Estimation of propagation speed during ventricular fibrillation from frequency-wavenumber power spectra

Propagation speed during ventricular fibrillation (VF) in pigs was estimated from the frequency-wavenumber power spectrum of epicardial mapping data. Data from a 22/spl times/23 array of epicardial electrodes were analyzed to obtain power as a function of wavenumber (spatial frequency). Temporal power spectra were found at each of 10 dominant wavenumbers. Wave speed (in units of mm/s) was estimated from a weighted average of temporal frequency (units of 1/s) divided by wavenumber magnitude (units of 1/mm). Speeds were estimated using data from simulations, sinus rhythm, and early VF. The effect of propranolol on conduction speed during VF was also investigated. The proposed method is useful for analyzing VF, when the existence of many incoherent wave fronts makes if difficult to measure wave speed.<>