Explore the performance of automated cough detection in real-world respiratory recordings.

CoughCheck™ is Strados Labs’ machine learning algorithm for automated cough detection and counting. Trained on more than 233,000 validated cough events, CoughCheck enables efficient, objective cough measurement across large datasets and extended monitoring periods.

Download the whitepaper to explore the performance of the latest CoughCheck algorithm using real-world, 24-hour RESP® Biosensor recordings from subjects with subacute and refractory chronic cough.

Download the Whitepaper

Inside the Whitepaper

Learn how CoughCheck performs across three key cough detection and measurement tasks:

Individual Cough Detection
Up to 94.3% average precision for detecting individual cough events.

Cough Presence Detection
Up to 0.993 AUROC for identifying whether one-minute respiratory recordings contain cough.

Cough Frequency Measurement
Strong agreement with manual cough counts, including an R² of 0.993 for 24-hour average cough frequency in the subacute cough cohort.

Built for Real-World Cough Monitoring

Performance was evaluated using real-world RESP® Biosensor recordings without special curation to remove challenging recording conditions such as periods of high background noise.

The results support automated cough detection for applications where fully manual cough counting may not be required, including:

Longitudinal and multi-day cough monitoring
High-volume clinical trial screening
Exploratory endpoints and feasibility studies
Population health research
Disease management

For clinical trial applications requiring gold-standard cough labeling, Strados also offers expert manual cough annotation.

Download the Whitepaper

Complete the form to access Performance Results of the CoughCheck™ Algorithm and learn more about the performance and potential applications of automated cough detection.