The concept of phase coupling has been widely discussed in the literature (e.g., Rosenblum et al., 1996; Pascual-Marqui, 2007a; Stam et al., 2007). In this work, the term “phase coupling” between two signals is meant as the presence of peaks in the distribution of the phase difference across time or signal realizations (e.g., trials, or different segments into which continuous signals can be divided), that reflect preferred values of the phase difference irrespective of the signal amplitudes (see Figure 3). To clarify this concept, we simulated two scenarios depicted in Figure 3. For scenario 1 (Figure 3A), we simulated one-thousand realizations, each of 1-s length and sampled at 512 Hz, of two phase-coupled signals i and j as described in the next sentences. The time series of a 5-Hz oscillator was generated by band-pass filtering white Gaussian noise around 5 Hz with 1 Hz bandwidth. The time series of the signal i, i.e., xi(t), was set to the time series of the 5 Hz oscillator while the time series of the signal j was set to a time-delayed copy of
