Exposure to an unpleasant odour increases the sense of Presence in virtual reality

Abstract

While olfactory cues affect the everyday human experience in the physical world, few studies have empirically examined the effect they could have on the human experience in virtual reality (VR). This project’s goal was to determine whether the exposure to olfactory stimuli would affect the senses of Presence (primary measure), Reality and Realism (exploratory measures) in VR. In a virtual kitchen devoid of obvious visual cues linking the visual scene to an odour, three groups of 20 randomly assigned participants (12 females and 8 males per group), unaware of the potential exposure to olfactory stimuli, were exposed to either ambient air, a pleasant odour, or an unpleasant odour. The results reveal that the unpleasant odour had a statistically significant effect on the sense of Presence (as measured by repeated brief measures of Presence and the Independent Television Commission Sense of Presence Inventory), but the pleasant one did not. The lower perceived intensity of the pleasant odour may have contributed to its lower detection rate which, in turn, may have contributed to the pleasant odour’s lack of effect on the sense of Presence. Neither of the olfactory stimuli had an effect on either the sense of Reality or the sense of Realism.

Appendix: Additional analyses

1.1 Characteristics of olfactory stimuli detected in virtuo

1.1.1 AMB group

For this group, although no olfactory stimulus was generated, the in virtuo detection rate of an olfactory stimulus in the virtual kitchen was 5 % (1 of 20). More specifically, this male participant reported that the odour was associated with a memory and rated the odour as follows: Pleasantness = 2.00, Intensity = 4.00, Familiarity = 3.00, and Concordance to the kitchen = 1.00.

1.1.2 PLE and UNP groups

Detection rates, rate of association with a memory, and presumed source of odours are shown in Table 2.

Table 2 Detection, association with memory, and source of odours detected in virtuo

1.1.3 Comparison between the characteristics of PLE and UNP odours

Due to the inequality of the size of the groups, and in order to compare the characteristics of the PLE and UNP odours, nonparametric comparisons were conducted. These confirmed that, as assessed by the PLE (n = 3) and UNP (n = 12) participants who detected the odour in virtuo, the UNP odours were statistically significantly more unpleasant (Mdn = 4.00) than the PLE odour (Mdn = 1.00; U = 0.00, z = −2.68, p = .01, d = 4.64, r = −.92). However, the differences between the PLE and UNP odours were not statistically different either in terms of Intensity (respectively, Mdns = 5.00 and 5.50; U = 7.50, z = −1.66, p = .10, d = 1.19, r = .51), Familiarity (respectively, Mdns = 4.00 and 2.50; U = 6.50, z = −1.71, p = .09, d = −1.11, r = −.49), or Concordance with the visual scene (respectively, Mdns = 3.00 and 2.00; U = 12.50, z = −.82, p = .41, d = −.58, r = −.58). Mean values and standard errors are shown in Fig. 8.

Fig. 8

Characteristics of olfactory stimuli detected in virtuo. These are based on the participants who reported to have detected an odour in the kitchen (for the pleasant odour, n = 3; for unpleasant odour, n = 12). Standard errors are represented in the figure by the error bars attached to each column

1.2 Characteristics of olfactory stimuli detected in vivo

1.2.1 PLE group

The post-experiment in vivo detection rate of the same olfactory stimulus these participants had been exposed to in the kitchen was 100 % (20 of 20). Fifteen of the group members (75 %) stated that the odour was associated with a memory.

1.2.2 UNP group

The post-experiment in vivo detection rate of the same olfactory stimulus these participants had been exposed to in the kitchen was 100 % (20 of 20). Eighteen of the group members (90 %) stated that the odour was associated with a memory.

1.2.3 Comparison between the characteristics of PLE and UNP odours

Due to the non-normal distribution of the characteristics of the odours, and in order to compare the characteristics of the PLE and UNP odours, nonparametric comparisons were conducted. These confirmed that, as assessed by the participants of the PLE (n = 20) and UNP (n = 20) groups when they were exposed to the odour in vivo (post-questionnaires), the UNP odour was statistically significantly less pleasant (Mdn = 0.00) than the PLE odour (Mdn = 4.00; U = 0.00, z = −5.54, p = .00, d = 4.42, r = .91), more intense than the PLE odour (respectively, Mdns = 6.00 and 5.00; U = 122.50, z = −2.20, p = .03, d = −.82, r = −.38), less familiar than the PLE odour (respectively, Mdns = 3.50 and 5.00; U = 113.00, z = −2.43, p = .02, d = .93, r = .42), and less concordant with the visual scene than the PLE odour (respectively, Mdns = 1.00 and 3.00; U = 121.00, z = −2.19, p = .03, d = .74, r = .35). Mean values and standard errors are shown in Fig. 9.

Fig. 9

Characteristics of olfactory stimuli detected in vivo. These are based on a 30-s exposure to the same stimulus the participants had been exposed to in virtuo. This time, however, participants had been made aware of the exposure to an olfactory stimulus (n = 20 for each of the groups). Standard errors are represented in the figure by the error bars attached to each column