Dear Yury,
On 3 Aug 2010, at 20:30, Yury Petrov <[log in to unmask]> wrote:
> Do I understand correctly that SPM8 EEG/MEG source localization tools do not have any way to incorporate an individual cortical mesh (created using Freesurfer, for example) and instead morph the MNI-based template cortex for the individual's head?
Individual meshes from freesurfer can be incorporated in theory, but
then it'll be difficult to do statistics across subjects as the
results for different subjects will not appear at the same location.
This is especially true for the latest release where there is much
less smoothing than before so the blobs became very focal. Also group
inversion will not work. Technically incorporating external meshes
will require some familiarity with the code and the data structures
but releasing a tool to make it simple is on my todo list.
> If this is the case, then what transformations are done to achieve this? Is it just an affine transformation based on the scalp shape, > or there is something more intricate involved, like morphing individual sulci and gyri to the MNI cortex, Freesurfer-wise?
>
There is indeed morphing of the brain based on normalization of the
individual structural to the template. See
J. Ashburner and K.J. Friston. Unified segmentation. NeuroImage,
26:839-851, 2005.
The advantage of morphing is that all the individual meshes have the
same vertices and activations can be easily mapped back to standard
space by just using the original mesh to write out an image.
Best,
Vladimir
|