Robust framework for quantitative analysis of optical mapping signals without filtering

Abstract

Experimental studies with in-vitro isolated hearts using optical mapping techniques have had a significant impact on our understanding of cardiac electrophysiology. The trans-membrane voltage Vm, and intracellular free calcium concentration [Cai]+2 signals obtained from optical mapping experiments can often be corrupted with noise. This is mostly due to the small light intensities and very short exposure times when high speed cameras are used at frames rates of 500-1000 fps. In addition, for small preparations or recordings of small areas, the noise floor levels are even greater and can be comparable to the amplitude of the signal (S/N ≈1). In general strong spatial and temporal filtering is necessary to remove the noise at the expenses of signal degradation and loss of critical information especially at high frequencies that are of a particular interest. In this paper we present and analyze an oversampling image processing technique where due to the cycling processes in cardiac activity during steady state dynamics we are able to stack (sum up) the images recorded at specific equidistant time intervals. The stacking process reduces the noise effectively as the square root of the number of stacked images used. We show that no spatial or temporal filtering is needed to obtain useful data with the stacking technique that allows us to resolve information on a time scale only limited with a sampling rate.