Social pressure, including both verbal and nonverbal persuasion, is known to create a high-risk situation that leads to relapse in alcohol dependence.¹ Cognitive-behavioral coping skill training (e.g., refusing drink offers and assertive communication skills) has been suggested to offer the best protection against social pressure-induced relapse.² Therefore, improvement of strategies for coping with drink offers is critical in reducing the risk of relapse in patients with alcohol dependence.³

While various efforts have been made to improve the efficacy of coping skill training, Monti et al.⁴ observed that drinkers who were exposed to alcohol cues performed poorly in drink refusal role-play situations and insisted that alcohol cues can disrupt an individual's ability to use coping skills. According to the social learning theory, alcohol cues can trigger cognitive and neurochemical reactions that may disrupt the individual's ability to use coping skills as well as the belief that one can effectively use those skills in the situation.⁵ There is now general agreement that coping skill training should be practiced in the presence of alcohol cues in order to increase both the effecttiveness and sense of self-efficacy in using coping skills when confronted with a similar real-life situation.^6,7 However, there is great individual variation in the contexts associated with drinking, and social pressure is a difficult component to construct in treatment settings.

We developed a virtual reality (VR) system in order to assess the interaction between alcohol cues and social pressure in the induction of alcohol craving.⁸ We took advantage of virtual reality techniques that provide immersion environments and enable social interactions with virtual avatars.^9,10 In our previous study,⁸ our VR system proved to induce craving in social drinkers through social pressure from virtual avatars. However, the presentation of an alcohol cue did not affect the social pressure-induced craving. In the present study, we applied our VR system to patients with alcohol dependence and hypothesized that patients with alcohol dependence would demonstrate different patterns of social pressure-induced craving according to the presence of alcohol cues.

In the VR environment without alcohol cues (virtual street), alcohol cravings were significantly induced by social pressure in both groups (control group: t=-3.565, df=13, p=0.003; alcoholic group: t=-2.664, df=13, p=0.019). In the VR environment with alcohol cues (virtual pub), alcohol craving were significantly induced by social pressure in the control group (t=-2.602, df=13, p=0.022). Meanwhile, the alcoholic group demonstrated a distinct pattern of craving. The craving was strongly induced upon exposure to the virtual pub, even before social pressure was introduced. Interestingly, social pressure did not lead to further increases in the level of alcohol craving in the alcoholic group (t=0.339, df=13, p=0.740) (Figure 2).

There was a significant group×alcohol cue×social pressure interaction effect for craving level (F=8.143, df=1, p=0.008). The difference in the induction of craving in the VR environment with alcohol cues should have contributed to the significant interaction (Figure 2).

This study is the first to use the VR technique to assess the interaction between alcohol cues and social pressure in alcohol craving. We observed that the presence of alcohol cues induced an extremely high level of craving in the alcoholic group, which might lead to relapse in situations with social pressure.

Our findings have implications for combining cue-exposure and coping skill training in the treatment of alcohol dependence. Treatment approaches that combine coping skill training and cue exposure therapy have received growing attention.⁶ However, the high-risk situations associated with drinking vary among patients, and it is quite difficult to provide various contexts in which to practice coping skills in a treatment setting. More critically, learning during cue exposure therapy does not generalize across contexts if training takes place in only one context.^12,13 Consistent with the findings of previous studies,^9,14 our study findings suggest that the methodological usefulness of virtual reality, which might play significant roles in increasing the efficacy of cognitive behavioral treatments by presenting various emotional and social contexts that mimic real-life situations.

A difference between previous VR studies and the present study is the block design of the present study. Lee et al.¹⁵ used only alcohol cues to induce alcohol craving. The VR system developed by Bordnick et al.⁹ is composed of a block that contains social interaction situations, but there was no control block with which to compare the results. Thus, Bordnick et al. could not analyze the effect of social pressure on alcohol craving. In contrast, a social pressure situation was used to induce alcohol craving in this study, and our two-by-two experimental model was focused on analyzing the interaction between social pressure and alcohol cues and their effects on alcohol craving.

We hypothesized that the pattern of social pressure-induced craving would differ between alcoholic patients and social drinkers according to the presence of alcohol cues. Although the level of craving was significantly increased in the control group by social pressure from virtual avatars, it is noteworthy that, in the alcoholic group, social pressure did not lead to additional increases in the level of craving induced by exposure to a cue-laden environment. There are several possible explanations for these findings. First, ceiling effects are known to be especially relevant to craving assessment.¹⁶ Substantial evidence demonstrates that exposure to alcohol cues increases the level of alcohol craving in abstinent alcoholics.^17,18 Alcoholic patients who rated their craving level near the endpoint of the scale when first exposed to the cue-laden VR environment could have failed to report the actual increase in the level of craving caused by social pressure. In addition, we cannot rule out the possibility that the strong cognitive and neurochemical reactions to alcohol cues should have interfered with the ability to accurately assess one's internal state.⁵ Because there was no difference in the level of craving between the two groups in the VR environment without alcohol cues, our findings indicate that patients with alcohol dependence are vulnerable to alcohol-cue laden environments. Limitations of the visual analogue scale should be considered. Fox et al.¹⁹ reported that alcohol cue-induced craving and stress-induced craving were associated with dissociable psychobiological states. It is plausible to assume that social pressure-induced craving and cue-induced craving are related to different components of craving, and therefore a single-item scale might have been insufficient to assess the synergistic interaction between alcohol cue and social pressure. Various measures (self-report, physiological, and behavioral) are required in order to track the different components of cue-reactivity.²⁰

In conclusion, our findings imply that the presence of alcohol cue-induced craving, regardless of social pressure, should interfere with the ability to use coping skill against social pressure in real-life situations. The role of virtual reality techniques in coping skill training should be expanded by providing various emotional and social contexts that mimic real-life situations.