Wavelet denoising of patch clamp signals for improved histogram analysis

This paper focuses on improving patch clamp signal histograms using wavelet denoising. The denoising method utilizes the one-dimensional discrete wavelet transform and hard thresholding. Two threshold criteria, the Stein's unbiased risk estimator and the Minimax, are investigated. The effects of the choice of mother wavelet and level of decomposition are studied. In order to measure the accuracy of the denoising methods, noise-free patch clamp signals are synthesized along with background thermal noise which is then added to the signal to form a synthesized noisy patch clamp signal. The histogram results of the denoised signals are compared to those of the original noise-free signals to show the benefits of the wavelet denoising.