Independent Component Analysis and Wavelet Decomposition Technique for the Detection of Motor Unit Action Potentials

This paper proposes the use of independent component analysis (ICA) and thresholding estimation calculated in wavelet transform for noise reduction in electromyographic (EMG) signals. In contrast to existing amplitude threshold detection scheme which either need to be participated by the operator or is time consuming, this method is more fast and completely automatic. The ICA is implemented by means of a fast and robust fixed-point algorithm. The basic tool is the method of power spectrum estimation, the Welch method, that allows us to analyze power spectral density of non-stationary signals