|
Article Excerpt
INTRODUCTION
Virtual environments (VEs) are available for countless applications in fields as diverse as medicine, training, and entertainment and other leisure-time activities (Stone, 2002). Formerly reserved for the military and scientific endeavors, this surge of new applications has caused a large number of people to be exposed to this technology, including children and civilian adults, who may not be aware of the potential adverse effects of VE exposure. With increased use of VE systems by the general public has come a rise in the number of users experiencing negative side effects (Greenfield, 1994). This situation is building the ground for potential social (Calvert, 2002), legal (Kennedy, Kennedy, & Bartlett, 2002), and economic (Swann & Stone, 2002) repercussions. Thus there is a great need to research and understand the causes and to establish safe parameters for VE exposure (Stanney et al., 1998).
Some of the most common adverse effects associated with VE exposure are dizziness, drowsiness, headache, nausea, fatigue, and general malaise (Kennedy, Lane, Berbaum, & Lilienthal, 1993). Collectively, these symptoms are often referred to as cybersickness (McCauley & Sharkey, 1992). In addition to these problems experienced during exposure, aftereffects often linger, including disturbed proprioception (Lampton et al., 1994; Rolland, Biocca, Barlow, & Kancherla, 1995; Stanney, Kennedy, Drexler, & Harm, 1999) and postural instability (DiZio & Lackner, 2002; Kennedy & Stanney, 1996).
Why are VE systems associated with such adverse effects? Although there is no exact science of cybersickness (also known as motion sickness), a few theories exist. The most widely accepted theory is the sensory conflict theory put forth by Reason (1978) and Reason and Brand (1975). This theory suggests that conflicts between sensory inputs--either immediately present to an observer (e.g., visual motion without concordant vestibular stimulation) or between current patterns of input and those anticipated based on experience (e.g., when a visual scene updates later than expected because of lag)--lead to conflict in the neural mechanisms responsible for interpreting and responding to orientation and self-motion (Money, 1990). According to Treisman (1977), such conflict triggers defense mechanisms that respond to minimal physiological disturbances, such as would be produced by an absorbed toxin (i.e., the poison theory). Thus motion sickness is seen as a reflex (i.e., nausea, vomiting) provoked by a response to an artificial stimulus (i.e., sensory rearrangements).
Based on the sensory conflict and poison theories, VE system designers should strive to reduce intersensory conflicts (i.e., those arising from missing or mismatched modalities) as well as those associated with sensory expectations established through experience (e.g., depth and distance distortions, form and size distortions, delays of sensory feedback; Welch, 2002). Although the sensory conflict and poison theories provide a conceptual framework within which to characterize motion sickness, both theories lack predictive power to indicate when sickness will occur and how severe it will be, and neither can account for individual susceptibility differences.
An alternative theory is the ecological theory of motion sickness (Riccio & Stoffregen, 1991). This theory suggests that motion sickness is caused by postural instability associated with environmental situations (i.e., low-frequency vibration, weightlessness, perturbed gravitoinertial force vectors, altered specificity) that destabilize the postural control system. Virtual environments are suggested to destabilize postural control through altered specificity (i.e., visually specified accelerations and rotations that lack correlated bodily forces). According to this theory, those who interact with a VE system should probably be seated or provided with a support bar to assist with maintaining postural control. Although the postural stability theory can account for individual susceptibility (i.e., those who can maintain postural control should avoid sickness), it too fails to have predictive power to indicate which VE systems will lead to postural instability.
From a theoretical perspective, because it is not currently possible to predict which VE systems will be the most disturbing (and thereby to remedy the cause), it may prove effective to predict those who are most susceptible to cybersickness and the conditions under which adverse effects are the most severe. Past research has indicated that individual susceptibility to motion sickness is influenced by gender, motion sickness history, prior experience, and overall state of health, among other factors (Kolasinski, 1995; McFarland, 1953; Mirabile, 1990; Reason & Brand, 1975; Kingdon, Stanney, & Kennedy, 2001; Stanney et al., 1998). In addition, recent research suggests that streamlining navigational control may reduce, by nearly half, the level of adverse effects associated with complete navigational control (Stanney & Hash, 1998): 20% or more of the variance in simulator sickness is governed by the kinematics of the visual scene (i.e., scene complexity), and an additional 20% or more of the variance may be determined by exposure duration (Kennedy, Berbaum, Dunlap, & Smith, 1995; Kennedy, Stanney, & Dunlap, 2000). Influences of other factors should also be considered, including technical system factors such as optical distortion, field of view. flicker, motion platforms, refresh rate, resolution, transport delays, and update rate; these have been reviewed elsewhere (Biocca, 1992; Kolasinski, 1995; Pausch, Crea, & Conway, 1992).
System Design and Usage Factors
Navigational control. Stanney and Hash (1998) presented empirical evidence indicating that the extent of motion sickness experienced by VE users will be directly related to the level of navigational (i.e., movement) control provided to users. Further, Hettinger and Riccio (1992) suggested that an examination of motion sickness would be incomplete without considering operator control behavior and the opportunity for self-initiated user interaction. These relations can be likened to a common experience--namely, that the driver of an automobile rarely if ever experiences motion sickness, whereas passengers often are afflicted (Casali, 1986). Several other studies (Held, 1965; Stott, 1990) have provided evidence suggesting that active and voluntary movements made when users have control over their own motion may provide the key to efficiently adapting to sensory rearrangements, such as those found in virtual environments. Further study is needed to determine the types of symptoms that different levels of navigational control cause, whether or not there are benefits to streamlining user control, and whether or not those exposed for long durations adapt to high levels of control, thus becoming less in over time.
Scene complexity. Although Pausch et al. (1992) suggested that scene complexity has minimal effects on motion sickness, several researchers (Hettinger, 2002: Kennedy & Fowlkes, 1992; McCauley & Sharkley, 1992) have suggested that the rate of visual flow (i.e., visual scene complexity) influences the incidence--and, more so, the severity--of motion sickness experienced by an individual, possibly because of positional data latencies of head-tracking hardware (DiZio & Lackner, 1997). This is probably attributable to the relationship between vection (i.e., illusory self-motion) and the spatial and temporal frequency of optical patterns (i.e., scene complexity; Hettinger, 2002). Vection, which is often associated with heightened motion sickness, generally increases with increased scene detail. More specifically, Howard (1986) indicated that vection is related to the optical texture density of a scene. Owen, Wolpert, and Warren (1985) further suggested that "edge rate" is more influential than global optical flow. Thus scenes with greater texture and more edges, such as might be found with high ceilings, as opposed to low ceilings, may produce high levels of vection and, in turn, sickness. Hettinger (2002) further suggested that the size of the visual field and the presence of movement in the background (i.e., periphery), as opposed to motion stimulation in the foreground, influence vection. It is important to further investigate such influences of scene complexity on VE sickness to determine if there are benefits to visually simplifying scenes.
Exposure duration. Exposure duration and number of repeat exposures have been shown to affect the level of motion sickness experienced. Kennedy et al. (2000) demonstrated that exposure duration is positively related and that repetition is negatively related to total sickness across a wide variety of simulators. Similarly, Fowlkes, Kennedy, and Lilienthal (1987) found that the intensity and duration of postural instability associated with exposure to a simulator increased with prolonged exposure. Because of these issues, the U.S. Army Research Institute (Knerr et al., 1998) has suggested that VE exposures should be limited to 15 rain, a period that may be too short for some training, educational, or analysis-based applications. Means of extending exposure duration while minimizing adverse effects are required. It may be, for example, that if navigational control is streamlined and visual scenes are simplified, exposure duration can be extended without adverse effects. These interrelationships between system design and usage factors need to be further examined.
Individual Factors
Gender. It is generally suggested that females experience greater motion sickness than do males (Kennedy, Lanham, Drexler, & Lilienthal, 1995). Biocca (1992) reported that this difference may be attributable to males being reticent to report sickness, as he suggested that males and females do not differ in their sensory response to motion stimuli. Kennedy and Frank (1985), however, found that females may have larger fields of view than males do, and thus their sensory experience may indeed be different from that of males. DoNe, May, McBride, and DoNe (2001) found that females experience significantly more motion sickness than do males on devices in which the groups had similar exposure history, regardless of age and level of physical activity, with no support for a lack of reporting from males. Further gender studies, specifically in VE systems, are required to clarify these differences.
Motion sickness histories. The incidence of motion sickness varies greatly among individuals; some appear immune, whereas others are highly susceptible (Kennedy, Hettinger, & Lilienthal, 1990). These differences are suggested to be attributable to individual factors, such as unstable binocular vision, individual variations in interpupillary distance, susceptibility to photic seizures and migraines, drug and/or alcohol consumption, health status, and ability to adapt to novel sensory environments (Stanney et al., 1998), as well as sensitivity to vection (i.e., illusory self-motion), optokinetic motion perception, transformations in optical flow patterns, monocular movement in depth, and other visual functions (Kennedy et al., 1990).
The Motion History Questionnaire (MHQ), which was developed 30 years ago to study airsickness and disorientation attributable to Coriolis stimulation, is often used to assess susceptibility differences based on past occurrences of sickness in inertial environments (Kennedy & Graybiel, 1965). Scores on the MHQ are generally predictive of an individual's susceptibility to motion sickness in physically moving environments. In a VE study (Kennedy, Stanney, Dunlap, & Jones, 1996), however, MHQ scores were not significantly correlated with preexposure, immediate postexposure, or 30-rain postexposure sickness reports. However, recent reports (Graeber, 2001a; Kennedy, Lane,...
|
