Development of MEG data analysis system by independent component analysis using parallel computing

摘要

ICA, which stands for Independent Component Analysis, has the capability of extracting statistically independent data streams from multiple MEG data streams. In this research. the application of ICA to actual MEG data and simulation data (sine and saw waves) has been explored. The results show that ICA is effective and so promising for extracting noise signals and actual brain signals from MEG data streams. Also, analysis and visualization system for MEG using ICA has been developed through this research. The system can offer not only the speedup in analysis time but also intuitive image of the results. The system would have the potential to go up the efficiency in evaluation for brain function.