1 Introduction

Massive advances in the area of virtual reality (VR) technologies in recent years have attracted the attention of not only regular users but also experts and researchers, who, with respect to its features, have started to use VR as a cost-effective, accessible, and user-friendly tool in research and many other areas of human endeavor (Boeldt et al. 2019). Nowadays, VR can efficiently and immersively generate virtual worlds possessing human-centered interactivity features and sensory feedback options (Sherman and Craig 2003). Additionally, VR can be combined with biosignal measurement tools, enabling researchers to gather valuable information on users' physiological, behavioral, cognitive, and emotional responses. This integration of technology expands the potential uses and research possibilities of VR. Therefore, VR is increasingly being used in various fields such as psychotherapy (Freeman et al. 2017; Emmelkamp and Meyerbröker 2021), where it is most often used in the form of virtual reality exposure therapy (VRET) (Chou et al. 2021). However, various aspects of VRET are still only barely explored, especially regarding the connections between psychological and physiological reactions during the process of exposure in various settings. This study explores the effects of psychological guidance on anxiety levels during VR exposure to a height environment in individuals with acrophobia (moderate fear of heights), utilizing both biosignal measurements and self-reported state anxiety questionnaires.

1.1 VR and its specifics

VR, a technology simulating a sense of presence in a virtual environment, has become increasingly accessible with advancements in immersive technologies such as head-mounted displays (HMDs) (Muhanna 2015; Lindner et al. 2020). Among HMDs strengths is the ability to dynamically modify, trigger, update, or adapt virtual models and possible collaboration among users, regardless of their physical location (Ugwitz 2019; Juřík et al. 2016). Immersive Virtual Reality (IVR) environments, ranging from photorealistic to simple vector environments, play a crucial role in enhancing immersion. These environments leverage advanced visualization capabilities to create a profound sense of presence and immersion for users, allowing them to engage deeply with the virtual world (see more in Alfaro et al. 2023; Alsaleh et al. 2021). Since iVR context seems to activate brain mechanisms similar to those that occur in the real world (Alcañiz et al. 2009; Slater 2018), it has been found to effectively simulate real-world scenarios (Kisker et al. 2021; Felnhofer et al. 2019). These aspects represent strong features applicable in VRET (Chou et al. 2021), that have great potential to be applied in psychotherapeutic practice.

1.2 VR as a tool for the therapy of acrophobia

iVR offers a distinct solution by transporting individuals into simulated environments that replicate difficult scenarios. This controlled environment allows for teaching and training of appropriate responses, as well as deeper comprehension of the specific disorder at hand (Freeman et al. 2017). A typical mental disorder for which VR has been applied is phobias. Among them is acrophobia—strong and irrational fear of heights that is often comorbid with other mental health problems, such as anxiety disorders or depression (Kapfhammer et al. 2016). It has the highest prevalence among other specific phobias characterized by the natural environment and is associated with increased sweating and heart rate (HR), chest pain, palpitations, nausea, tremors, trembling, dizziness, or loss of balance (Kapfhammer et al. 2016, 2015). CBT is the most common psychotherapeutic method for acrophobia because it acts on a person's psyche in several aspects—thinking, emotions, and behavior (Beck 2011). The most common CBT intervention for acrophobia is exposure therapy (Chou et al. 2021; Arroll et al. 2017). Exposure assumes that the individual can overcome fear by repeatedly interacting with it. The use of VR in exposure therapy is unique as it offers repeated exposure in a virtual environment, potentially without the need for direct real-life experiences, as demonstrated by previous research (Ong et al. 2022). In a recent meta-analysis by Chou et al. (2021), which compared the effectiveness and acceptability of different interventions for acrophobia, the benefits of VR technology were particularly highlighted. Virtual environment can trigger realistic behavioral and physiological responses, and participants in a VR situation exhibit behaviors similar to those in a real-life setting (Kisker et al. 2021; Morina et al. 2015), and even a low-cost VR intervention can help to improve negative symptoms of acrophobia (Donker et al. 2019). The use of VRET in exposure therapy has also been supported by a systematic review and meta-analysis of studies ranging from 2012 to 2017 by Botella et al. (2017). However, there are some limitations to consider. One limitation could be the low adoption rate among therapists, which may be due to concerns about provoking distress, drop out, and logistical barriers (Demír and Köskün 2023). Another limitation may present the need for further research to optimize the treatment, such as conducting studies on VRET's real-world effectiveness and treatment optimization trials (Wray et al. 2023).

When evaluating VRET, an important consideration is whether the presence of a psychologist is crucial for its efficacy and influences the therapy session. While earlier research on VR interventions for phobia sufferers concentrated on therapist-assisted exposure (Kampmann et al. 2016; Anderson et al. 2013), recent studies have investigated the possibility of automated psychological therapy, thanks to technological advancements. (Freeman et al. 2018; Miloff et al. 2019; Lindner et al. 2020). Conversely, studies indicate that a psychologist-led VRET session places a higher priority on establishing a robust therapeutic relationship by mutually agreeing on tasks and goals, resulting in better treatment results, including improved adherence (Buchholz and Abramowitz 2020). The importance of psychological guidance during exposure therapy in VR cannot be overstated, yet it remains an under-researched area despite the significant potential of this technology as an intervention tool. While some scholars concentrate on automated interventions, the value of basic human guidance is still uncertain.

1.3 Physiological responses to anxiety and stress

Examining the physiological manifestations associated with acrophobia can provide insight into the experience of anxiety. The most common non-invasive physiological measures of stress and anxiety are HR and HRV (Ihmig et al. 2020; Giannakakis et al. 2022).

HR is the number of beats per minute and can be used to objectively assess stress, as it correlates with the increased metabolic demands of the organism when coping with stressful stimuli by increasing the oxygen consumption in the brain due to greater mental effort. One explanation for why HR increases briefly during the onset of anxiety is that in defensive and startle responses, the initial increased HR serves to reduce sensory input (Fiľo and Janoušek 2022a). Such an individual's adaptation is referred to as a proactive coping style (Koolhaas et al. 1999; Andringa and Lanser 2013), active behavior modification (Obrist 1981), and the rejection hypothesis (Gang and Teft 1975).

HRV, on the other hand, represents the time variation between consecutive heartbeats. Changes in parasympathetic and sympathetic nervous system activity result in beat-to-beat fluctuations in HR (Duong et al. 2020) known as HRV. These fluctuations occur simultaneously on different time scales (Shaffer and Ginsberg 2017), reflecting the adaptation strategies of the organism (Fiľo and Janoušek 2022b). High and low levels of HRV may reflect different stress coping strategies. In case of multiple or repeated exposures of the organism to anxiety-inducing stimuli, HR becomes a complex biosignal that combines responses to individual stimuli at different time scales into a single biosignal. An analysis of the HRV provides a detailed view of such a complex signal that a person's mental state can be objectively assessed (Ihmig et al. 2020). Such measures can provide crucial data on the VRET process, where specific mechanisms and processes are vaguely mapped in various settings. Available technologies for VR and HR can be, therefore, easily implemented to effectively enhance CBT practice in VRET.

1.4 Aims of the present study

It is only vaguely studied whether therapists play an important role in VRET. At the same time, however, many studies have demonstrated the utility of VR as a potential therapeutic tool (Coelho et al. 2009; Ose et al. 2019; Maples-Keller et al. 2017). In this study, using the subjective experience of anxiety as measured by the State-Trait Anxiety Inventory (STAI-Y1; Spielberger 1989) and measurements of HRV as an objective indicator of experienced anxiety (Felnhofer et al. 2014; Wiederhold et al. 2002), we study individuals' experience in the cases where specific psychological guidance during the VRET is present compared to those where it is not. Psychological guidance in our research refers to the guidance provided by the psychologist during the exposure—communicating throughout the exposure, attempting to regulate anxiety, and presenting supportive phrases to make the participant feel that they can cope better and are not alone in the situation. Because this support could not be considered standard CBT psychotherapy, the term psychological guidance better reflects the reality of the condition. Although the combined approach of measuring objective and subjective data during exposure has been addressed by a small number of researchers (Felnhofer et al. 2014), direct comparison and investigation of possible discrepancies between these approaches during exposure to a phobic stimulus have not been properly addressed. Therefore, we focus our research on a deeper exploration of their role in experiencing anxiety. Next, we are interested in examining how the presence of psychological guidance influences the experience of anxiety. Although findings in the area of psychotherapist presence during exposure emphasize the need for therapists to pay special attention to patients' interpersonal behavior during treatment, there are no precise conclusions for VRET (Maiwald et al. 2019).

The primary aim of this study is to investigate the anticipated patterns in self-reported anxiety levels among two cohorts of individuals exposed to heights in a simulated reality setting. The first group receives psychological guidance during the height exposure (PG), while the second group does not (no psychological guidance; NPG). Additionally, the expected trends are evaluated and discussed concerning the fluctuation in HRV of the participants.

2 Methods

2.1 Participants

The target population were individuals with a moderate fear of heights, who were chosen based on their score in the Heights Interpretation Questionnaire (HIQ; Steinman and Teachman 2011) (see more details below). Regarding the use of VR technology, our study focused on young adults, primarily college students and graduates recruited from Masaryk University. All participants were Czech speakers. Inclusion criteria required that participants have no serious health or visual disorders or any other medical impediments that would disqualify them from the study, ensuring that all participants were healthy young adults. Wearing prescription glasses was not considered an impediment, but participants were encouraged to use contact lenses if possible. Before the research began, participants were inquired for visual impairments and other medical conditions that might prevent their participation, particularly visual-sensitive epilepsy. Participants were informed about the possibility of experiencing motion sickness while using HMDs and were assured they could withdraw from the experiment at any time without needing to provide a reason. The total research sample consisted of 36 people (24 females; 66.7%), ages ranging from 19 to 36 years (m = 24.6; SD = 4; med = 24). We approached the potential candidates through various groups on the social network Facebook and flyers posted around the Faculty of Arts of Masaryk University in Brno. Participants were randomly assigned to either an experimental group or a control group. Participants did not receive any material compensation for their participation.

2.2 Ethics

The study followed the necessary ethical guidelines/standards, and it was assessed and approved as ethically indisputable by the Ethical panel of the Department of Psychology, Faculty of Arts, Masaryk University Brno. The study was conducted in accordance with relevant guidelines and regulations following the principles of the Declaration of Helsinki. Participation was voluntary and informed consent was presented to all participants to be read and signed before the beginning of the experimental session. Participants were explicitly informed that they could retrieve from the experimental session at any time without any negative consequences or loss of promised remuneration. All relevant aspects of the study were explained in detail to the participants before the start of the study, together with the study design and data processing. The anxiety experienced did not cause any long-term or even irreversible changes in HR.

2.3 Materials and Setting

Each research exposure took place in the Grey Lab held at the Department of Psychology, Faculty of Arts, Masaryk University in Brno. Participants received the visual stimulus using HTC Vive Pro, which consists of two 3.5-inch (8.89 cm) AMOLED screens with a resolution of 1440 × 1600 pixels per eye. The refresh rate allows for 90 Hz. A desktop PC equipped with an Intel i7 8700 K CPU and NVIDIA GTX1070 GP was used as an input device. Richie's Plank Experience app (RICHIE'S PLANK EXPERIENCE https://toast.games/) was used to represent the virtual height environment. In this app, the users find themselves in a bustling city center, where a tall 80-storey skyscraper is located. An elevator takes the individual to the top where a wooden plank awaits when the doors are opened. In order to create the greatest sense of realism of the application, we decided to place a real wooden plank on the ground, according to whose size the virtual plank was precisely calibrated (see Fig. 1). The fact that the virtual and real planks are placed in parallel makes this connection interactive and convincing. Besides city background noise, the sound of the virtual board cracking when stepped on also stimulates the auditory modality. In order to relax the participants after being exposed to a height situation in VR, we decided to use the Guided Meditation VR application (Guided Meditation VR https://store.steampowered.com/app/397750/Guided_Meditation_VR/). In the app, the user can choose between 27 environments designed for meditation.

Fig. 1
figure 1

The experimental setup using a VR headset and a real wooden plank as a haptic interface

Full size image

2.4 Instruments and measures

2.4.1 Instruments

Multiple questionnaires were used to capture participants' experiences during exposure to a height situation in a virtual environment. The translation of the questionnaires, previously unused in the Czech environment, was carried out in accordance with the standard procedure–two translations from English into Czech were produced, followed by synthesis and finally, back-translation into English for validation (Behling and Law 2000).

2.4.2 Heights interpretation questionnaire (HIQ)

HIQ (Steinman and Teachman 2011) is a self-report questionnaire consisting of 16 items. The scale of this questionnaire strongly predicts fear, anxiety, and avoidant behaviors in the case of actual heights. Participants were asked to rate their anxiety fears (e.g., of falling or hurting themselves) when imagining two different height-related situations (being on a ladder or on a balcony) on a five-point Likert scale ranging from 1 (unlikely) to 5 (likely). The HIQ was used in this study as both an inclusion and exclusion criterion for participation, and as an important indicator of fear of heights. The study was originally intended to include only participants scoring between 29 and 55 points, indicating an approximately moderate fear of heights and consistent with the selection criteria of other studies (Freeman et al. 2018; Arroll et al. 2017). However, due to the lower number of participants, we decided to include three participants whose HIQ scores approached the criterion of moderate distress, and thus our final range of HIQ scores is in the range between 26 and 55 points. Inclusion of participants with moderate fear of heights in our study prompts consideration of implications for those with higher fear scores, particularly individuals diagnosed with phobias. Previous studies have noted that individuals with anxiety disorders often exhibit reduced HRV (Held et al. 2021).

2.4.3 State-trait anxiety inventory (STAI)

STAI (Spielberger 1989) represents the most widely used method of measuring anxiety and is often used to measure subjectively experienced anxiety in the context of VR exposure (Ling et al. 2014). It consists of two subtests, each comprising 20 items. For the purposes of our research, the subtest focusing on state anxiety (STAI-Y1) was used, in order to capture their immediate feelings after exposure to a height situation. The task was to express their attitude towards several statements, such as "I feel tense", "I feel excited" etc. by answering "not at all" (1), "a little" (2), "rather yes" (3) or "very" (4). The STAI-Y1 questionnaire was used due to (a) inspiration from past studies with a similar intention to measure state anxiety in VR (Concannon et al. 2020; Felnhofer et al. 2014); (b) although it would have been ideal to capture actual fear instead of anxiety, that was also chosen due to the lack of relevant questionnaires capturing the current level of fear.

2.4.4 Acquisition system for HR monitoring

The HR was continuously monitored with a Polar H10 chest strap (Go Wireless® Exercise Heart Rate https://www.vernier.com/product/go-wireless-exercise-heart-rate/) with remote data recording using the Elite HRV application. It is a non-invasive, precise (sampling frequency of 1000 Hz), safe, and easy-to-use HR monitoring tool that connects via Bluetooth wireless technology. The device is depicted in Fig. 2.

Fig. 2
figure 2

Polar H10 chest strap

Full size image

2.5 Procedure

Participants were asked to complete the HIQ questionnaire and based on their scores were invited into the laboratory experiment. Participants were randomly divided into two groups: the PG group (psychological guidance during the height exposure) and the NPG group (no psychological guidance), during the exposure to the height situation in VR. To validate the random assignment, we inspected both groups' HIQ test scores to ensure that the groups had similar HIQ score distributions.

First, participants were asked to deploy the Polar H10 device and HR recording started with the aim of capturing the participants' basal HR for at least five minutes. Then, participants were asked to look at a screenshot from Richie's Plank Experience app and then complete the first STAI-Y1 questionnaire (STAI-Y1-bef). Participants could then deploy the HDM and we gave them a few minutes to adapt to Richie's plank experience environment—take a walk around (ground floor—street), and look in the elevator so we could spot possible complications (e.g. cybersickness) before the exposure. In the experimental scenario participants were instructed to enter the elevator in the virtual environment and take it to the top floor by pressing the elevator button. Participants were informed that they would ideally arrive at the end of the board, turn around, and return to the elevator, but they were reminded to only take as many steps as their physical condition would allow. After returning to the elevator, participants rode back down to the ground floor. The whole procedure of the experiment is depicted in Fig. 3.

Fig. 3
figure 3

Scheme of the experiment

Full size image

Before the exposure, all members of the PG group were briefed on what to expect, along with psychological guidance—we assured them that we would be communicating with them throughout the process, and aid in reducing their fear of the situation, as through the exposure we want to focus mostly on perceived fear. In order to maintain a therapeutic standard (the client must be informed about the procedure in advance), this was a crucial part of the pre-exposure preparation process.). All formulas and training for psychological guidance (PG group) were prepared in active and demonstrative collaboration with the expert CBT therapist. The therapist provided thorough leadership and mentoring to the researcher regarding the exposure process, including the appropriate language to be used and how to respond to clients, even during critical situations. For example, the researcher was advised on effective methods to induce calmness in the client, emphasizing that anxiety is a temporary state, and that relief can be achieved by persevering through the situation. The psychologist's guidance was influenced by both biosignal measurements (live observation of HR data via the display) and visual observations of the participant, so that the exposure could be continuously modified as the participant's condition changed.

The last step for participants was to take off the VR goggles and once again complete the STAI-Y1 (STAI-Y1-aft). In the meantime, the researchers monitored the participants' HR for five minutes after exposure, which should provide the data on the process of calming down. After this part, the Polar H10 device was removed and HRV recording stopped. In the last part of the session, participants were given time to relax through the "Compassionate Body scan" mode of the Guided Meditation app. The experiment ended after the meditation was completed.

2.6 Data analysis

The data were analyzed with the use of JASP (version 0.16.1) and Python (version 3.6, Pandas 1.5.0). The analyzed psychological variables were STAI-Y1-bef score captured before the height exposure and STAI-Y1-aft score captured after the height exposure. Homogeneity tests allowed for using repeated measures ANOVA for analyzing data from these psychological questionnaires. The correlation between STAI-Y1-aft and the mean HR at the time of exposure (HRV2) was calculated using Spearman's correlation coefficient. Regarding HR, the HR series were transformed into NN series (normal-to-normal intervals) and linearly detrended by subtracting the mean NN interval. The standard HRV parameters (according to the Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology 1996) were computed using pyHRV: open-source python toolbox for HRV (Gomes et al. 2019). Time domain, frequency domain, and nonlinear parameters were computed, namely mean HR (hr_mean), minimal and maximal value or HR (hr_minimal_value, hr_maximal_value), standard deviation of instantaneous HR values (hr_std), mean interbeat difference (nni_diff_mean), standard deviation of interbeat intervals (sdnn), square root of the mean squared difference between successive interbeat intervals (rmssd), standard deviation of interbeat differences (sdsd), number of interbeat differences greater than 50 ms (nn50), ratio between number of interbeat differences greater than 50 ms and total number of interbeats interval (pnn50), baseline width of the interpolated triangle (tinn), triangular index (tri_index), LF/HF ratio computed by Welch’s method (fft_ratio), Lomb-Scargle periodogram (fft_lomb) and autoregressive method (ar_ratio), standard deviation of the major axis (SD1) and minor axis (SD2) of Poincaré diagram and its ratio (sd_ratio), sample entropy of the interbeat series (sampen), alpha value of the short-term fluctuation (dfa_alpha1) and long-term fluctuation (dfa_alpha2). All parameters were statistically evaluated using ANOVA and post hoc Tukey’s test.

3 Results

Initially, we verified the equivalence of the groups (PG and NPG) regarding their HIQ scores and physiological states before the experiment by calculating the mean HR and standard deviation of HR during the 30 s prior to the start of the exposure, when all participants were seated without moving and waiting for the 5 min time window to elapse. In terms of HR mean values, the groups were equivalent (89 for PG and 87 for NPG). As for standard deviations, the PG group flew more around the mean than the NPG group (18 vs. 13). However, we can consider the groups equivalent, the general mood of participants in each group was similar before the experiment began.

3.1 Psychological response

Self-reported levels of anxiety were analyzed for each condition in two phases (before and after VR exposure), measured scores are summarized in Table 1.

Table 1 Measures of central tendency for STAI-Y1 measures before and after VR exposure
Full size table

Assumptions checked by Levene's test allowed for using repeated measures ANOVA with additional analysis of experimental conditions as between-subject factors. Regarding self-reported anxiety before and after VR exposure, repeated measures ANOVA identified that there was a significant increase in the level of anxiety in both conditions (see Fig. 4), which was found significant with a large effect (F = 395.636; p < 0.001; ηp2 = 0.921). Post hoc comparisons revealed significant differences in anxiety increase in both the PG group (t = − 15,164; p < 0.001) as well as in the NPG group (t = − 12,966; p < 0.001). However, the degree of experienced state anxiety captured by the STAI-Y1 questionnaire did not statistically significantly differ between the PG and NPG groups (F = 0.153; p = 0.698) with negligible effect (ηp2 = 0.004). This was observed in both phases (see Fig. 4), where even post hoc comparisons did not show any interaction effects. The expectations about the different levels of experienced anxiety between PG and NPG groups were not captured with the use of STAI-Y1 questionnaires.

Fig. 4
figure 4

Comparison of anxiety level reported by STAI-Y1 before and after the VR exposure according to PG and NPG groups; 95% CI error bars

Full size image

3.2 Physiological response

RR interval analysis using standard HRV parameters was used to objectively assess stress during the experiment. The main trends are represented by the observed differences in dynamics for PG and NPG groups in the experimental session. The degree of anxiety in the anticipation phase (i.e., before the administration of the VR stress stimulus) sets the participants' nervousness equally both for psychologically guided and non-guided participants. As mentioned above, the groups were equal in the initial mean HRV ratio (anticipation). During the stress phase (stress), HRV increased for both groups. After the VR exposure (catharsis), for the PG group, the HR values strongly decreased to values even lower than values in the anticipation phase of the experiment. However, for the NPG group, HR values fall slowly and steadily until they reach the anticipation phase values (see Fig. 5).

Fig. 5
figure 5

HR in PG (left) and NPG (right) groups during anticipation, stress, and catharsis. The mean HR is shown by a dotted line, and the range of HR is presented as a grey area

Full size image

Further visual inspection of Fig. 5 suggested that psychologically guided participants experienced anxiety intensely just after the administration of a stressful stimulus, but then the level of anxiety gradually decreased with minor oscillations. Participants without psychological guidance, on the other hand, experienced anxiety in unequally intense repeated, and periodically varying waves. This quasi-periodic way of experiencing anxiety manifests itself mainly in the NPG group. More detailed analyses provided additional information. To describe a more complex understanding of the underlying trends, selected parameters (mean HR; standard deviations of the HR; entropy) were further analyzed (see Fig. 6 and Tables 2, 3 and 4). Considering the small size of the research sample, non-parametric methods (Mann–Whitney U tests for group comparisons and Wilcoxon Signed-Rank Tests for phase comparisons) were applied to the data.

Fig. 6
figure 6

Difference between experimental phases (anticipation, stress and catharsis) in mean heart rate (left), standard deviation of the heart rate (middle) and sample entropy (right)

Full size image
Table 2 Mean heart rate measures and tests results for individual conditions and phases
Full size table
Table 3 Standard deviation of the HR measures and tests results for individual conditions and phases
Full size table
Table 4 Sample entropy measures and tests results for individual conditions and phases
Full size table

Mean HR was analyzed to assess the intraindividual level of the stress caused by the experimental stimulus. The mean HR significantly increased during the stress phase of the experiment both in the PG and NPG group (see Fig. 6, left). The mean HR significantly decreased in the catharsis phase for both the PG and NPG group to values even lower than values in the anticipation phase of the experiment (see Table 2). No differences were identified between PG and NPG groups in their mean HR levels. More details can be seen in Table 2.

The consistency level of the reaction to the stress event can be expressed by the standard deviation of the HR, where a higher SD indicates greater variability in HR measurements and a lower SD indicates a more consistent HR. The magnitude of the standard deviation reflects the extent to which participants compensated for a stress reaction. The standard deviation of the HR measured in the research sample increased statistically significantly in the stress phase of the experiment among participants in both the PG and NPG group (Fig. 6, middle). Based on the recorded variability of the HR standard deviation measures, we can observe that in both groups, the stress compensation reaction was continuously and repetitively affected by experimental stimulus, however, this compensation was more prominent in the NPG group. This difference between experimental groups in the third phase of catharsis was identified as statistically significant (U = 88; p = 0.02); more details can be found in Table 3.

The complexity of the HR signals can be described by the sample entropy. The analysis of entropy of the NN interval series suggested that there was generally a similar effect of the VR exposure on stress reaction in both conditions since VR exposure caused a significant physiological response in both groups (see Fig. 6, right). The complexity of HR series differs statistically significantly between the stress phase and other phases. For more details see Table 4.

4 Discussion

This study aimed to evaluate the responses of individuals who experience moderate fear of heights in an iVR environment. Specifically, we sought to explore the changes in their subjective anxiety and physiological responses while exposed to the iVR environment. The primary objective was to compare the reactions of two groups with and without psychological guidance: PG and NPG groups. The groups' stress levels were assessed using the STAI-Y1 questionnaire and heart rate variability (HRV) analysis. Regarding psychological self-reported anxiety levels, repeated measures ANOVA revealed a significant increase in the level of participants' anxiety after VR exposure compared to the phase before the exposure. The effect of VR-based height exposure on the anxiety increase was large and present in both the PG and NPG groups. Corresponding to results from self-reported questionnaires, the HRV rate in both groups indicated increased anxiety in the phase of the VR exposure. The mean HR dramatically increased during the stress phase of the experiment. This observation further underlines the expected effect of VR exposure within the context of VRET and supports the view that the immersive virtual exposure in the context of a perfectly safe laboratory context can induce objective changes in HR and related physiological parameters. We conclude that in this aspect the iVR exposure has the potential to activate the stress reaction in a similar way as it would in the reality, which supports the previous claims about VR's suitability for therapeutic purposes (Kisker et al. 2021; Morina et al. 2015).

Interestingly, the general expectation about differences in anxiety levels among PG and NPG groups was not supported. The experienced anxiety captured by the STAI-Y1 questionnaire did not statistically differ with negligible effect. No differences were observed in either phase. Regarding this, expectations about the different levels of subjectively experienced anxiety between PG and NPG groups were not captured with the use of STAI-Y1 questionnaires. The HRV rate, on the other hand, suggested several interesting trends regarding PG and NPG groups allowing a better understanding of anxiety dynamics during the VRET.

Data evidence supported our expectation that the VR exposure significantly increases the mean HR in both groups. Also, mean HR values significantly decreased when the stress stimulus vanished in both groups. Remarkably, for both the PG and NPG group, mean HR values decreased significantly to values even lower than values in the anticipation phase of the experiment. As shown in the Fig. 5, the decline in HR was very steep, taking only a few heartbeats to get to the baseline. For a NPG group, mean HR values fall slowly and steadily until they reach the anticipation phase values. Closer inspection of this trend in the third catharsis phases (sse Fig. 5) suggested that in psychologically guided participants who in general experienced anxiety intensely just after the administration of a stressful stimulus, anxiety was not experienced constantly, but the level of anxiety gradually decreased with small oscillations. Participants without psychological guidance, on the other hand, experienced anxiety in unequally intense repeated waves. Regarding this observation, the statistical analysis of the HR standard deviation was conducted to statistically test the consistency of the reaction to the stress. In this measure, a higher SD indicates greater variability in HR activity and a lower SD indicates a more consistent HR, i.e., it expresses a level of participants' compensation for a stress reaction. Based on the observed HR SD in the third phase of catharsis, the compensation for the stress was found to be more prominent in the NPG group with the statistically significant difference in comparison to PG condition. These findings are consistent with the habituation principle during the phobic stimulus exposure technique, in which the stress level does not increase indefinitely – on the contrary, at a certain point, it decreases (Benito and Walther 2015; Beck 2011; Ihmig et al. 2020). This may be since participants do not perceive anxiety as a constant variable but as more regular bouts of anxiety. Another explanation is that although the participants felt anxious constantly, their bodies responded by periodically suppressing it. An important finding is that the quasi-periodic way of experiencing anxiety manifests itself mainly in the group of participants without psychological guidance. Thus, it seems that psychological guidance can mobilize the autonomous nervous system so that both the effect of the sympathetic nervous system involved in the fight-or-flight reaction and the effect of the parasympathetic nervous system establishing relaxation is more effective. Individuals without psychological guidance, on the other hand, could be overwhelmed by feelings of chaotic quality and intensity. This further promotes the potential of therapeutic support during the process of VRET as discussed above (Freeman et al. 2018; Maiwald et al. 2019). From the perspective of heart entropy, we can speak of an equivalent course in both groups, where the role of the psychologist is eliminated because the fundamental functioning of the heart was the same for both groups.

This research has several limitations which should be discussed. As the main limitation of the study, we consider the limited research sample which was caused by the increased logistical and technical requirements of the study in the time of the global pandemic, and which may partially downplay the reliability of conducted statistical inferences. In this manner we consider the study combining the psychological and physiological measurement approach as an exploratory probe into the exciting topic of VRET possibilities. Also, we anticipate that the scope of this study was limited in the context of actual therapeutic processes where a specific long-term relationship between the client and the psychologist plays a significant role. It also applies to the specific procedure of psychological guidance used in this study, which was based on the CBT methods, but cannot be considered as a full psychotherapeutic intervention. However, regarding trends suggested in this study, we believe that the VRET processes discussed in this study may be even more amplified in the context of real psychotherapy. Another limitation of our study is that participants may have had pre-existing expectations about the session, potentially affecting their emotional responses. These expectations, influenced by past experiences or external information, could introduce bias into our findings despite efforts to mitigate it through clear communication and guidance. In future studies, we should pay closer attention to participants' previous experiences, including VR exposure, passive exposure to similar situations and collect questionnaire data also on cybersickness. This should be reflected in participant selection criteria to maintain a more homogeneous sample, helping us better understand and address the potential impact of pre-existing expectations on emotional responses. We can not forget to mention limitations related to the way subjective anxiety is captured in our research. While the STAI-Y1 is reliable for pre- and post-exposure anxiety assessment, it doesn't capture stress fluctuations during exposure. For this purpose, a simple self-report scale (for example A Subjective Units of Distress Scale—SUDS, for example see Takac et al. 2019) during exposure sessions would be appropriate, but since we utilized a commercial app without such functionality, it couldn't be integrated (we did not want to administer SUDS verbally because we did not intervene at all in the NPG group). Including a self-report scale for assessing stress fluctuations during exposure sessions would provide a more comprehensive understanding of participants' experiences. Addressing this limitation could enhance the validity of the study's findings and offer valuable insights into the dynamics of anxiety within exposure therapy contexts. As a final point, the absence of pre-testing for negative emotions specifically induced by the VR exposure environment is also a limitation. For instance, we did not gauge distress levels through exposure to an unrelated stressful environment. Incorporating pre-testing with diverse scenarios in future research could enhance our understanding of emotional reactions and improve the validity of our findings.

5 Conclusions

In this research paper, the anxiety experience of participants with a moderate fear of heights when exposed to a height situation in VR was explored, given the growing trend of the use of VR in the psychotherapy of acrophobia. The main aim was to investigate differences in HRV in two groups of participants–with and without psychological guidance when exposed to the heights in an immersive virtual environment. We also wanted to explore if the groups would differ in scores on the STAI-Y1 questionnaire. Both items of interest were assessed with the intent to learn about participants' experience of anxiety in a height’s situation presented in VR and to study the effect of the presence of psychological guidance on such an experience. Psychological measures (STAI-Y1) captured an expected increase in experienced anxiety of participants after the height exposure; however, no differences were identified regarding experimental conditions. The most significant findings based on objective physiological measures indicate that, at the physiological level, VR exposure elicits responses similar to those observed in real contexts. Additionally, participants who received psychological guidance demonstrated a greater ability to compensate for anxiety compared to those without such support. Although several limitations may have influenced this research, our findings may contribute to a deeper understanding of the capturing of experienced anxiety during VRET, and thus approach appropriate tools in future investigations of its effectiveness.