Human Odor: Sniffing Out Identity and DeceptionApril 2012
Topics: Criminology, Prevent Terrorism
Security efforts at airports and other protected areas are designed to determine if people are using a false identity and/or are hiding the intent to perform a malicious act. Studies seem to suggest that an individual's odor could both uniquely identify him and give an indication as to whether he is engaged in an act of deception.
The Nose Knows
The famed ability of dogs to identify people simply by their scent is remarkable. Studies have shown that some trained dogs can distinguish readily between any two individuals, with the exception of identical twins on the same diet. In several European countries, canine identification of objects handled by suspects can be considered as evidence in criminal cases.
While we have some understanding of how a dog's nose identifies a human odor, we don't know exactly what makes a person's odor unique or even if people have a unique, unchanging odor. However, a growing body of research supports the idea that odor is unique to the individual and relatively stable over time. This suggests that variations in human odor may be due to genetic differences. And, indeed, it has been shown that differences in the genes that code for the immune system can be traced to odor differences. If odor is a physiologically unique characteristic, it could serve as a useful and reliable biometric identifier.
Research on human odor stretches beyond individual identity. Studies show that human odor contains information related to an individual's physical health, mental health, gender, ethnic group, and stress level. This last is of particular interest because of the likelihood that individuals who are committing deception experience heightened stress. So while the stress level of many people going through a security checkpoint will rise, the individual hiding an intention to do harm may experience measurably higher stress than those with no such bad intentions. There is not yet a truly reliable physiological indicator of deception, so even if odor proves to be only moderately accurate in predicting deception, it might still be useful in conjunction with other indicators.
National security efforts would certainly benefit from a simple odor-based test to determine whether people are who they claim to be and if they are hiding malicious intent. MITRE recently conducted a study, funded by the Department of Homeland Security's Science & Technology Directorate, to look for evidence for our dual hypotheses: that human odor may serve as a unique biometric identifier, and that an individual's odor may indicate an act of deception or intent to deceive.
Sixty male subjects participated in an experiment in which half of them, randomly selected, were required to behave deceptively. The researchers collected underarm odor samples on cotton pads from each individual before and after the experiment. After analyzing the odor samples, we found promising results both for human identification and for detecting deception. We have had some success in matching different odor samples taken from the same individual by comparing their chemical signatures. We also found evidence that there are deception indicators in human odor, as we were also able to classify deceptive versus non-deceptive individuals with approximately 80-percent accuracy.
In addition, we identified two compounds that turned up in different amounts in the deceptive and non-deceptive samples. As it happened, the compounds expressed more strongly in nondeceptive individuals, so these particular compounds are not useful for deception detection. (It's not practical to build a sensor to detect the absence of something.) But since we found compounds expressed more strongly in non-deceptive people, it's reasonable to guess that we might eventually also identify compounds expressed more strongly during the act of deception.
Some roadblocks exist along the route to using odor as a reliable biometric or security warning. A human odor sample contains hundreds of volatile chemicals. With current technology, it's difficult to separate them in order to obtain a clean odor profile. Gas chromatography has been widely used for separating chemicals and mass spectrometry for identifying each chemical. However, separating several hundreds of trace volatile chemicals may be beyond the capability limit of gas chromatography. And since mass spectrometry identifies only pure chemicals, the identification may fail if the chemical separation by gas chromatography is not complete.
We also don't have detailed enough knowledge about the odorants that would serve as the core information in an individual's odor signature. Nor do we know the whole set of odorants associated with deception. These uncertainties, when compounded with the difficulty of chemical separation, pose a great challenge to research in this area.
Future research will surely develop more refined tools for odor analysis and uncover additional information on the compounds that could serve as a "bar code" for human odor or signal deceptive intent. If so, the poor bloodhound may find itself collecting unemployment.
—by Brigitte Rolfe and Sichu Li