Three functions that estimate the complex coherency coefficients using Fourier decomposition of vector X and Y. Functions differ in the way the the cross and auto spectra are smoothed.
-tfcohf: auto and cross spectra are smoothed with identical smoothing kernels.
-tfcohf2: only the auto spectra are smoothed ( ensemble averaged spectra)
-tfcohf3: auto and cross spectra are smoothed with different smoothing kernels.
Using identical smoothing operators (tfcohf) is the gold standard and this estimator is bound to [0,1]. Using non-identical smoothing operators (tfcohf2 and tfcohf3), the estimator is not bound to [0,1], but may be useful for specific purposes. Gaussian smoothing kernels can be 1D (only temporal) or 2D (smoothing in time and frequency domain).
Functions have similar setup and variable names as functions from the Signal Processing Toolbox. Functions are very basic and can be applied to different types of signals.
The package is associated with the following reference:
Mehrkanoon S, Breakspear M, Daffertshofer A, Boonstra TW (2013). Non-identical smoothing operators for estimating time-frequency interdependence in electrophysiological recordings. EURASIP Journal on Advances in Signal Processing 2013, 2013:73. doi:10.1186/1687-6180-2013-73
In the case you found the code useful for your work, please cite the above reference in your publications. |