Frequency Spectrum of Each Signal (SEMG and ECG) and Their Respective Means

First, to decompose the signals in their spectrum of frequencies, we used a Wavelet code from a MATLAB environment (R2015b; Natick, MA, United States), Wavelet Packet Decomposition 1-D (“wpec” function, level 6, and wavelet mother symlet14). Symlet14 is a wavelet mother associated with 28 low-pass decomposition filters and 28 high-pass decomposition filters. The decomposition generated a matrix with 64 columns matching the scales and the rows correspondent to the length of the signal vector. Subsequently, we applied the “wpspectrum” function from MATLAB, which returned a matrix of Wavelet Packet Power spectrum based on the Wavelet Packet Transform. Then, for each column containing the scale with their respective frequencies, we calculated the mean of these frequencies that resulted in a row vector 1 × 64. Finally, we included the frequencies on that row vector in ascending order (from scale 1 to scale 64) and determined the center frequency as the mean of this vector (SEMG_MNF). We considered the mean value as representative of the frequency spectrum.

From the ECG signal, the HR was calculated from the reciprocal of R–R time in Hertz.