Related Experiment Videos
Application of empirical mode decomposition to heart rate variability analysis
J C Echeverría1, J A Crowe, M S Woolfson
1School of Electrical & Electronic Engineering, University of Nottingham, UK. jcea@xanum.uam.mx
Medical & Biological Engineering & Computing
|August 29, 2001
Summary
Empirical mode decomposition (EMD) offers advanced analysis for heart rate variability (HRV) by overcoming limitations of traditional methods. EMD accurately isolates autonomic components in HRV data, improving non-stationary and non-linear time series analysis.
Area of Science:
- Physiology
- Biomedical Engineering
- Signal Processing
Background:
- Traditional spectral estimation methods for heart rate variability (HRV) analysis have limitations in capturing autonomic modulation dynamics.
- Existing methods provide time-averaged power estimates, failing to address non-stationary and non-linear characteristics of HRV.
- There is a need for advanced techniques to analyze complex HRV data, especially under varying autonomic conditions.
Purpose of the Study:
- To evaluate the efficacy of empirical mode decomposition (EMD) and Hilbert spectra for analyzing short-term HRV data.
- To assess the capability of EMD in isolating autonomic components in both simulated and real HRV signals.
- To determine the accuracy of EMD-derived Hilbert amplitude and instantaneous frequency for tracking changes in non-stationary HRV.
Main Methods:
- Application of empirical mode decomposition (EMD) to simulated (chirp, integral pulse frequency modulation model) and real short-term HRV data.
- Analysis of stationary and non-stationary HRV conditions under controlled breathing maneuvers.
- Assessment of Hilbert amplitude component ratio and instantaneous frequency for time-tracking amplitude and frequency variations.
Main Results:
- EMD successfully isolated key components in simulated signals, including two components in a chirp series and three in IPMF signals.
- EMD consistently identified at least four components in real HRV signals, localized within known autonomic frequency bands.
- Frequency tracking error was minimal (<0.22%) for simulated signals and a chirp series, demonstrating high accuracy.
Conclusions:
- Empirical mode decomposition (EMD) provides superior capabilities for analyzing heart rate variability (HRV) compared to traditional spectral methods.
- EMD effectively decomposes complex HRV signals into intrinsic mode functions, revealing underlying autonomic modulation patterns.
- The Hilbert-Huang Transform, coupled with EMD, offers accurate time-frequency analysis for non-stationary and non-linear HRV data.