Estimation of high-frequency components of the brain electrical activity by inhomogeneity indices

The qualitative techniques are considered of mathematical description of curve inhomogeneity which adequately reveal non-stationary fragments in EEG samples with a pseudo-random process. The introduced inhomogeneity coefficient turned to be a sufficiently adequate indicator of the presence of short-term high-frequency bursts in the electrical activity of dog's brain and a sensitive index for discrimination of quasihomogeneous fragments by the level of ''general activation''. It is shown that the values of this coefficient reflect the individual features of the brain electrical activity of the animals and its dynamics in the process of alimentary motor conditioning. At the same time, this index is ineffective in revealing the regional differences. A novel approach is developed for signal inhomogeneity description which is based on the traditional harmonic analysis (Fourier transform).