New Artificial Intelligence-based Device Enables Highly Accurate Cough Monitoring

A new multi-centre study led by ISGlobal, an institution supported by ”la Caixa” Foundation, validates the accuracy and reliability of the Hyfe CoughMonitor, an innovative cough monitoring system based on artificial intelligence (AI). The results, published in Scientific Reports, show that this device passively and continuously monitors cough episodes with a sensitivity of over 90% and a low false positive rate.

Cough is one of the most common symptoms in healthcare and can be associated with a wide range of respiratory diseases such as COVID-19, chronic obstructive pulmonary disease (COPD), or tuberculosis. However, to date there has been no automated, clinically validated system that can continuously monitor it in real-life situations.

“Currently, diagnostic accuracy is key, which is why the objective quantification of symptoms is essential”, explains Carlos Chaccour, lead author and researcher at ISGlobal at the time of the study. “Efficient, objective, and long-term monitoring could provide a valuable database to improve diagnosis, prognosis, treatment response, and even symptom surveillance for multiple diseases”, he adds.

An algorithm built into a smartwatch

The device, which is worn on the wrist in the form of a smartwatch with an integrated application, uses an artificial intelligence algorithm to analyse sounds in real time in order to accurately monitor coughs. The study compared the device’s detection with manual counting of cough episodes by experts. A total of 546 hours of continuous recordings were collected from 23 participants with various respiratory conditions, more than half of whom were recruited in the United States, and a 99% correlation between the device and manual detection was achieved.

The Hyfe CoughMonitor is based on a “Convolutional Neural Network” (CNN), a machine learning programming model. The monitoring application processes sounds directly on the device, ensuring user privacy while performing daily activities. Only the necessary data, such as cough frequency and duration, is stored and securely synchronised with the cloud for further analysis.

The future of respiratory disease management

The results of the study suggest that the monitor could be very useful in understanding the causes of cough, as well as in improving the diagnosis and management of the diseases that cause it. “By providing quantitative data, continuous monitoring would facilitate the study of respiratory diseases while helping to overcome current statistical limitations in clinical trials of new antitussive drugs”, says Chaccour, researcher at the ICS of the University of Navarra.

Future studies should focus on the potential clinical utility of this monitoring in diagnosis, prognosis, and assessment of treatment response. Examples include its use in ambulatory monitoring of COPD patients to predict exacerbations and in the follow-up of asthma patients when starting a new treatment.

Reference

Chaccour, C., Sánchez-Olivieri, I., Siegel, S. et al. Validation and accuracy of the Hyfe cough monitoring system: a multicenter clinical study. Sci Rep 15, 880 (2025). https://doi.org/10.1038/s41598-025-85341-3