In Russia, an Algorithm to Detect Stress Through Breathing Patterns Is Being Developed

Researchers at Petrozavodsk State University and Binali Yıldırım University in Erzincan, Turkey, have developed a computer algorithm that can detect signs of stress by analyzing a person’s breathing patterns. The algorithm is based on WESAD, an open international dataset of physiological data, and achieved a diagnostic accuracy of 96%.

The system analyzes more than breathing rate. It also takes into account inhalation duration, breathing patterns, and changes in rhythm across different frequencies. The researchers used these features to train the system to recognize various markers of stress.

Building on the study’s findings and the features they identified, the researchers developed BreathCouple, a free Android app. Unlike the experiment that used a chest-worn sensor, the app records breathing with a compact external microphone, such as one connected to a smartphone.

The researchers are continuing to refine the system so it can be used in everyday practice.