1) Matlab files (.mat) may be read in Python using the Scipy library, loadmat method : scipy.io.loadmat(filename).

2) In order to build the spectrogram images shown in Fig. 3, the following Matlab function was used:

% create a time vector that has the same number of point (90000)
% Sampling rate = 2000Hz
timevec = -5:1/2000:40-(1/2000)

%load the file
load('f03_B_001_Time-frequency-matrix-global-WT.mat')

% generate the spectrogram
contourf(timevec, linspace(0, 200, 200), tf_WT, 60, 'linecolor', 'none');