Sleep Apnea Events Recognition Based on Polysomnographic Recordings: A Large-Scale Multi-Channel Machine Learning approach
2024

Detecting Sleep Apnea Events with Machine Learning

Sample size: 1587 publication 10 minutes Evidence: high

Author Information

Author(s): Nicolò La Porta, Stefano Scafa, Michela Papandrea, Filippo Molinari, Alessandro Puiatti

Primary Institution: Università della Svizzera Italiana (USI)

Hypothesis

Can machine learning models effectively recognize apneic events using low-invasive signals from polysomnographic recordings?

Conclusion

The study demonstrates that a machine learning approach can accurately detect sleep apnea events, improving the efficiency and accessibility of diagnosis.

Supporting Evidence

  • The machine learning model achieved an overall accuracy of 87.2% in detecting apneic events.
  • The study utilized a large dataset of 1587 polysomnographic recordings.
  • Different types of apnea were distinguished with an accuracy of 62.9%.

Takeaway

This study shows that computers can help doctors find breathing problems during sleep without needing complicated equipment.

Methodology

The study used machine learning algorithms to analyze polysomnographic data from the Wisconsin Sleep Cohort database to detect apneic events.

Potential Biases

Potential bias due to the variability in scoring by different experts over a long period.

Limitations

The study is limited by inter-operator variability and the use of hardware that may have affected signal quality.

Participant Demographics

The sample included 848 males and 739 females, with an average age of 58 years.

Statistical Information

P-Value

p<0.001

Statistical Significance

p<0.001

Digital Object Identifier (DOI)

10.1109/OJEMB.2024.3508477

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication