Proc. of the Workshop on Information-Based Induction Sciences (IBIS), pp. 223-228, 2003 (in Japanese)
Evaluation for decomposed brain activities of single-trial and averaged multiple-trials MEG data
Y. Konno, J. Cao, T. Arai, T. Takeda and H. Endo
Abstract: Independent component analysis (ICA) has been applied to electroencephalographic (EEG) or magnetoencephalographic (MEG) data to determine the behavior and localization of brain sources. In this paper, to investigate the performance of source decomposition of single-trial and averaged multiple-trials MEG data, we apply our ICA algorithm to synthesized data. The main advantage of our synthesized data set is that dipole location of evoked responses and its dynamics are known in advance, we demonstrate and evaluate the results of source decomposition of averaged MEG data.