Study, Based On The Music Eeg Inverse Problem Method And The Cross Path Audio-visual Perception Induced Eeg Experiment

The MUSIC (MUltiple SIgnal Classification) algorithm and its variants are generalized, which attract much attention in the brain source localization problem in recent years. The principles of delay second-order MUSIC and fourth-order cumulant based MUSIC are presented and their better performances of suppressing the noise than canonical second-order MUSIC are shown through simulation.Based on fourth-order MUSIC, a fourth-order cumulant diagonal-slice matrix based MUSIC is developed, which not only enjoys the advantage of suppressing the Gaussian noise but also evidently reduces the calculation burden. The effectiveness of the algorithm is attested through simulation. And with the analysis of the real EEG data from the Neuroscan EEG system involved in MMN problem, its validity is further proved. M7MN problem is therefore discussed. The experimental results show that MIIvIN reflects the partly automatic processing and there exists visual MMN.