Parameter Estimation from the Cross-Spectrogram Reassignment Vectors

In this paper we propose a novel technique to estimate the parameters of two Gaussian envelope oscillatory signals, with the same time-locations and oscillatory frequencies but possibly different phases. The phase difference and the length of the Gaussian envelope are estimated directly from the slopes of the corresponding cross-spectrogram reassignment vectors. Including the phase difference and the envelope length in the scaled reassignment of the cross-spectrogram will give a perfectly concentrated time-frequency spectrum where the location of the maximum gives estimates of the time and oscillatory frequency parameters. The proposed method is evaluated for different SNRs and is also compared to state-of-the-art techniques for phase estimation of oscillatory electrical activity measured from the brain.