Simulated central apnea detection using the pressure variance.

Area of Science:

• Biomedical Engineering
• Respiratory Medicine
• Sensor Technology

Background:

• Central sleep apnea is a serious condition affecting breathing during sleep.
• Accurate detection of central apnea events is crucial for effective treatment.
• Current detection methods can be invasive or lack precision.

Purpose of the Study:

• To evaluate an unobtrusive pressure sensor array for simulating central apnea detection.
• To compare single-sensor versus multisensor approaches for apnea event classification.
• To investigate the impact of body position on detection accuracy.

Main Methods:

• Collected breathing data from seven volunteers performing breathing and breath-holding exercises.
• Utilized epoch-based variance calculations for feature extraction from breathing signals.
• Developed and compared single-sensor and multisensor (voting) algorithms.
• Modified the multisensor approach to mitigate false positives from ballistocardiogram signals.

Main Results:

• Breathing events were successfully differentiated using epoch-based variance.
• The multisensor voting approach demonstrated superior performance over the single-sensor method.
• The modified multisensor approach enhanced sensitivity and maintained a low false positive rate, accounting for body position.
• Low variability was observed in intersubject classification results.

Conclusions:

• An unobtrusive pressure sensor array shows promise for central apnea detection.
• A multisensor voting strategy improves classification accuracy and robustness.
• The system's performance is adaptable to variations in body position, crucial for real-world application.