Model-based detection of heart rate turbulence

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Heart rate turbulence (HRT) and ventricular ectopic beats (VEBs) are important indicators of cardiac health.
• Current methods for detecting HRT, such as turbulence onset (TO) and turbulence slope (TS), have limitations in accuracy and efficiency.
• The integral pulse frequency modulation model provides a framework for understanding heart rate variability.

Purpose of the Study:

• To extend the integral pulse frequency modulation model to incorporate ectopic beats and HRT.
• To develop and validate a novel statistical approach for detecting and characterizing HRT.
• To compare the performance of the new HRT detection method against existing parameters.

Main Methods:

• The study extended the integral pulse frequency modulation model.
• A new detector was developed using Karhunen-LoEve basis functions and a generalized likelihood ratio test statistic T(x).
• Detector performance was evaluated using simulated ECG data and real ECG signals from hemodialysis patients.

Main Results:

• The T(x) statistic demonstrated superior performance compared to TO and TS in simulations, requiring fewer VEBs for accurate detection.
• Simulations showed the influence of signal-to-noise ratio (SNR), QRS jitter, and sampling rate on detector performance.
• ECG analysis revealed significantly higher T(x) values in hypotension-resistant (HtR) patients compared to hypotension-prone (HtP) patients.

Conclusions:

• The proposed T(x) statistic offers a more robust and accurate method for HRT detection and characterization than conventional parameters.
• HRT appears to be more prevalent in HtR patients undergoing hemodialysis.
• The T(x) statistic provides better discrimination between HtR and HtP groups compared to TO and TS.