fMRI-constrained MEG source imaging and consideration of fMRI invisible sources

Authors

  • Chang-Hwan Im,

    Corresponding author
    1. School of Electrical Engineering and Computer Science, Seoul National University, Korea
    2. Brain Information Group, National Institute of Information and Communications Technology (NICT), Japan
    • School of Electrical Engineering, and Computer Science, Seoul National University, San 56-1, Shillim-dong, Kwanak-gu, Seoul, 151-744, Korea
    Search for more papers by this author
  • Hyun-Kyo Jung,

    1. School of Electrical Engineering and Computer Science, Seoul National University, Korea
    Search for more papers by this author
  • Norio Fujimaki

    1. Brain Information Group, National Institute of Information and Communications Technology (NICT), Japan
    Search for more papers by this author

Abstract

Recent studies on multimodal brain source imaging have shown that the use of functional MRI (fMRI) prior information could enhance spatial resolution of magnetoencephalography (MEG), while MEG could compensate poor temporal resolution of fMRI. This article deals with a multimodal imaging method, which combines fMRI and MEG for enhancing both spatial and temporal resolutions. Recent studies on the combination of fMRI and MEG have suggested that the fMRI prior information could be very easily implemented by just giving different weighting factors to the diagonal terms of source covariance matrix in linear inverse operator. We applied the fMRI constrained imaging method to several simulation data and experimental data (Japanese language lexical judgment experiment), and found that some MEG sources may be eliminated by the introduction of the fMRI weighting and the eliminated sources may affect source estimation in fMRI activation regions. In this article, in order to check whether the eliminated sources were fMRI invisible ones or just spurious ones, we placed small numbers of regional sources (rotating dipoles) around all possible activation regions and investigated their temporal changes. By investigating the results carefully, we could evaluate whether the missed sources were real or not. Hum Brain Mapp, 2005. © 2005 Wiley-Liss, Inc.

Ancillary