A new method for the quantification of amplitude variations in biomedical signals through moving approximate entropy is presented. Unlike the usual method to calculate the approximate entropy (ApEn), in which the tolerance value (r) varies based on the standard deviation of each moving window, in this work ApEn has been computed using a fixed value of r. We called this method, moving approximate entropy with fixed tolerance values: ApEnf. The obtained results indicate that ApEnf allows determining amplitude variations in biomedical data series. These amplitude variations are better determined when intermediate values of tolerance are used. The study performed in diaphragmatic mechanomyographic signals shows that the ApEnf curve is more correlated with the respiratory effort than the standard RMS amplitude parameter. Furthermore, it has been observed that the ApEnf parameter is less affected by the existence of impulsive, sinusoidal, constant and Gaussian noises in comparison with the RMS amplitude parameter.

Peer Reviewed