Matrix of regularity for improving the quality of ECGs

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• The 12-lead electrocardiography (ECG) is crucial for diagnosing heart abnormalities.
• ECG signals are prone to artifacts, potentially causing misdiagnosis and improper treatment.
• Distinguishing between ECG artifacts and genuine disease patterns presents a significant clinical challenge.

Purpose of the Study:

• To introduce a novel computational framework, the matrix of regularity, for evaluating ECG signal quality.
• To develop a method for accurately differentiating between ECG artifacts and disease-related signals.
• To provide a continuous quality grade for ECGs, enhancing diagnostic reliability.

Main Methods:

• Developed a computational framework named the matrix of regularity.
• The matrix of regularity integrates results from multiple signal quality assessment tests.
• The framework generates a continuous quality score for ECG recordings.

Main Results:

• The algorithm achieved up to 95% accuracy in classifying ECG quality on a benchmark dataset.
• The area under the receiver operating characteristic curve reached 0.97, indicating high performance.
• The matrix of regularity provides a nuanced, continuous measure of ECG quality.

Conclusions:

• The matrix of regularity framework effectively evaluates ECG signal quality and differentiates artifacts.
• This computational approach has the potential to improve the reliability of ECG diagnostics.
• The developed framework and software can enhance ECG data quality from both conventional and portable devices.