Course:COGS200/2017W1/Project-Group24

Introduction

We are proposing to conduct a study on the effectiveness of remote virtual reality exposure therapy, or VRET, for social phobias in comparison to standard VRET. Virtual reality, as described by the Virtual Reality Society is a, “three-dimensional, computer generated environment, which can be explored and interacted with by a person.” Virtual reality therapy, in turn, is the use of therapeutic practices or techniques within a virtual reality space. In the context of our study, this would entail the combination of psychological therapy techniques used in treating social phobias and the setting of virtual reality. VRET, specifically, combines exposure therapy practices with the virtual environment with which to play out the exposure. Exposure therapy is a psychological treatment most commonly used to help those with phobias confront their fears (American Psychological Association’s PTSD Clinical Practice Guideline), with frequent use in treatment of social phobias. A social phobia, as defined by the DSM-5, is a “persistent fear of one or more social or performance situations in which the person is exposed to unfamiliar people or to possible scrutiny by others.” Furthermore, exposure to these situations “almost invariably provokes anxiety,” for the person, leading them to avoid these feared situations “or else [endure them] with intense anxiety and distress,” (DSM-5).

Virtual reality exposure therapy has frequently been shown to work in assisting those with social phobias (Parsons & Rizzo, 2008). Furthermore, studies have shown that participants with phobias preferred VRET as opposed to traditional exposure therapy (Garcia-Palacios et. al, 2007). Remote VRET opens new doors of treatments to patients suffering from social phobias, as it would allow them to have access to exposure therapy without having to leave the home. Furthermore, we are looking at aspects of VR that haven’t been thoroughly explored, such as the ways in which it provides a way of returning data or information to the experimenter via various settings and responses, including in how it can combine with and react to galvanic skin response sensors.These sensors would allow us to measure the electrical conductance of the skin, and act as an indicator of psychological or physiological arousal. With such feedback, we could have the program adjust to the anxiety levels of patients in real time, similar to the way a therapist can gauge and modify exposure therapy when treating a patient in person. Ultimately, we are looking at the possibilities of self-guided remote VRET.

This proposal includes a multi-phase, multi-disciplinary study that looks at 1) the efficacy of remote VRET, 2) the possible creation of software that would allow for self-guided remote VRET (i.e. without therapist intervention), and 3) a further study to then test this product. We would complete these studies with an end-goal of defining the needs for creating a more effective VRET tool for patients to use within the comfort of their home. This proposal is ambitious for an undergraduate proposal and, if carried out, would need expert collaborators to see it succeed, particularly in the area of software/hardware creation.

Method

This will be a multi-phase research project, with each phase building on the results of the prior phase(s). In the first phase, we will conduct a study on the effectiveness of therapist-guided remote VRET compared to standard VRET. If the results of this study indicate that remote VRET with a therapist is effective in reducing patients’ stress levels, the next question to address is the feasibility of self-guided remote VRET. The second phase of the project will be product development, with the goal of implementing galvanic skin response sensors in VRET technology in such a way that the software can automatically adjust the stress level of the virtual environment based on biofeedback from the sensors about a patient’s stress levels. If a functioning product is developed and tested, we will conduct another study to test the effectiveness of this self-guided remote VRET compared to therapist-guided remote VRET. These three phases are summarized below:

Phase I: Experimental study to investigate the effectiveness of therapist-guided remote VRET vs. standard VRET. If therapist-guided remote VRET is found to be similarly or more effective than standard VRET, proceed to Phase II.

Phase II: Development of VRET technology that adjusts the immersive virtual environment based on the stress level of the patient. If successful in developing a working prototype/product, proceed to Phase III.

Phase III: Experimental study (using prototype/product developed in Phase II) to investigate the effectiveness of self-guided remote VRET vs. therapist-guided remote VRET.

The end goal of the project is defining the needs for developing a more effective VRET tool for people with social phobias. The ideal final product would be a software module that enables remote, self-guided VRET. We envision this VRET product incorporating galvanic skin response (GSR) sensors to regulate the stress level of the virtual environment. GSR sensors measure electrical conductance in the skin, which increases with emotional arousal. We plan on implementing these sensors in our VRET product to monitor the user’s stress level as they experience exposure therapy through a virtual environment. This will provide real-time biofeedback to the software, and the software will use this feedback to adjust the stress level of the VR environment. For example, the environment could be a crowded street; if the sensors identify that the user is experiencing a level of stress beyond some threshold, the number of people in the street will be reduced. Once the user’s stress level has stabilized below a specified level, the software will add back some people to the environment and repeat the feedback process. As the user’s tolerance for the stimuli gradually increases, the software will adjust the stress level of the environment appropriately. Ideally, phobia patients will be able to use this product at home on their own. Therefore the product should be relatively easy to use, and not require overly complex or expensive equipment such as large, specialized computers.

We believe the potential advantages of remote and self-guided VRET are most applicable to people with social phobias, so this is our target audience for our proposed product. Social phobias affect millions of people, but seeking treatment can be difficult and stressful; seeing a therapist involves interacting with someone unfamiliar, and likely feeling scrutinized by them—things that someone with a social phobia would want to avoid. A 2007 Anxiety and Depression Association of America survey found that 36% of people with social phobia (waited at least 10 years after they first started experiencing symptoms before seeking treatment ([Social Anxiety Disorder]). Garcia-Palacios et al. (2007) suggest that more phobia patients may seek out exposure treatment if VR treatment is more accessible. By being remote and self-guided, our proposed VRET system would reduce or eliminate several factors that may prevent someone with a social phobia from seeking treatment. It would also make VRET more accessible to people who cannot leave their house (due to their phobia, physical limitations, or any other reason), or do not currently have access to VRET where they live. Ideally, our final product will not only be an effective VRET tool, but will enable more people with social phobias to seek treatment.

Phase I

Study Design:

Our hypothesis is that therapist-guided remote VRET will be more effective in reducing anxiety than standard, in-person VRET. To test this hypothesis, our experimental study will randomly assign participants (who must have been previously diagnosed with a social phobia) to one of three groups. The first group will be a control group, who will receive no treatment; the results from this group will provide a baseline against which to compare the results from the treatment groups. The second group will participate in virtual reality exposure therapy in the presence of a therapist. The third group will participate in remote VR exposure therapy sessions. These remote sessions will still be guided by a therapist, but the therapist will not be in the same physical location as the participant. The two treatment groups will allow us to examine the efficacy of both standard VRET and remote VRET. We will analyze how effective remote VRET is in reducing anxiety compared to standard VRET, as well as how effective each treatment is compared to no treatment. We will aim to have eight participants in each group. In Levy et al.’s pilot study on remote VRET (2015), the remote VRET sessions took place in a hospital. This was to control the conditions of the treatment and avoid having to lend participants equipment. However, we believe a major benefit to remote VRET, and the main reason it may be more effective than in-person VRET for those with social phobias, is that patients can undergo treatment in whatever environment they feel most comfortable. Therefore, in this study we will lend participants the necessary equipment and allow them to choose where they do their remote sessions. Participants receiving remote VR therapy will have an initial meeting with a therapist to receive training in the technology. Both the in-person and remote treatment groups will receive six VRET sessions - two sessions a week for three weeks. Every participant will undergo a clinical screening where a therapist will assess the patient’s baseline anxiety level. Participants will also complete an initial questionnaire where they will self-report their anxiety levels and the phobia symptoms they are experiencing. During the VRET sessions, we will measure participants’ anxiety levels by measuring skin conductance with the GSR-Grove unit. In both treatments, the therapist will be able to see GSR data in real-time and use this to adjust the stress level of the VR environment. In the remote VRET treatment, a webcam will be set up so that the therapist can also monitor the subject’s behaviour during treatment. The day after every session, participants will complete the anxiety questionnaire again. The same therapist that performed the initial screenings will also reassess the participants at the end of the three weeks of treatment. This will provide us with both quantitative and qualitative data to assess the effectiveness of remote VRET compared to standard VRET. We will use a crowded street as the virtual scene in the VRET for this study. We are using this scene because it applies to several different social phobias, such as agoraphobia, claustrophobia, and social anxiety disorder. There are several measurable parameters that will need to be considered when designing the VR experience, including the number of people in the scene, the spacing between people, the “confines” provided by the buildings, etc. These parameters need to be adjustable so that the stress level of the environment can change as the patient goes through the therapy. We will work with a therapist who currently administers VRET to determine the appropriate parameters and settings. The virtual world will be created with the free, open-source software Blender version 2.79. We will use TeamViewer 8 as the remote control software, and the HTC Vive as the VR headset.

Anticipated results:

We expect our study to support external findings that VRET is effective compared to no treatment, since this result already has been replicated in several studies (Morina et al., 2015 and Parsons & Rizzo, 2008). The effectiveness of remote VRET vs. standard VRET is less certain. While research shows that remote VRET is feasible (Levy et al., 2015), very little research has been done on its efficacy. However, we believe that the positive effects of VRET are the result of exposure in the immersive virtual environment, rather than the direct interaction with a therapist during the treatment. Therefore, as long as the remote VRET is administered successfully, we expect our results to show that remote VRET is as or more effective than standard VRET.

Phase II

Phase II of our multiphase study will consist of the development and testing of a VR software that we are creating that will alter the virtual environment based on a participant’s anxiety levels. The software that we are creating will take in data from the GSR-Grove and based on that data will modify the virtual reality as needed. The GSR-Grove will be used to measure skin conductivity. The device comes with a GSR probe that applies a small voltage to two points on the skin while the current of the resulting circuit loop is measured. Units for conductance are Mhos and Seimens however the skin's conductance, while very sensitive to emotion, is very small and is outputted from the microcontroller to the USB port in units of microSiemens. Both positive and negative stimuli can result in an increased skin conduction.

We are using the HTC Vive headset which will be connected to a VR software that we create. The software will be used to monitor the microSiemen value on the USB port in a continuous loop while performing a rolling average. We can for example average the skin conductance over a one second period flushing out the oldest value and the newest value is received while calculating a new average and standard deviation with each new read. The software will also create the graphics that will be outputted to the headset.

The software can lock a baseline average upon the user's command during a time when no intentional stimuli is being given to the subject being tested. Some characterization may be required for tuning parameters of the software however we can define a new reading that is say 1.5 sigma higher from the moving average as a significant change and thereby a significant emotional response or arousal. A detection of a significant emotional response can execute a method to send a flag or indication to the engine. Once the engine receives this warning it will alter the virtual environment to reduce the negative stimuli. To be able to alter the virtual environment a HDMI cable will be attached to the HTC Vive headset which will take in data about video and sound and output the changes to the environment. The software will also continuously indicate how far the skin conductance is away from the baseline average in units of sigma. This data will be saved and be automatically sent to the subject’s therapist. Throughout Phase II we will be continuously be testing the software to see whether or not it functions as it should. We want to continuously test that the software accurately detects the indication that there is a significant change in average and will thereby appropriately modify the virtual environment.

Phase III

Assuming that our device works, we will be using it to test whether or not using this device will be more effective in aiding people with social phobias than using therapist-guided remote VRET. To test this, we will be recruiting volunteers to participate in our study that will test which method is more effective. The recruitment process will consist of screening for participants who have social phobias and are deemed as qualifying candidates by their therapists. First we would contact clinics and speak to specific therapists who would then recommend patients who they think would really benefit from this study. Once participants have been selected they will be split into two groups. The first group will be using our self-guided device and the second group will be with a therapist in a therapist-guided VRET. The participants in group one will be in a location that they deem comfortable and safe. The entire purpose of the development of our device is to have people who have social phobias to be able to experience exposure therapy without having to leave their safety zone. The device is meant to help people who would in any other case not want to experience exposure therapy because their social phobia will not allow them to leave the comforts of their home. Participants will then put on the Vive headset which will be connected to the software that we developed. The software we created will be able to track the emotional state of a person and will send that data to the therapist. The data will be interpreted as the average skin conductance a participant had while experiencing self-guided VRET. Participants in group two will be using therapist-guided remote VRET. Similar to Phase I, the GSR-Grove device will be used to to measure the skin conductance of participants. The data will be sent to therapist in real-time and therapist will interpret the values of skin conductance as anxiety levels. If the therapist deems that the participant is experiencing levels of anxiety that are too high, then the therapist has the power to change the virtual environment that the participant is experiencing. The average skin conductance will be saved and will then be compared to the data from group one. We will then analyze the data to determine which one of the two groups show a lower average in skin conductance. The group with a lower skin conductance average indicates participants in that group had lower anxiety levels during VRET. The therapist will have to compare each participant's’ anxiety levels were before the study and at the end of the study to see if overall anxiety levels improved. However to assess which one of the methods is more effective overall in aiding people with social phobias then this each participant must attend several sessions.

Discussion

We started this project knowing we wanted to look at how the world around us could be improved by virtual reality and we believe that attempting to help address mental health issues absolutely fulfills that goal. Traditional VRET has been referred to as an intermediate step between “imagined and direct exposure,” and remote VRET still carries this quality (Brakoulias, 2013). However, remote VRET only becomes truly available to those who suffer from extreme social anxiety once it becomes self-guided. Our system could reduce or eliminate factors that would otherwise prevent patients from attempting or accepting treatment of exposure therapy. Therefore, by choosing to investigate the possibilities of remote VRET, we are taking the next plausible step in researching this field.

We hope to get a direct measure of the value of remote VRET. These findings are currently undetermined: this form of therapy might be better, the same, or worse than current VRET practice. If it is better, the project could proceed by sending the software and project design to other clinicians to replicate our findings. If it is the same as current VRET therapy practices, we feel we have still learned something important and could look at the feedback data to help develop a quantitative scale for social phobia responses. If, against our expectations, the remote VRET is worse than current practices, we can try to find out what is limiting it — whether it is the virtual reality aspect, the exposure therapy itself, or the remote setting.

We expect, if this project were to be carried out, that we would receive help from colleagues with clinical trial experiences who can provide the appropriate statistical tests to validate our findings and determine which are statistically significant. For example, some characterization may be required for tuning parameters of the software. However, as this is such a late phase in the study, we do not have any current findings.

This proposal uses the basis of already existing wearable tech, including VR headsets and GSR sensors. Computers with the specifications to handle VR are not the cheapest technology today, with a VR-ready PC system costing a minimum of $999 USD (Ackerman, 2016), in addition to the cost of the headset itself, an HTC Vive, which retails at $599 USD. We would use GSR-Grove unit for our GSR sensor, which retails at approximately $10 USD each. If remote VRET use does seem more effective than traditional VRET, we would then proceed to create software that could produce changeable virtual environments in relation to the data compiled in Java. This would require software developers, coders, and most likely some form of UX design. A team of technicians trained on the production of the product would also be necessary. As this is a multi-phase study, it is hard to calculate the cost of future products that would only be designed following the initial study. However, the base cost of the technology necessary would be approximately $1400 USD per unit.

Our initial study would have very few risks involved. Following the rule of “first, do no harm,” we have researched both potential physical harm and mental harm. In the physical category are the concerns relevant to any electronic device. As VR equipment has become commercialized, it meets various safety codes, though patients would be warned about possible physical side effects of VR (such as headaches, dizziness, etc.). The measurement of skin conductance is inherently a low risk procedure. However, all therapists or psychologists of the patients used in the study would need to be trained on how to best support their patients’ use of VRET. Furthermore, we would need to create some form of safe monitoring system to use in what would be the focus group following the study. With the idea behind remote VRET being the ability to use it in the safety of the patient's’ own home, it does not account for a patient putting themselves in danger while using a remote VRET system in their own home. Therefore, a way to monitor the patient and have some form of aide or emergency assistance on standby is important to the patient’s safety. We must also consider risks to the patient’s mental health. We would insist that all VR material be screened by trained health professionals and control groups in hospital or therapy settings before proceeding with the remote experiments. As a further safeguard, we would release the source code to any other research groups so that they could replicate our experiment, identify bugs or errors, or suggest improvements.

If this works, there is a long list of social phobias that could be helped by this technology; however, they would require detailed planning which we have not done for this proposal. Remote VR opens an avenue to provide treatment to millions of people at a much lower cost. We have not researched the real cost of social phobias in society, and the medical cost of treating them, but to us remote VR offers real potential at both effectiveness and cost savings.

Conclusion

From this project we have gained considerable insight into the amount of planning and work that goes into creating a research proposal. This project required us to draw from external research to design our own studies, consider how we could build on our research results to develop a product, and think critically about the feasibility and risks of our proposal. The more we thought about each phase of the project, the more we realized how many details needed to be considered. The scope of our proposal may have been too large for this project – however, pushing ourselves to complete this ambitious proposal was a great learning experience. We learned how to critically evaluate scientific studies and apply research from different disciplines into one coherent project. From thinking about how we could extend the use of VRET and make it more effective and convenient for people with social phobias, we learned a lot about the current use of VRET and the future possibilities of this technology. We believe that remote, self-guided VRET that implements biofeedback from GSR sensors has the potential to become a reality and help people with social phobias. More generally, this project has shown us the exciting possibilities of combining disciplines to tackle big ideas. There seems to be great potential in integrating computer science in approaching psychology problems. Developing a multi-discipline research proposal has given us a better understanding of the field of cognitive systems and the kinds of research needs it can address.

Effective remote, self-guided VRET that incorporates biofeedback may still be a long way off. However, we believe our proposal outlines a reasonable starting-off point for achieving such a product. VRET has already been established as an effective treatment method for social phobias, and research has proven that remote VRET is feasible. The efficacy of remote VRET has not yet been established, so testing this is the first step of our project. If the results from our first study are as anticipated and indicate that remote VRET is effective enough to be worth pursuing, we will move on to developing a product. It should be feasible to develop a remote VRET system that incorporates biofeedback from GSR sensors, since this will mostly involve integrating several existing technologies into one product rather than developing something completely new. If a working product can be developed, our second study will examine whether self-guided remote VRET is effective compared to therapist-guided remote VRET. The results from this study will provide insight into how well our product is performing compared to our goals. Further adjustments to the product may be needed, and more studies will have to be conducted before self-guided remote VRET can be widely accepted as an effective treatment method. Regardless of the final results, embarking on this project will bring us closer to more effective and accessible VRET.

Bibliography

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author
American Psychological Association. (n.d.). Clinical Practice Guideline for the Treatment of PTSD. Retrieved from https://www.apa.org/ptsd-guideline/ptsd.pdf
Brakoulias, V. (2013). Virtual reality can help to develop a new reality. US National Library of Medicine National Institutes of Health. https://doi.org/10.1177/0004867412455852
Garcia-Palacios, A., Hoffman, H., Carlin, A., Furness, T., III, & Botella, C. (2002). Virtual reality in the treatment of spider phobia: a controlled study. US National Library of Medicine National Institutes of Health. https://doi.org/10.1016/S0005-7967(01)00068-7
Levy, F., Leboucher, P., Rautureau, G., & Jouvent, R. (2015). E-virtual reality exposure therapy in acrophobia: A pilot study. Levy F1, Leboucher P2, Rautureau G2, Jouvent R3. US National Library of Medicine National Institute of Health. https://doi.org/10.1177/1357633X15598243
Page, S., & Coxon, M. (2016). Virtual Reality Exposure Therapy for Anxiety Disorders: Small Samples and No Controls? US National Library of Medicine National Institute of Health. https://doi.org/10.3389/fpsyg.2016.00326
Parsons, T., & Rizzo, A. (n.d.). Affective outcomes of virtual reality exposure therapy for anxiety and specific phobias: a meta-analysis. US National Library of Medicine National Institutes of Health. https://doi.org/10.1016/j.jbtep.2007.07.007
[Social Anxiety Disorder]. (n.d.). Retrieved from Anxiety and Depression Association of America website: https://adaa.org/understanding-anxiety/social-anxiety-disorder Parsons, T. D., & Rizzo, A. A. (2008). Affective outcomes of virtual reality exposure therapy for anxiety and specific phobias: A meta-analysis. Journal of Behavior Therapy and Experimental Psychiatry, 39(3), 250-261. https://doi.org/10.1016/j.jbtep.2007.07.007