An Analysis of the Effects of Noisy Electrocardiogram Signal on Heartbeat Detection Performance

Area of Science:

• Biomedical Engineering
• Cardiovascular Technology
• Signal Processing

Background:

• Ambulatory cardiac monitoring faces challenges from high noise and artefacts during daily activities.
• Understanding the impact of specific noise types on electrocardiogram (ECG) beat detection is crucial for improving monitoring systems.

Purpose of the Study:

• To investigate the relationship between ECG noise characteristics and beat detection performance in ambulatory settings.
• To evaluate the effectiveness of established beat detection algorithms under various noise conditions.

Main Methods:

• Re-implementation of three established beat detection algorithms.
• Validation using the MIT-BIH Arrhythmia Database and simulated noise-contaminated ECG signals (MIT-BIH Noise Stress Test Database).
• Analysis of noise types including baseline wander (BW), muscle artefact (MA), and electrode motion (EM) artefact at different intensities.

Main Results:

• Noise and artefacts significantly degrade beat detection performance in ambulatory ECG signals.
• Electrode motion (EM) artefacts had the most substantial negative impact, causing the highest number of misdetections and false detections.
• No algorithm achieved perfect QRS complex detection at the highest noise levels.

Conclusions:

• Existing beat detection algorithms struggle with high levels of noise and artefacts common in ambulatory monitoring.
• Electrode motion artefacts pose the greatest challenge to accurate heartbeat detection compared to muscle artefacts and baseline wander.
• Further advancements are needed to enhance the robustness of cardiac monitoring systems against environmental noise.