Quantitative measurement of sleep quality using cardiopulmonary coupling analysis: a retrospective comparison of...

The purpose of this study was to determine the utility of a new operator-independent, automated measure of sleep physiology based on cardiopulmonary coupling (CPC) analysis in subjects with primary insomnia vs. good sleepers.

The polysomnograms of 50 subjects with primary insomnia and 36 good sleepers were summarized and analyzed from a consecutive two-night protocol. The electrocardiograms (ECG) from adaptation and baseline night polysomnograms were analyzed using CPC analysis. This Fourier-based technique uses heart rate variability and ECG R wave amplitude fluctuations associated with respiration to generate frequency maps of coupled autonomic-respiratory oscillations. The resulting sleep spectrogram is able to categorize sleep as "stable" (high-frequency coupling [HFC], 0.1-0.4 Hz) and "unstable" (low-frequency coupling [LFC], 0.1-0.01 Hz), independent of standard sleep stages. Wake and rapid eye movement sleep exhibit very low-frequency coupling (VLFC, 0.0039-0.01 Hz). Elevated LFC (e-LFC) is a subset of LFC that is associated with fragmented sleep of various etiologies.

CPC variables showed a significant multivariate analysis of variance group, night, and group × night main effect, except for HFC by night. Relative to good sleepers, primary insomnia patients on adaptation night had lower HFC, a putative biomarker of stable sleep, and HFC/LFC ratio, an indicator of sleep quality. The primary insomnia group also had higher LFC, an index of unstable sleep, and an increase in VLFC and e-LFC compared to good sleepers on adaptation night. On baseline night, the primary insomnia group had increased LFC, VLFC, and e-LFC and a lower HFC/LFC ratio. Except for HFC, good sleepers had larger CPC variable differences between adaptation and baseline nights compared to the primary insomnia group.

Primary insomnia subjects have a marked worsening of sleep quality on the adaptation night, which is well captured by both conventional and ECG-derived sleep spectrogram techniques. The larger improvement of sleep quality was found among good sleepers and captured only by CPC analysis. The operator-independent, automated measure of sleep physiology demonstrated functionality to differentiate and objectively quantify sleep quality.