Virtual Training For Fear Of Public Speaking – Design Of an Audience For Immersive Virtual Environments Sandra Poeschl * Ilmenau University of Technology Research Group Media Psychology & Media Design Nicola Doering + Ilmenau University of Technology Research Group Media Psychology & Media Design ABSTRACT Virtual Reality technology offers great possibilities for Cognitive Behavioral Therapy on Fear of Public Speaking: Clients can be exposed to virtual fear-triggering stimuli (exposure) and are able to role-play in virtual environments, training social skills to overcome their fear. This poster deals with the design of a realistic virtual presentation scenario based on an observation of a real audience. INDEX TERMS: J.4 [Computer Applications]: Social and Behavioral Sciences—Psychology; I.3.7 [Computing Methodologies]: Three-Dimensional Graphics and Realism— Virtual Reality 1 INTRODUCTION Fear of Public Speaking (or glossophobia or speech anxiety) is a common psychological disorder and characterized by anxiety prior to or at the thought of having to communicate verbally with any group of people. It leads to avoidance of those situations, physical distress and even panic. Treatment typically involves Cognitive Behavioral Therapy (CBT), including exposure to fear- triggering stimuli (e. g., speaking in front of a group), social skills training and relaxation training [1]. Virtual Reality technology as a therapeutic or training tool offers great possibilities for CBT: Clients can be exposed to virtual fear-triggering stimuli (exposure) and are able to role-play in virtual and different scenarios, training social skills to overcome their fear [2]. Usually, prototypical (neutral, social and anti-social) audience behavior serves as stimulus in virtual training sessions, although there is significant lack of theoretical basis on typical and realistic audience behavior. Presence (a user’s subjective psychological response to a VR system) is related to the experience of emotions [3], like fear. Therefore, using prototypical behavior – instead of realistic audience behavior – may lead to lower experiences of presence and to lesser performance in VR training applications. The work presented in this poster aims at overcoming this problem by designing an application including realistic audience behavior. We describe an explorative observation study on realistic audience behavior. The results obtained are used to design a virtual training scenario for fear of public speaking in a CAVE. 2 RELATED WORK Related work shows that virtual presentation environments can induce fearful experiences [4], especially when featuring negative audience behavior [5]. Also, VR exposure for Fear of Public speaking is more effective than waiting list control groups [6, 7]. In the studies mentioned, prototypical audience behavior was simulated. However, to further enhance transfer from virtual training to real life performance, modeling realistic audience behavior may be a promising next step. Our explorative study deals with this problem by gathering observational data on natural nonverbal audience behavior. 3 METHOD A real audience (consisting of n = 18 men and women) in an undergraduate seminar was observed in a structured, non- participant overt observation. We used event samples of three frontal lecture sessions, taping the lectures on video and analyzing the video material (see figure 1). Behavior frequency of four nonverbal dimensions (eye contact, facial expression, gesture, and posture; N = 5916 behavioral actions in total, coded into 35 categories) was rated by means of a quantitative content analysis, in regard to frequency and positioning across three rows of seats within the lecture room. Further, we analyzed the first, middle and last 15 minutes of the lecture sessions (with a duration of 90 min. each), as we assumed that nonverbal audience behavior may change over time (for example packing away things towards the end of a session). Figure 1. Screenshot of analyzed video material on audience behavior 4 RESULTS Selected behavior patterns (facial expressions and gesture) are presented by means of excerpts of the accomplished design manual. The audience to be designed will show behavioral actions according to a database that includes mean frequencies for one person per row and per minute. These are used to present our findings. Due to the multitude of facial expressions and gestures, findings are highlighted by prototypical behavior actions. 4.1 Facial expression As table 1 shows, friendly and neutral face expressions are rather common. Also, they are closely related, as a joyful facial expression changes into a neutral one when coming to an end. Social facial expressions increase with distance to the presenter, maybe to establish a closer contact. Further, neutral and social expressions increase with time. This can be explained by the fact that at the end of a lecture session, more interactions and * sandra.poeschl@tu-ilmenau.de + nicola.doering@tu-ilmenau.de 101 IEEE Virtual Reality 2012 4-8 March, Orange County, CA, USA 978-1-4673-1246-2/12/$31.00 ©2012 IEEE