How you breathe is like a fingerprint that can identify you

Summary

Every breath you take ... could add to a breathing pattern that is unique to you, a study finds.Credit: Anusak Laowilas/NurPhoto via GettyLike the swirls in fingerprints, a person’s breathing pattern might be unique to them — offering a way not only to identify individuals, but also to identify some of their physical and mental traits.A team of researchers measured the breathing of 97 healthy people for 24 hours, and found that they could identify participants with relatively high accuracy from their breathing pattern alone. What’s more, they found that these patterns can be correlated with body-mass index (BMI) and signs of depression and anxiety.“In a way, we’re reading the mind through the nose,” says co-author Noam Sobel, a neurobiologist at the Weizmann Institute of Science in Rehovot, Israel. “This could be a very powerful diagnostic tool.” The team published its study today in Current Biology1.Taking a breathBreathing is deeply connected to the brain. Every inhalation and exhalation is coordinated to supply the oxygen needed for the brain to manage the body’s systems. Sobel and his team wondered: if every brain functions differently, shouldn’t every person’s breathing be unique, too?To test this, the researchers developed a custom, wearable device that records airflow through each of a person’s nostrils. Mounted on the back of the neck, the device, which has tubes fitted under the nose, tracks people’s breathing during their everyday routines, both while they are awake and while they are asleep.Researchers measured study participants’ breathing patterns over 24 hours, using a custom device that sits on the back of the neck.Credit: Soroka et al., Current BiologyTo characterize a person’s breath pattern, the team extracted 24 parameters from the airflow data, including duration of inhalation and exhalation and airflow asymmetry between nostrils. They separated the periods when participants were awake and asleep, and trained a machine-learning algorithm with the...