The Chemistry of Stress: What Dogs May Smell in Us

Humans tend to interpret stress through visible behavior: a tense face, a shaky voice or restless movement. Dogs inhabit a different sensory world. Research led by Queen's University Belfast suggests that acute psychological stress also alters human odor in a way trained dogs can distinguish with striking accuracy. [1]

In the 2022 study, researchers collected combined breath and sweat samples from participants at baseline and immediately after a mental-arithmetic task designed to induce stress. The human response was checked through self-report and physiological measures, reducing the chance that researchers were simply labeling a difficult task as stressful without evidence.

Thirty-six participants' samples were ultimately presented to four trained dogs. During the key discrimination trials, the animals chose among a stress sample, the same person's baseline sample and a blank. Across 720 trials, their combined accuracy was 93.75 percent, far above chance. [1]

The experiment does not establish that dogs identify an abstract concept called "stress." Rather, it provides evidence that acute stress produces changes in volatile compounds from breath and/or sweat that are detectable by the canine olfactory system. The sample also involved only four carefully trained dogs, so the study was a proof-of-principle rather than a claim about every household pet. [2]

A later University of Bristol study moved from detection to possible emotional consequences. Eighteen dogs were exposed to odors collected from unfamiliar people in stressed or relaxed states while completing a cognitive-bias task involving food bowls. Under some conditions, the stress odor made dogs less likely to approach an ambiguous location associated with a possible reward, a response researchers interpreted cautiously as more risk-averse or "pessimistic." [3][4]

That distinction is important. One study asks whether stress has a smell dogs can detect; the other asks whether that smell can influence canine behavior. Neither result proves that dogs read human minds, and both require replication and broader samples.

Even with those limits, the work complicates our idea of communication. Humans may believe an emotion is private if we control our expression, yet physiological changes can broadcast information through channels we barely perceive. For an animal whose nose is a primary source of knowledge, our bodies may be saying far more than our faces reveal.