Myocardial infarction evaluation from stopping time decision toward interoperable algorithmic states in reinforcement...

Area of Science:

• Cardiology
• Machine Learning
• Signal Processing

Background:

• The Elliot wave principle is applied to analyze impulsive and corrective trends in financial markets and electrocardiograms (ECGs).
• Pathological ECG waveforms, including those indicative of myocardial infarction (MI), present challenges for current commercial electrocardiographs.
• Accurate differential diagnosis for right ventricle infarction is crucial to avoid contraindicated medications like nitroglycerin.

Purpose of the Study:

• To develop a novel algorithm for accurate myocardial infarction detection using ECG analysis.
• To apply stopping time theory and reinforcement learning for analyzing impulsive wave trends in ECGs.
• To support clinical interpretation of 12-channel ECGs, particularly for right ventricle infarction diagnosis.

Main Methods:

• Implementing stopping time theory within a reinforcement learning framework to identify impulsive wave trends.
• Utilizing least-first-power approximation and approximate entropy to evaluate impulsive waveform shapes.
• Employing neural networks to approximate isoelectric baseline amplitude and detect myocardial infarction conditions.

Main Results:

• The algorithm achieved high accuracy in discerning myocardial infarction: 99.2754% for CSV data and 99.3579% for representative beats.
• The clinical dataset comprised 276 ECGs from CSV files and 623 representative beats.
• The developed method provides a robust approach for analyzing ECGs and identifying pathological waveforms.

Conclusions:

• The proposed algorithm enhances the accuracy of myocardial infarction detection from ECGs.
• This work supports differential diagnosis for right ventricle infarction, preventing adverse medication events.
• The algorithm's ability to analyze impulsive waveforms aids in understanding ECG abnormalities related to MI.