Related Experiment Videos
Computationally efficient noninvasive cardiac activation time imaging
G Fischer1, B Pfeifer, M Seger
1Institute for Biomedical Signal Processing and Imaging, University for Health Sciences, Medical Informatics and Technology (UMIT), Innsbruck, Austria. gerald.fischer@umit.at
Methods of Information in Medicine
|January 10, 2006
Summary
A new computational method significantly speeds up cardiac activation sequence mapping, reducing computation time from nearly an hour to minutes. This advance aids in guiding ablation therapy for cardiac arrhythmias.
Area of Science:
- Computational electrophysiology
- Cardiac electrophysiology modeling
Background:
- Computer models for cardiac activation sequence computation are clinically validated.
- Current computation times (nearly 1 hour) limit clinical applicability for ablation therapy guidance.
Purpose of the Study:
- To develop a computational method for cardiac activation sequence mapping with a run time of a few minutes.
- To improve the efficiency of cardiac activation sequence computation for clinical use.
Main Methods:
- Optimized the computationally intensive product of the lead field matrix and source pattern matrix.
- Utilized biophysical properties of matrices to accelerate computations by over an order of magnitude.
- Developed a conjugate gradient optimizer in C++ for activation map computation.
Main Results:
- Achieved a 30-fold increase in speed compared to the previous Fortran 77 implementation.
- Maintained comparable quality of results with the accelerated method.
- Demonstrated that a coupled regularization strategy reduced sensitivity to the regularization parameter.
Conclusions:
- The developed software provides diagnostically valuable information in a significantly shorter time.
- The method is applicable for localizing focal arrhythmogenic substrates, as shown in a WPW patient case.
- The accelerated computation enables more practical clinical application of cardiac activation sequence mapping.