Speeding-Up MEG Beamforming Source Imaging by Correlation between Measurement and Lead-Field Vector

Abstract

Beamforming technique is promising in that it can easily deal with spatiotemporal multi-dipole problems and is robust to noise. One of weak points is to scan the whole source region, thereby requiring some amount of computing time. We introduce the method to speed-up beamforming source imaging. In order to speed-up localization as well as to keep good spatial resolution, we propose the way to reduce the number of scanning points by using the correlation between measurement and lead-field vector. It is found that this idea improves beamforming speed up to about 63% in computation time without sacrificing accuracy even though computing correlation requires some computing time. Through Monte Carlo simulation, correlation threshold depending on signal-to-noise ratio (SNR) is reasonably estimated in a least square sense.