Is this from the Normalise to MNI space option. If so, then the fix is
to change line 240 of spm8/toolbox/DARTEL/spm_dartel_norm_fun.m from:
scal = det(NI.mat(1:3,1:3))/det(NO.mat(1:3,1:3)); %
Account for vox sizes
abs(det(NI.mat(1:3,1:3))/det(NO.mat(1:3,1:3))); % Account for vox sizes
The reason for this (I hope) is that the original images have a
voxel-to-world transform with a positive determinant (so a positive x
voxel size when displayed), whereas the spatially normalised images have
a negative determinant for the voxel-to-world transform (the same as the
template data). Your suggestion therefore makes sense.
The negative value thing will be fixed in the next set of updates.
On Tue, 2009-07-07 at 19:18 +0000, cliff macnab wrote:
> Hi experts,
> I got negative modulated data from DARTEL. Their values at background are close to zero, but negative at brain area. Does the negative value come from negative voxel size or something wrong? (I found the nii images have negative x-direction voxel size).
> Thanks for any explanation,
John Ashburner <[log in to unmask]>