homework exercises consist of assessing thoughts, behavior
and emotions of participants themselves and others in their daily environments.

Module 3: Social Interaction Training (sessions 10-16)
In the final module, participants learn to apply the techniques and strategies
they acquired in the first two modules. They practice in one-on-one role play
interactions in the virtual environment. The therapist interacts with the
participant through an avatar, using a transformed voice. Since the participant
is wearing the VR headset and noise cancelling headphones, it appears to them
that they are interacting with a virtual person. Role play situations are
tailored as much as possible to the participant*s goals; however, occasionally,
standard scenarios are used, for example, in case participants do not
contribute situations they*d like to practice with. To help participants
determine how to react in a social situation, participants use a series of
steps: 1) determine your own thoughts, emotions and behavior (GGG); 2)
determine the other person*s thoughts, emotions and behavior (GGG); 3) list the
possible ways you could react; 4) evaluate how desirable each option is (and,
if possible, practice it by role playing); and 5) choose the reaction with the
most desirable outcome. In the final part of the intervention, participants use
this series of steps in difficult social situations they encounter in their
daily lives.


The control condition also consists of 16 weekly one-on-one sessions of 60
minutes with a therapist. It is matched to the DiSCoVR condition for therapist
contact and use of VR. VRelax is relaxation therapy using 360 degrees videos in
VR (www.vrmentalhealth.org). The videos are played using a Samsung Gear VR
head-mounted display, powered by a Samsung Galaxy smartphone. Several
environments are available, including swimming with dolphins, a beach and a
forest. Relaxation exercises such as progressive muscle relaxation and guided
meditation are embedded in the videos. Participants can look around in 360
degrees by moving their head. The Gear VR registers head movement and adjusts
the image accordingly. Participants can navigate between environments and
activate relaxation exercises by looking at hotspots within the videos.

Participation in the study implies the following time investment:
- Three measurements of approximately two hours each
- 30 minutes per day of experience sampling, for a week, at all three
measurement occasions.
- 16 hours of therapy
- For DiSCoVR participants, an extra assessment of 30 minutes.

We do not anticipate any adverse side effects of participation. Some people may
experience cyber sickness when they use VR. The symptoms of this resemble
motion sickness, including dizziness and nausea. After removing the VR headset,
cybersickness usually passes quickly. Importantly, cybersickness also appears
to be caused or exacerbated by stress and / or anxiety. With repeated (gradual)
exposure, it is less likely that participants will suffer from cybersickness.
To prevent cyber sickness, the protocols are structured in such a way that the
time in VR is gradually increased.