Phase shifting prior to spatial filtering enhances optical recordings of cardiac action potential propagation

Optical imaging of cardiac electrical activity using a voltage-sensitive dye provides high spatial resolution maps of action potential propagation and repolarization. Charge-coupled-device (CCD) camera-based imaging systems, however, are limited by their low signal-to-noise ratio. We have developed an image processing method to enhance the quality of optical signals recorded using a CCD camera. The method is based on the observation that within a small neighborhood of adjacent pixels, the morphology of the optical action potential varies little except for a phase shift in time resulting from the propagation of the wavefront. The method uses a phase-correlation technique to first correct for this time shift before spatially filtering with a 5 x 5 Gaussian convolution kernel (sigma = 1.179). A length 5 median filter is then applied to further reduce noise by filtering in the temporal domain. The image-processing scheme allows for more accurate extraction of maps of electrical activation, repolarization, and action potential duration.