Spatiotemporal MEG pattern analysis using a virtual beamformer

In research into induced neuromagnetic fields using multichannel magnetoencephalography (MEG), methods for extracting spatiotemporal information regarding neuronal activity related to a stimulus from the neuromagnetic field data obtained are sought. In this paper, we propose the virtual beamformer as one method for estimating the contribution of specific neuronal activity in neuromagnetic field data. The virtual beamformer method consists of applying directional beamforming to spatiotemporal features of neuromagnetic field data from the observation of specific neuronal activity. We apply this method to auditory evoked magnetic field data resulting from perceptual primitives of sound source movement and confirm that this method can estimate the contribution of specific neuronal activity from such data. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 2, 88(6): 27–33, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjb.20188