Virtual reality techniques, involving three-dimensional imaging and surround sound, are increasingly being used in health care education and research (1,2). Virtual reality has been used to provide medical education about health care responses to emergencies, such as earthquakes, plane crashes, and fires; to treat phobias, such as the fears of heights, flying, and spiders; and to treat posttraumatic stress disorder (3,4). Virtual reality has been used to simulate the effects of stroke to facilitate a greater understanding and empathy in caregivers (5). Virtual reality has also been investigated for post-brain injury rehabilitation, with treatment occurring in the patient’s home (6,7), for virtual psychiatric treatment (8) and for pain control (9). Integration of virtual reality, simulation and modeling into educational curricula has been identified as a goal by a recent large survey of information technology use by students (10).
Schizophrenia is a severe mental illness with a lifetime prevalence of approximately 1%. Most patients with schizophrenia experience auditory hallucinations, particularly hearing voices, and approximately one-quarter of patients experience visual hallucinations (11). A common cause of frustration and alienation for sufferers of schizophrenia is that their therapists, family members, and caregivers cannot really understand their experiences.
In previous work, Dr. Yellowlees has been involved in creating room-sized virtual reality environments to simulate the experience of psychosis. That project began in 2001 and has been described in detail elsewhere (12, 13). In brief, patients with stable, well-treated schizophrenia were extensively interviewed about their hallucinations. Programmers and artists were employed to create a virtual psychiatric ward and to create simulations of the auditory and visual hallucinations using a visualization laboratory driven by an SGI Onyx computer. This construction process took two full-time programmers approximately 4 months to create the custom software required, and the modeling of three sets of patient experiences took about 2 years in total. The final three-dimensional product was an effective recreation of the hallucinatory phenomena but could only be experienced within the expensive visualization laboratory and was therefore relatively inaccessible to potential students.
Janssen Pharmaceutica, manufacturer of risperidone, has also created several versions of a virtual reality hallucinations environment. Their system, which has been shown at conferences and described in the popular press, uses a custom software environment that runs on a laptop computer with stereo display goggles (14).
We were interested in assessing whether we could deliver a virtual psychotic environment over the Internet in order to increase its educational reach. We also wanted to see if personal computer graphics systems could reproduce hallucinations sufficiently well that users would feel they learned something from the experience.
We built a virtual educational environment in the form of a hospital ward using the commercial virtual world system Second Life (Linden Lab, San Francisco, Calif.). Second Life is a three-dimensional online computer simulation of the real world. Other examples of virtual worlds include There.com and ActiveWorlds (15,16). Users of Second Life download a 20-megabyte client program directly from the developer which allows them to create their own avatar (or “virtual person”) and move around the world. They can chat with other users, play simple games, build buildings, design clothing, buy and sell items created by other users, design their own three-dimensional objects, and write programs to imbue their objects with behavior. Approximately 100,000 people have accounts in Second Life, and over 10,000 individuals log in every day. Over 1,000,000 unique objects have been created in the system. For our purposes, Second Life provided a simple three-dimensional construction environment, as well as a way to distribute educational content over the Internet.
We took photographs of an inpatient ward and hospital furnishings at the University of California, Davis, Medical Center and created a virtual inpatient ward in the Second Life system. We then constructed simulations of auditory and visual hallucinations from recorded audio samples and digital images based directly on the interview transcripts from the schizophrenia patients. Some anonymous descriptions of auditory and visual hallucinations that had been collected in earlier studies (12,13) were also reused with the consent of previous research collaborators. We inserted the hallucinations as individual objects throughout the virtual ward so that they appeared automatically, triggered by an avatar’s presence as the avatar toured the environment. The hallucinations included:
Multiple voices, occasionally overlapping, criticizing the user
A poster that would change its text to obscenities
A newspaper where the word “death” would stand out of a headline
A floor that would fall away, leaving the user walking on stepping stones above a bank of clouds
Books on bookshelves with titles related to fascism
A TV that would play a political speech, then criticize the user and encourage suicide
A gun which would appear under a cone of light and pulse, with associated voices telling the user to take the gun and commit suicide
A mirror in which a person’s reflection would appear to die, becoming gaunt with bleeding eyes
Examples of the virtual hospital ward and the hallucinations are shown in Figure 1.
The tour began with the participant taking a virtual “badge” that provided the audio track. This allowed us to count the number of users who toured the environment. The virtual clinic created in Second Life and viewed on a standard personal computer provided equivalent or improved graphics and sound fidelity compared to the version displayed in a supercomputer data visualization facility in 2001. The off-the-shelf construction environment enabled much more rapid development with the entire environment, including creating the hallucinations, built in less than a month.
A survey collection system was placed at the end of the ward. This allowed users to respond to questions immediately after their tour and while they were still inside the environment by simply clicking on the response of their choice. Their answers were automatically e-mailed to us and downloaded directly to a spreadsheet for analysis. We used a Likert-like scale from 1 to 5 for the subjective questions described in Appendix 1, with responses “Strong No,” “No,” “Maybe,” “Yes,” and “Strong Yes.” To simplify the online survey object, ages were presented in ranges <18, 18–25, 25–35, 35–45, and >45. Salaries were presented in US$ as <$30,000, $30–$60,000, $60–$90,000,>$90,000. A survey response was considered valid if at least seven of the nine questions were answered. In cases where a user submitted more than one survey, only the first submission was analyzed.
Survey-takers were recruited from the general Second Life population. Because Second Life requires a computer with three-dimensional graphics capability and a broadband Internet connection, we knew our population would not be representative of the general United States population. In particular, Second Life users are younger, more educated and more likely to be male than the general population (R. Harper, personal communication). However, for the first formal feedback on the project, such a convenience sample was felt to be appropriate.
Second Life users were offered L$100 (Linden dollars) for completing a valid survey. Linden dollars are a form of currency within the Second Life virtual world system, and at the time of the study L$100 was worth approximately US$0.75 on the virtual-world to real-world currency exchange (17).
An online journalist took an interest in our project and wrote an article about it which was posted in the official Second Life blog several days after the environment was made public (18). The piece was referenced by several other popular blogs over the next week (19, 20). Based on narrative comments from users, some survey-takers signed up for Second Life for the sole purpose of experiencing the virtual psychotic environment. The UC Davis Institutional Review Board granted permission for this study.
Over a period of 2 months, the virtual psychosis environment was toured 836 times with 579 (69%) valid survey responses received. The demographics of the survey responders closely mirrored the demographics of the Second Life system at large. Most responders were men (N=366, 63%), but this represented a more balanced gender distribution than other online communities. The most common age range was 25 to 35 years old (N=186, 32%) followed by 18 to 25 years old (N=182, 31%). Most responders worked outside the home (N=316, 55%) or were students (N=113, 20%). Most responders earned less than US$30,000 per year (N=203, 35%) or US$30,000 to $60,000 per year (N=170, 29%). The results of the educationally focused questions are presented in Appendix 1.
Subpopulations were analyzed for differences using two-tailed t tests. No significant differences were found when replies were analyzed by gender, occupation, or income group.
Although the survey object in this first version had no formal mechanism for narrative comments or feedback from users, many spontaneously sent comments to us. Some of their comments are recorded in Appendix 2.
Using modern visualization technology, we have created an Internet accessible three-dimensional simulation of the hallucinations of psychosis. Users of the environment felt it improved their understanding of the experience of hallucinations.
As a pilot project, this study has several important limitations. First, the population surveyed is a significantly biased convenience sample. Existing users of an online virtual world system are not a representative sample of the general population. Second, there was no pre-test given, so we cannot prove that participants improved their knowledge. We relied on the participants’ own perceptions of educational value. Finally, the environment focuses only on the hallucinations of patients with schizophrenia. Users might therefore give inappropriate weight to these symptoms of the disease, rather than having a more balanced view incorporating other symptoms, such as delusions, disordered speech and behavior, and negative symptoms.
In spite of these limitations, we feel this approach to describing phenomenology is exciting. We are using the virtual reality scenario as a teaching tool in our medical school program and will perform a more formal evaluation of its effectiveness in teaching students about the lived nature of psychotic symptoms compared to traditional teaching approaches. We also intend to use this model as an educational tool for caregivers attending an early intervention program for patients experiencing a first episode of psychosis.
We feel that commercial virtual world systems offer significant advantages for the development of educational environments. Compared with custom software development, using an existing software system cut development time for the most recent version of the psychosis environment by a factor of 10. While some of this may be attributed to experience gained with the previous iteration of the project, we feel a large part of the improvement was due to the tools.
In addition, because Second Life can be accessed by anyone with a Windows or Macintosh computer and a broadband Internet connection, a substantial number of people can experience an educational three-dimensional graphics environment. Interested readers can find instructions on creating a Second Life account and accessing the virtual hallucinations environment on our Web site at www.ucdmc.ucdavis.edu/ais.
The direct collection of survey data from within the three-dimensional environment is novel and has implications for other educational environments. Though we only attempted a simple survey, this type of assessment process could be further developed to allow online students to answer examination questions within a three-dimensional environment. It would be straightforward, for example, for us to develop specific questions for students about aspects of abnormal phenomena in the environment we have already developed, and thereby create a combined virtual educational and assessment process that could be used in a wide variety of other educational areas. It is possible that virtual reality may well become an integral part of medical practice and training as is suggested by a number of other studies in the disciplines of Obstetrics (21), Neurology (22), and Family Practice (23).
It is difficult to teach about the internal phenomena of mental illness, such as hallucinations, using traditional educational methods. In the past, virtual reality systems have been used to create simulations of these phenomena for educational purposes, but these systems have generally used custom computer hardware that has limited their educational reach. This project has demonstrated that an Internet-based virtual reality system for personal computers can be used to create simulations of hallucinations and that it is possible to integrate evaluation tools into a virtual reality environment. Furthermore, users of the Internet-based simulation felt that it improved their understanding of the experience of hallucinations. More comprehensive studies of this approach to online education are warranted.
FIGURE 1. Inside the Virtual Reality Program
(Top left) The outside of the virtual hospital ward building. (Top right) Inside one of the hospital ward rooms. (Bottom left) Clinic hallway showing stepping stones over abyss. (Bottom right) Survey collection at the end of the tour.
We would like to thank the team led by Professor Kevin Burrage at the University of Queensland for their work on the original version of this project, and Dr. Jasmine Banks for the assistance of migrating it to the Second Life environment.
We would also like to thank the members of the Second Life community who gave feedback on the initial versions of the environment and participated in the survey.
Dr. Cook is a paid consultant to Linden Lab and owns stock in the company.