Abstract. The multichannel signal analysis method, based on precise Fourier transformation, was used in study of magneto encephalography data, acquired in different experimental conditions, and electroencephalography data. Same experiment with the same person was conducted on two different types of magnetometers and on multichannel electroencephalograph. All experimental time series was processed, using the same data analysis technique, based on Fourier transformation of long time series and fine frequency grid tuning to the known external stimulus frequency. Such approach made it possible to determine the set of basis functions, optimal for the particular experiment, and maximize the signal to noise ratio. Magnetic field patterns, calculated at harmonics of stimulus frequency are almost noise-free and allows us to solve the inverse problem of filed sources spatial structure reconstruction. It was found, that the usage of proposed data analysis method makes it possible to conduct encephalography experiments in conditions of strong external noises.

 

Key words: electroencephalography, magnetoencephalography, multichannel signal processing, signal quality, signal identification algorithms.