Using analytical methods and Monte Carlo simulations, we analyze new statistics designed to detect isolated step-like discontinuities which are coherent over large areas of Cosmic Microwave Background (CMB) pixel maps. Such coherent temperature discontinuities are predicted by the Kaiser-Stebbins effect to form due to long cosmic strings present in our present horizon. The background of the coherent step-like seed is assumed to be a scale invariant Gaussian random field which could have been produced by a superposition of seeds on smaller scales and/or by inflationary quantum fluctuations. We find that the proposed statistics can detect the presense of a coherent discontinuity at a sensitivity level almost an order of magnitude better compared to more conventional statistics like the skewness or the kurtosis.

10 pages, 3 Figures, Use RevTeX