When people watch the same video, changes in their brain waves over time tend to resemble one another to some extent. In a 2020 study, that similarity was greater in the presence of eucalyptol than under an odorless condition. Limonene, also considered an “energizing” scent, did not produce the same result.
The study focused on a methodological question: could the effects of scents be assessed through the brain’s responses while watching videos, rather than only through impressions such as “stimulating” or “calming”? Across three experiments, the researchers explored the potential of this measure for comparing fragrance materials, along with questions that remain unresolved.
Measuring similarity in EEG responses to videos
The measure used was inter-subject correlation, abbreviated ISC. By combining EEG recordings from multiple electrodes on the scalp, the researchers calculated how closely the changes in one person’s signal over time resembled those of other people watching the same video.
The idea is to capture how much people’s brains respond in common to the same moments in a story. Previous research had linked ISC to attention and engagement with videos. The authors therefore tested whether scents could change this measure.
The three experiments involved separate groups of 19, 14, and 20 young adults. A device delivered scent for 6 seconds in synchrony with inhalation, followed by a 24-second break, in repeated cycles. EEG was recorded at 32 electrode locations, and the same participants experienced both scented conditions and an odorless control containing only the solvent. Scent intensities were adjusted beforehand to make them similar, and participants’ intensity ratings showed no significant differences between fragrance materials.
Eucalyptol increased ISC in two experiments
In the first experiment, 19 participants watched six short animated narratives under eucalyptol, linalool, and odorless conditions. Drawing on earlier reports, the authors selected eucalyptol as a candidate stimulating scent and linalool as a candidate calming scent.
ISC was higher with eucalyptol than under the odorless condition and with linalool. The difference from the odorless condition was p = .0019, and the difference from linalool was p = .007. This suggests that, while participants smelled eucalyptol, the time course of their brain responses to the videos became more aligned with that of other viewers.
The next experiment, involving 14 participants, replaced linalool with limonene. ISC with eucalyptol was again higher than under the odorless condition, but the difference between limonene and the odorless condition was not significant, and ISC was higher with eucalyptol than with limonene. Participants classified the two scents as “energizing” or “calming” in similar proportions, so these broad impressions alone could not explain the changes in ISC.
In this experiment, the odorless condition was placed between the two fragrance conditions. In addition to combining approximately 18 minutes of video data, the researchers tried analyzing individual short videos of about 3 minutes each. Although ISC differed substantially between videos, the interaction testing whether scent effects varied by video was not significant. However, the same ordering did not appear in every short segment, leaving variability to consider when using brief measurements.
Expanding to 10 scents did not establish a ranking
The third experiment examined 10 scents, including peppermint oil, rosemary oil, vanillin, rose-like phenylethyl alcohol, eugenol, and fragrance blends. Participants, experimenters, and analysts were not told the names or identities of the fragrance materials. Across two days, the same videos were compared with scent present and under an odorless condition.
The methods describe 10 pairs of videos, each pair totaling at least 6 minutes, assigned to avoid a bias in the combinations of scents and videos. Across all 10 scents, ISC was higher under scented than odorless conditions. However, a test comparing how much each scent changed ISC from its odorless baseline found no significant difference among the 10 scents (p = .73).
Individual comparisons reported increases for phenylethyl alcohol and eugenol, but no correction for testing multiple scents was described. These results are best read as identifying candidates for further study. The ordering of the means does not provide a basis for ranking the scents that “improve concentration the most.”
Although this experiment is described as involving 20 participants, the main t-test is reported with 18 degrees of freedom, and the text does not clearly explain how this relates to the participant count. A larger sample and clear reporting of the number included in each analysis are needed to distinguish fragrance effects in greater detail.
ISC provided information different from the immediate response to a scent
The researchers examined not only ISC but also EEG responses immediately after scent presentation. In the first experiment, no significant difference between scented and odorless conditions was found in responses within 1 second of onset. No clear differences were observed in oscillatory power, including alpha activity, although some of these observations were not formally tested statistically.
The scalp distributions of the signals contributing to ISC resembled those found in previous audiovisual studies. The authors considered the possibility that scents altered responses to sights and sounds, rather than simply strengthening a brief response to the odor itself. They also argued that stimulation of the trigeminal nerve, which contributes to sensations of nasal irritation, could not straightforwardly explain the results for different fragrance materials. However, the experiments did not directly separate these neural pathways.
The study also did not include behavioral attention tasks, memory tests, or measures of learning performance. An increase in ISC is consistent with a change in attention, but could also arise from physiological arousal or changes in sensory responses themselves. Clarifying these distinctions is the next step toward moving from a fragrance-assessment measure to applications in everyday learning or work.
Summary
Presenting eucalyptol increased inter-subject correlation in EEG during video viewing relative to an odorless condition, and this result was replicated in a separate group. Limonene did not show the same increase. A classification such as “energizing scent” therefore does not necessarily correspond directly to the brain changes being measured.
In the experiment testing 10 scents, ISC was higher overall when a scent was present, but differences among the scents could not be established. The findings offer evidence that scent can affect the brain’s responses to videos and suggest a way to assess those effects using ISC. They highlight the importance of carefully matching impressions of a fragrance to the functions actually measured.
The research was funded by International Flavors & Fragrances (IFF). Two authors employed by the company at the time contributed to the intensity-matching study, supplied materials, advised on delivery methods, and helped write the manuscript. Another author was employed by Neuromatters at the time. These relationships are disclosed in the paper.
Reference
DeGuzman, P., Jain, A., Tabert, M. H., & Parra, L. C. (2020). Olfaction Modulates Inter-Subject Correlation of Neural Responses. Frontiers in Neuroscience, 14, 702. Original research
