Immersion and Presence

In a Virtual Environment (VE), Immersion, defined in technical terms, is capable of producing a sensation of Presence, the sensation of being there (part of the VE), as regards the user (Ijsselsteijn & Riva, 2003). Presence is indeed, historically, at the core of Virtual Reality (VR). Presence has often be conceived as a sign of "ecological validity" of VR devices, also as a sign of potential positive transfer of skills or knowledge learned in a VE to the real world. 1Immersion and Presence Virtual reality technology does allow users to have unique experiences, such as standing inside a molecule or foresee a future vehicle's interior before production has even started. In this sense, VR (and more generally computerized devices) really acts as a problem-solving device, transforming enormous quantities of mind-breaking data into "graspable illusions" [1]. In this sense, Rheingold [2] draws a naive history of computers. He particularly shows that, from the beginning, computers were conceived as "mind-amplifying" devices, helping the human operator process complex data . He also shows that a decisive step was made when researchers had the idea to connect a television screen to a computer, then a keyboard and a mouse, instead of having punched cards as inputs and number on a sheet of paper as output [3]. In this sense, the primary characteristic distinguishing VEs from other means of displaying information is the focus on immersion. In a technical acceptation of the term, immersion is achieved by removing as many real world sensations as possible, and substituting these with the sensations corresponding to the VE. Immersion is by essence related to the multi-modal nature of the perceptual senses, and also to the interactive aspects of a VR experience. From this viewpoint, immersion is intuitively related to the resemblance of the VR devices with human characteristics. These include the size of the human visual field, the stereoscopic aspects of the simulation, the "surround" aspects of the sound, that is the extent to which the computer displays are extensive, surrounding, inclusive, vivid and matching. The term 'immersion thus stands for what the technology delivers from an objective point of view. The more that a system delivers displays (in all sensory modalities) and tracking that preserves fidelity in relation to their equivalent real-world sensory modalities, the more that it is 'immersive' [4]. From this technological standpoint, immersion is intended to instill a sense of belief that one has left the real world and is now "present" in the virtual environment. This notion of "being present" in the virtual world has been considered central to VE [5]. Thus, whereas immersion is a "technology-related", objective aspect of VEs, presence is a psychological, perceptual and cognitive consequence of immersion. Presence is thought of as the psychological perception of "being in" or "existing in" the VE in which one is immersed [6-9]. To date, the utility of the presence construct, either to enhance interactive design or human performance, is not clearly established. In fact, the most direct evidence for a positive role of presence in the efficacy of VR comes from therapy applications of VR [10, 11]. Before discussing more precisely the potential interests of the quest for presence in industrial applications, we would like now to draw a current state-of-theart of current measures of presence, which will eventually clarify a little the concept itself. We will retrain our discussion to the (initial) concept of spatial presence (i.e. self-orientation and self-location with respect to a media environment, not the real environment). One has certainly to recognize that, beside spatial presence, social presence is certainly something users of VE (and more generally video games) are looking for. Social presence is defined as the sensation of interacting with other forms of intelligent agents in the VE [12]. It is in particular related to the presence of avatars in the VE, which is obviously part of the future developments of VR, and beyond the scope of this manuscript. We will thus use hereafter "presence" to mean "spatial presence". 2Measuring Presence 2.1Questionnaires and subjective measures In the 90's, researchers in the field of VR have developed questionnaires, trying to evaluate the subjective degree of presence [13]. Interestingly, these authors refer to presence as a “psychological state experienced as a consequence of focusing one’s energy and attention on a coherent set of stimuli”. In this definition, two points are mainly relevant. First, the "coherent set of stimuli" is a reference to the immersive characteristics of the VE (high fidelity graphics and sound, interaction, ..., with the general assumption that the more senses are stimulated and the more interaction in the EV, the more Immersive is the VE). Secondly, the focus is clearly put on cognitive processes, such as attention and situation awareness [14]. Indeed, Prothero et al. [15] argue that presence and situation awareness are related, since they both imply that observers perceive their self-location and self-orientation with respect to an environment. Presence can thus be defined as a special case of situation awareness, in which self-orientation and self-location are defined with respect to a media environment, not the real environment. In this conception , most of the questionnaires used to a access presence are using Visual Analog Scales, asking the subject to rate dimensions such as degree of control, sensorial immersion or realism (figure 1). Figure 1. An exemple of Visual Analog Scale. The subject is asked to rate his/her sensation on a 7-step scale, with opposite descriptors at each extremity of the scale (adapted from [13]) More precisely, Witmer & Singer [13] first systematized this methodology. They first suggested that presence was dependant not only on immersion [16], but also of the user's involvment (defined as "a consequence of focusing one’s energy and attention on a coherent set of stimuli or meaningfully related activities and events" [13]). Witmer & Singer list 4 categories of factors susceptible of contributing to the sensation of presence: Control factors (dealing with the user's interaction with the VE), Sensory factors (various aspects of the sensorial stimulation, such as multimodality), Distraction factors and Realism factors. Whereas the first two factors are mainly objective aspects of the VE, the last two are related to both objective and subjective determinants. For instance, Distraction factors include selective attention and Realism factors include the "meaningfulness of experience", which itself refers to the user's previous experience with the situation, among other factors. From this framework, the authors build 2 questionnaires: The Presence Questionnaire (PQ), which directly measures the degree to which subjects are sensitive to the factors presented above see figure 1, for example), and the Immersive Tendencies Questionnaire (ITQ), which tries to evaluate their tendency to become involved or immersed. In the ITQ, questions like "Are you easily disturbed when working on a task?" are presented. Obviously, such questionnaires are introspective, "post-immersive", tentative evaluations, and one might argue that they measure more of the subject's perception of the system's properties than presence per se [17, 18]. These questionnaires are certainly valuable, notably given their ease of use and the fact that they do not interfere with the user's experience while in the VE. They can be however be criticized, mainly because they do not clearly define what they are supposed to measure. Another inherent problem is the fact that they are insensitive, by essence, to time-varying qualities of presence during the exposure itself. In this matter, Slater & Steed [19] introduced a methodology, based on the idea that presence is by essence a bi-stable phenomenon, the subject oscillating between feeling "in the real world" and "in the virtual world" (due to various factors, such as temporal delays or poor realism in the VE). These authors thus designed experiments in which users had to report "breaks in presence". However, they can be used in conjunction with other measures, such as the subject's behavior and /or performance in a VE, as well as physiological measurements. This point is notably emphasized in [20], noting in particular that body movement have to be coherent with the spatial structure of the environment the subject experiences as "reality". 2.2Physiological measures To pursue a little on Slater and col. work, a famous experiment of this group involves a "pit room". In this "pit room" (figure 2) the participant walks into a virtual room with an 18m precipice at its center. In fact, the person stands and walks in a CAVE [21] system, in which a wooden ledge is positioned at the place where the virtual edge lies in the VE. This meant to corroborate visual signals, indicating that they are standing in a room with a dangerous precipice. Figure 2. The "visual pit" virtual environment, with the "pit" on the left and the "training" room on the right. Slater [20] interestingly note that heart rate increases when subjects reach the virtual precipice, suggesting that physiological measures might be an indicator of presence (see also [22]). Indeed, a number of physiological indicators including skin conductance and temperature, muscular tension, cardiovascular responses, and pupilometry have been suggested as presence measures [18]. The idea is that these indicators can deliver continuous information regarding the effects of specific environmental stimuli or events experienced in a VE. For example, Strickland and Chartier [23] illustrate the feasibility of recording EEG (electroencephalogram) signals in a head mounted display. Measuring and interpreting the differences in cortical responses in real and virtual environments may lead to a better understanding of the effects of various software and hardware influences in a virtual environment. For example, change in heart rate measures the increase or decrease in the number of heartbeats per minute, and can be quite easily measured with an electrocardiogram (ECG). Skin conductance changes when sweat is produced in the palm of the hand, and is commonly associated with stress and reaction to unexpected stimuli. These measures are now widely used, in connection with subjective ratings of presence [22, 24]. For example, Meehan [22] used heart rate, skin temperature, and galvanic skin response in the "pit room" experiment (figure 2). He showed that the physiological measures reliably distinguished between the training room (visually safe) and the pit room. Recently, studies have started using fMRI (functional magnetic resonance) to study brain activity associated to the sensation of presence [25]. For now, it appears that valuable information can be collected using physiological signals. However, beside the fact that some methods might be intrusive and/or relatively unreliable, it seems important to distinguish two levels of analysis. In Meehan's study [22], average heart rate is compared between the time the subject is in the training (control) room and in the pit room, and results are quite clear. However, this is quite a large temporal-scale level of analysis. It might be interesting to try to go down to a more fine-grained temporal level of analysis. This is especially important when one wants to correlate physiological data with actions or events in the VE. And this is where things get more complicated to analyze. Many events (internal and external) can affect a person’s heart rate, for instance: stress, fear, exertion, emotion, etc. Heart increases when a person is under stress. Heart rate is also affected by emotions, by unexpected stimuli, ... Nevertheless, for validity and usefulness, one would like to correlate events happening in the VE with changes in physiological signals. In other terms, an eventrelated analysis is required. It appears that things are not there yet, although moving in this direction. This is an important issue, since physiological data appear as an objective indicator, which has to be related to behavioral data, in order to be fully functional in the evaluation of presence, notably in relation with the subject's performance in the VE. 2.3Behavioral measures