When you spatially normalise, there is an option on the user interface where you can define the voxel sizes and bounding box of your spatially normalised images.  The default settings of 2mm isotropic voxels with a range from -78mm to 78mm in the x direction, -112mm to 76mm in the y direction and -70mm to 85mm in the z direction (all relative to the AC) give spatially normalised images of size 79 x 95 x 79 voxels.

If you want to have the same dimensions as the template data to which the images are matched, then you can set the voxel sizes to [NaN NaN NaN] and the bounding box to ones(2,3)*NaN.  When values of "not a number" are specified, SPM will try to do something sensible in terms of defining a bounding box.

Alternatively you could specify the voxel sizes and bounding box of the template data:

>> [bb,vx]=spm_get_bbox(fullfile(spm('dir'),'tpm','TPM.nii'))
bb =
   -90  -126   -72
    90    90   108
vx =
   -1.5000    1.5000    1.5000

Best regards,

On 8 September 2015 at 15:23, SPM Anonymous <[log in to unmask]> wrote:

I've EPI-images with the dimension [88 100 40].
When I normalize these data with spm12 sets I get [79 95 79]
and if I want to have them in the template (ICBM-space) I've to reslice
them in order to get [121 145 121].

What space is [79 95 79]? Is there a template/brain-mask for this space?