Every person has a unique pattern of breathing through their nose, which is stable over time and acts like a biological signature. By tracking how individuals inhale and exhale over a 24-hour period, new research has demonstrated that these unique breathing patterns can identify people with near-perfect accuracy, while also predicting an individual’s levels of anxiety, depression, and weight. The study was published in the journal Current Biology.
Breathing frequently seems like an automatic and simple physical process. Many people only become aware of their breathing when they are out of breath or after intense exercise. However, the act of drawing air into and pushing it out of the body is controlled by an extensive and complex neural network.
This neural network operates primarily from the brainstem. It acts as a biological pacemaker that continuously adjusts human breathing to meet physiological needs. This system takes in vast amounts of sensory information from throughout the body and controls the rate and depth of inspiration and exhalation.
Because the human brain exhibits individual uniqueness in its wiring and activity, the researchers suspected that the biological output produced by these local brain networks might also reflect a high level of individuality. To pursue this idea, researchers at the Weizmann Institute of Science in Israel designed an experiment to precisely track how air moves through the nose over long periods of time.
The project’s lead researchers, Timna Soroka and Noam Sobel, decided to focus on the nose rather than the mouth. The nasal passages have a special connection to the brain and are filled with sensory nerves that send constant feedback about air flow. The brain actively controls this process, systematically alternating which nostril does most of the work of breathing.
To capture these long-term breathing patterns, the team developed a specialized wearable device. This small tracker was placed on the back of the volunteer’s neck and connected to a nasal cannula. A nasal cannula is a thin plastic tube with two small prongs that sit just inside your nostrils.
Unlike standard medical tests, which monitor breathing for just a few minutes and check lung capacity, this setup recorded breathing continuously throughout the day and night. The device contained highly sensitive pressure sensors that measured airflow in the left and right nostrils separately in real time. It recorded data six times per second, capturing small dynamic fluctuations in air movement.
The study involved about 100 healthy participants, mostly in their 20s. Each person wore the tracker 24 hours a day as they went about their daily lives, recording their basic activities and sleep schedule on a provided smartphone app.
The research team repeated the process in its entirety on some of the more than 40 participants. These people again wore recording devices for 24 hours. The period between the first and second recording sessions ranged from a few days to nearly two years.
When researchers fed raw breathing data into a computational model, they found that it could identify individuals with surprising accuracy. Based solely on breathing patterns during wakefulness, the system accurately identified a specific individual from a group with 96.8% accuracy.
The success rate of this identification process puts breathing patterns on roughly the same level as established biometric markers such as voice recognition. The data showed that human breathing is not just a general mammalian rhythm, but a characteristic of individual behavior.
This ability to recognize people based on air currents holds true even after a long time. Once the computer model learned a person’s breathing patterns on the first day of testing, it was able to successfully pick that person out of the crowd using data collected up to 23 months later.
To ensure that the computer was not only monitoring distinct patterns of physical movement, but also the actual act of breathing, the researchers also analyzed data from motion sensors embedded within the device. Although some discrimination was possible through body movements, it was significantly inferior to the accuracy achieved by analysis of nasal airflow.
The researchers evaluated dozens of different parameters within the respiratory data to validate these results. They grouped the data into metrics such as the amount of air inhaled, the length of the pause between breaths, and the asymmetry of airflow between the left and right nostrils.
There was no single characteristic that could stand alone to distinguish one person from another. For overall identification accuracy to be high, the computational model needed to examine approximately 20 to 100 different respiratory characteristics working together.
The researchers evaluated whether these respiratory signs reveal anything about a person’s physical condition beyond simple human identification. They dissected the raw data to look at physiological markers such as transitions from wakefulness to sleep.
The analyzed data showed dramatic changes between wakefulness and sleep states. As participants fell asleep, the total amount of air they breathed decreased and the shift of dominance between the right and left nostrils increased. By analyzing exactly five minutes of a person’s breathing data, the model can easily classify whether the person is asleep or awake.
Continuous airflow data also mathematically matched the participants’ body mass index, a standard calculation based on human height and weight. The research team noted a mathematical relationship between a person’s body weight and certain aspects of their nasal cycle, suggesting that the neurodynamics that drive breathing interact directly with body composition.
In addition to monitoring physiological functions, the researchers wanted to investigate whether these breathing patterns reflected certain aspects of human cognition or emotion. All participants completed standard psychological questionnaires to assess baseline levels of anxiety, depressive symptoms, and behavioral traits associated with the autism spectrum.
Even though the study group consisted of typical adults without severe clinical diagnoses, their measured breathing patterns correlated with survey scores. Researchers found that an individual’s score on a depression inventory could be predicted, in part, based entirely on breathing characteristics, such as peak inspiratory velocity during waking hours.
A similar predictive relationship was revealed for general anxiety. Participants with higher scores on trait anxiety ratings tended to have slightly shorter inhalation times during sleep. Small changes in the length of pauses between breaths were also associated with differences in levels of self-reported anxiety.
When looking at autism spectrum questionnaires, the data once again highlighted mathematical connections related to how participants breathe. Subtle changes in the amount of time a person paused during inhalation corresponded to different behavioral scores. These findings show that emotional and cognitive states leave subtle but readable biological traces of how the brainstem regulates breathing.
Although this study introduces a new way to measure basic human biology, this experimental method has some discernible limitations. The nasal cannula occasionally became dislodged while participants slept, interrupting nighttime data collection.
Additionally, while pressure-based sensors placed in the nose are good at measuring the precise timing of breathing, they lack integrity when it comes to calculating the total amount of air entering the lungs. The physical appearance of the device can also limit its everyday use, as wearing a medical tube on the face is conspicuous.
In the future, researchers plan to expand this testing method to a wider range of people. Because breathing patterns provide direct insight into brain function, the research team envisions applying this approach to studying a variety of diseases. Monitoring a patient’s unique respiratory fingerprint over time could ultimately serve as a passive, non-invasive tool to track general neurological health.
The study, “Humans have nasal respiratory fingerprints,” was authored by Timna Soroka, Aharon Ravia, Kobi Snitz, Danielle Honigstein, Aharon Weissbrod, Lior Gorodisky, Tali Weiss, Ofer Perl, and Noam Sobel.
