Partial (coherence & correlation) estimates of brain autorhythmicity

These experiments try to exhibit the interactions of the electrical activity of the microenvironment of a neuronal population spatially distant from the microenvironment of a different neuronal population. These experiments involve 18 cats with chronic electrodes surgically implanted into cerebral cortex, septal nuclei and amygdala. The Fast Fourier Transform (FFT) algorithm is done to process frequency domain data. In order to compare the phase relations between two channels, x(t) and y(T), use is often made of the coherence function. Computing partial coherence implies first eliminating from each two signals, that part which can be considered as being determined by, or predictable, on the basis of a third signal. It is assumed that there are three stochastic signals Y1, Y2, Y3. These variables can be considered as spectral components for a particular frequency. We assume that the correlations can be interpreted as square root of the coherence. For simplicity we use, in the partial correlations analyses.