Algorithm for the classification of multi-modulating signals on the electrocardiogram

Area of Science:

• Biomedical Engineering
• Signal Processing
• Medical Diagnostics

Background:

• Sleep apnoea diagnosis typically requires polysomnography.
• Electrocardiogram (ECG) signals contain information about cardiorespiratory interactions.
• Developing non-invasive diagnostic methods from ECG is of clinical interest.

Purpose of the Study:

• To present a novel algorithm for simultaneous measurement of ECG and respiration from ECG recordings.
• To evaluate the diagnostic potential of this algorithm for sleep apnoea.

Main Methods:

• The algorithm combines three scale transforms (a(j)(t), u(j)(t), o(j)(a(j))) with statistical Fourier transform (SFT).
• Scale transforms convert the source signal into a periodic form, confining harmonics.
• SFT decomposes the multi-modulating source, followed by inverse transform to reconstruct ECG and respiration signals.

Main Results:

• The algorithm successfully decomposes and reconstructs ECG and respiration signals from a single ECG recording.
• It enables the estimation of partial ventilation and heart rate variability.
• The method demonstrates high potential for clinical application in sleep apnoea diagnosis.

Conclusions:

• The developed algorithm offers a promising approach for simultaneous cardiorespiratory signal measurement using ECG.
• This technique has significant potential for non-invasive sleep apnoea diagnosis from ECG recordings.
• Further clinical validation is warranted to confirm its diagnostic accuracy.