Virtual Reality (VR) technologies are advancing in their application to mental health research. Virtual environments have been employed in clinical applications such as treatment of acrophobia [1], agoraphobia [2], and the fear of flying [3]with research studies attempting to verify the effectiveness of these treatments. Other studies have attempted to show improvement in body satisfaction for patients suffering from eating disorders [4], increased engagement in autism [5], and improved sexual function in men suffering from impotence [6]. Despite such broad and promising clinical uses, several factors currently limit how effectively VR can be applied to mental health research. In general, these fall into two broad categories. First are the technical limitations that make VR technology expensive and difficult to employ in effectively engaging the viewer. Second are the characteristics of the conduct of mental health research itself which impede the use of virtual reality. This chapter will describe each of these limitations in depth as well as discuss potential solutions.

At the current time, virtual environments are limited by the technical complexity required to build them. We can see examples of this as we follow the steps of virtual environment construction. First, objects need to be defined in three dimensions using 3D modeling software. Not only does this bring up the difficulty of learning how to use 3D software or finding someone with these skills, but this also highlights the laborious nature of building every object in an environment. Shapes need to be defined, oriented to one another, and mapped into specific locations in three-dimensional space. Textures need to be defined as well as characteristics such as degree of transparency, shininess, reflectance, etc. Even with experts available, this process consumes time, and therefore money. Further programming is needed to make these models appropriately interactive, defining whether or not you can pick up an object or whether you can walk through it. Currently, we have to define rules to make objects behave as we assume they do in real life (cast shadows, fall at a constant acceleration, shatter or splinter if broken.) "Physics based models" which calculate how objects deform or shatter under pressure are computationally expensive and not easily integrated into virtual environments. Even more complex is the idea of the "virtual person." Although discussed in the literature [7], working systems have barely moved beyond basic simulation of facial expressions and body mechanics [8, 9]. Extensive programming will be needed to simulate other aspects of voice, inflection, reactions to verbal or physical actions, etc. In sum, high degrees of realism can be modeled but require high degrees of work. Many details of the physical world that we take for granted cannot yet be practically generated by the average VR system.

We can also see technical limitations in the simulation equipment used to experience virtual environments. The most common device is the Head Mounted Display (HMD) which is bulky and often has a limited field of view. Such attributes provide a distraction from the ideal visual and auditory stimulus that we might expect a virtual environment to provide. Although some state of the art systems like the CAVE (CAVE Automatic Virtual Environment) reduce these problems by using screen projection technology, we continue to have limitations in other senses. Independent systems that provide tactile stimuli are usually limited to small specific objects with a restricted range of motion such as a pencil that provides some resistance when grasped. Other systems to simulate smell or vibration of floor or air have only been created as prototypes or for specific experiments. The lack of these modalities of experience limits the range of potential experiences that mental health researchers might want to simulate.

Although these difficulties seem extensive, VR technology is rapidly growing, leading to a likely solution to many of these problems. Software advances have lead to the development of more intuitive user interfaces and design systems to make it easier to create VR objects. Libraries of textures or even complete 3D objects are commercially available, and some companies now specialize in the commissioned design of specific environments. Just as one can contract for the architectural design and construction of a building, one can now contract for the construction of virtual buildings or furniture. Free models are also appearing on the World Wide Web, some distributed as commercial promotions and demonstrations, and others supplied by government agencies. Many of these changes are occurring because of the development of standards for virtual models. Virtual Reality Modeling Language (VRML) was originally designed to allow Internet transmission of 3D objects and "worlds." The newest version of this standard (VRML 97) establishes a protocol so that all VR models use the same framework for describing their appearance and how they interact [10]. The existence of such standards permits more rapid development as communications between different research groups improve and the confidence of industry increases in new product development and acceptance.

Not only will technological development help solve the technical impediments to the use of virtual environments, but we may find that some of these "impediments" are actually useful in mental health research. Indeed, in exploring basic phenomena of perceptual, emotional, or other brain function, we often seek to remove extraneous details that provide variation in comparison populations. VR equipment was used in one PET study to alter the prominence of environmental landmarks in navigating a virtual environment, demonstrating metabolic changes in the right parahippocampal gyrus when landmarks were present [11]. The fact that a virtual environment cannot replicate all the subtleties that are present in a real environment simplifies the process of controlling for environmental variables in experiments. A full replication of reality is not always needed for a particular application if focused definitions of the goal make a sense of reality or actual "presence" irrelevant [12].

The question of how to properly establish the goal of research leads us to a second barrier limiting the application of virtual reality to mental health research. We argue that the current construction of mental health research itself limits how virtual reality can be applied. To explain this argument, we first define the term "mental health research" by restricting it to research that explores mental illness as defined by standardized classifications of mental disorders such as the Diagnostic and Statistical Manual (DSM-IV). We make this restriction because although these efforts have clear critics and limitations, they represent a widely recognized standard that is advanced through a defined and recognized method and process. It is this standardization process that has made these systems widely recognized despite their weaknesses, because standardization improves communication and creates sufficient stability of terminology so that learning and comparative research can occur. As a by-product, this standardization process has created groups of recognized experts on each type of mental disorder and a set of cultures with their own expectations of what types of research instruments, techniques, and analysis possess acceptable levels of scientific rigor. In many ways, one could say that it is the culture of the mood disorders research community or the culture of the borderline personality disorder research community that truly defines these disorders more than the limited descriptions they have distilled for the diagnostic manuals. Such manuals, nonetheless, are the final common point for communication.

Diagnostic manuals like the DSM-IV contain only phenomenological criteria and provide no guidelines to establish or even extrapolate what environmental features are critical for any particular disorder. With such definitions, the sole possible guide to deciding what should go into a virtual environment would be to present the hypothetical stimuli to a clinical population. In acrophobia, for example, many virtual reality projects have created virtual environments and documented effective treatment of subjects, yet much of this research did not apply DSM or ICD frameworks. Indeed, the DSM and ICD do not provide specific guidelines to direct the choice of environmental elements for inclusion or the perceived qualities that these elements would need to possess [13]. Clinical and subjective experience of the designers guided decisions about emphasizing the change in angle in looking down upon a tree or lamppost or adding features to highlight the size of a window or other types of barriers perceived between a subject and the visible edge. In other disorders, such lists of decisions grow even longer and more complex. In major depression, if one is to interaction with a virtual human, what adjustments in the tempo of speech are appropriate? What changes in word choice should reflect changes in cognitive capacity or desire for social interaction? In borderline personality disorder, what shifts in body position create desired changes in comfort or discomfort? What environmental factors regarding the realism or "presence" in the situation will encourage dissociation or losses of reality testing? The decisions made to answer such questions need to be consonant with the types of decisions that would be acceptable to the "research culture" for each of these disorders.

Finding such a level of specificity in our diagnostic descriptions is many years or even decades away. Yet, as implied above, much of this knowledge is already contained in the "research culture." In many cases, sufficient information and expertise is available to guide the development of virtual reality experiments in a manner acceptable to the research mainstream. The true barrier is the limited speed and facility with which members of a particular "culture" can be educated and started in the process of virtual environment development for experimental purposes. In order to overcome this barrier, we need to build connections between the virtual reality research community and the mental health research communities. We need to have researchers laboring together to advance their research and to discover how to make this interface work. We will also need to create broad standards which can simplify and regularize the process of virtual environment design for mental disorders, so that interested researchers can have a template of how to proceed in a fashion that will be complete and effective.

Virtual reality technology has immense potential for applications to the field of mental health research. It provides a complex, replicable stimulus for testing a broad range of hypotheses from function of perceptional apparatus to function of the individual in specific interpersonal and social situations. Current technical limitations make it complex and expensive to devise appropriate virtual environments for experimentation. Current limitations in the construction of the mental health research mainstream make it difficult for virtual reality based research to be easily accepted. Although technical difficulties and the need for expertise will likely decrease with continuing improvement in hardware and software, development of virtual reality in mental health will continue to be limited by our ability to build collaborations between virtual reality researchers and the experts who define mental health research standards. Although virtual reality research can continue without acceptance from such experts, it is the creation of standards that will advance our work. Just as the VRML97 standard advanced virtual environment development by creating a guideline to unify the work done in dozens of individually developed virtual reality development systems, new standards can help us advance virtual experimentation in mental health by creating guidelines to encourage and unify diverse mental health research efforts into effectively using virtual reality.

[1] Rothbaum, B.O., Hodges, L.F., Kooper, R., Opdyke, D., Williford, J. and North, M.M. Effectiveness of computer-generated (virtual reality) graded exposure in the treatment of acrophobia. American Journal of Psychiatry. Vol 152 (4), pp. 626-628, April 1995.
[2] North MM, North SM, Coble JR: Effectiveness of virtual environment desensitization in the treatment of agoraphobia. Presence: Teleoperators and Virtual Environments. 5(3): 346-352.
[3] North MM, North SM, Coble JR. Virtual Environments Psychotherapy: A Case Study of Fear of Flying Disorder. Presence: Teleoperators and Virtual Environments. Vol 6 (1), pp. 127-132, February 1997.
[4] Riva G: The virtual environment for body image modification (VEBIM): development and preliminary evaluation. Presence: Teleoperators and Virtual Environments. Vol 6 (1), February 1997.
[5] Strickland D: A virtual reality application with autistic children. Presence: Teleoperators and Virtual Environments. Vol 5 (3), pp. 319-329, 1996.
[6] Optale G, Munari A, Nasta A, Pianon C, Baldaro Verde ,. Viggiano G. Multimedia and virtual reality techniques in the treatment of male erectile disorders. International Journal of Impotence Research. (December 1997); 9(4):197-203.
[7] Glantz K, Durlach NI, Aviles WA: Virtual Reality (VR) and Psychotherapy: Opportunities and Challenges. Presence: Teleoperators and Virtual Environments. Vol 6 (1), pp. 87-105, February 1997.
[8] NM Thalmann, D Thalmann: Towards Virtual Humans in Medicine: A Prospective View. Computerized Medical Imaging and Graphics. 18(2), (1994), 97-106.
[9] I Pandzic, NM Thalmann, TK Capin, D Thalmann: Virtual Life Network: A Body-Centered Networked Virtual Environment. Presence: Teleoperators and Virtual Environments. 6(6), (December 1997), 676-686.
[10] Virtual Reality Modeling Language ISO/IEC Draft International Standard 14772-1 (December 1997).
[11] EA Maguire, CD Frith, N Burgess, JG Donnett, J O'Keefe. Knowing where things are: Parahippocampal involvement in encoding object locations in virtual large-scale space. Journal Of Cognitive Neuroscience. 10 (1) (Jan 1998), 61-76.
[12] SR Ellis. Presence of Mind: A Reacation to Thomas Sheridan's "Further Musings on the Psychophysics of Presence." Presence: Teleoperators and Virtual Environments. 5 (2), (Spring 1996) 247-259.
[13] MP Huang, J Himle, K Beier, NE Alessi. Comparing Virtual and Real Worlds for Acrophobia Treatment. In: JD Westwood, HM Hoffman, D Stredney, and SJ Weghorst, (eds.) Medicine Meets Virtual Reality: Art Science, Technology: Healthcare (R)evolution. IOS Press, Amsterdam, 1998.