Reliable P wave detection in pathological ECG signals.

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Accurate P wave detection in ECGs is crucial for diagnosing cardiac arrhythmias.
• Current software often fails with pathological ECGs, limiting automated analysis.
• Existing methods lack validation on diverse pathological signals.

Purpose of the Study:

• To develop a novel, accurate, and reliable method for P wave detection in ECGs.
• To improve P wave detection specifically in pathological cardiac conditions.
• To validate the method's performance across multiple ECG databases, including those with pathologies.

Main Methods:

• Utilized phasor transform of ECG signals.
• Developed innovative decision rules based on cardiac arrhythmia manifestations.
• Validated the algorithm on three diverse, annotated ECG databases (MITDB, QT, BUT PDB).

Main Results:

• Achieved high sensitivity (Se) and positive predictive value (PP) for normal ECGs (e.g., MITDB: Se=98.56%, PP=99.82%).
• Demonstrated superior performance on pathological ECGs (e.g., MITDB: Se=96.40%, PP=91.56%; BUT PDB: Se=93.07%, PP=88.60%).
• Outperformed existing methods on pathological signals.

Conclusions:

• The novel method provides accurate and reliable P wave detection in both normal and pathological ECGs.
• This advancement significantly enhances the potential for fully automated ECG analysis in clinical practice.
• The approach represents a substantial step towards improving cardiac arrhythmia diagnosis and treatment strategies.