| |
Abstract:
The interpretation of magneto- and electroencephalography
(MEG, EEG) data is limited by the difficulty of determining the
source currents in the brain. Functional MRI can be used to inform
the MEG/EEG analysis by suggesting likely locations of the sources.
However, because of the possibility of a mismatch in the activity
seen by these methods, fMRI cannot be used as an exclusive
constraint [1,2]. We propose a novel MEG/EEG source estimate, which
is constructed by selecting from the set of all MEG/EEG compatible
solutions the one that minimizes the distance to an fMRI-defined
subspace of the source space. This method was evaluated by
simulations and compared with a generalized Wiener estimate, in
which fMRI modifies the prior source covariance matrix [2]. The
methods performed similarly when fMRI foci were at the correct
locations of the MEG sources. However, the new method was found to
be less sensitive to simultaneous false positive and false negative
fMRI activations. The new method appears potentially useful in
combined fMRI and MEG/EEG studies of the dynamical patters of
activation in the human brain. [1] Ahlfors, S.P., et al. J.
Neurophysiol. 82, 2545 (1999). [2] Liu, A.K., Belliveau, J.W. &
Dale, A.M. PNAS 95, 8945 (1998).
|
