Maximum entropy estimation of spatial patterns of activation rate during ventricular fibrillation

Philip V. Bayly, Bruce H. KenKnight, Nipon Chattipakorn, Stephan Windecker, Masahiro Usui, Jack M. Rogers, Chad R. Johnson, Raymond E. Ideker, William M. Smith

Research output: Contribution to journalConference articlepeer-review

Abstract

In order to investigate spatial variations in ventricular fibrillation (VF), activation rates during VF in pigs were estimated on a grid of 504 epicardial sites using the maximum entropy (ME) method. ME spectral estimates were obtained from unipolar electrograms recorded with a 21 × 24 array of epicardial electrodes (1 mm spacing) during electrically induced VF. The frequency associated with the most power in each ME spectrum was denoted the peak frequency (PFreq). We compared ME estimates of mean activation interval (1/PFreq) to mean activation intervals (MAI) obtained by picking the local minimum dV/dt. The Class III drug artilide was used to extend the range of activation rate. A significant positive correlation was found between 1/PFreq and MAI. ME estimates of 1/PFreq did not correlate better with MAI than previous estimates obtained using the Fast Fourier Transform. The spatial distribution of activation rate varied dynamically during the first 12 seconds of VF.

Fingerprint

Dive into the research topics of 'Maximum entropy estimation of spatial patterns of activation rate during ventricular fibrillation'. Together they form a unique fingerprint.

Cite this