Discrimination of multiple sources of alpha brain activity with 3-D MEG measurement

We carried out measurement of magnetic alpha waves using a 195-channel whole-head vector magnetoencephalogram (MEG) system for detecting three magnetic field components Br, B/sub /spl theta//, B/sub /spl phi//. In order to discuss multiple sources of the magnetic alpha wave, a source-estimation algorithm with singular value decomposition (SVD) was used. The SVD method was applied to spatio-temporal data with time window of every 100 ms, and two dominant singular values were obtained for each of the three components. Two singular values with score proportion > 80% were also obtained for both the right and the left hemisphere. With an isofield contour map of the tangential component of the magnetic alpha wave, two extreme values existed. Tangential components (B/sub /spl theta//, B/sub /spl phi//) with SVD were useful for discriminating multiple sources of magnetic alpha waves overlapping in time rather than using only the radial component Br.