PNAS: Chao Liu’s Research Group Uses VR Technology to Study the Psychological Mechanisms and Influencing Factors of Heroic Behavior

On April 16, 2024, the research team led by Liu Chao from the State Key Laboratory of Cognitive Neuroscience and Learning, as well as the IDG/McGovern Institute for Brain Science at Beijing Normal University, published a paper online in the journal PNAS titled “Observing Heroic Behavior and its Influencing Factors in Immersive Virtual Environments.” The study used immersive virtual reality (VR) technology to investigate the psychological mechanisms and influencing factors behind people’s heroic behaviors.

Heroic behaviors are widely regarded as the pinnacle of human morality and altruism. In essence, they refer to the voluntary actions people take to help others in need, at the risk of their own safety. Despite the importance of this research theme, progress has been limited due to ethical challenges in manipulating the risk level to elicit heroic behaviors in laboratory environments. Previous studies have largely relied on self-report scales and survey studies to research heroic behaviors, which may lead to discrepancies between reported attitudes and actual behaviors.

To address the inherent challenges in researching heroism, our study used immersive virtual reality technology to induce participants’ proactive heroic actions. Specifically, participants, wearing VR headsets, carried out a routine VR task (playing the role of an environmental volunteer, using a pole to clean garbage in a pond) but unexpectedly became witnesses to a sudden criminal event (a thief stealing and then fleeing). During the incident, a female victim and police officer were chasing the fleeing thief, and as the thief ran past, the participants had the option to proactively use their poles to stop the thief (Figure 1). Since such behavior might bear certain safety risks in real life, the study considered the participants’ proactive intervention as heroic behavior.

Figure 1. The processes and scenarios in Study 1 and Study 2

Through this experimental setup, we examined whether there were individual differences in the propensity to engage in this heroic behavior among 397 participants in VR across three studies (comprising six experimental conditions), and further explored whether situational factors (voluntariness, presence of authority, risk level) or personal factors (gender, impulsivity, empathy, social value orientation) could explain the individual differences in this heroic behavior.

Study 1 served as a preliminary study with the aim of verifying the feasibility of the VR experimental setup and exploring potential gender differences. During the experiment, the victim and the police officer would directly call out to participants to help intercept the thief. The results indicated that participants generally performed the interception behavior under this condition, and the proportion, response time, and heart rate variability were insensitive to gender (Figure 2a, d). Study 2 tested the influence of voluntariness (by comparison with Study 1) and the presence of authority (police officer) as well as their possible interaction with gender. Firstly, the experiment removed the requests for help from the victim and police officer, ensuring that the participants’ interception actions were voluntary. Secondly, the task scenarios in two conditions of Study 2 included either the presence or absence of a police officer, thereby manipulating the factor of “presence of authority”. The results showed that the proportion of participants performing interception in Study 2 was significantly lower than in Study 1, highlighting the impact of voluntariness on heroic behavior. We also used logistic regression analysis, taking the proportion of intercepting behavior as the dependent variable and gender, scenario (whether authority was present), and their interaction as independent variables. No significant main effects or interactions were found, suggesting that the presence of authority had little influence on heroism (Figure 2b).

Figure 2. Proportions and response times of the heroic behavior by males and females under different scenarios

Study 3 continued to explore the gender difference in heroic behavior in situations with higher perceived risk level and investigated its potential relationship with social conformity behaviors (Figure 3a), with three experimental conditions set up: “unrelated startled bystanders,” “daytime verbal threats from the thief,” and “nighttime verbal threats from the thief” (Figure 3b, c, d). The results showed that the proportion of participants performing interceptions in Study 3 was significantly lower than in Study 2, indicating that the higher risk level reduced the occurrence of heroic behavior. Likewise, using logistic regression analysis with the proportion of intercepting behaviors as the dependent variable and gender, scenario, and their interaction as independent variables, gender’s main effect was significant (Figure 2c), with males significantly more likely than females to volunteer to intervene in “catching the escapee.” Additionally, compared to females, males required shorter response times (Figure 2f) to perform the interception behavior and exhibited lower heart rate variability.

Figure 3. The processes and scenarios in Study 3

Additionally, we found that the capacity for pain empathy and gender significantly interact in deciding whether to engage in the heroic behavior. For men, stronger individual pain empathy capacity was associated with a lower probability of heroic behavior, while for women, the reverse was true; the stronger the individual’s pain empathy, the higher the likelihood of heroic behavior (Figure 4).

Figure 4. Significant gender difference in the relationship between individual’s pain empathy capacity and heroic behavior

Lastly, we also found a significant negative correlation between the tendency to engage in the heroic behavior and social conformity – participants who had exhibited heroic behavior were less willing to follow the crowd in action that violate the moral norm (e.g., jaywalking in a group) (Figure 5).

Figure 5. The relationship between heroic behavior and conformity in study 3

Finally, we also conducted pre-registered online survey studies in Study 4 and Study 5, exploring the effectiveness of the VR experimental condition setup and validating the results of the VR experiments by describing the experimental scenarios. These two online studies included samples from both China (Study 4: n = 528; Study 5: n = 71) and overseas (Study 4: n = 541; Study 5: n = 71). The results indicated that almost all the main findings of the VR experiments were replicated in the Chinese participant group.

The results enhance our understanding of the personal and situational factors that affect heroic behavior and demonstrate the vast potential of immersive VR technology in the field of social psychology.

The co-first authors of the paper are Liu Chao’s former master’s students Kelou Jin, Jie Wu, and current master’s student Ran Zhang, with Researcher Ruolei Gu from the Institute of Psychology, Chinese Academy of Sciences as the co-corresponding author. This work was supported by the Scientific and Technological Innovation 2030-Major Projects (2021ZD0200500), the National Natural Science Foundation of China (no. 32271092, 32130045, 32071083, 32020103008), and the Major Project of National Social Science Foundation (19ZDA363).

Publication details:
Jin K, Wu J, Zhang R, Zhang S, Wu X, Wu T, Gu R, Liu C. Observing heroic behavior and its influencing factors in immersive virtual environments. Proc Natl Acad Sci U S A. 2024 Apr 23;121(17):e2314590121.
doi: 10.1073/pnas.2314590121.

Lab website: Welcome to Social & Affective Neuroscience lab (bnu.edu.cn)