Multimodal heart beat detection using signal quality indices.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiovascular Monitoring

Background:

• Accurate R peak detection in electrocardiogram (ECG) signals is crucial for deriving vital parameters like heart rate.
• Signal corruption from movement or sensor failure can impede R peak detection, leading to inaccurate parameter estimation.
• Synchronous measurement of multimodal signals offers a potential solution for robust R peak detection in compromised ECG recordings.

Purpose of the Study:

• To develop and evaluate a novel method for fusing R peak detections from ECG and arterial blood pressure (ABP) waveforms.
• To enhance the robustness of R peak localization in the presence of signal noise and artefacts.
• To reduce false alarms in real-time monitoring by improving the accuracy of derived cardiovascular parameters.

Main Methods:

• A fusion method combining R peaks detected on ECG via an energy detector and on ABP via the length transform was proposed.
• A signal quality index (SQI) was derived for both ECG and ABP signals to assess data integrity.
• ECG SQI relied on the agreement between two peak detectors, while ABP SQI ensured physiological plausibility of blood pressure values.

Main Results:

• The fusion algorithm achieved a sensitivity of 95.1% and a positive predictive value of 89.3% on an external evaluation dataset.
• The method demonstrated excellent performance across diverse signal morphologies encountered in clinical practice.
• The algorithm attained a competitive score of 91.5% in the Physionet/Computing in Cardiology Challenge 2014.

Conclusions:

• The proposed multimodal signal fusion method effectively enhances R peak detection accuracy, even with noisy or absent ECG signals.
• This approach has the potential to significantly improve real-time cardiovascular monitoring and reduce false alarms.
• Freely available source code facilitates the adoption and further development of this robust R peak detection technique.