> Currently I am working on a VBM project comparing
> childhood brain tumor survivors with normal
> controls. The lesion region resulting from resection
> caused distortion during spatial normalization.
> To solve this problem, one methos is to use
> affine registration only and the other is to mask
> out the lesion region. I am wondering if there
> is any other solution for this problem. I
> searched the SPM achives, but very few
> documents were found for this issue.
> Where can I find more sources for this problem?
There's no solution yet, but I suspect that if the spatial normalisation used
an objective function that was more like mutual information, then it would be
more able to cope with tumours and other lesions.
Best regards,
-John
|