Performance evaluation of three methods for respiratory signal estimation from the electrocardiogram

Area of Science:

• Biomedical Engineering
• Physiological Signal Processing

Background:

• Electrocardiogram (ECG) signals contain information beyond cardiac activity.
• Estimating respiratory signals from ECG (ECG-derived respiration or EDR) offers a non-invasive monitoring approach.

Purpose of the Study:

• To comparatively evaluate three distinct EDR methods.
• To determine the most accurate EDR estimation techniques against real respiratory signals.

Main Methods:

• Analyzed three EDR methods: R wave area, R peak amplitude, and heart rate variability (HRV).
• Computed cross-correlation coefficients and spectral coherence up to 0.5 Hz.
• Validated EDR signals against oronasal and two inductance plethysmography recordings (chest and abdominal).

Main Results:

• EDR methods utilizing R wave area and HRV demonstrated superior correlation with actual respiratory signals.
• These methods also exhibited wider spectral coherence compared to the R peak amplitude method.
• The R peak amplitude method showed less agreement with the reference respiratory signals.

Conclusions:

• ECG-derived respiration estimation is feasible using R wave area and HRV.
• These methods provide more reliable respiratory monitoring than the R peak amplitude approach.
• R wave area and HRV are recommended for developing robust EDR monitoring systems.