Level-set method for robust analysis of optical mapping recordings of fibrillation

We have developed a level-set based approach that can accurately and reliably extract the temporally and spatially resolved positions of wavefronts, wavebacks, and phase singularities (PSs) from noisy optical mapping data. The algorithm consists of three steps: pointwise phase reconstruction, computation of smooth level sets of the phase field, and identification of phase singularities as the level set intersections. This method was illustrated by performing an analysis of optical mapping voltage data during ventricular fibrillation in a Langendorff-perfused pig heart. Observed behavior of the PSs was in excellent agreement with previous experimental studies: PSs appeared and disappeared in pairs of opposite topological charge, their lifetimes ranged from milliseconds to seconds, and the longest-lived PSs became "pinned" on structural inhomogeneities of the heart tissue.