News

Hexoskin Unveils Breakthrough Cough Detection AI Algorithm in Computers in Biology and Medicine Journal

Hexoskin’s Artificial Intelligence team has published this week a breakthrough study in Computers in Biology and Medicine, introducing a novel algorithm for automatic cough detection in real‑world settings. Unlike traditional methods, our approach preserves patients’ speech privacy and excels even in noisy environments.

This novel AI detection model was developed using physiological recordings from Hexoskin smart garments, a Digital Health Technology (DHT) wearable cardiac, respiratory, and activity sensor. The model delivers best‑in‑class performance with an area under the ROC curve (AUC) of 95.2 %, sensitivity of 98.5 %, and specificity of 91.9 %. This level of accuracy meets the stringent requirements for cough clinical trial endpoints.

Passive and objective cough measurement devices like Hexoskin smart garments can report outcomes for patients living with respiratory conditions such as asthma and COPD. Because Hexoskin Smart Clothing can be used in clinical trials at home, enabling remote patient monitoring, patients can be monitored for their respiratory condition during daily activities and sleep. AI-based cough detection can provide a detailed analysis of the frequency of cough events, combined with other clinical trial endpoints.

“Detecting coughs accurately without capturing speech content is essential for both patient privacy and clinical utility,” said Dr. Philippe Dixon, lead author of the study and assistant professor at McGill University. “Our model’s robustness to ambient noise makes it an ideal tool for home‑based monitoring and telemedicine applications, key aspects of remote respiratory monitoring.”

This announcement represents the latest advance in Hexoskin’s expanding clinical AI portfolio. In parallel with ongoing clinical trials, Hexoskin leverages the world’s largest physiological database—accumulated through years of wearable sensor deployments—to accelerate the development of AI‑driven tools for patient monitoring and diagnostics.

Here's an excerpt from the paper's abstract:

Coughing behavior is associated with conditions such as sleep apnea, asthma, and chronic obstructive pulmonary disorder (COPD) and can severely affect quality of life in those affected. In this context, coughing quantification is often important, but routinely performed via questionnaires. This approach is dependent on patient compliance or recall, which may affect validity and be especially difficult for nocturnal coughs. Manual review of audio recordings is potentially more accurate, but raises adherence and privacy concerns due to the collection and review of sensitive audio-data by a human annotator. Today, machine learning approaches are increasingly used to quantify coughs; however, algorithms often rely on microphone recordings, resulting in the same privacy issues, especially if data are sent to a remote server for analysis.

Hexoskin's open platform is used by hundreds of researchers for data collection for new digital biomarkers and AI development. Hexoskin is also working with foundations, pharmaceuticals, and biotechnology to speed the clinical development of respiratory diseases and other cardiac, neurological, and rare diseases, developing primary and secondary endpoints for clinical development. We also collaborate on digital biomarkers for patient monitoring and find new cures.

Contact our team today to collaborate on the future of clinical development and AI in healthcare.