A method for suppressing the interference by biased nonlinearities is presented. The memoryless nonlinearity taken into account is of the form \(y(x) = (|x |- \xi)^\nu\) for \(|x |> \xi\), and \(y(x) = 0\) for \(|x |\leq \xi\), where \(\xi > 0\) and \(\nu\) a positive number greater than \(-1\). For this case \(y\) can be represented in terms of \(x\) by \[ y(x) = {2 \over \pi} \Gamma (\nu + 1) \int^\infty_0 \cos \left( v \xi + {\nu \pi \over 2} \right) \sin (vx) {dv \over v^{\nu + 1}}, \] where \(\nu >-1\) and \(\Gamma\) is the Gamma function. This representation was first given by S. O. Rice and has the following computing advantage: when \(x\) is a sum of sinusoids, the integral can be developed into an (infinite) sum of sinusoids of combination frequencies and the weighting coefficients are Bessel depending on the weighting coefficients entering the sum representing the input signal \(x\). Several interesting relations are derived showing the possibility of reducing interference in some cases. The paper is of interest for researchers working in the areas of interference and spread spectrum communication and represents a fair performance in handling special functions and their integral representations.