Usability and Affects Study of a Virtual Reality System Toward Scorpion Phobia Exposure Therapy

Abstract

In this study, we present a framework to develop and evaluate a virtual reality exposure therapy system with biofeedback toward scorpion phobia treatment. The system is developed based on the methodology for the development of virtual reality educational environments; usability is evaluated with the System Usability Scale (SUS), the affects are measured with the Positive and Negative Affect Schedule (PANAS), and the biofeedback heart rate is measured in real time using a wearable device and the HypeRate app. A descriptive study was conducted with a non-probabilistic convenience sample of undergraduate students. The non-clinical sample consisted of 51 participants (11 women and 40 men) (mean = 20.75, SD = 2.42 years). The system usability score was 75.49, higher than the average of 68. For positive affects, the average value of the overall sample was 28.18, while for negative affects it was 13.67. The results of this preliminary study, while not determining that the system could currently be applied in clinical settings, demonstrate however that the system can initially be considered as a pre-feasibility study, and if the limitations of the unbalanced non-clinical sample are addressed, it could be used in the future for this purpose. The main contribution is the proposed framework to integrate usability and affects evaluation, as well as biofeedback in a VRET system toward scorpion phobia treatment.

1. Introduction

A phobia is characterized by the presence of uncontrollable anxiety that is not consistent with real risk and affects the daily life of a person [1,2]. Specific phobias are considered the most common anxiety disorders, and their causes, consequences, and treatments require further research [3]. The estimated lifetime prevalence of these disorders ranges from 3% to 15%, with phobias of heights and animals being the most common [4,5]. One of the most feared animals by humans are spiders and, although their medical importance is indisputable, there is a significant amount of misinformation about this and other arachnids, such as scorpions [6,7,8]. Hence, arachnophobia has been the most common specific phobia for research topics, being the most studied [9].

Cognitive behavioral therapy (CBT) is the current gold standard for treating specific phobias, especially those related to insects [10,11]. CBT has been used with graded exposures to the phobic stimulus in different contexts and non-aversives. On the other hand, exposure therapy (ET) is a type of behavioral therapy, also considered a standard treatment for specific phobias such as arachnophobia, and it is estimated that 75% of patients receive clinical benefit compared to placebo [12]. ET extinguishes the conditioned fear by creating a safe environment, in which the patient is exposed to the feared stimulus repeatedly over time, to decrease the fear [13]. However, even though ET is considered the most effective way to treat specific phobias, there are patients who do not respond to ET, and many phobic patients refuse treatment or discontinue treatment during the process [14]. Virtual reality exposure therapy (VRET) is an emerging technology that psychotherapists have recently used in therapy designs and settings using sensory stimulation to generate an interactive perception of a realistic, immersive, and controlled virtual world [15,16,17]. VRET, employing various scenarios and integrating monitoring equipment, has provided evidence of its effectiveness, reliability, and validity [18].

Several VRET systems for the treatment of spider phobia have been described in the literature over time [19,20,21,22], highlighting advantages such as continuous exposure of individuals to the object of their fear [23] and better control of anxiety due to the absence of a real threat [24]. In some cases, participants showed a reduction in both subjective anxiety and spider cognitions in a more comfortable way than in vivo exposure therapy (IVET). In the same way, it is complicated to have a real spider for IVET, as it is a process that consumes the time and resources of the therapist [25]. For the purposes of realism, controlled interaction, and time-consuming reduction, in some cases VRET has presented highly effective treatments over CBT or IVET for a broad spectrum of mental health conditions [26,27,28]. In addition, the use of VRET allows for the exposure of patients to the phobic environment to be adjusted to reduce the additional harm that can be caused by stress and panic attacks resulting from overexposure [29].

Recently, it has been suggested that spider fear may be the consequence of a more general fear of chelicerates, of which the scorpion is the main model [30,31]. Scorpion venom represents a health problem in several parts of the world and it can cause serious medical complications and untimely death if it is injected into the human body [32]. Mexico is home to the greatest diversity of scorpions in the world (around 12%), including some of the most medically important species [33]. Although efficient results and advances in these systems have been documented, most have focused on spiders, as opposed to scorpions, as we have undertaken in this work. In [34], we found a study in which the behavioral and psychophysiological influence of conscious access and recognition of fear of scorpions was studied. Similarly, augmented reality has also been used as an approach to support the treatment of specific phobias in small animals such as scorpions [35]. However, most arachnophobia VRET studies lack usability evaluations and biofeedback. Some of the works related to usability studies are presented in [36,37], where the use of a spider instead of a scorpion makes a difference. On the other hand, biofeedback, including the integration of real-time biofeedback [38] of physiological variables such as electrodermal activity and heart rate, ensures effective implementations and applications [39], but these works are not related to scorpion phobia.

To address the limitations and opportunities mentioned above, in this work our aim was to develop and evaluate the usability and affect impact of a virtual reality exposure therapy system toward scorpion phobia, using a framework based on the Methodology for the Development of Virtual Reality Educational Environments (MEDEERV), which was originally published in 2018 [40]. The research questions that we addressed for this purpose were as follows: 1.—How does the proposed framework make it possible to develop and evaluate a virtual reality exposure therapy toward the treatment of scorpion phobia? and 2.—What is the result of the evaluation of the usability and emotional impact of the virtual reality system toward scorpion phobia exposure therapy? The main contribution of this work is a framework that integrates usability testing, affect measurement, and real-time heart rate biofeedback in a virtual reality system toward exposure therapy for scorpion phobia.

2. Materials and Methods

In this section, we describe the methodology for the development and evaluation of virtual reality exposure therapy toward the treatment of scorpion phobia, as well as the experimental setting.

2.1. The Scorpion Phobia Virtual Reality Exposure Therapy System Methodology

Virtual reality exposure therapy is used for the treatment of scorpion phobia in a playful and interactive way using an improved version of the Methodology for the Development of Virtual Reality Educational Environments [41]. Mainly, we add a usability testing stage, which contains, in addition to the usability test, the biofeedback and psychometric measurement, as well as the feedback loop to the functional design stage. Figure 1 shows each of the stages of the improved version of the MEDEERV.

The methodology consists of seven main stages: the systematic instructional design, the functional design, the virtual world modeling, the environmental effects, the implementation, the support, and the usability testing, and other secondary blocks. Each of these stages contains a series of phases that must be broken down to create an optimal virtual environment.

2.1.1. Systematic Instructional Design

In this case, we consider instructional design as the process of creating an effective, interactive, and engaging virtual environment that focuses on helping people overcome their fear of scorpions. In addition, the results that individuals must obtain by exposure to scorpions in a virtual controlled environment are defined. In the first stage, the expected results of this project are focused on determining the usability, emotions, and proof of concept of the VRET. As a next step, the expected results are reduction or elimination of scorpion fear or phobia in a clinical sample.

2.1.2. Functional Design

In this phase, the functional requirements and actions that the system develops to ensure its proper functioning are indicated, with the aim of making it intuitive and dynamic [42]. The behavioral avoidance test (BAT) has often been used in studies of spider phobia [43]. Based on [44,45], we design 3 scenarios related to exposure levels. The distance in meters from the participant to the scorpion was rated in the 3 levels, level 1 being the lowest anxiety level. The expected behavior of the three levels is described as follows:

• Level 1: the individual can remain calm in the presence of a scorpion moving between 5 and 6 m away.
• Level 2: the individual can remain calm in the presence of a scorpion moving between 3 and 4 m away.
• Level 3: the individual can remain calm in the presence of two scorpions moving between 1 and 2 m away.

Figure 2 shows a diagram of the script of the behavior of the application including the different scenarios.

2.1.3. Virtual World Modeling

The process of developing the virtual world through scripting and artistic design represents the first step of the third phase of MEDEERV. In this stage, detailed observation, modeling technique, use of color and perspective, and composition are considered. We focused on the need to make a realistic environment, and for this we designed a scenario as a psychotherapeutic room. The room must have adequate lighting, preferably natural. In addition, it should be painted with light colors that reflect light and provide a warm and friendly environment for its users. It must ensure the privacy, both visual and auditory, of the patients. The main object of the virtual world is the scorpion; the model was inspired by the Diplocentrus zacatecanus specie, which is endemic in central Mexico [46], the place where the study was performed. Two standard primitives were used for their modeling: cubes and spheres. Figure 3 shows a perspective orthographic view of the scorpion.

After the modeling of the objects, the texturing phase continues with two different techniques. The first is the unwrap (uvw) modifier in 3D Studio Max, applied to each modeled polygon, and the second consists of repetitive textures (tileable). 3D Studio Max 2023 versión 25.3.4.84 was used for its features to create textures, animations, its integration with unity, and the plugins for 3D glasses and biofeedback [47,48].

2.1.4. Environmental Effects

The environmental effects are additional animation techniques that add realism to the scene; these are usually applied after the modeling and texturing of the elements in the environment. In this case, the model animated with the action of walking as well as opening and closing the pincers is the scorpion. In a realistic and immersive environment, the right lighting creates a sense of familiarity with the real world. In this application, several volumetric lighting techniques were used, such as spot light, light to highlight, and directional light. Figure 4 shows the difference in visualization between the application of spot light and directional light.

The light to highlight has a specific position and range over which the light strikes; however, it generates a shaped region of illumination due to the angle of incidence and directional light. Meanwhile, directional light is very useful for creating distant light effects.

2.1.5. Implementation

The implementation phase focuses on transforming the conceptual design into a functional interactive experience. In this stage, by integrating physical behaviors, navigation mechanisms, and multimedia components, a realistic and immersive experience is generated between the user and the virtual world. Physical behaviors refer to the interaction of interface elements with user actions [41]. One of these behaviors was the consideration of gravity acceleration which, to contribute with realism, was set at 9.81 m/s². For the movement of the scorpion, a script was made for the trajectory tracking and speed control of the displacement that the scorpion performs, depending on the selected level. The adjustable dimensions (length = 8 cm, width = 3.5 cm and height = 1.6 cm) and the average speed (0.30 m/s) of the 3D scorpion model are based on the real values of these parameters [49,50]. Figure 5 shows the trajectory that the scorpion follows during the path of level 2.

However, in the implementation phase, navigation mechanisms are also developed to intuitively guide the user through the virtual world [41]. The buttons used to navigate the three levels, the menu, and the other screens are controlled with a pointer located in the center of the screen, synchronized with the head movement of the user. Finally, in this implementation phase, multimedia, such as images, videos, sounds, and animations, are integrated and optimized, including a heart rate monitor. The end result of this phase is the virtual world.

2.1.6. Support

The support phase is useful to ensure that the virtual environment developed in Unity works optimally. To this end, the following configurations were made to guarantee a better compatibility with devices.

• The graphics API was limited to OpenGL to avoid rendering conflicts with Vulkan [51].
• The minimum version of the smartphone operating system is Android 8.0 Oreo, and the maximum set of versions is version 14 [52].
• The system has been compiled for two types of processors; ARMv7 (32-bit) and ARM64 (64-bit) [53].
• For multiplatform compatibility, we use IL2CPP as the intermediary language between C# and C++ [54].
• The smartphone processor must be equal or superior to the Qualcomm Snapdragon 778 G, Mediatek Dimensity 1080 or Samsung Exynos 990.
• The smartphone must have a display with a minimum refresh rate of 90 Hz with Full HD+ resolution. However, a screen with a rate greater than 120 Hz, with QHD+ resolution, is recommended [55].
• For heart rate measurement, you must use a device with bluetooth connection compatible with HypeRate technology, for which there are multiple smartwatches and arm bands [56].

2.1.7. Virtual Reality Environment

The virtual reality environment proposed in this work, which can be used as a generic framework, is composed of the user, the virtual world, the evaluation stages, and the clinical and computational specialists. Figure 6 presents a block diagram of the interaction flow of the elements.

2.1.8. Usability Testing

Usability testing is the method that aims to test the functionality of VRET for scorpion phobia, observing real users as they attempt to complete the virtual trip for the three levels. It is very important to focus on the user and not the product, because you know what works for your users. In this section, we describe only some of the essentials for performing a usability test [57].

The next step in usability testing is to determine how the product is tested and how user groups are established. There is a debate on how many participants are needed in a reliable usability test; however, in [58] has been concluded that a majority or about 80% (given a probability of detection 30%) of usability issues is observed with the first five participants [57]. Many usability experience professionals only test with five or six participants, and typically testing 10–12 users provides the same results as testing with 25 users. Figure 7 shows a real example of the post-task questionnaire to obtain immediate feedback from the participants after the virtual trip at the three levels. This is a scale of 10 items with a 5-point Likert-type response.

In this study, the internal consistency of the SUS questionnaire using the alpha of Cronbach is alpha = 0.74, so the coefficient value is acceptable [59].

2.1.9. Positive and Negative Affect Schedule

The Positive and Negative Affect Schedule (PANAS) developed by [60] is one of the most widely used scales to measure dominant affects [61]. The PANAS has been validated in several languages, such as Spanish [62,63]. This scale is composed of 20 items with a 5-point Likert-type response scale, ranging from 1 (very little or not at all) to 5 (extremely), distributed in two factors: Positive Affect (PA) and Negative Affect (NA). The general internal reliability of both dimensions was similar to the original scale at the times asked “last week” and “usually” (AP $\alpha$ = 0.86 and 0.90, respectively, and AN $\alpha$ = 0.84 and 0.87, respectively). In order to measure the positive and negative affects generated in users with the use of VRET in the moment approach, we apply PANAS. The alpha of Cronbach for the positive affects construct was $\alpha =0.86$ and for negative affects was $\alpha =0.81$. Alpha values are considered appropriate for the consistency of this applied research.

2.1.10. Biofeedback

Biofeedback is a technique based on physiological variable sensors for both the patient and the clinical specialist to monitor in real time the physiological variables that describe the functioning of the body [39]. In this case, during the intervention, visual heart rate biofeedback is offered directly to the patient on the smart phone screen, as well as to the specialist via the control panel. We used HypeRate technology, an initiative for the integration of a heart rate meter with various development environments such as Unreal Engine, Unity, and Godot. The Unity Plugin is an easy to integrate add-on, which makes it possible to receive the measured data directly on the device on which the app is running. Although HypeRate allows for higher sample rates (Fs), in this case we have used a higher $Fs$ = 1 Hz.

2.2. Study Design

In this work, our objective is to conduct an experimental study to evaluate the usability and affects of a VRET with biofeedback for scorpion phobia exposure therapy in a general sample of undergraduate Mexican students. To do this, we use the SUS [64] and PANAS [65], respectively. The intervention consisted of a virtual exposure task in a virtual world of scorpions, supervised by a clinical specialist. This is a minimal risk study; Some of the contraindications to the use of VRET in general are the possible risk of dizziness, nausea, and dizziness.

2.3. Ethical Considerations

This study was approved by the Ethics and Research Committee of the Autonomous University of the State of Hidalgo with protocol 257/2024. This study was carried out according to the Declaration of Helsinki for medical research involving human subjects [66] and the general law on health research in force at the national level [67].

2.4. Experimental Setting

The experiment was carried out in a virtual reality environment consisting of (i) a virtual reality headset, (ii) a smartphone showing the virtual tridimensional scorpion, (iii) a smartband for heart rate biofeedback, and (iv) a smartphone to answer the post-task questionnaires. In detail, following the workflow and specifications of the functional design (Figure 2), the user enters the virtual world and a menu appears with access to the three levels, exit and about-object through interactive objects by fixing the view (Figure 8).

Each level is related to the distance to the approach of the scorpion. Figure 9 shows the virtual world from the user perspective, once you have selected level 1.

2.5. Participants

A total of 54 participants (41 men (71.9%) and 13 (28.1% women ranging from 18 to 25 years) were randomly recruited face-to-face by invitation. They were undergraduate students in computer science and electronics engineering. The imbalance in favor of female participants is related to the engineering profile. Inclusion criteria: participants enrolled at the undergraduate level, in face-to-face mode, between 18 and 25 years of age, who wish to be part of the study and sign the informed consent. Exclusion criteria: students with a nationality other than Mexican, as well as students who state that they are or suspect they are pregnant. Students who expressly notify them that they have a diagnosed and untreated phobia of scorpions. Elimination criteria: students who decide not to complete their participation in the study. Among the 54 volunteers initially recruited, 3 (2 women and 1 man) were excluded based on exclusion criteria. The final sample considered for the analyses was then reduced to 51 participants (mean age = 20.75 years, SD = 2.4), of which 11 women (mean age = 19.6 years, SD = 1.8) and 40 men (mean age = 21.4 years, SD = 2.1). The highest rate of men (78.4% of the final sample) is consistent with the higher prevalence of men in engineering programs in electronics, telecommunications, and computer science. Thirty-five of the participants reported living in an urban ( 8 women, 27 men) area and 16 in a rural (3 women, 13 men) area. We develop a general descriptive analysis by groups of sex and living area.

2.6. Procedure

The study included three stages: (1) participant recruitment; (2) virtual scorpion exposure; (3) post-task usability and affects assessment. After expressing the intention to participate and signing informed consent, participants were asked to complete a preliminary questionnaire to collect demographic data such as sex, age, and living area (rural or urban), as well as the fear of scorpions that was ranked as low (1 to 5) or high (6 to 10). The participants were received in an experimental room and were seated in a comfortable chair. Subsequently, the objective and procedure of the study were described to them and, if they approved, they were asked to sign the informed consent form. The experimental test consisted of a virtual tour through three scenarios referring to three levels of exposure to scorpions, using virtual reality goggles and a smartphone. During the intervention, the user wears a smartband that allows them to measure and visualize heart rate in real time. The experimental test was developed in the presence of study technical managers and a psychologist for the management of any adverse anxiety event.

Finally, participants answered the usability test questionnaire and the PANAS questionnaire on the Google Forms platform through the corresponding link and QR code. Participants entered the questionnaire through a mobile device. The duration of the intervention was approximately 30 min.

2.7. Data Analysis and Presentation

The data capture of the usability test and the PANAS questionnaire was carried out using the Google forms online questionnaire and was stored in a database in Google Drive. To facilitate visualization, processing, and analysis of descriptive data, we used Matlab, which is optimized to solve scientific and engineering problems [68]. The level of statistical significance was established at $p<$ 0.05. The reliability of the instruments was determined using alpha of Cronbach ($\alpha$) [69], usability was determined using the SUS methodology, and emotional impact was determined using the Positive and Negative Affect Schedule.

3. Results

3.1. Usability Testing Findings

Related to self-reported scorpion fear, 9.8% of participants reported considerable fear, while 90.2% reported a low level of fear. The average usability score based on the System Usability Scale (SUS) [57] for all participants using the VRET system toward scorpion phobia was 75.49. Although an average value for the SUS score for virtual environments has not been determined, previous research studies indicate an average SUS score of about 68 [64]. Then, if your score is less than 68, there are probably serious problems with the app. In this study, the participants reflected a generally positive perception of usability. To detail the perceptions of the participants for each question,

Table 1. Mean and standard deviation (SD) of the Likert scale answers for the SUS questionnaire by questions ($Q_n$), where n is the question number.

Question	Mean	SD
$Q_1$	3.16	1.05
$Q_2$	1.94	1.02
$Q_3$	3.76	1.37
$Q_4$	1.82	0.81
$Q_5$	3.63	1.25
$Q_6$	2.02	1.00
$Q_7$	4.16	1.14
$Q_8$	1.29	0.53
$Q_9$	4.12	1.25
$Q_{10}$	1.55	0.77

shows the raw value (Likert scale) answered by all participants.

Regarding the frequent use of the system ($Q_1$) as possible, in general, participants show a neutral position, and women tend to increase the likelihood. In relation to complexity, the participants disagree that the system is unnecessarily complex ($Q_2$). Most of the participants agree that the system is easy to use and that no technical help is needed to use it ($Q_3$ and $Q_4$). The integration of functions of the system is assessed with $Q_5$ and $Q_6$, where the general experience tends to agree that the system is well-integrated. In addition, the answers concerning the system can be quickly learned by other people and the confidence in the system ($Q_7$ and $Q_9$) is biased toward strongly agreeing. Finally, $Q_8$ and $Q_{10}$ are geared toward investigating difficulties and need to learn many things before using the system. The participants strongly disagree that the system is cumbersome to use and requires a significant amount of prior knowledge. Although the sex-imbalanced sample does not allow us to make a comparative study, we present the values by sex descriptively in the results section. An overview of the results of the usability study follows, with recommended changes for areas where users found advantages and recommendations.

Favorable findings

• The experience is interesting; it feels like you are there, as if everything is real.
• They all found the drop-down menus and multimedia objects attractive.
• The system is easy to use and easy to understand.
• Users commented that the virtual world fit well on the screen.
• The size and speed of the scorpion is adequate.

Recommendations

The following is a list of the top user recommendations, and some of them were considered to reengineer this version.

• More levels in the menu are required.
• Implement other places or cases to be able to navigate.
• Add instructions so that the user knows what to do when it is running.
• They could add more and bigger scorpions for each level.
• Use of a haptic device for selection and navigation.

3.2. Affects Findings

Table 2. Means and standard deviations for the PANAS at the time approach, where PA: Positive Affects and NA: Negative Affects.

Affect	Mean	SD
PA	28.18	8.33
NA	13.67	4.25

presents the basic descriptive statistics for PANAS in general and by sex.

Inspecting Table 2, in general, participants report more positive than negative affects, with most of the answers on a moderate scale (3) with a high spread of the data distribution.

3.3. Biofeedback Findings

Biofeedback is used in this study as a noninvasive monitoring of the physiological signal of the heart rate to determine the progress of the intervention, as well as as a stop criterion. In normal resting conditions, as in the experiment, the heart rate of adults ranges from 60 to 100 beats per minute (bpm). During the experimental test, the participant uses a smartband to measure heart rate through the HypeRate app. To illustrate the behavior of the signal during the test, Figure 10 shows a real heart rate signal for a man and a woman who participated in the intervention at the three levels.

In Figure 10, the domain of the function corresponds to the timeline of the virtual world, at the beginning and when the scorpion appears, hides, and reappears, while the codomain is related to the bpm. In level 1, in general, the woman presents a higher HR level than the man, and in contradiction the appearance of the scorpion provokes the decrement in the HR of the woman and increment in that of the man. Level 2 implies that the scorpion approaches at 3 or 4 m, so the distance is still considerable, and in that case there is no difference in the average HR between the female and the male. However, at level 3, the scorpion approaches between one and two meters close to the bed, making the exposure more perceptive to the alertness of the senses and increasing the HR. In this way, there is a different behavior, since when the scorpion hides for the second time the male HR increases and the female HR decreases. Although we only present the signals of two participants, a more comprehensive study could be conducted with all participants.

Table 3. Mean and standard deviation (SD) biofeedback heart rate (BPM) for the three levels for the participants.

Level	Mean	SD
L1	84.08	10.10
L2	84.07	8.32
L3	84.33	8.59
Total	84.01	7.81

summarizes the mean and standard deviation of the heart rate for the three levels for the participants.

In general, in the three levels the HR remained at 84 bpm, approximately within the normal range, with greater dispersion at level 1. However, in levels 2 and 3, women had a higher average HR than men. VR technology can provide instant feedback on physiological signals of the patient and proprioception, The biofeedback from therapists and test subjects is very positive regarding the current concept and realization.

Finally, in the VRET framework, the information generated in the three evaluations (usability, affects, and biofeedback) of the virtual reality environment is fed back into the continuous improvement of the next version of the system (Figure 1).

4. Discussion

Part of the main contribution of this work is the usability and affects study to examine user experiences of undergoing virtual reality system toward scorpion phobia exposure therapy with real-time biofeedback. Using the System Usability Scale and the Positive and Negative Affect Schedule, we evaluated a non-clinical sample of undergraduate students. The system is proposed as a support tool for psychologists to optimize resources and take advantage of the benefits of VRET. This work provides evidence that it is possible to realistically replace live exhibits with virtual models for specific phobias. Furthermore, this research aims to systematically develop a VRET platform using a MEDEERV-based framework [41]. This framework could used to develop scalable virtual environments; not only VRET but virtual reality environments in general.

Several papers have presented studies on VRET for specific phobias, with some studies related to VRET for arachnids. PhoByeVR is an interesting innovative educational tool for managing fear that targets arachnophobia, ophidiophobia, and acrophobia [20]. For arachnophobia, they manage the levels of exposure depending on the scene. The first level is a room with pictures on the wall and videos of spiders, in level 2 there is a VR head set with PC components to display the use of ${360}^{\circ }$ videos, level 3 is a room with a terrarium, and level 4 has a miniature jungle. In our case, we use a virtual psychotherapeutic room that has three levels according to the BAT based on the proximity distance, between the participant and the scorpion. In terms of testing, PhoByeVR was tested within a single case study with a single participant who reported a medium intensity of phobia, making it necessary to test with a larger population. For our part, we tested with a nonclinical sample of more than 50 participants.

In association with other works, Think+ is a mobile game application to treat different phobias [70]. The three level games of Think+ are related to the collection of stars containing spiders. Although they report that the game can be applied successfully on a psychotherapeutic basis, its main limitation is its lack of realism. Quite a complete work has been presented in [25], of which the objective was to check, among other things, whether VR participants were able to approach a real tarantula to a greater extent than non-VR participants. There was no indication that participants in the VR condition were able to approach a real-life spider to a greater extent than the non-VR group. Although it was not a treatment study as in this work, one limitation is that fear of spiders was not reassessed at a later stage, which does not present evidence of the treatment effect.

The VRET-AP is a recent tool for the one-session treatment of arachnophobia [19]. VRET-AP has been used as a feasibility study for exposure therapy, but with spiders and not scorpions. A limitation of this work is that the system was tested with 11 participants. In a novel way, ref. [71] propose a framework for VRET utilizing procedural content generation (PCG) and reinforcement learning (RL), demonstrating a higher performance of the system compared to more common rule-based VRET methods. Two scenarios of exposure to relaxing and anxious spiders were used. Furthermore, its implementation within controlled conditions does not indicate that the system is currently ready for clinical applications. Furthermore, the relatively small sample size may not fully capture the variety of responses.

Few studies have developed direct experiments to evaluate usability, focusing more on the functionality of VRET than its usability. Some of these papers, such as those presented in [35,36,37], have limitations related to sample size, lack of clinical samples, and analysis of the explanation of the usability results. Finally, in terms of biofeedback, although there are several that have integrated biofeedback or sensory feedback [38,39,72], it has not been applied in VRET systems for arachnophobia so far, as we have investigated. In our study, real-time heart rate biofeedback has been included, but other physiological measurements such as electrodermal activity [39] would benefit this study in the automatic anxiety measure.

Strengths and Limitations

To the best of our knowledge, the proposed system is one of the first works to present a framework for VRET toward scorpion phobia treatment using heart rate biofeedback and usability and affects evaluation. Compared to most studies on the use of VRET for arachnophobia, this study has a relatively large sample (54), where 30 is recommended [25]. When the sample increases, the opportunity to generalize the findings increases the reliability. So, for future studies, we plan to increase the sample.

First, the proposed framework allows one to develop a virtual environment with biofeedback and usability and affects assessment. Then, the VRET could be extended and scaled to other specific phobias or other arachnids. Although the system has several levels according to the exposure distance, it is necessary to increase the number of levels, the exposure scenarios, and the inclusion of haptic devices for user guidance. In addition, this study includes heart rate measures that allow the monitoring of decisions during interventions. In the future, studies could include other physiological variables, such as electrodermal activity, using a fusion sensor approach.

The system was tested in an unbalanced sample, both in terms of gender and living area (rural/urban). Although there are studies that report that women are generally more afraid of arachnids, such as spiders [43,73], in this study we did not address this question, nor did we address the difference depending on the area of living. Future research needs to test with a larger balanced sample, also the sex group and living area group, to probe some hypothesis related to scorpions phobia.

Another limitation is that the experiment was carried out in a non-clinical sample instead of people with diagnosed scorpion phobia, using initially to this end the widely used spider phobia questionnaire (SPQ) or the fear of spiders questionnaire (FSQ) [25]. However, the inclusion criteria study was not related to a sample diagnosed with arachnophobia. Among participants with scorpion phobia, real findings are expected [23]. This study could have benefited from initially using an adaptation of the fear spider questionnaire, and by using the specialized psychological diagnosis, it could avoid using self-reported fear of scorpions [74].

5. Conclusions

In this work, we present a framework for developing and assessing a novel virtual reality system based on the Methodology for the Development of Virtual Reality Educational Environments (MEDEERV) that includes real-time biofeedback toward scorpion phobia exposure therapy. We evaluated usability and affects in an unbalanced nonclinical sample, which are some of the most important limitations. However, for future research, we could test the system in a larger balanced clinical sample. In addition to favorable findings, there is evidence that the user experience tends to make the system acceptable. Although the average for positive affect tends toward moderate excitement, the average for negative affect generalizes somewhat. That makes it possible to determine the system as preliminarily technologically feasible and favorable for the proposed application.

The biofeedback measure that we used was heart rate, but electrodermal and sensor fusion techniques can be used. As a preliminary study, even though the system was tested in a laboratory with a controlled environment, our results demonstrate that our framework can be considered as a pre-feasibility study toward the scorpion phobia exposure therapy. The structured framework system could be extended to other specific phobias, to other arachnids, or including haptic user guidance. Finally, due to the large, multidimensional, multivariate nature of the data for the treatment of specific phobias, both in the development and the analysis of virtual reality exposure therapy (VRET) data, AI-based algorithms can be applied.