The impact of virtual reality scenes on stress response characteristics of individuals with different personality traits

Abstract

Virtual reality based physical stress (VRPS) paradigms could eliminate the influence of social factors on participants, and it may be a desirable tool to explore the impact of personality traits on stress levels. In this study, we attempt to explore the effects of VRPS on stress response among individuals with different personality traits. Forty male participants with an average age of 22.79 ± 0.41 years were divided into two groups based on Harm Avoidance (HA) scores of Tridimensional Personality Questionnaire (TPQ), referred to as the Low-HA group and the High-HA group. The stress levels of the participants were assessed using salivary α-amylase (sAA) activity and heart rate variability (HRV) indices pre- and post-stress. The influence of personality traits on stress response among different groups was analyzed. VRPS significantly affected the sAA activity and HRV indicators of both groups. During and after stress, there were significant differences in sAA activity and HRV indicators between the two groups. The sAA levels and HRV indices of the Low-HA group were lower than those of the High-HA group. Furthermore, sAA levels and HRV indices were correlated with the scores of TPQ. VRPS scenarios elicit different stress responses on individuals with different harm avoidance personality traits. Stress evaluation based on VR scenarios presents potential in personality trait assessments, particularly for distinguishing between individuals with low and high HA tendencies.

Keywords:

• Stress
• VRPS
• TPQ
• personality traits
• sAA
• HRV

1. Introduction

Stress is a nonspecific reaction, which can activate the hypothalamic-pituitary-adrenal (HPA) axis, autonomic nervous system, and cholinergic system, leading to the release of glucocorticoids, catecholamines, neuropeptides, and other substances that exert regulatory effects (Karanikas, 2022). Currently, there are several widely used stress paradigms employed in laboratory settings, such as the Trier Social Stress Test (TSST), Montreal Imaging Stress Task (MoCA), and Socially Evaluative Cold Pressor Test (SECPT) (Dedovic et al., 2005; Kirschbaum et al., 1993; Schwabe et al., 2008). Nonsocial stressors refer to stressors that do not involve social interactions or social evaluation, which can encompass environmental pressures, work or academic stress, and physiological stress, and so on. Nonsocial stressors more accurately reflect the genetic underpinnings, such as the stress regulation of neuroendocrine substances involved in sensory seeking, harm avoidance, and similar stress response patterns (Lü et al., 2022).

Stress responses exhibit individual differences, and some studies differentiate between different personality types using personality scales, such as Big Five Personality Traits (Big-5), Tridimensional Personality Questionnaire (TPQ), and Eysenck Personality Questionnaire (EPQ), and so on (Malin & Littlejohn, 2012). The TPQ scale was developed by Cloninger based on a comprehensive review of numerous studies in neurophysiology, biochemistry, anatomy, and human and animal learning behavior. It is considered a personality assessment tool with a neurogenetic foundation (Huang et al., 2013). The TPQ scale consists of three independent dimensions: Novelty Seeking (NS), Harm Avoidance (HA), and Reward Dependence (RD), which are associated with three neurotransmitters, namely serotonin, norepinephrine, and dopamine. Research proved that personality development is formed through the interaction of the individual’s genetic disposition on their environmental domains (Hopwood et al., 2011). The HA dimension of TPQ reflects changes in the brain’s “punishment” or behavioral inhibition system. Individuals with low HA tend to be more extroverted, optimistic, and willing to explore stimulating experiences compared to individuals with high HA (Huang et al., 2013). Harm avoidance is a trait that involves the reactivity to negative stimuli, and individuals with this trait learn to inhibit their behavior to avoid punishment, novelty, and non-rewarding situations. Previous studies have found that participants in the Low-HA group exhibited significantly lower levels of stress compared to those in the High-HA group (Ma et al., 2018). Therefore, In the present study, the TPQ was selected as a tool to explore the biological perspectives of personality (Cloninger, 1987; Hofmann & Loh, 2005; Wetzel et al., 1992).

Biomarkers can be used for the quantitative analysis of stress levels (Aversa-Marnai et al., 2022). It includes corticosteroids (such as cortisol, COR), catecholamines (CA), other adrenal medulla-released hormones, and so on. Indeed, each biological marker responds to stress at different rates, reflecting the complexity of the body’s response systems. The temporal dynamics of biomarkers are crucial for interpreting the physiological stages and severity of the stress response. Here is an overview of the response times for several key stress biomarkers: For instance, the peripheral blood cortisol response reflects the organism’s stress response within a few minutes (Walther et al., 2022), while the CA response in morning urine reflects the situation within a few hours, and cortisol levels in hair can provide an average measure over several months (Abdelwahed et al., 2022,Ewing-Cobbs et al., 2023). In recent years, salivary alpha-amylase (sAA) has also been recognized as an important biomarker associated with autonomic nervous system stress response (Kang, 2010; Matsuura et al., 2012; Myers et al., 2023; Sahu et al., 2014; Tanaka et al., 2012; Tecles et al., 2014). The secretion of salivary proteins is the result of coordinated activity between the sympathetic and parasympathetic nervous systems, and sAA concentration is the most sensitive indicator of sympathetic nervous system activity, exhibiting excellent temporal sensitivity. Due to the effectiveness and reliability of sAA, it has been widely used as a biomarker in research on psychological stress and disease. Assessing stress in different populations through sAA detection provides supplementary information and offers temporal advantages over assessing methods such as salivary cortisol and catecholamines (Glier et al., 2022, Abdelwahed et al., 2022).

Heart rate variability (HRV) reflects the changes in cardiac autonomic nervous system function, indicating the dynamic balance between the sympathetic and parasympathetic branches. The high-frequency component (HF) and low-frequency component (LF) of HRV signals respectively reflect the activity of the parasympathetic and sympathetic branches of the autonomic nervous system. HRV is a noninvasive measure that assesses autonomic activity (Whitcomb et al., 2017). Short-term HRV analysis can serve as an effective indicator for measuring acute psychological stress and quantifying the level of stress experienced by individuals (Li et al., 2021).

Virtual Reality (VR) technology utilizes computer-generated virtual environments to engage participants in immersive experiences (Jessica et al., 2023). VR scenarios can be designed to induce psychological and physiological stress responses in participants through stress-inducing tasks (Powers & Emmelkamp, 2008). It has been widely utilized as a stressor (Silva et al., 2022,Geslin et al., 2020). Additionally, VR technology allows for the creation of more thrilling scenarios, such as high-altitude or flying scenes, that cannot be achieved by conventional social stress induction paradigms (Dedovic et al., 2005; Kirschbaum et al., 1993; Schwabe et al., 2008,Dammen et al., 2022). Currently, the Virtual Reality Trier Social Stress Test (TSST-VR) has been established as an effective and widely used tool for inducing laboratory-induced stress responses (Halbeisen et al., 2023; Zimmer et al., 2019). Various stress assessment methods based on VR are becoming increasingly popular (Dammen et al., 2022). However, conventional VR stress response paradigms like TSST-VR still fall under the category of social stress response paradigms and cannot eliminate the influence of social components on participants. Virtual reality based physical stress (VRPS) paradigms could eliminate the influence of social factors on participants, and it may be a desirable tool to explore the impact of personality traits on stress levels.

In recent years, the impact of different personality traits on stress levels has become a research focus, and the relationship between sAA levels, HRV indicators and stress response of participants with different personality trait require further exploration (Afrisham et al., 2015). In this study, we hypothesize that the High-HA group has more pronounced stress responses, and a correlation exists between sAA levels, HRV indicators and the scores of the TPQ scale. The results of this study will lay the foundation for establishing a quantitative model of stress responses in individuals with different personality traits under VRPS.

2. Research method

2.1. Research participants

The study recruited participants through social media, such as school forum, WeChat, and poster, and we will give each participant 30 yuan. All the participants were male students with an average age of 22.79 ± 0.41 years. Female participants were excluded because individual cortisol response in stressful situations can be influenced by hormonal fluctuations during the menstrual cycle (Weiss et al., 1999). There were no significant between-group differences among the participants in terms of marital status, body mass index (BMI), blood pressure (BP), and exercise levels. Prior to the testing, participants were required to sign an informed consent form. They were also required to have no history of working at heights, having less than three experiences with virtual reality, be in good physical health, and have no psychiatric disorders. During the preparation phase of the experiment, participants were not allowed to take any medications. More than 60 individuals responded to the experiment, and we selected 40 eligible participants among them. The 40 participants were sorted based on their HA scores, with the top 20 participants forming the High-HA group and the bottom 20 participants forming the Low-HA group.

2.2. VRPS

VRPS paradigm was used to eliminate the influence of social factors on participants. Participants were required to experience the VR roller coaster for 5 min, and the VR roller coaster scenarios are divided into three stages, ranging from gentle to intense. The VR device utilized in this study was specifically designed to meet the experimental requirements. We employed the HTC VIVE PRO head-mounted display, known for its high-quality visual display capabilities. Additionally, a small fan was used to simulate the high-speed winds experienced during a roller coaster ride.

2.3. TPQ

This study assessed individual personality differences among the participants using the TPQ scale (Huang et al., 2013). The internationally standardized version of the TPQ scale used in this research consists of 100 short yes-or-no questions and is divided into three independent dimensions: NS, HA, and RD. Prior to completing the questionnaire, participants were in a stable emotional state. The time required for each participant to complete the personality questionnaire was within 10 minutes, ensuring reliable and authentic results.

2.4. sAA activity level

Before collecting the first saliva sample from each participant, all the participants were instructed to rinse their mouth with water, and participants were not allowed to consume any beverages or food other than water throughout the experiment. The saliva sample was stored in a light-protected environment at −20 °C until fully frozen, and then transferred to a −70 °C ultra-low temperature freezer for further analysis. To avoid any potential interference with the tests, participants were instructed to refrain from consuming stimulant drinks or food, refrain from smoking, and maintain a standardized diet starting from the day before the testing. The collected saliva was divided into two portions and placed into two plastic vials with caps (one for immediate analysis and the other as a backup for retesting in case of measurement errors), and each sample was promptly labeled with a unique identification number. The sAA response to stress was assessed using a self-developed flow injection method (Chen et al., 2012).

2.5. Experimental design

All participants were divided into two groups based on the HA dimension of the TPQ scale. VRPS was used as a stressor. The stress response levels of participants were assessed through the comprehensive evaluation of sAA vitality and HRV indicators, to evaluate the impact of virtual reality technology on stress reactions among diverse individuals.

2.6. Test procedure

In the experiment, participants are required to wear a wireless heart rate monitoring device (Polar RX800). Firstly, the participants undergo the first saliva collection and electrocardiogram (ECG) recording. Subsequently, they are asked to complete the TPQ questionnaire, with a time limit of 10 minutes for completion. After a 15-minute rest, the participants are informed that the stress test is about to commence, and the second saliva collection and ECG recording are conducted. The participants are then exposed to a 5-minute stimulation using a VR roller coaster scenario. After stimulation, the third saliva collection and ECG recording take place. A 15-minute rest is provided before the fourth saliva collection and ECG recording. The testing procedure is illustrated in Figure 1.

Figure 1. Test procedure.

2.7. Statistic analysis

All data in this study were processed and analyzed using SPSS 26. T-test was used to evaluate the differences in TPQ between two groups, RM-ANOVA and Wilcoxon rank-sum test were conducted to examine group differences in sAA activity levels and HRV indices, and Spearman correlation analysis was employed to assess the relationships between TPQ and sAA activity levels, as well as HRV indices (Table 1).

Table 1. Descriptive data and statistical indices for sample characteristics as well as stress induction measures (n = 40).

Variable	Low-HA (n = 20)		High-HA (n = 20)		Statistical indices	p-value	Cohen’s d
	Mean	SD	Mean	SD
Sample characteristics
Age	22.96	0.39	22.73	0.44	t(38)=1.72	p = 0.09	0.54
BMI	23.86	1.01	23.70	1.00	t(38)=0.49	p = 0.63	0.15
Weight	73.25	3.73	72.99	4.20	t(38)=0.20	p = 0.84	0.06
Height	174.94	2.64	174.57	2.45	t(38)=0.45	p = 0.66	0.14
TPQ
NS	22.84	2.54	18.94	1.29	t(38)=5.98	p < 0.001	1.89
HA	13.77	0.38	19.23	0.37	t(38)= −11.42	p < 0.001	−3.61
RD	19.83	1.4	18.31	2.17	t(38)=2.57	p < 0.05	0.81
sAA
sAA1	1700.84	104.16	1644.28	95.65	W = 344	p = 0.07	–
sAA2	1799.82	171.62	2281.22	152.03	W = 223	p < 0.001	–
sAA3	1844.69	361.30	2148.06	445.64	W = 335	p < 0.05	–
sAA4	1496.09	119.01	2060.40	105.88	W = 210	p < 0.001	–
HRV
HFnu1	29.10	3.73	31.81	3.81	W = 327	p < 0.05	–
HFnu2	39.69	3.04	45.00	3.27	W = 251	p < 0.001	–
HFnu3	27.59	1.51	38.64	5.39	W = 210	p < 0.001	–
HFnu4	22.63	3.37	35.64	5.68	W = 214	p < 0.001	–
LFnu1	76.63	9.62	75.38	8.68	t(38)=0.42	p = 0.68	0.13
LFnu2	87.50	6.62	102.10	7.33	t(38)= −6.44	p < 0.001	−2.04
LFnu3	72.22	5.83	83.84	4.99	t(38)= −6.59	p < 0.001	−2.09
LFnu4	82.22	5.83	94.14	6.66	t(38)= −5.87	p < 0.001	−1.86

Sample characteristics, TPQ, and LFnu follow a normal distribution, and the independent samples t-test is used. sAA and HFnu do not follow a normal distribution, and the Wilcoxon rank-sum test is used. The numbers after sAA, HFnu, and LFnu respectively represent the first to the fourth measurement.

3. Result

3.1. TPQ

In this study, 40 representative participants were selected and divided them into two groups: the Low-HA group and the High-HA group, each consisting of 20 individuals.

The Low-HA group of participants exhibited high levels of novelty seeking and reward dependence, as well as low levels of harm avoidance. The High-HA group displayed low levels of novelty seeking, high levels of reward dependence, and high levels of harm avoidance. According to the independent samples t-test, there were significant differences in TPQ scores between the Low-HA group and the High-HA group (p < 0.05*).

3.2. sAA activity level

3.2.1. Variation trend of sAA activity value in different populations

The sAA activity data for both groups of participants 25 mins before stress, before stress, after stress and 15 mins after stress are denoted as $\mathit{\text{sA}}{A}_i(i=1,2,3,4)$ and are presented in Figure 2.

Figure 2. sAA levels 25 mins before stress, before stress, after stress and 15 min after stress. (A) Low-HA; (B) High-HA. The “*” indicates a significant difference between the current data and the sAA level 25 minutes before the stress.

According to the Shapiro-Wilk test, sAA₂, sAA₃, and sAA₄ do not follow a normal distribution (p < 0.05*). Based on the Wilcoxon rank-sum test, significant differences were found in sAA₂, sAA₃, and sAA₄ between the two groups (p < 0.05*).

For participants in the Low-HA group, there is a slight increase in sAA activity before stress. After stress, there is individual variability in sAA activity. 15 minutes after the stress ends, sAA activity returns to a low level. Participants in the High-HA group exhibit a significant increase in sAA activity before stress. After stress, there is substantial individual variability in sAA activity, and the degree of difference is greater compared to the Low-HA group. 15 minutes after the stress ends, sAA activity remains at a higher level.

3.2.2. Differences in stress levels among different populations

The sAA activity levels of the two groups of participants were relatively similar 25 minutes before stress. To assess the different stages of stress response, the difference between the sAA activity values of the subsequent three measurements and the first measurement was calculated and denoted as $\Delta \mathit{\text{sA}}{A}_i(i=1,2,3)$. The stress levels of the two groups are depicted in Figure 3.

Figure 3. ΔsAA levels. (A)Low-HA; (B)High-HA.

The ΔsAA levels of participants in the Low-HA group are lower than those in the High-HA group before stress, after stress, and 15 minutes after stress. Both groups exhibit substantial individual differences in ΔsAA₂ levels, with the High-HA group showing even greater individual variability.

3.2.3. Correlation analysis of ΔsAA

To gain a better understanding of the relationship between personality traits and ΔsAA activity levels, Spearman correlation analysis was conducted to examine the correlation between personality traits and ΔsAA activity levels, as depicted in Figure 4.

Figure 4. The correlation analysis results between ΔsAA and TPQ. (A)NS and ΔsAA₁; (B) NS and ΔsAA₂; (C) NS and ΔsAA₃; (D)HA and ΔsAA₁; (E) HA and ΔsAA₂; (F) HA and ΔsAA₃; (G) RD and ΔsAA₁; (H) RD and ΔsAA₂; (I) RD and ΔsAA₃.

The NS dimension of TPQ is moderately correlated with ΔsAA₁ and ΔsAA₂ and has a low correlation with ΔsAA₂. The HA dimension is moderately correlated with ΔsAA₁ and ΔsAA₃ and has a low correlation with ΔsAA_2. The RD dimension shows a weaker correlation with ΔsAA₁, ΔsAA₂, and ΔsAA₃. Due to significant individual differences in sAA₃ for both groups of participants, the correlations between NS, HA, RD, and ΔsAA₂ are relatively weak.

3.3. HRV

3.3.1. Variation trend of HRV index in different populations

To further investigate the impact of VR scenes on the stress response characteristics of individuals with different personality traits, HRV analysis was conducted to examine the stress reactivity levels of participants. The HRV indicators for both groups of participants 25 min before stress, before stress, after stress and 15 mins after stress are denoted as ${\mathit{\text{HFnu}}}_i$ and ${\mathit{\text{LFnu}}}_i(i=1,2,3,4)$, the results are presented in Figure 5.

Figure 5. HRV of the two groups; (A) HFnu of Low-HA; (B) HFnu of High-HA; (C) LFnu of Low-HA; (D) LFnu of High-HA. The “*” indicates a significant difference between the current data and the HRV indicators 15 minutes before the stress.

According to the Shapiro-Wilk test, HFnu₂, HFnu₃, and HFnu₄ do not follow a normal distribution (p < 0.05*), while LFnu₂, LFnu₃, and LFnu₄ do follow a normal distribution (p > 0.05). Based on the Wilcoxon rank-sum test, significant differences were found in HFnu₂, HFnu₃, and HFnu₄ between the two groups (p < 0.05*). By using RM-ANOVA, significant differences were found in LFnu₂, LFnu₃, and LFnu₄ between the two groups (p < 0.001***).

For participants in the Low-HA group, there is a slight increase in HFnu and LFnu before stress. After stress, HFnu and LFnu return to normal levels. 15 minutes after stress ends, HFnu is at a lower level, and LFnu shows a slight increase. In the High-HA group, participants exhibit a significant increase in HFnu and LFnu before stress. After stress and 15 minutes after stress ends, HFnu and LFnu remain at higher levels. Overall, participants in the High-HA group have higher HRV indicators in the last three measurements, and the individual differences in the High-HA group are greater. Both groups have LFnu values higher than HFnu, and the individual differences in LFnu are greater than in HFnu.

3.3.2. Correlation analysis of HRV

To enhance our comprehension of the relationship between personality traits and HRV indicators HFnu and LFnu, a comparative analysis of the various groups was conducted using Pearson correlation analysis, as depicted in Figures 6 and 7.

Figure 6. The correlation analysis results between HFnu and TPQ. (A) NS and HFnu₂; (B) NS and HFnu₃; (C) NS and HFnu₄; (D) HA and HFnu₂; (E) HA and HFnu₃; (F) HA and HFnu₄; (G) RD and HFnu₂; (H) RD and HFnu₃; (I) RD and HFnu₄.

Figure 7. The correlation analysis results between LFnu and TPQ. (A) NS and LFnu₂; (B) NS and LFnu₃; (C) NS and LFnu₄; (D) HA and LFnu₂; (E) HA and LFnu₃; (F) HA and LFnu₄; (G) RD and LFnu₂; (H) RD and LFnu₃; (I) RD and LFnu₄.

The NS dimension of TPQ is weakly correlated with HFnu₂, moderately correlated with HFnu₃ and HFnu₄, and weakly correlated with LFnu₂, LFnu₃, and LFnu₄. The HA dimension is moderately correlated with HFnu₂ and HFnu₃, highly correlated with HFnu₄, and moderately correlated with LFnu₂, LFnu₃, and LFnu₄. The RD dimension is weakly correlated with HFnu₂ and HFnu₄, has poor correlation with HFnu₃, and poor correlation with LFnu₂, LFnu₃, and LFnu₄. Overall, the correlations between NS and HA dimensions with HRV indicators are higher than those with the RD dimension. The correlations between TPQ and HFnu are higher than those with LFnu.

4. Discussion

With the advancements in neuroscience technology, researchers have discovered that psychological stress responses have a biological basis (Liu et al., 2019). The investigation of stable biological indicators associated with these responses is currently a prominent research focus. Immersive stress-inducing methods have been identified as effective in intensifying stress effects on participants, offering a viable approach for stress-related scientific research. This approach is particularly desirable for investigating the quantitative relationship of biomarkers.

In this study, VRPS was used as a stressor to extensively investigate the stress response levels among participants with different harm avoidance personality traits. Previous research has indicated that the scores of TPQ are associated with dopamine, serotonin, and norepinephrine (Rademaker et al., 2009). These neurotransmitters influence the activity of the sympathetic nervous system and the secretion of salivary alpha-amylase (Drummond & Clark, 2023). In this study, the TPQ was used as an assessment tool for evaluating the participants’ personality profiles. Participants were grouped based on their scores of HA dimension. We collected saliva samples from the participants before and after the stress induction, while simultaneously monitoring their heart rate using electrocardiography. sAA activity values, as well as the HRV indicators HFnu and LFnu, were employed as physiological and biochemical markers of the stress response.

Individuals with high harm avoidance are likely to show strong reactions and attentional biases toward stressors (Cloninger, 1987). Some evidence proved that high harm avoidance is related with increased stress, cortisol level, maladaptive coping style, and maladaptive emotional regulation as well as depression and anxiety (Chae et al., 2019; Tyrka et al., 2008; van Berkel, 2009). Those with high harm avoidance might inevitably experience more stress in their daily life, and those with low harm avoidance may exhibit contrasting characteristics. The results of TPQ indicate a bimodal distribution of HA scores among all participants. Consequently, participants were divided into Low-HA and High-HA groups based on their HA scores. This categorization helps us comprehend the differences in stress responses between participants in the Low-HA and High-HA groups. Additionally, we observed significant variations in the NS dimension between the two groups. The average scores on the NS and RD dimensions of participants in the Low-HA group were higher than those in the High-HA group. This indicates that individuals with high harm avoidance personality traits tend to exhibit a tendency to seek novelty. This finding is consistent with our previous research results (Ma et al., 2018).

During stress, the sympathetic and parasympathetic nervous systems may experience activation or inhibition to varying degrees, indicating a dynamic interplay between these two branches of the autonomic nervous system (Bian et al., 2022). It is generally believed that activation of the sympathetic nervous system leads to an increase in salivary alpha-amylase (sAA) concentration, while activation of the parasympathetic nervous system leads to an increase in saliva secretion (Nater & Rohleder, 2009). In this study, participants were subjected to strong stimuli, the sympathetic and parasympathetic nervous systems may be in a state of certain dysregulation. The results indicate a significant increase in sAA activity for all participants, suggesting that VRPS has a good stress-inducing effect. There is a correlation between TPQ scores and sAA activity. Besides. Compared to previous research, the increase in sAA activity induced by VRPS is more remarkable (Ma et al., 2018). Therefore, the immersive stress paradigm has advantages such as good stress-inducing effects, controllable intensity, and safe operation (Halbeisen et al., 2023). The sAA activity of Low-HA group is significantly lower than the High-HA group during and after stress, and the recovery period of sAA activity for the Low-HA group participants is shorter than that of the High-HA group. These results suggest that individuals with a higher tendency for harm avoidance may exhibit more pronounced physiological reactions when facing stress, and the time to recover from a stressed state to a normal state may be longer. This might be related to stress vulnerability as well as adverse psychological outcomes such as depression and anxiety (Krebs et al., 1998,Chen et al., 2015). Therefore, sAA activity levels may, to some extent, reflect TPQ personality traits.

The current research regarding the physiological significance of HRV indicators LF and HF remains contentious. However, it seems that the HF indicator is more sensitive to vagal nerve activity than sympathetic nervous activity (Thomas et al., 2019). There are some debates regarding the LF component, with some studies suggesting that LF reflects both sympathetic and parasympathetic nervous activity (Geng et al., 2022). The HRV results in this study indicate a significant increase in HRV indicators (HFnu, LFnu) for all participants, reiterating the effect of the immersive stress paradigm. The scores of TPQ are correlated with HRV indicators HFnu and LFnu. The levels of HFnu and LFnu in the Low-HA group are lower than those in the High-HA group, and the recovery period of HRV indicators in the High-HA group is significantly longer than that in the Low-HA group, consistent with the sAA results. Therefore, TPQ personality traits can be partially clarified through the analysis of HRV indicators.

In summary, the results of this study indicate that the stress paradigm we established (VRPS) has a good stress-inducing effect, and both sAA activity and HRV indicators can reflect TPQ personality traits. Participants in the Low-HA group show less sympathetic and parasympathetic nervous activity compared to the High-HA group. The sympathetic system in the Low-HA group is relatively more stable, allowing for better coordination of physiological activities. In contrast, participants in the High-HA group show relatively poorer stability in the sympathetic system, which is associated with sympathetic nervous system activity surpassing parasympathetic activity when exposed to stimuli (Kao et al., 2016).

The innovation of this study lies in the following aspects: Firstly, VRPS elicits a strong stress effect, reducing potential confounding variables (Dedovic et al., 2005; Kirschbaum et al., 1993; Schwabe et al., 2008). Secondly, compared to traditional physical stress paradigms such as the cold pressor test, VRPS are easier to carry out (Hamburger et al., 2022). Thirdly, the affordability of VR devices enables researchers to conduct studies with larger sample sizes, which can enhance the statistical power and reliability of study findings.

This study has several limitations: First, we should consider additional factors, such as sleep patterns and smoking status, although this may introduce more confounding variables (Sa et al., 2020). Second, we should collect repeated measurements to enhance the accuracy of the results, despite the increased difficulty of the task. Third, despite annual health checkups for each participating students, we were unable to fully control for potential psychological issues among individuals. Due to the self-discipline and cooperative behavior of the students participating in the study, the potential errors during the saliva sampling process have been mitigated.

Therefore, our results partially indicate that individuals with different personality traits exhibit differences in sAA and HRV responses during stress tasks, which contributes to the verification of the influence of personality on stress reactivity and the prediction of stress dysregulation (Gottesman & Gould, 2003). The assessment method of evaluating stress response levels in different populations using VRPS can be employed in psychological stress research. VRPS have strong stimulus effects and serve as effective stressors, providing a viable research tool for stress science.

5. Conclusion

VRPS presents differential stress responses among individuals with different harm avoidance personality traits. The assessment of stress response levels through VRPS holds significant potential in the context of personality trait evaluations, particularly in distinctions between individuals exhibiting low and high harm avoidance tendencies. VRPS is a desirable tool for exploring the impact of personality traits on stress levels.

Acknowledgements

The authors thank all the research assistants and study participants for their support. Lei Ma and Zhaoxin Wang contributed equally to this work and should be considered co-first author.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Additional information

Funding

This work was supported by Humanity and Social Science Youth foundation of Ministry of Education of China (17YJC890022) and Jiangsu Province Industry University Research Cooperation Project for Lei Ma.

Notes on contributors

Lei Ma

Lei Ma, doctor, his research interests include stress psychology and medical signal processing.

Zhaoxin Wang

Zhaoxin Wang, graduate student, his research interests include virtual reality and stress psychology.

Wenwen Yang

Wenwen Yang, doctor, his research interests include stress psychology and behavioral science.