The wavelet-based phase coherence classification decomposes signals such as the local field potential in electrophysiological recordings into subsignals depending on the pairwise phase-coherence relations. It then computes the power spectral densities of these subsignals thus allowing a detailed spectral analysis. For a detailed description see von Papen et al., Phase-coherence classification: a new wavelet-based method to separate local field potentials into local (in)coherent and volume-conducted components, Journal of Neuroscience Methods, 2017.
The wavelet transform is applied using a modified code originally provided by Torrence & Compo. The modified code allows to perform the wavelet analysis for any given frequency instead of a logarithmically equidistant set of frequencies.
plot_synth_data.m generates a PCC of synthetic data as presented in Fig. 7 of von Papen et al. (2017). The example requires the mseb plotting function by Andreas Trier Poulsen.
Michael von Papen, Felix Gerick
GNU