An electrocardiogram-based technique to assess cardiopulmonary coupling during sleep

Area of Science:

• Cardiorespiratory Physiology
• Sleep Medicine
• Signal Processing

Background:

• Cardiopulmonary coupling (CPC) dynamics during sleep are crucial for understanding physiological regulation.
• Current methods for assessing CPC often require complex polysomnography.
• Automated analysis of CPC from single-lead electrocardiograms (ECG) could offer a more accessible approach.

Purpose of the Study:

• To evaluate a novel automated method for measuring cardiopulmonary coupling during sleep.
• To assess the utility of an ECG-derived sleep spectrogram for tracking cardiopulmonary interactions.

Main Methods:

• Utilized training and test datasets (35 polysomnograms each) from a sleep disorders center.
• Extracted normal-to-normal sinus interbeat intervals and ECG-derived respiration from single-lead ECG.
• Employed Fourier-based techniques to generate a spectrographic representation of CPC dynamics.

Main Results:

• The ECG-derived sleep spectrogram revealed distinct high- and low-frequency CPC regimes during non-rapid eye movement sleep.
• Agreement with standard sleep staging was poor (43.9%-62.7%).
• Agreement with cyclic alternating pattern (CAP) scoring was higher (74%-77.3%).

Conclusions:

• A sleep spectrogram from a single-lead ECG can dynamically track cardiopulmonary interactions.
• The identified CPC regimes correlate better with CAP states than standard sleep stages.
• This technique offers a complementary method for characterizing sleep stages, particularly non-rapid eye movement sleep.