Extinguish burnout: a computer algorithm recognizes stress by changing breathing

Scientists have developed a computer algorithm that recognizes stress by changes in breathing rhythm with an accuracy of more than 96%. In the future, the technology can be used to monitor drivers, dispatchers, pilots and operators of complex systems, as well as in sports, education and medicine. Researchers are also considering the possibility of tracking breathing without chest sensors. At the same time, experts interviewed by Izvestia warn that for healthy people whose work is not associated with increased responsibility, constant monitoring can become an additional source of anxiety. To whom such a system is suitable and in what cases it can harm — in the Izvestia article.

Measuring stress by breathing rhythm

Researchers at Petrozavodsk State University, together with colleagues from Binali Yildirim University of Turkey in Erzincan, have taught a computer algorithm to detect signs of stress based on breathing parameters.

As the scientists explained, breathing varies depending on a person's condition. When agitated, it may become more frequent and less regular, the duration of inspiration and the shape of the respiratory wave, a signal reflecting the movement of the chest, change. However, the respiratory rate alone is not enough to reliably determine stress: the normal rhythm varies significantly from person to person, so a separate indicator does not allow us to unambiguously assess a person's condition.

The authors analyzed the WESAD open international physiological dataset, which includes information on respiration, heart function, skin electrical conductivity, and movements of volunteers. In the new work, the researchers used only data on breathing.

— The results of the study can be used to monitor stress in highly responsible professions — drivers, dispatchers, pilots, operators of complex systems, as well as in sports, education and medicine. They can help track a person's reaction to stress, observe the dynamics of recovery after physical or psychoemotional stress and notice the state of increased tension in time," Andrey, associate professor of the Department of Information and Measuring Systems, Electronics and Automation, head of the educational and scientific laboratory for the development of electronic component base based on micro— and nanostructures of Petrozavodsk State University, told Izvestia. Velichko.

Labor in the colony: victory over the "egoism" of bacteria will help create new medicines

Scientists have found a way to make every microorganism in the population work together.

During the experiment, the volunteers were consistently in different states: resting, watching entertainment videos, meditating, or taking a stress test. In the latter case, participants were asked to tell a three-person panel about their strengths and weaknesses, and then count them out loud. During the experiment, breathing was recorded using a chest sensor that recorded chest movements.

The received signals were divided into minute segments and machine learning algorithms were applied. The researchers compared the two approaches. In the first case, a compact neural network analyzed the respiratory wave directly. In the second case, respiration characteristics were previously extracted from the signal — for example, features of the waveform, changes in rhythm at different frequencies, and its stability over time. These parameters were then used to determine a person's condition.

The best result was shown by a compact neural network that analyzed the original signal.: she correctly identified the presence or absence of stress in more than 96% of cases. At the same time, not only the respiratory rate was associated with stress, but also other features such as the duration of inspiration, the shape of the respiratory wave, and changes in rhythm over time. Thus, the algorithm relied on a combination of several features rather than a single indicator.

Based on the results of the study, scientists have developed a free Android application — BreathCouple. Unlike the experiment, which used a chest sensor, the app registers breathing using a compact external microphone connected to a smartphone. It should be placed next to your nose to record your inhales and exhales. Based on this data, the program builds a respiratory wave and allows you to track the frequency and fluctuations of the breathing rhythm.

— We have seen that stress manifests itself not only in how often a person breathes. The whole breathing pattern is more informative.: how is the energy of the signal distributed, how stable is the ratio of the duration of inspiration and the entire respiratory cycle, how does the waveform change and how long does its temporal structure last," said Andrei Velichko.

Monitoring the condition of pilots and controllers

At the same time, the authors emphasize that the accuracy obtained from experimental data cannot yet be expected from the application. The microphone and chest sensor register different physical signals, and their readings have not yet been compared simultaneously on the same group of participants. Therefore, the BreathCouple is intended for personal observation and pilot studies and is not considered a medical device for diagnosis.

In the future, the scientists plan to repeat the experiment on a larger group of volunteers and test the method not only in the laboratory, but also in everyday conditions. Participants will walk, work at the table, relax and sleep. This will allow you to evaluate how movement, body position, conversation, and other natural factors affect the accuracy of stress recognition.

Signal with help: the "commander" molecule may be the key to restoring the brain

Researchers have found a new target for the treatment of Alzheimer's and Parkinson's diseases

The respiratory rhythm can really serve as an informative and reliable marker of stress, Ivan Plushaev, psychologist, director of ANO Strategies of Opportunities, market expert at NTI Neuronet, told Izvestia. However, stress is caused by both external and internal factors, and it manifests itself not only in changes in breathing or sleep disorders, but also in other emotional and cognitive symptoms. Therefore, it is necessary to assess a person's condition comprehensively, without relying solely on one physiological indicator.

— For a specialist psychologist, it is not the fact of physiological arousal itself that is important, but its context and subjective experience. At the same time, the approach has significant advantages: the implementation in the form of an application makes the technology potentially widespread and accessible, which distinguishes it from bulky chest sensors. In addition, unlike subjective questionnaires, physiological markers (respiratory wave) are more difficult to fake or distort due to socially desirable responses," the expert clarified.

Participant of the scientific project Mehmet Tahir Khuyut in the laboratory

Photo: Mehmet Tahir Huyut's personal archive

This study opens up promising opportunities for psychological practice, says Olga Valaeva, head of the Center for the Development of Digital Clinics at Moscow State University.

"Studying breathing patterns and their relationship to human psychophysiological conditions will improve the effectiveness of working with stress, anxiety, and the regulation of emotional reactions," she said.

For most people, such an algorithm is redundant, says Marina Chumakova, a leading market expert at NTI Helsnet. According to her, a healthy adult usually notices signs of overload himself: he sleeps worse, gets tired faster and gets irritated more often. A device that reports this with 96% accuracy adds little to a person's own feelings.

"A justified niche for such technologies is professions with a high cost of error, such as the work of dispatchers and drivers, as well as clinical research, in which objective data is important," she noted.

According to the expert, a separate risk is associated with people who are prone to increased anxiety. For them, constant monitoring can become an additional source of anxiety: a notification provokes a new wave of anxiety, a person begins to monitor their breathing more closely, and check their indicators over and over again. Doctors are already observing a similar effect in users of fitness trackers and smart watches.

The results of the study, supported by a grant from the Russian Science Foundation (RSF), are published in the journal Biomedical Signal Processing and Control.