Related Experiment Videos
Criteria for local myocardial electrical activation: effects of electrogram characteristics
K P Anderson1, R Walker, M Fuller
1Cardiology Division, University of Utah, Salt Lake City 84132.
IEEE Transactions on Bio-Medical Engineering
|February 1, 1993
Summary
Identifying local myocardial activation from complex electrograms is challenging. A statistical approach using signal characteristics effectively distinguishes local from distant electrical activity in the atria and ventricles.
Area of Science:
- Cardiology
- Electrophysiology
- Biomedical Engineering
Background:
- Detecting local myocardial electrical activation via extracellular recordings is often hindered by polyphasic electrograms.
- Distinguishing local activation from nonlocal activity is crucial for accurate electrophysiological assessment.
Purpose of the Study:
- To compare the efficacy of various signal variables in differentiating unipolar deflections caused by local activation versus nonlocal activity.
- To identify optimal criteria for detecting local activation in different cardiac signal populations.
Main Methods:
- Utilized a model of polyphasic deflections derived from atrial recordings during reentrant tachycardia.
- Assessed variable performance using areas under receiver operating characteristic curves.
- Identified optimal thresholds by maximizing statistics that corrected for pretest probability of local activation.
Main Results:
- The greatest negative first derivative of unipolar potential best discriminated local from distant ventricular signals.
- The ratio of the first derivative to the potential outperformed the greatest negative first derivative for distinguishing local atrial from distant ventricular signals.
- A linear combination of potential and its first derivative-to-potential ratio showed robust performance across all signal types.
Conclusions:
- Optimal criteria for detecting local myocardial activation are signal population-dependent.
- A statistical approach can successfully identify optimal detection criteria for specific signal populations.