Segmenting cardiac-related data using sleep stages increases separation between normal subjects and apnoeic patients

Area of Science:

• Cardiovascular Physiology
• Sleep Medicine
• Biomedical Signal Processing

Background:

• Inter-patient comparisons of cardiovascular metrics are limited by individual variations in physical and mental activity.
• Assessing activity during conscious states is challenging, necessitating objective measures during sleep.
• Heart rate variability (HRV) analysis offers potential for objective central nervous system activity assessment.

Purpose of the Study:

• To develop an objective scale for measuring central nervous system activity during sleep.
• To compare the effectiveness of different spectral estimation techniques (FFT vs. LSP) for HRV analysis in sleep.
• To differentiate between apnoeic and healthy subjects using HRV metrics during sleep stages.

Main Methods:

• Dividing heart rate (RR) interval time series into sleep stage segments.
• Estimating the LF/HF-ratio (low to high frequency power balance) within these segments.
• Comparing the performance of Fast Fourier Transform (FFT) and Lomb-Scargle periodogram (LSP) for spectral estimation.

Main Results:

• The Lomb-Scargle periodogram (LSP) demonstrated superior separation of patients by condition compared to the Fast Fourier Transform (FFT).
• Patient separation was most pronounced during slow wave sleep.
• The LF/HF-ratio analysis, applied to sleep segments, reduced activity-based noise for more effective HRV comparison.

Conclusions:

• The LSP method provides a more effective tool for analyzing HRV during sleep compared to FFT.
• Objective HRV analysis during sleep, particularly in slow wave sleep, can enhance patient stratification.
• This approach offers a promising method for objective assessment of central nervous system activity during sleep.