Can You Smell a Forest in Unscented VR?

If you watched a forest in VR and were told that the device also reproduced smells, would you smell the forest even without fragrance reaching your nose? In 2025, Anton Berg and Pentti Henttonen investigated this question using footage of the Amazon rainforest and a device that did not actually release fragrance.

People told that odors would also be reproduced were more likely to report olfactory experiences than people not given that explanation. The study focused on how short-term expectations about what one is about to perceive, alongside molecules entering the nose, relate to odor experiences and their reporting.

Creating odor expectations with a convincing-looking device

Thirty adults, including university students and staff, participated; 21 were women. The researchers assigned 16 to an expectation-manipulation group and 14 to a control group. Allocation is described as pseudorandom in the paper and cannot be treated as equivalent to a strictly randomized trial. The experimental plan and main hypotheses were preregistered.

The expectation group was told that odors had been modeled alongside the images and sounds and that vapor would be released into the room. They were also shown an apparatus resembling an odor-simulation device, combining a water-vapor diffuser with cables and a small computer. The control group received neither this explanation nor the demonstration.

Both groups viewed 360-degree footage of the Amazon rainforest through a VR headset. A one-minute video was shown twice, for two minutes in total. No forest fragrance was released from the device. However, because the experiment took place in an ordinary room, it did not chemically establish the complete absence of odorant molecules in the air.

What did the researchers ask about the olfactory experience?

After viewing the footage, participants rated how realistic and convincing their olfactory experience had been on a scale of 1–10. They also rated the images, sounds, bodily sensations, and overall VR experience. This was neither a scale of odor intensity itself nor a measure of accuracy in identifying actual fragrances.

Participants also freely described the odors they had perceived and moments when they had felt them particularly strongly. Researchers classified these answers into three categories: certain they had not experienced odors, uncertain whether they had, and certain they had.

Questionnaires also assessed tendencies to become absorbed in imagination, immersion in VR, the importance attached to smell, and socially desirable responding. These measures helped examine whether such factors alone could explain differences in odor reports.

Ten of 16 people in the expectation group clearly reported odors

In the classification of free-text responses, 10 of the 16 expectation-group participants were certain they had experienced odors. Three were uncertain and three reported none. Among the 14 controls, two were certain they had experienced odors, one was uncertain, and 11 reported none.

This distribution differed statistically between groups. Whether uncertain responses were counted as experiencing odors or not, the group given the explanation and device demonstration remained more likely to report them. Still, not everyone given the explanation experienced an odor, and not everyone without it reported none.

Mean olfactory-realism ratings were 3.56 in the expectation group and 1.86 in the control group. The difference was significant on the 1–10 scale, but even the expectation group’s average was far from the maximum. Ratings of images, sounds, bodily sensations, and the overall experience did not differ significantly between groups.

The association between expectation manipulation and olfactory ratings remained in analyses including questionnaire scores and previous VR experience. The authors emphasize that the findings were not simply due to people who rated the entire VR experience highly giving high scores to everything.

How expectations may contribute to perception

The authors drew on predictive processing: the idea that the brain does not simply receive external information but predicts what should be present based on experience and context. Experience is thought to emerge from the combination of those predictions and incoming information.

Here, the forest footage, explanation that odors would be released, and appearance of the device supplied grounds for expecting forest smells. The authors interpreted the change in reported olfactory experiences after even a brief introduction as supporting a framework in which top-down predictions also contribute to smell. They propose VR as a tool for investigating this process because it can vary expectations while providing a shared scene.

Did perception change, or did the way people answered change?

This study cannot fully separate changes in perception itself from changes in responses prompted by an understanding that the experiment expected reports of odors. Including socially desirable responding in the analysis does not completely resolve this problem. The authors acknowledge that the design could not distinguish having an experience from reporting one.

Moreover, the researchers who classified free-text responses knew which group participants belonged to, allowing expectations to influence classification. There was also no measured baseline of participants’ actual olfactory ability. This was a small experiment with 30 people and only one combination of footage and expectation manipulation.

The authors call for future classification by independent raters unaware of group assignment, behavioral and physiological measurements, and comparisons with conditions presenting real odors. They discuss possible applications such as rehabilitation for olfactory impairment, but this was not a patient treatment trial and provides no evidence of restored smell.

Summary

Even without fragrance release, adding an explanation that odors would be delivered and demonstrating a device increased reports of smelling the forest in VR. This change appeared in olfactory ratings, rather than as a uniform increase in ratings of the whole audiovisual experience.

The findings suggest that understanding odor experiences requires examining not only external stimulation but also what people expect to perceive. Manipulating expectations in VR offers an experimental entry point. Distinguishing changes in perception from changes in reporting is the next important step.

References

Berg, A., & Henttonen, P. (2025). Believing in simulated virtual scents produces reported olfactory sensations. Scientific Reports, 15, 41583. DOI: 10.1038/s41598-025-25418-1 / Full original article.

Original article: © The Author(s) 2025. CC BY-NC-ND 4.0.