- Email: [email protected]
Do conversations with virtual avatars increase feelings of social anxiety?
Journal of Anxiety Disorders 27 (2013) 398–403
Contents lists available at SciVerse ScienceDirect
Journal of Anxiety Disorders
Do conversations with virtual avatars increase feelings of social anxiety? Mark B. Powers a,∗ , Nicole F. Briceno a , Robert Gresham a , Ernest N. Jouriles a , Paul M.G. Emmelkamp b,c , Jasper A.J. Smits a a
Southern Methodist University, Dallas, TX, United States The University of Amsterdam, The Netherlands c King Abdulaziz University, Saudi Arabia b
a r t i c l e
i n f o
Article history: Received 27 August 2012 Received in revised form 1 March 2013 Accepted 10 March 2013 Keywords: Virtual reality New technology Social anxiety Conversation Fear Immersion
a b s t r a c t Virtual reality (VR) technology provides a way to conduct exposure therapy with patients with social anxiety. However, the primary limitation of current technology is that the operator is limited to preprogramed avatars that cannot be controlled to interact/converse with the patient in real time. The current study piloted new technology allowing the operator to directly control the avatar (including speaking) during VR conversations. Using an incomplete repeated measures (VR vs. in vivo conversation) design and random starting order with rotation counterbalancing, participants (N = 26) provided ratings of fear and presence during both VR and in vivo conversations. Results showed that VR conversation successfully elevated fear ratings relative to baseline (d = 2.29). Participants also rated their fear higher during VR conversation than during in vivo conversation (d = 0.85). However, in vivo conversation was rated as more realistic than VR conversation (d = 0.74). No participants dropped out and 100% completed both VR and in vivo conversations. Qualitative participant comments suggested that the VR conversations would be more realistic if they did not meet the actor/operator and if they were not in the same room as the participant. Overall, the data suggest that the novel technology allowing real time interaction/conversation in VR may prove useful for the treatment of social anxiety in future studies. © 2013 Elsevier Ltd. All rights reserved.
1. Introduction Social anxiety disorder (SAD) is a debilitating disorder that affects approximately 12% of the United States population (Kessler et al., 2005). Currently the most researched and effective treatment for SAD is exposure-based cognitive behaviour therapy (CBT) (Hofmann & Smits, 2008; Powers, Sigmarsson, & Emmelkamp, 2008). However, given the feared situation in SAD involves interacting with or performing in front of other people, creating realistic exposure scenarios in clinical practice is challenging and not often achieved in clinical practice. Indeed, most therapists do not conduct any exposure therapy. For example, bibliotherapy, medication, dynamic therapy, and cognitive therapy are all more commonly used than exposure (Freiheit, Vye, Swan, & Cady, 2004; Goisman, Warshaw, & Keller, 1999). Among the few therapists that utilize exposure therapy, they often rely on prescribing exposures as homework (risking poor compliance) or, if time permits, meeting the patient at various locations to conduct exposure exercises. These methods come with increased cost, ethical concerns,
∗ Corresponding author. Tel.: +1 2145945520. E-mail address: [email protected] (M.B. Powers). 0887-6185/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.janxdis.2013.03.003
and difficulty controlling stimuli. In response to these concerns, researchers have developed alternative methods of exposure delivery including over the internet and virtual reality exposure therapy (VRET) environments. Anxiety treatment via the internet is helpful in reaching clients who may otherwise not have access to trained exposure therapists (Carlbring et al., 2006; Carlbring et al., 2007; Lange et al., 2000; Lange, Van de Ven, Schrieken, & Emmelkamp, 2001; Lange et al., 2003). However, exposure is still limited to email/texting, a webcam where the therapist still needs to recruit confederates to interact with, or an avatar that is not fully controlled in real time by the therapist. Exposure through virtual reality has proven quite successful across the anxiety disorders (Emmelkamp, 2005; Opris et al., 2012; Powers & Emmelkamp, 2008). A noteworthy benefit of using (VRET) is that it easily enables the therapist to manipulate the feared situation and environment for public speaking anxiety (Anderson, Zimand, Hodges, & Rothbaum, 2005; Safir, Wallach, & Bar-Zvi, 2012; Wallach, Safir, & Bar-Zvi, 2009). In addition, VRET is considered more tolerable and acceptable to patients (Emmelkamp, 2005). Unfortunately, VRET technology does not yet allow the patient to have conversations with the people (avatars) in the virtual environments in real time. For example, it is possible to hear a voice over the internet while looking at an avatar. However, the
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403
avatar’s lips and movements/gestures are not in sync with the voice. Trials to date have only included exposure to such avatars with limited response options and speech either through typing or auditory output without linked facial movement and expressions (Harris, Kemmerling, & North, 2002; Klinger et al., 2005). This is unfortunate given findings that show involvement in the VRET is crucial for success (Price, Mehta, Tone, & Anderson, 2011) and that involvement is enhanced when avatars are visually responsive to the participant (Garau, Slater, Pertaub, & Razzaque, 2005). Thus, interactive realtime conversations still need to be conducted in vivo; VRET would be far more useful for treating SAD if the therapist were able to also manipulate the avatar to converse with the patient in real time. This would allow therapists to conduct in-session exposure exercises aimed at reducing patients’ fears of performance situations involving social interaction and conversation. Two recent studies piloted new virtual reality (VR) technology in the context of training women to resist unwanted sexual advances. Jouriles, McDonald, Kullowatz, Rosenfield, Gomez, and Cuevas (2009) randomized 62 undergraduate students to a standard faceto-face role-play with a male actor or a virtual conversation with a male avatar (fully controlled by the actor) in a virtual environment. Consistent with predictions, VR conversations were rated as more realistic and participants showed greater negative affect compared to the face-to-face condition. Similar findings were observed in the second study of 48 women randomized to VR or face-toface conditions (Jouriles, Simpson Rowe, McDonald, Platt, & Gomez, 2011). Although preliminary, findings from these studies support the validity of virtual conversation for assessing undergraduate women’s reactions in sexually threatening situations and underscore the potential utility of this VR technology for simulating social interaction situations. However, this technology has not yet been extended to the treatment of social anxiety. As a first step in this process, we tested the ability of this technology to elicit social anxiety in a college population. This proof-of-concept pilot study investigated the use of VR interactive conversation technology in college populations. At this stage, we define “proof” as activation of the fear structure (operationalized as SUDs during VR conversation). Although we anticipated SUDs to be higher in the in vivo comparison condition, we wanted to see if VR conversation significantly raised SUDs relative to baseline. Participants discussed controversial topics with a actor/operator in either VR or in vivo environments. Based on previous findings (Barlow, Leitenberg, Agras, & Wincze, 1969; Dyckman & Cowan, 1978; Emmelkamp & Wessels, 1975; Jouriles et al., 2009; Jouriles et al., 2011; Litvak, 1969; Powers & Emmelkamp, 2008; Sherman, 1972; Watson, Mullett, & Pillay, 1973), we expected that: (a) SUDs would be significantly higher during VR conversation than at baseline, (b) participants would report higher anxiety during the in vivo conversation compared to the VR conversation, and (c) participants would rate the in vivo conversation as more realistic than the VR conversation. 2. Methods 2.1. Participants Participants were recruited from an undergraduate course in psychology, and received extra credit for their participation in the study. The 26 undergraduate participants were primarily female (73.1%). Most were Non-Hispanic White (76.9%, 11.5% Black, 7.7% Hispanic, 3.8% Asian) with a mean age of 20.42 (SD = 0.45). Participants in this trial scored similar to college students in previous studies on measures (described below) of social anxiety (LSAS M = 36.12, SD = 18.61) and general anxiety (STAI Trait M = 35.46, SD = 7.04). However, participants did not complete a structured diagnostic interview.
399
2.2. Design All study participants took part in both an “In Vivo” and a “VR” conversation condition. The order in which a participant engaged in one conversation before the other (i.e. In Vivo then VR or VR then In Vivo) was determined by counterbalancing the condition order prior to participants’ assignment. When a participant arrived for their scheduled time they were assigned to the next available condition. Half of the participants (n = 13) engaged in the In Vivo conversation first and the other half (n = 13) participated in the VR conversation first.
2.3. Procedure Procedures were modified (to include controversial topics) from previous trials (Jouriles et al., 2009; Jouriles et al., 2011; Smits, Powers, Buxkamper, & Telch, 2006). All procedures were approved by the university Institutional Review Board. Upon arrival, participants signed informed consent, after which they completed a battery of self-report measures. After completing the measures the participants were told they would be engaging in two 5min conversations. They were instructed that they could stop the conversations at any time if they became too uncomfortable. Depending on their assignment into the rotated order of conversation, they were informed they would be doing either In Vivo followed by VR or VR followed by In Vivo. In Vivo was explained to the participants as; “Right here, just as we are now,” and VR as “In virtual reality, where we will talk to each other, but while wearing ‘this’ headset and ‘these’ earphones.” Participants were instructed by the facilitator, a male research staff member with graduate training (blind to study hypotheses), would pick the topic of both conversations, leave the room for 3 min so that they might prepare any thoughts on the topic, then reenter, at which point the conversation would begin. Although participants were not given specifics regarding the facilitator’s dialogue, they were told they would be asked their opinions on a topic, as well as questions regarding any opinions or conclusions they might express. All sessions were video recorded for adherence coding. The adherence coding form was adapted from our previous trials in prolonged exposure therapy for posttraumatic stress disorder and exposure with ritual prevention for obsessive compulsive disorder (Foa et al., 2013; Gilboa-Schechtman et al., 2010). The form included 2 questions on essential elements, 6 questions on essential but not unique elements, and 2 questions on adherence overall. The rater then took into account the Likert items to make a categorical adherence rating. After providing a baseline fear rating (0 [no fear] to 100 [most fear imaginable]), participants were given the list of possible conversation topics for the In Vivo and VR conversations (same sex marriage, abortion, Iraq war, Evolution/Creationism in schools, torture to gather intelligence, homosexuals in the military, most embarrassing moment). The topics were similar to those used in previous social anxiety studies (Smits et al., 2006). To increase the likelihood of participant engagement in a dialogue with the facilitator, topics were contemporary and relatively controversial. Participants were instructed to rate their highest predicted level of fear during a 5-min conversation for each of the seven topics, with ratings to be given for conversation in both the In Vivo and in the VR environment. The topic with the highest predicted SUDs rating was selected as the topic of conversation. The same topic was used for both conversations. In the event of equivalent ratings across multiple topics the facilitator exercised his discretion in selecting a topic. The participant was then given the topic and the facilitator left the room for 3 min while the participant prepared their thoughts on the topic.
400
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403
Fig. 2. VR conversation avatar.
Fig. 1. Participant and therapist converse in VR.
In the VR conversation condition the facilitator introduced the VR Head Mounted Display and headphones, and instructed/aided the participant in putting them on (see Fig. 1). In the In Vivo condition the facilitator simply sat to the left of the participant on a couch. Their distance from the participant was matched to the distance depicted in the VRET scene. In both the VR and In Vivo conditions, the facilitator began timing the conversation and stated, “Okay, now we’re going to have (selected topic inserted). To begin, what a conversation about do you personally think about ?” During the course of the conversation the participant was asked several open-ended questions from a predetermined list to ensure the conversation would continue for the 5-min duration (e.g. “And why do you believe that? How did you come to believe that?). In addition, the facilitator was instructed to voice a viewpoint in opposition to whatever position the participant chose for a given topic. After 5-min the conversation with the participant was stopped, and they were asked to rate their highest level of fear during the 5 min conversation. Immediately after completing the first conversation, procedures began for the second conversation (i.e. 3-min for the participant to prepare their thoughts on the topic). After completing both conversations, participants were thanked for their participation, and the session was concluded. Given the aim of this study, participants were also invited to comment on the experience after debriefing. 2.4. Materials The hardware and software used in this study was the same as that used in previous studies of this technology (Jouriles et al., 2009; Jouriles et al., 2011). The VR conversation was facilitated through the use of a virtual avatar interacting with the participant in a virtual environment (VE). The Virtual environment was created from modification (mod) templates based around the video game HalfLife 2® . The avatar was a modified character from the game, and designed to look like a typical male university student. The virtual environment was designed to look like a college dorm or apartment bedroom (e.g. bed, television, movie posters, computer desk, couch). From the participants virtual perspective the avatar was seated to their left on a couch (same position as in vivo condition) (see Fig. 2). The participant interacted with the virtual environment via the use of an eMagine z800 stereovision head mounted display (HMD). The HMD allowed the participant to view the virtual environment. Furthermore, head tracking hardware built into the HMD allowed the VR software to sense the participants head movements and adjust their view accordingly, thus giving participants fully immersive control of where and what they viewed within the virtual environment.
In addition to the HMD, in the VR conversation condition, the participant wore a pair of over-the-ear noise cancellation headphones. The facilitator used a Microsoft Lifechat LX-2000 microphone headset to speak with the participant. The microphone-headphone communication was relayed through the VR software, so that as the facilitator spoke, the virtual avatars mouth would move thereby approximating the look of normal speech. Furthermore, the facilitator, via keyboard command, was also able to change the avatar’s facial expressions (e.g. happy, sad, angry), as well as a number of gestures (e.g. nodding, pointing at the participant, shrugging, throwing his hands in air). All hardware and software was run through a personal computer using a Pentium IV® CPU, 512 megabytes of RAM, and a 6800 Graphics processor (Jouriles et al., 2009). 2.5. Measures 2.5.1. Liebowitz Social Anxiety Scale (LSAS) The LSAS is a widely used 24-item interviewer-rated instrument that assesses fear and avoidance of social situations. Participants rate fear items on a 4-point scale ranging from 0 (none) to 3 (severe). Likewise, avoidance is also rated on a 4-point scale ranging from 0 (never) to 3 (usually). The LSAS takes approximately 20–30 min to complete. The LSAS shows good reliability (˛ = 0.81–0.92) and validity (Cox, Ross, Swinson, & Direnfeld, 1998; Safren et al., 1999). Scores may be interpreted as: 55–65 Moderate, 65–80 Marked, 80–95 Severe, >95 Very Severe. The mean for participants with social anxiety disorder is 67.2 (SD = 27.5) (Heimberg et al., 1999). 2.5.2. State-Trait Anxiety Inventory (STAI) The state-trait anxiety inventory is a 20-item self-report measure of state and trait anxiety (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983). The measure takes about 10 min to complete. The STAI shows good reliability (˛ = 0.86–0.95) and validity (Spielberger et al., 1983). Interpretation of scores: Adults (M = 35 SD = 10), Generalized Anxiety Disorder (47–61), Panic Disorder with Agoraphobia (51–54), Panic Disorder (44–46). 2.5.3. Subjective Units of Disturbance (SUDs) The SUDs scale is a measure of subjective anxiety (0 = No Disturbance to 100 = Most Disturbance) that was rated by participants after the conversations (Wolpe, 1958; Wolpe, 1990). Participants reported the highest level that their SUDs reached. 2.5.4. Presence/immersion The measure of presence/realism was a 9-item immersion questionnaire used in our two previous similar studies with women resisting unwanted sexual advances (Jouriles et al., 2009; Jouriles et al., 2011). Items were rated on a 0 (Not at All) to 5 (Very Much)
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403
401
3.5. Feasibility, acceptability/tolerability, and qualitative responses
Fig. 3. Presence and fear as a function of VR or in vivo conversation.
scale (e.g. Did you ever feel that you might actually be defending your position to a stranger? Were you involved in the role play to the extent that you lost track of time?, etc.). The scale showed excellent reliability with this sample (Cronbach’s ˛ = 0.82; Guttman Split-half = 0.84).
No participants dropped out of the trial early and 100% completed both conversations. Participants comments after debriefing included: (a) the VR condition would be enhanced by having the facilitator/actor sit in another room rather than controlling the VR from the couch next to the participant, (b) realism and fear would be greater if participants also did not meet the facilitator/actor before the conversation in VR, and (c) the controversial conversation was always easier the second time through the same topic (suggesting a between-subjects design in the future). However, post hoc analysis with a paired samples t-test suggested that fear ratings were still significantly higher in the VR condition when only analysing the first conversations (p = 0.024). Also, a post hoc paired samples t-test showed enhanced fear activation was not due solely to differences in predicted fear between the VR (M = 34.36) and in vivo (M = 31.91) conditions (t = −0.88, p = 0.405). Finally, 10% of videos (n = 3) were observed for adherence to the pre-determined probes stated earlier. Ratings suggested that actor performance was consistent in both the VR and In vivo conditions.
4. Discussion 3. Results There were no dropouts and no data missing in this small pilot trial. Thus all results represent completer analyses. Overall, fear ratings ranged between 5 and 90 and correlations between the outcome measures ranged between −0.49 and 0.59. None of these correlations were significant except for the correlation between the rated presence in the VR and in vivo conditions (r = 0.53, p = 0.006). 3.1. Randomization check ANOVAs showed there were no significant differences in participants randomized to VR then in vivo compared to those randomized to in vivo then VR (age p = 0.42, LSAS total p = 0.14, VR presence rating p = 0.92, in vivo presence rating p = 0.67, baseline SUDs p = 0.81, VR peak SUDs p = 0.42, & in vivo peak SUDs p = 0.71). 3.2. Order effects An ANOVA showed there was no main effect for order (VR first or in vivo first; F(1,25) = 0.17, p = 0.68) or condition by order interaction (F(1,24) = 1.35, p = 0.26). 3.3. Fear ratings Hypothesis (a) paired samples t-tests and Fig. 3 show that consistent with prediction, participants rated their peak fear significantly higher during VR conversation (M = 38.46, SD = 16.78) than at baseline (M = 16.73, SD = 11.99, t(25) = 8.26, p = 0.000, d = 2.29). Hypothesis b) Contrary to prediction, participants rated their peak fear significantly higher during conversations in the VR condition (M = 38.46, SD = 16.78) compared to the in vivo condition (M = 29.23, SD = 10.36, t(25) = 3.08, p = 0.005, d = 0.85). 3.4. Presence (realism) ratings Hypothesis (c) Consistent with prediction, a paired samples ttest showed that participants rated conversations as significantly more realistic (presence) in vivo (M = 29.35, SD = 5.73) compared to in VR (M = 26.27, SD = 5.58, t(25) = 2.67 p = 0.013, d = 0.74).
Virtual reality technology increases access to exposure therapy for patients with social anxiety. However, the primary limitation of current technology is that the operator is limited to pre-programed avatars that cannot be controlled to interact/converse with the patient in real time. The current study piloted new technology allowing the operator to directly control the avatar (including speaking) during VR social interactions. We used a within-subjects incomplete repeated measures design with random starting order with rotation counterbalancing. Results showed that: (a) in vivo conversation was rated as more realistic than VR conversation, (b) fear ratings were higher during VR conversation compared to in vivo conversation, (c) no participants dropped out, and 100% completed both VR and in vivo conversations, and (d) qualitative participant comments suggested that the VR conversations would be more realistic if they did not meet the actor/operator and if they were not in the same room as the participant. Although in vivo conversation was rated as more realistic, participants reported higher fear ratings during the VR conversation condition. Previous studies suggest that for reducing anxiety with exposure therapy, fear ratings are more predictive of efficacy than realism ratings (Foa & Kozak, 1986; M. Krijn, P. M. G Emmelkamp, R. Biemond, C. de Wilde de Ligny, M. J. Schuemie, & C. A. P. G. van der Mast, 2004; M. Krijn, P. M. G Emmelkamp, R. P. Olafsson, & R. Biemond, 2004; Powers & Emmelkamp, 2008; Price et al., 2011). Realism or presence is the extent to which one interprets the virtual environment as if it were real (Lee, 2004). Consistent with our findings, studies that have manipulated presence found that it did not significantly affect treatment outcome (M. Krijn, P. M. G Emmelkamp, R. Biemond, et al., 2004; M. Krijn, P. M. G Emmelkamp, R. P. Olafsson, et al., 2004; Price & Anderson, 2007; Price et al., 2011). We can only speculate why fear ratings were higher in the VR conversation. Potential reasons could include: context (e.g. party atmosphere vs couch in psychology department; greater activation of the fear structure through a higher number of fear relevant stimuli), discomfort of simply being in any VR environment under any circumstance, less prior experience with the avatar vs. the actor (actor also conducted consent), or other reasons. Future studies may help clarify why fear ratings are higher in VR compared to in vivo conversations. The dropout and completion rates are also encouraging for further development of this technology for social anxiety applications. These initial pilot data are encouraging for
402
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403
exposure applications. In fact, if these findings are replicated and extended, VR exposure could even prove more effective as a clinical tool than in vivo exposure. Several limitations deserve comment. First, participants were not selected based on social anxiety scores for this feasibility pilot study. Future studies should determine if realism and fear ratings generalize to those with high social anxiety. Second, participants stated that presence would be enhanced if they did not meet the facilitator/actor controlling the avatar prior to the conversation and if the actor also controlled the avatar from another room. Although the participants reported that the noise cancellation headphones and head mounted display were effective, they were still aware of the presence of the actor on the couch next to them and that this thought was at times distracting from the realism of the VR conversation. Third, our small sample does not permit analysis of multiple dimensions of presence (spatial presence, involvement, and realness) (Price et al., 2011). Future studies could help determine which dimensions are enhanced through avatar interaction. Fourth, this study does not indicate if real time interaction is superior to standard VRET for social anxiety. Future studies should compare VRET with and without real time avatar interaction to determine the relative efficacy. Fifth, participants were given course credit based on their participation. Thus, finding a 100% completion rate may not generalize to other populations. Sixth, although the actor was instructed to only use facial expressions, gestures, and gaze consistent with the avatar condition, this control was not verified. Several lines of research suggest the importance of gestures, facial expressions, gaze, and mimicry in social anxiety (J. N. Vrijsen, W. G. Lange, E. S. Becker, & M. Rinck, 2010; J. N. Vrijsen, W. G. Lange, R. Dotsch, D. H. J. Wigboldus, & M. Rinck, 2010; Wieser, Pauli, Grosseibl, Molzow, & Muhlberger, 2010). Future studies should also code and rate in vivo actors on gestures and facial expressions to match the VRET condition. Seventh, although SUDs fear ratings are often used as outcome data, future studies should incorporate additional measures of distress (e.g. behavioural approach, psychophysiology, etc.). Eighth, although content choices for the conversation topics have been used previously (e.g. Smits et al., 2006), future studies may help determine the optimal content choices for social anxiety exposures. Finally, participants only completed each conversation condition (VR & in vivo) once. Thus, habituation/extinction could not be measured. Overall, the data suggest that the novel technology allowing real time interaction/conversation in VR may prove useful for the treatment of social anxiety in future studies. Future studies should include participants with high social anxiety, not have the participant meet the actor/therapist prior to the session, and have the therapist control the avatar from a different room.
References Anderson, P. L., Zimand, E., Hodges, L. F., & Rothbaum, B. O. (2005). Cognitive behavioral therapy for public-speaking anxiety using virtual reality for exposure. Depression and Anxiety, 22(3), 156–158. http://dx.doi.org/10.1002/da.20090 Barlow, D. H., Leitenberg, H., Agras, W. S., & Wincze, J. P. (1969). The transfer gap in systematic desensitization: an analogue study. Behaviour Research and Therapy, 7(2), 191–196. Carlbring, P., Bohman, S., Brunt, S., Buhrman, M., Westling, B. E., Ekselius, L., et al. (2006). Remote treatment of panic disorder: a randomized trial of internet-based cognitive behavior therapy supplemented with telephone calls. American Journal of Psychiatry, 163(12), 2119–2125. http://dx.doi.org/10.1176/appi.ajp.163.12.2119 Carlbring, P., Gunnarsdottir, M., Hedensjo, L., Andersson, G., Ekselius, L., & Furmark, T. (2007). Treatment of social phobia: randomised trial of internet-delivered cognitive-behavioural therapy with telephone support. British Journal of Psychiatry, 190, 123–128. Cox, B. J., Ross, L., Swinson, R. P., & Direnfeld, D. M. (1998). A comparison of social phobia outcome measures in cognitive-behavioral group therapy. Behavior Modification, 22(3), 285–297.
Dyckman, J. M., & Cowan, P. A. (1978). Imaging vividness and the outcome of in vivo and imagined scene desinsitization. Journal of Consulting and Clinical Psycghology, 46(5), 1155–1156. Emmelkamp, P. M. G. (2005). Technological innovations in clinical assessment and psychotherapy. Psychotherapy and Psychosomatics, 74(6), 336–343. Emmelkamp, P. M. G., & Wessels, H. (1975). Flooding in imagination vs flooding in vivo: a comparison with agoraphobics. Behaviour Research and Therapy, 13(1), 7–15. Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: exposure to corrective information. Psychological Bulletin, 99(1), 20–35. Foa, E. B., Simpson, H. B., Liebowitz, M. R., Powers, M. B., Rosenfield, D., Cahill, S. P., et al. (2013). Augmenting SRI treatment with exposure and ritual prevention (EX/RP) for obsessive-compulsive disorder: long-term outcome. Journal of Clinical Psychiatry, in press. Freiheit, S. R., Vye, C., Swan, R., & Cady, M. (2004). Cognitive-behavioral therapy for anxiety: is dissemination working? Behavior Therapist, 27(2), 25–32. Garau, M., Slater, M., Pertaub, D. P., & Razzaque, S. (2005). The responses of people to virtual humans in an immersive virtual environment. Presence, 14, 104–116. Gilboa-Schechtman, E., Foa, E. B., Shafran, N., Aderka, I. M., Powers, M. B., Rachamim, L., et al. (2010). Prolonged exposure versus dynamic therapy for adolescent PTSD: a pilot randomized controlled trial. American Academy of Child and Adolescent Psychiatry, 49, 1034–1042. Goisman, R. M., Warshaw, M. G., & Keller, M. B. (1999). Psychosocial treatment prescriptions for generalized anxiety disorder, panic disorder, and social phobia, 1991–1996. American Journal of Psychiatry, 156(11), 1819–1821. Harris, S. R., Kemmerling, R. L., & North, M. M. (2002). Brief virtual reality therapy for public speaking anxiety. Cyberpsychology and Behavior, 5(6), 543–550. Heimberg, R. G., Horner, K. J., Juster, H. R., Safren, S. A., Brown, E. J., Schneier, F. R., et al. (1999). Psychometric properties of the Liebowitz Social Anxiety Scale. Psychological Medecine, 29(1), 199–212. Hofmann, S. G., & Smits, J. A. (2008). Cognitive-behavioral therapy for adult anxiety disorders: a meta-analysis of randomized placebo-controlled trials. Journal of Clinical Psychiatry, 69, 621–632. Jouriles, E. N., McDonald, R., Kullowatz, A., Rosenfield, D., Gomez, G. S., & Cuevas, A. (2009). Can virtual reality increase the realism of role plays used to teach college women sexual coercion and rape-resistance skills? Behavior Therapy, 40(4), 337–345. http://dx.doi.org/10.1016/j.beth.2008.09.002 Jouriles, E. N., Simpson Rowe, L., McDonald, R., Platt, C. G., & Gomez, G. S. (2011). Assessing women’s responses to sexual threat: validity of a virtual role-play procedure. Behavior Therapy, 42(3), 475–484. http://dx.doi.org/10.1016/j.beth.2010.11.005 Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62(6), 593–602. Klinger, E., Bouchard, S., Legeron, P., Roy, S., Lauer, F., Chemin, I., et al. (2005). Virtual reality therapy versus cognitive behavior therapy for social phobia: a preliminary controlled study. CyberPsychology and Behavior, 8(1), 76–88. Krijn, M., Emmelkamp, P. M. G., Biemond, R., de Wilde de Ligny, C., Schuemie, M. J., & van der Mast, C. A. P. G. (2004). Treatment of acrophobia in virtual reality: the role of immersion and presence. Behavior Research and Therapy, 42(2), 229–239. Krijn, M., Emmelkamp, P. M. G., Olafsson, R. P., & Biemond, R. (2004). Virtual reality exposure therapy of anxiety disorders: a review. Clinical Psychology Review, 24(3), 259–281. Lange, A., Rietdijk, D., Hudcovicova, M., van de Ven, J. P., Schrieken, B., & Emmelkamp, P. M. G. (2003). Interapy: a controlled randomized trial of the standardized treatment of posttraumatic stress through the internet. Journal of Consulting and Clinical Psychology, 71(5), 901–909. Lange, A., Schrieken, B., Van De Ven, J. P., Bredeweg, B., Emmelkamp, P. M. G., Van Der Kolk, J., et al. (2000). ‘Interapy’: the effects of a short protocolled treatment of posttraumatic stress and pathological grief through the internet. Behavioural and Cognitive Psychotherapy, 28(2), 175–192. Lange, A., Van de Ven, J. P., Schrieken, B., & Emmelkamp, P. M. G. (2001). Interapy. Treatment of posttraumatic stress through the Internet: a controlled trial. Journal of Behavior Therapy and Experimental Psychiatry, 32(2), 73–90. Lee, K. M. (2004). Presence, Explicated. Communication Theory, 14, 27–50. Litvak, S. B. (1969). A comparison of two brief group behavior therapy techniques on the reduction of avoidance behavior. The Psychological Record, 19, 329–334. Opris, D., Pintea, S., Garcia-Palacios, A., Botella, C., Szamoskozi, S., & David, D. (2012). Virtual reality exposure therapy in anxiety disorders: a quantitative meta-analysis. Depression and Anxiety, 29(2), 85–93. http://dx.doi.org/10.1002/da.20910 Powers, M. B., & Emmelkamp, P. M. G. (2008). Virtual reality exposure therapy for anxiety disorders: a meta-analysis. Journal of Anxiety Disorders, 22, 561–569. Powers, M. B., Sigmarsson, S. R., & Emmelkamp, P. M. G. (2008). A meta-analytic review of social phobia treatments. International Journal of Cognitive Therapy, 1, 94–113. Price, M., & Anderson, P. (2007). The role of presence in virtual reality exposure therapy. Journal of Anxiety Disorders, 21(5), 742–751. http://dx.doi.org/10.1016/j.janxdis.2006.11.002 Price, M., Mehta, N., Tone, E. B., & Anderson, P. L. (2011). Does engagement with exposure yield better outcomes? Components of presence as a predictor of treatment response for virtual reality exposure therapy for social phobia. Journal of Anxiety Disorders, 25(6), 763–770. http://dx.doi.org/10.1016/j.janxdis.2011.03.004 Safir, M. P., Wallach, H. S., & Bar-Zvi, M. (2012). Virtual reality cognitive-behavior therapy for public speaking anxiety: one-year follow-up. Behavior Modification, 36(2), 235–246. http://dx.doi.org/10.1177/0145445511429999
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403 Safren, S. A., Heimberg, R. G., Horner, K. J., Juster, H. R., Schneier, F. R., & Liebowitz, M. R. (1999). Factor structure of social fears: the Liebowitz Social Anxiety Scale. Journal of Anxiety Disorders, 13(3), 253–270, doi: S0887-6185(99)00003-1 [pii]. Sherman, A. R. (1972). Real-life exposure as a primary therapeutic factor in the desensitization treatment of fear. Journal of Abnormal Psychology, 79(1), 19–28. Smits, J. A., Powers, M. B., Buxkamper, R., & Telch, M. J. (2006). The efficacy of videotape feedback for enhancing the effects of exposure-based treatment for social anxiety disorder: a controlled investigation. Behaviour Research and Therapy, 44(12), 1773–1785. Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the state-trait anxiety inventory. Palo Alto, CA: Consulting Psychologists Press. Vrijsen, J. N., Lange, W. G., Becker, E. S., & Rinck, M. (2010). Socially anxious individuals lack unintentional mimicry. Behaviour Research and Therapy, 48(6), 561–564. http://dx.doi.org/10.1016/j.brat.2010.02.004 Vrijsen, J. N., Lange, W. G., Dotsch, R., Wigboldus, D. H. J., & Rinck, M. (2010). How do socially anxious women evaluate mimicry? A Virtual reality study. Cognition and Emotion, 24, 840–947.
403
Wallach, H. S., Safir, M. P., & Bar-Zvi, M. (2009). Virtual reality cognitive behavior therapy for public speaking anxiety: a randomized clinical trial. Behavior Modification, 33(3), 314–338. http://dx.doi.org/10.1177/0145445509331926 Watson, J. P., Mullett, G. E., & Pillay, H. (1973). The effects of prolonged exposure to phobic situations upon agoraphobic patients treated in groups. Behavior Research and Therapy, 11(4), 531–545. Wieser, M. J., Pauli, P., Grosseibl, M., Molzow, I., & Muhlberger, A. (2010). Virtual social interactions in social anxiety – the impact of sex, gaze, and interpersonal distance. Cyberpsychology, Behavior and Social Networking, 13(5), 547–554. http://dx.doi.org/10.1089/cyber.2009.0432 Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press. Wolpe, J. (1990). The practice of behavior therapy. New York: Pergamon Press.
Contents lists available at SciVerse ScienceDirect
Journal of Anxiety Disorders
Do conversations with virtual avatars increase feelings of social anxiety? Mark B. Powers a,∗ , Nicole F. Briceno a , Robert Gresham a , Ernest N. Jouriles a , Paul M.G. Emmelkamp b,c , Jasper A.J. Smits a a
Southern Methodist University, Dallas, TX, United States The University of Amsterdam, The Netherlands c King Abdulaziz University, Saudi Arabia b
a r t i c l e
i n f o
Article history: Received 27 August 2012 Received in revised form 1 March 2013 Accepted 10 March 2013 Keywords: Virtual reality New technology Social anxiety Conversation Fear Immersion
a b s t r a c t Virtual reality (VR) technology provides a way to conduct exposure therapy with patients with social anxiety. However, the primary limitation of current technology is that the operator is limited to preprogramed avatars that cannot be controlled to interact/converse with the patient in real time. The current study piloted new technology allowing the operator to directly control the avatar (including speaking) during VR conversations. Using an incomplete repeated measures (VR vs. in vivo conversation) design and random starting order with rotation counterbalancing, participants (N = 26) provided ratings of fear and presence during both VR and in vivo conversations. Results showed that VR conversation successfully elevated fear ratings relative to baseline (d = 2.29). Participants also rated their fear higher during VR conversation than during in vivo conversation (d = 0.85). However, in vivo conversation was rated as more realistic than VR conversation (d = 0.74). No participants dropped out and 100% completed both VR and in vivo conversations. Qualitative participant comments suggested that the VR conversations would be more realistic if they did not meet the actor/operator and if they were not in the same room as the participant. Overall, the data suggest that the novel technology allowing real time interaction/conversation in VR may prove useful for the treatment of social anxiety in future studies. © 2013 Elsevier Ltd. All rights reserved.
1. Introduction Social anxiety disorder (SAD) is a debilitating disorder that affects approximately 12% of the United States population (Kessler et al., 2005). Currently the most researched and effective treatment for SAD is exposure-based cognitive behaviour therapy (CBT) (Hofmann & Smits, 2008; Powers, Sigmarsson, & Emmelkamp, 2008). However, given the feared situation in SAD involves interacting with or performing in front of other people, creating realistic exposure scenarios in clinical practice is challenging and not often achieved in clinical practice. Indeed, most therapists do not conduct any exposure therapy. For example, bibliotherapy, medication, dynamic therapy, and cognitive therapy are all more commonly used than exposure (Freiheit, Vye, Swan, & Cady, 2004; Goisman, Warshaw, & Keller, 1999). Among the few therapists that utilize exposure therapy, they often rely on prescribing exposures as homework (risking poor compliance) or, if time permits, meeting the patient at various locations to conduct exposure exercises. These methods come with increased cost, ethical concerns,
∗ Corresponding author. Tel.: +1 2145945520. E-mail address: [email protected] (M.B. Powers). 0887-6185/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.janxdis.2013.03.003
and difficulty controlling stimuli. In response to these concerns, researchers have developed alternative methods of exposure delivery including over the internet and virtual reality exposure therapy (VRET) environments. Anxiety treatment via the internet is helpful in reaching clients who may otherwise not have access to trained exposure therapists (Carlbring et al., 2006; Carlbring et al., 2007; Lange et al., 2000; Lange, Van de Ven, Schrieken, & Emmelkamp, 2001; Lange et al., 2003). However, exposure is still limited to email/texting, a webcam where the therapist still needs to recruit confederates to interact with, or an avatar that is not fully controlled in real time by the therapist. Exposure through virtual reality has proven quite successful across the anxiety disorders (Emmelkamp, 2005; Opris et al., 2012; Powers & Emmelkamp, 2008). A noteworthy benefit of using (VRET) is that it easily enables the therapist to manipulate the feared situation and environment for public speaking anxiety (Anderson, Zimand, Hodges, & Rothbaum, 2005; Safir, Wallach, & Bar-Zvi, 2012; Wallach, Safir, & Bar-Zvi, 2009). In addition, VRET is considered more tolerable and acceptable to patients (Emmelkamp, 2005). Unfortunately, VRET technology does not yet allow the patient to have conversations with the people (avatars) in the virtual environments in real time. For example, it is possible to hear a voice over the internet while looking at an avatar. However, the
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403
avatar’s lips and movements/gestures are not in sync with the voice. Trials to date have only included exposure to such avatars with limited response options and speech either through typing or auditory output without linked facial movement and expressions (Harris, Kemmerling, & North, 2002; Klinger et al., 2005). This is unfortunate given findings that show involvement in the VRET is crucial for success (Price, Mehta, Tone, & Anderson, 2011) and that involvement is enhanced when avatars are visually responsive to the participant (Garau, Slater, Pertaub, & Razzaque, 2005). Thus, interactive realtime conversations still need to be conducted in vivo; VRET would be far more useful for treating SAD if the therapist were able to also manipulate the avatar to converse with the patient in real time. This would allow therapists to conduct in-session exposure exercises aimed at reducing patients’ fears of performance situations involving social interaction and conversation. Two recent studies piloted new virtual reality (VR) technology in the context of training women to resist unwanted sexual advances. Jouriles, McDonald, Kullowatz, Rosenfield, Gomez, and Cuevas (2009) randomized 62 undergraduate students to a standard faceto-face role-play with a male actor or a virtual conversation with a male avatar (fully controlled by the actor) in a virtual environment. Consistent with predictions, VR conversations were rated as more realistic and participants showed greater negative affect compared to the face-to-face condition. Similar findings were observed in the second study of 48 women randomized to VR or face-toface conditions (Jouriles, Simpson Rowe, McDonald, Platt, & Gomez, 2011). Although preliminary, findings from these studies support the validity of virtual conversation for assessing undergraduate women’s reactions in sexually threatening situations and underscore the potential utility of this VR technology for simulating social interaction situations. However, this technology has not yet been extended to the treatment of social anxiety. As a first step in this process, we tested the ability of this technology to elicit social anxiety in a college population. This proof-of-concept pilot study investigated the use of VR interactive conversation technology in college populations. At this stage, we define “proof” as activation of the fear structure (operationalized as SUDs during VR conversation). Although we anticipated SUDs to be higher in the in vivo comparison condition, we wanted to see if VR conversation significantly raised SUDs relative to baseline. Participants discussed controversial topics with a actor/operator in either VR or in vivo environments. Based on previous findings (Barlow, Leitenberg, Agras, & Wincze, 1969; Dyckman & Cowan, 1978; Emmelkamp & Wessels, 1975; Jouriles et al., 2009; Jouriles et al., 2011; Litvak, 1969; Powers & Emmelkamp, 2008; Sherman, 1972; Watson, Mullett, & Pillay, 1973), we expected that: (a) SUDs would be significantly higher during VR conversation than at baseline, (b) participants would report higher anxiety during the in vivo conversation compared to the VR conversation, and (c) participants would rate the in vivo conversation as more realistic than the VR conversation. 2. Methods 2.1. Participants Participants were recruited from an undergraduate course in psychology, and received extra credit for their participation in the study. The 26 undergraduate participants were primarily female (73.1%). Most were Non-Hispanic White (76.9%, 11.5% Black, 7.7% Hispanic, 3.8% Asian) with a mean age of 20.42 (SD = 0.45). Participants in this trial scored similar to college students in previous studies on measures (described below) of social anxiety (LSAS M = 36.12, SD = 18.61) and general anxiety (STAI Trait M = 35.46, SD = 7.04). However, participants did not complete a structured diagnostic interview.
399
2.2. Design All study participants took part in both an “In Vivo” and a “VR” conversation condition. The order in which a participant engaged in one conversation before the other (i.e. In Vivo then VR or VR then In Vivo) was determined by counterbalancing the condition order prior to participants’ assignment. When a participant arrived for their scheduled time they were assigned to the next available condition. Half of the participants (n = 13) engaged in the In Vivo conversation first and the other half (n = 13) participated in the VR conversation first.
2.3. Procedure Procedures were modified (to include controversial topics) from previous trials (Jouriles et al., 2009; Jouriles et al., 2011; Smits, Powers, Buxkamper, & Telch, 2006). All procedures were approved by the university Institutional Review Board. Upon arrival, participants signed informed consent, after which they completed a battery of self-report measures. After completing the measures the participants were told they would be engaging in two 5min conversations. They were instructed that they could stop the conversations at any time if they became too uncomfortable. Depending on their assignment into the rotated order of conversation, they were informed they would be doing either In Vivo followed by VR or VR followed by In Vivo. In Vivo was explained to the participants as; “Right here, just as we are now,” and VR as “In virtual reality, where we will talk to each other, but while wearing ‘this’ headset and ‘these’ earphones.” Participants were instructed by the facilitator, a male research staff member with graduate training (blind to study hypotheses), would pick the topic of both conversations, leave the room for 3 min so that they might prepare any thoughts on the topic, then reenter, at which point the conversation would begin. Although participants were not given specifics regarding the facilitator’s dialogue, they were told they would be asked their opinions on a topic, as well as questions regarding any opinions or conclusions they might express. All sessions were video recorded for adherence coding. The adherence coding form was adapted from our previous trials in prolonged exposure therapy for posttraumatic stress disorder and exposure with ritual prevention for obsessive compulsive disorder (Foa et al., 2013; Gilboa-Schechtman et al., 2010). The form included 2 questions on essential elements, 6 questions on essential but not unique elements, and 2 questions on adherence overall. The rater then took into account the Likert items to make a categorical adherence rating. After providing a baseline fear rating (0 [no fear] to 100 [most fear imaginable]), participants were given the list of possible conversation topics for the In Vivo and VR conversations (same sex marriage, abortion, Iraq war, Evolution/Creationism in schools, torture to gather intelligence, homosexuals in the military, most embarrassing moment). The topics were similar to those used in previous social anxiety studies (Smits et al., 2006). To increase the likelihood of participant engagement in a dialogue with the facilitator, topics were contemporary and relatively controversial. Participants were instructed to rate their highest predicted level of fear during a 5-min conversation for each of the seven topics, with ratings to be given for conversation in both the In Vivo and in the VR environment. The topic with the highest predicted SUDs rating was selected as the topic of conversation. The same topic was used for both conversations. In the event of equivalent ratings across multiple topics the facilitator exercised his discretion in selecting a topic. The participant was then given the topic and the facilitator left the room for 3 min while the participant prepared their thoughts on the topic.
400
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403
Fig. 2. VR conversation avatar.
Fig. 1. Participant and therapist converse in VR.
In the VR conversation condition the facilitator introduced the VR Head Mounted Display and headphones, and instructed/aided the participant in putting them on (see Fig. 1). In the In Vivo condition the facilitator simply sat to the left of the participant on a couch. Their distance from the participant was matched to the distance depicted in the VRET scene. In both the VR and In Vivo conditions, the facilitator began timing the conversation and stated, “Okay, now we’re going to have (selected topic inserted). To begin, what a conversation about do you personally think about ?” During the course of the conversation the participant was asked several open-ended questions from a predetermined list to ensure the conversation would continue for the 5-min duration (e.g. “And why do you believe that? How did you come to believe that?). In addition, the facilitator was instructed to voice a viewpoint in opposition to whatever position the participant chose for a given topic. After 5-min the conversation with the participant was stopped, and they were asked to rate their highest level of fear during the 5 min conversation. Immediately after completing the first conversation, procedures began for the second conversation (i.e. 3-min for the participant to prepare their thoughts on the topic). After completing both conversations, participants were thanked for their participation, and the session was concluded. Given the aim of this study, participants were also invited to comment on the experience after debriefing. 2.4. Materials The hardware and software used in this study was the same as that used in previous studies of this technology (Jouriles et al., 2009; Jouriles et al., 2011). The VR conversation was facilitated through the use of a virtual avatar interacting with the participant in a virtual environment (VE). The Virtual environment was created from modification (mod) templates based around the video game HalfLife 2® . The avatar was a modified character from the game, and designed to look like a typical male university student. The virtual environment was designed to look like a college dorm or apartment bedroom (e.g. bed, television, movie posters, computer desk, couch). From the participants virtual perspective the avatar was seated to their left on a couch (same position as in vivo condition) (see Fig. 2). The participant interacted with the virtual environment via the use of an eMagine z800 stereovision head mounted display (HMD). The HMD allowed the participant to view the virtual environment. Furthermore, head tracking hardware built into the HMD allowed the VR software to sense the participants head movements and adjust their view accordingly, thus giving participants fully immersive control of where and what they viewed within the virtual environment.
In addition to the HMD, in the VR conversation condition, the participant wore a pair of over-the-ear noise cancellation headphones. The facilitator used a Microsoft Lifechat LX-2000 microphone headset to speak with the participant. The microphone-headphone communication was relayed through the VR software, so that as the facilitator spoke, the virtual avatars mouth would move thereby approximating the look of normal speech. Furthermore, the facilitator, via keyboard command, was also able to change the avatar’s facial expressions (e.g. happy, sad, angry), as well as a number of gestures (e.g. nodding, pointing at the participant, shrugging, throwing his hands in air). All hardware and software was run through a personal computer using a Pentium IV® CPU, 512 megabytes of RAM, and a 6800 Graphics processor (Jouriles et al., 2009). 2.5. Measures 2.5.1. Liebowitz Social Anxiety Scale (LSAS) The LSAS is a widely used 24-item interviewer-rated instrument that assesses fear and avoidance of social situations. Participants rate fear items on a 4-point scale ranging from 0 (none) to 3 (severe). Likewise, avoidance is also rated on a 4-point scale ranging from 0 (never) to 3 (usually). The LSAS takes approximately 20–30 min to complete. The LSAS shows good reliability (˛ = 0.81–0.92) and validity (Cox, Ross, Swinson, & Direnfeld, 1998; Safren et al., 1999). Scores may be interpreted as: 55–65 Moderate, 65–80 Marked, 80–95 Severe, >95 Very Severe. The mean for participants with social anxiety disorder is 67.2 (SD = 27.5) (Heimberg et al., 1999). 2.5.2. State-Trait Anxiety Inventory (STAI) The state-trait anxiety inventory is a 20-item self-report measure of state and trait anxiety (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983). The measure takes about 10 min to complete. The STAI shows good reliability (˛ = 0.86–0.95) and validity (Spielberger et al., 1983). Interpretation of scores: Adults (M = 35 SD = 10), Generalized Anxiety Disorder (47–61), Panic Disorder with Agoraphobia (51–54), Panic Disorder (44–46). 2.5.3. Subjective Units of Disturbance (SUDs) The SUDs scale is a measure of subjective anxiety (0 = No Disturbance to 100 = Most Disturbance) that was rated by participants after the conversations (Wolpe, 1958; Wolpe, 1990). Participants reported the highest level that their SUDs reached. 2.5.4. Presence/immersion The measure of presence/realism was a 9-item immersion questionnaire used in our two previous similar studies with women resisting unwanted sexual advances (Jouriles et al., 2009; Jouriles et al., 2011). Items were rated on a 0 (Not at All) to 5 (Very Much)
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403
401
3.5. Feasibility, acceptability/tolerability, and qualitative responses
Fig. 3. Presence and fear as a function of VR or in vivo conversation.
scale (e.g. Did you ever feel that you might actually be defending your position to a stranger? Were you involved in the role play to the extent that you lost track of time?, etc.). The scale showed excellent reliability with this sample (Cronbach’s ˛ = 0.82; Guttman Split-half = 0.84).
No participants dropped out of the trial early and 100% completed both conversations. Participants comments after debriefing included: (a) the VR condition would be enhanced by having the facilitator/actor sit in another room rather than controlling the VR from the couch next to the participant, (b) realism and fear would be greater if participants also did not meet the facilitator/actor before the conversation in VR, and (c) the controversial conversation was always easier the second time through the same topic (suggesting a between-subjects design in the future). However, post hoc analysis with a paired samples t-test suggested that fear ratings were still significantly higher in the VR condition when only analysing the first conversations (p = 0.024). Also, a post hoc paired samples t-test showed enhanced fear activation was not due solely to differences in predicted fear between the VR (M = 34.36) and in vivo (M = 31.91) conditions (t = −0.88, p = 0.405). Finally, 10% of videos (n = 3) were observed for adherence to the pre-determined probes stated earlier. Ratings suggested that actor performance was consistent in both the VR and In vivo conditions.
4. Discussion 3. Results There were no dropouts and no data missing in this small pilot trial. Thus all results represent completer analyses. Overall, fear ratings ranged between 5 and 90 and correlations between the outcome measures ranged between −0.49 and 0.59. None of these correlations were significant except for the correlation between the rated presence in the VR and in vivo conditions (r = 0.53, p = 0.006). 3.1. Randomization check ANOVAs showed there were no significant differences in participants randomized to VR then in vivo compared to those randomized to in vivo then VR (age p = 0.42, LSAS total p = 0.14, VR presence rating p = 0.92, in vivo presence rating p = 0.67, baseline SUDs p = 0.81, VR peak SUDs p = 0.42, & in vivo peak SUDs p = 0.71). 3.2. Order effects An ANOVA showed there was no main effect for order (VR first or in vivo first; F(1,25) = 0.17, p = 0.68) or condition by order interaction (F(1,24) = 1.35, p = 0.26). 3.3. Fear ratings Hypothesis (a) paired samples t-tests and Fig. 3 show that consistent with prediction, participants rated their peak fear significantly higher during VR conversation (M = 38.46, SD = 16.78) than at baseline (M = 16.73, SD = 11.99, t(25) = 8.26, p = 0.000, d = 2.29). Hypothesis b) Contrary to prediction, participants rated their peak fear significantly higher during conversations in the VR condition (M = 38.46, SD = 16.78) compared to the in vivo condition (M = 29.23, SD = 10.36, t(25) = 3.08, p = 0.005, d = 0.85). 3.4. Presence (realism) ratings Hypothesis (c) Consistent with prediction, a paired samples ttest showed that participants rated conversations as significantly more realistic (presence) in vivo (M = 29.35, SD = 5.73) compared to in VR (M = 26.27, SD = 5.58, t(25) = 2.67 p = 0.013, d = 0.74).
Virtual reality technology increases access to exposure therapy for patients with social anxiety. However, the primary limitation of current technology is that the operator is limited to pre-programed avatars that cannot be controlled to interact/converse with the patient in real time. The current study piloted new technology allowing the operator to directly control the avatar (including speaking) during VR social interactions. We used a within-subjects incomplete repeated measures design with random starting order with rotation counterbalancing. Results showed that: (a) in vivo conversation was rated as more realistic than VR conversation, (b) fear ratings were higher during VR conversation compared to in vivo conversation, (c) no participants dropped out, and 100% completed both VR and in vivo conversations, and (d) qualitative participant comments suggested that the VR conversations would be more realistic if they did not meet the actor/operator and if they were not in the same room as the participant. Although in vivo conversation was rated as more realistic, participants reported higher fear ratings during the VR conversation condition. Previous studies suggest that for reducing anxiety with exposure therapy, fear ratings are more predictive of efficacy than realism ratings (Foa & Kozak, 1986; M. Krijn, P. M. G Emmelkamp, R. Biemond, C. de Wilde de Ligny, M. J. Schuemie, & C. A. P. G. van der Mast, 2004; M. Krijn, P. M. G Emmelkamp, R. P. Olafsson, & R. Biemond, 2004; Powers & Emmelkamp, 2008; Price et al., 2011). Realism or presence is the extent to which one interprets the virtual environment as if it were real (Lee, 2004). Consistent with our findings, studies that have manipulated presence found that it did not significantly affect treatment outcome (M. Krijn, P. M. G Emmelkamp, R. Biemond, et al., 2004; M. Krijn, P. M. G Emmelkamp, R. P. Olafsson, et al., 2004; Price & Anderson, 2007; Price et al., 2011). We can only speculate why fear ratings were higher in the VR conversation. Potential reasons could include: context (e.g. party atmosphere vs couch in psychology department; greater activation of the fear structure through a higher number of fear relevant stimuli), discomfort of simply being in any VR environment under any circumstance, less prior experience with the avatar vs. the actor (actor also conducted consent), or other reasons. Future studies may help clarify why fear ratings are higher in VR compared to in vivo conversations. The dropout and completion rates are also encouraging for further development of this technology for social anxiety applications. These initial pilot data are encouraging for
402
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403
exposure applications. In fact, if these findings are replicated and extended, VR exposure could even prove more effective as a clinical tool than in vivo exposure. Several limitations deserve comment. First, participants were not selected based on social anxiety scores for this feasibility pilot study. Future studies should determine if realism and fear ratings generalize to those with high social anxiety. Second, participants stated that presence would be enhanced if they did not meet the facilitator/actor controlling the avatar prior to the conversation and if the actor also controlled the avatar from another room. Although the participants reported that the noise cancellation headphones and head mounted display were effective, they were still aware of the presence of the actor on the couch next to them and that this thought was at times distracting from the realism of the VR conversation. Third, our small sample does not permit analysis of multiple dimensions of presence (spatial presence, involvement, and realness) (Price et al., 2011). Future studies could help determine which dimensions are enhanced through avatar interaction. Fourth, this study does not indicate if real time interaction is superior to standard VRET for social anxiety. Future studies should compare VRET with and without real time avatar interaction to determine the relative efficacy. Fifth, participants were given course credit based on their participation. Thus, finding a 100% completion rate may not generalize to other populations. Sixth, although the actor was instructed to only use facial expressions, gestures, and gaze consistent with the avatar condition, this control was not verified. Several lines of research suggest the importance of gestures, facial expressions, gaze, and mimicry in social anxiety (J. N. Vrijsen, W. G. Lange, E. S. Becker, & M. Rinck, 2010; J. N. Vrijsen, W. G. Lange, R. Dotsch, D. H. J. Wigboldus, & M. Rinck, 2010; Wieser, Pauli, Grosseibl, Molzow, & Muhlberger, 2010). Future studies should also code and rate in vivo actors on gestures and facial expressions to match the VRET condition. Seventh, although SUDs fear ratings are often used as outcome data, future studies should incorporate additional measures of distress (e.g. behavioural approach, psychophysiology, etc.). Eighth, although content choices for the conversation topics have been used previously (e.g. Smits et al., 2006), future studies may help determine the optimal content choices for social anxiety exposures. Finally, participants only completed each conversation condition (VR & in vivo) once. Thus, habituation/extinction could not be measured. Overall, the data suggest that the novel technology allowing real time interaction/conversation in VR may prove useful for the treatment of social anxiety in future studies. Future studies should include participants with high social anxiety, not have the participant meet the actor/therapist prior to the session, and have the therapist control the avatar from a different room.
References Anderson, P. L., Zimand, E., Hodges, L. F., & Rothbaum, B. O. (2005). Cognitive behavioral therapy for public-speaking anxiety using virtual reality for exposure. Depression and Anxiety, 22(3), 156–158. http://dx.doi.org/10.1002/da.20090 Barlow, D. H., Leitenberg, H., Agras, W. S., & Wincze, J. P. (1969). The transfer gap in systematic desensitization: an analogue study. Behaviour Research and Therapy, 7(2), 191–196. Carlbring, P., Bohman, S., Brunt, S., Buhrman, M., Westling, B. E., Ekselius, L., et al. (2006). Remote treatment of panic disorder: a randomized trial of internet-based cognitive behavior therapy supplemented with telephone calls. American Journal of Psychiatry, 163(12), 2119–2125. http://dx.doi.org/10.1176/appi.ajp.163.12.2119 Carlbring, P., Gunnarsdottir, M., Hedensjo, L., Andersson, G., Ekselius, L., & Furmark, T. (2007). Treatment of social phobia: randomised trial of internet-delivered cognitive-behavioural therapy with telephone support. British Journal of Psychiatry, 190, 123–128. Cox, B. J., Ross, L., Swinson, R. P., & Direnfeld, D. M. (1998). A comparison of social phobia outcome measures in cognitive-behavioral group therapy. Behavior Modification, 22(3), 285–297.
Dyckman, J. M., & Cowan, P. A. (1978). Imaging vividness and the outcome of in vivo and imagined scene desinsitization. Journal of Consulting and Clinical Psycghology, 46(5), 1155–1156. Emmelkamp, P. M. G. (2005). Technological innovations in clinical assessment and psychotherapy. Psychotherapy and Psychosomatics, 74(6), 336–343. Emmelkamp, P. M. G., & Wessels, H. (1975). Flooding in imagination vs flooding in vivo: a comparison with agoraphobics. Behaviour Research and Therapy, 13(1), 7–15. Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: exposure to corrective information. Psychological Bulletin, 99(1), 20–35. Foa, E. B., Simpson, H. B., Liebowitz, M. R., Powers, M. B., Rosenfield, D., Cahill, S. P., et al. (2013). Augmenting SRI treatment with exposure and ritual prevention (EX/RP) for obsessive-compulsive disorder: long-term outcome. Journal of Clinical Psychiatry, in press. Freiheit, S. R., Vye, C., Swan, R., & Cady, M. (2004). Cognitive-behavioral therapy for anxiety: is dissemination working? Behavior Therapist, 27(2), 25–32. Garau, M., Slater, M., Pertaub, D. P., & Razzaque, S. (2005). The responses of people to virtual humans in an immersive virtual environment. Presence, 14, 104–116. Gilboa-Schechtman, E., Foa, E. B., Shafran, N., Aderka, I. M., Powers, M. B., Rachamim, L., et al. (2010). Prolonged exposure versus dynamic therapy for adolescent PTSD: a pilot randomized controlled trial. American Academy of Child and Adolescent Psychiatry, 49, 1034–1042. Goisman, R. M., Warshaw, M. G., & Keller, M. B. (1999). Psychosocial treatment prescriptions for generalized anxiety disorder, panic disorder, and social phobia, 1991–1996. American Journal of Psychiatry, 156(11), 1819–1821. Harris, S. R., Kemmerling, R. L., & North, M. M. (2002). Brief virtual reality therapy for public speaking anxiety. Cyberpsychology and Behavior, 5(6), 543–550. Heimberg, R. G., Horner, K. J., Juster, H. R., Safren, S. A., Brown, E. J., Schneier, F. R., et al. (1999). Psychometric properties of the Liebowitz Social Anxiety Scale. Psychological Medecine, 29(1), 199–212. Hofmann, S. G., & Smits, J. A. (2008). Cognitive-behavioral therapy for adult anxiety disorders: a meta-analysis of randomized placebo-controlled trials. Journal of Clinical Psychiatry, 69, 621–632. Jouriles, E. N., McDonald, R., Kullowatz, A., Rosenfield, D., Gomez, G. S., & Cuevas, A. (2009). Can virtual reality increase the realism of role plays used to teach college women sexual coercion and rape-resistance skills? Behavior Therapy, 40(4), 337–345. http://dx.doi.org/10.1016/j.beth.2008.09.002 Jouriles, E. N., Simpson Rowe, L., McDonald, R., Platt, C. G., & Gomez, G. S. (2011). Assessing women’s responses to sexual threat: validity of a virtual role-play procedure. Behavior Therapy, 42(3), 475–484. http://dx.doi.org/10.1016/j.beth.2010.11.005 Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62(6), 593–602. Klinger, E., Bouchard, S., Legeron, P., Roy, S., Lauer, F., Chemin, I., et al. (2005). Virtual reality therapy versus cognitive behavior therapy for social phobia: a preliminary controlled study. CyberPsychology and Behavior, 8(1), 76–88. Krijn, M., Emmelkamp, P. M. G., Biemond, R., de Wilde de Ligny, C., Schuemie, M. J., & van der Mast, C. A. P. G. (2004). Treatment of acrophobia in virtual reality: the role of immersion and presence. Behavior Research and Therapy, 42(2), 229–239. Krijn, M., Emmelkamp, P. M. G., Olafsson, R. P., & Biemond, R. (2004). Virtual reality exposure therapy of anxiety disorders: a review. Clinical Psychology Review, 24(3), 259–281. Lange, A., Rietdijk, D., Hudcovicova, M., van de Ven, J. P., Schrieken, B., & Emmelkamp, P. M. G. (2003). Interapy: a controlled randomized trial of the standardized treatment of posttraumatic stress through the internet. Journal of Consulting and Clinical Psychology, 71(5), 901–909. Lange, A., Schrieken, B., Van De Ven, J. P., Bredeweg, B., Emmelkamp, P. M. G., Van Der Kolk, J., et al. (2000). ‘Interapy’: the effects of a short protocolled treatment of posttraumatic stress and pathological grief through the internet. Behavioural and Cognitive Psychotherapy, 28(2), 175–192. Lange, A., Van de Ven, J. P., Schrieken, B., & Emmelkamp, P. M. G. (2001). Interapy. Treatment of posttraumatic stress through the Internet: a controlled trial. Journal of Behavior Therapy and Experimental Psychiatry, 32(2), 73–90. Lee, K. M. (2004). Presence, Explicated. Communication Theory, 14, 27–50. Litvak, S. B. (1969). A comparison of two brief group behavior therapy techniques on the reduction of avoidance behavior. The Psychological Record, 19, 329–334. Opris, D., Pintea, S., Garcia-Palacios, A., Botella, C., Szamoskozi, S., & David, D. (2012). Virtual reality exposure therapy in anxiety disorders: a quantitative meta-analysis. Depression and Anxiety, 29(2), 85–93. http://dx.doi.org/10.1002/da.20910 Powers, M. B., & Emmelkamp, P. M. G. (2008). Virtual reality exposure therapy for anxiety disorders: a meta-analysis. Journal of Anxiety Disorders, 22, 561–569. Powers, M. B., Sigmarsson, S. R., & Emmelkamp, P. M. G. (2008). A meta-analytic review of social phobia treatments. International Journal of Cognitive Therapy, 1, 94–113. Price, M., & Anderson, P. (2007). The role of presence in virtual reality exposure therapy. Journal of Anxiety Disorders, 21(5), 742–751. http://dx.doi.org/10.1016/j.janxdis.2006.11.002 Price, M., Mehta, N., Tone, E. B., & Anderson, P. L. (2011). Does engagement with exposure yield better outcomes? Components of presence as a predictor of treatment response for virtual reality exposure therapy for social phobia. Journal of Anxiety Disorders, 25(6), 763–770. http://dx.doi.org/10.1016/j.janxdis.2011.03.004 Safir, M. P., Wallach, H. S., & Bar-Zvi, M. (2012). Virtual reality cognitive-behavior therapy for public speaking anxiety: one-year follow-up. Behavior Modification, 36(2), 235–246. http://dx.doi.org/10.1177/0145445511429999
M.B. Powers et al. / Journal of Anxiety Disorders 27 (2013) 398–403 Safren, S. A., Heimberg, R. G., Horner, K. J., Juster, H. R., Schneier, F. R., & Liebowitz, M. R. (1999). Factor structure of social fears: the Liebowitz Social Anxiety Scale. Journal of Anxiety Disorders, 13(3), 253–270, doi: S0887-6185(99)00003-1 [pii]. Sherman, A. R. (1972). Real-life exposure as a primary therapeutic factor in the desensitization treatment of fear. Journal of Abnormal Psychology, 79(1), 19–28. Smits, J. A., Powers, M. B., Buxkamper, R., & Telch, M. J. (2006). The efficacy of videotape feedback for enhancing the effects of exposure-based treatment for social anxiety disorder: a controlled investigation. Behaviour Research and Therapy, 44(12), 1773–1785. Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the state-trait anxiety inventory. Palo Alto, CA: Consulting Psychologists Press. Vrijsen, J. N., Lange, W. G., Becker, E. S., & Rinck, M. (2010). Socially anxious individuals lack unintentional mimicry. Behaviour Research and Therapy, 48(6), 561–564. http://dx.doi.org/10.1016/j.brat.2010.02.004 Vrijsen, J. N., Lange, W. G., Dotsch, R., Wigboldus, D. H. J., & Rinck, M. (2010). How do socially anxious women evaluate mimicry? A Virtual reality study. Cognition and Emotion, 24, 840–947.
403
Wallach, H. S., Safir, M. P., & Bar-Zvi, M. (2009). Virtual reality cognitive behavior therapy for public speaking anxiety: a randomized clinical trial. Behavior Modification, 33(3), 314–338. http://dx.doi.org/10.1177/0145445509331926 Watson, J. P., Mullett, G. E., & Pillay, H. (1973). The effects of prolonged exposure to phobic situations upon agoraphobic patients treated in groups. Behavior Research and Therapy, 11(4), 531–545. Wieser, M. J., Pauli, P., Grosseibl, M., Molzow, I., & Muhlberger, A. (2010). Virtual social interactions in social anxiety – the impact of sex, gaze, and interpersonal distance. Cyberpsychology, Behavior and Social Networking, 13(5), 547–554. http://dx.doi.org/10.1089/cyber.2009.0432 Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press. Wolpe, J. (1990). The practice of behavior therapy. New York: Pergamon Press.