Can Tension Make Your Hands Smell Like Stir-Fried Leeks?

When questions from someone you have just met make you nervous, does the odor of the gas emitted from your hands also change? In a 2022 study, experts detected an odor they described as resembling stir-fried leeks in skin gas collected during interviews. Analysis identified allyl mercaptan and dimethyl trisulfide as sulfur compounds associated with this characteristic.

The researchers not only looked for odor changes accompanying tension, but also investigated whether smelling a model odor made from the identified components changed the recipient’s mood. It is important to read these as two separate experiments.

Comparing skin gas during an interview and on a day spent reading magazines

The main trial involved 40 healthy Japanese women aged 35–44. To avoid influences such as smoking, perfume, and strongly scented foods, participants followed restrictions and washed their hands with unscented soap before their nondominant hand was enclosed in a bag to collect the gas.

On the first day, an unfamiliar interviewer questioned them for 20 minutes. On another day, they relaxed in a chair while reading magazines. The order was fixed, with the interview first and reading afterward, so the trial compared two situations within the same people.

Electrocardiograms covering the periods before and after the interview and salivary cortisol were also measured. To account for cortisol’s daily variation, baseline samples collected while resting at home were taken at the same time of day on a different day.

Physiological responses associated with tension, and a leek-like odor

During the interview, LF/HF, an index derived from heart-rate variability, was higher than before and afterward. Salivary cortisol after the interview was also higher than at the home baseline and after reading. These findings support the interview having served as a tension-inducing condition. LF/HF is a heart-rate-variability index, not a direct recording of sympathetic nerve activity.

Four experts who were not told the collection conditions assessed the odors. A characteristic odor reminiscent of stir-fried leeks was consistently recorded in interview samples. Its intensity was positively associated with the rate of increase in LF/HF during the interview.

Additional small-scale observations reported a similar odor during tension in nine people: four Japanese men, three Chinese men and women, and two French men. In a separate exercise comparison, this characteristic odor was not detected when cycling increased heart rate. This suggests that increased heart rate alone may not explain it, but further comparisons are needed before generalizing to broader populations or everyday situations.

Tracing small amounts of sulfur compounds with noses and instruments

To identify the components, the researchers separated substances in the gas by gas chromatography and sniffed the output. This linked the instrument’s chemical information to the points at which an odor was actually perceived.

The first compound was dimethyl trisulfide (DMTS). It appeared at the same position as a reference standard, and its mass-spectral features and odor matched. The second was allyl mercaptan (AM), with a stronger sulfurous, onion-family odor. It was identified by matching its separation position, mass spectrum, and odor to a reference standard.

According to the authors, AM alone did not adequately reproduce the collected gas’s characteristic odor; adding DMTS brought it closer to the observed stir-fried-leek odor. Identifying these two components made it possible to create a model odor approximating that character and investigate its effects separately.

Tension, confusion, and fatigue scores rose after smelling the model odor

The psychological-effects trial involved 33 Japanese women aged 35–44. They smelled a model odor made by mixing AM and DMTS, not skin gas collected directly from other people.

First, the researchers determined the concentration at which each person began to detect the odor. A sample adjusted to that individual’s threshold was then placed beneath the nose and smelled for two minutes. The POMS 2 mood questionnaire administered before and afterward showed increases on three scales: tension–anxiety, confusion, and fatigue–inertia.

This trial had no control condition in which participants smelled an odorless sample or another odor using the same procedure. The findings can therefore be reported as changes before and after smelling the model odor, but they cannot separate effects specific to the two compounds from unpleasantness, expectations, or repeated questionnaire administration. The authors also note that earlier answers might influence later ones.

Identifying components and explaining emotional transmission are different stages

The achievement here was recording a characteristic odor in skin gas collected during a tension-inducing situation and identifying specific molecules as candidate components. The relationship with tension warrants testing under different conditions, given the fixed interview–reading order and the limited participant population.

The mechanism by which psychological tension leads to the release of these components also remains unclear. A before-and-after comparison with a model odor alone cannot establish that stress spreads through odor during actual human interactions. Because the research addresses both skin gas and mood, its findings and the next questions to test should be distinguished.

Summary

Skin gas collected from the hand during interviews was reported to have a characteristic stir-fried-leek odor different from that during reading and rest, and AM and DMTS were identified. Odor intensity was also associated with the rate of increase in a heart-rate-variability index during the interview.

In a separate trial using a model odor made from the two components, tension, confusion, and fatigue scores were higher afterward. However, without a control condition, the study does not establish emotional transmission between people. It provides a concrete starting point for investigating links between psychological states and molecules emitted through the skin.

The paper reports no specific funding for this work and declares no competing interests. It also lists Shiseido, Takasago International Corporation, and Takata Koryo among the authors’ affiliations at the time of the research.

Reference

Katsuyama M, et al. (2022). How emotional changes affect skin odor and its impact on others. PLOS ONE, 17(6), e0270457. Original paper and open full text