MNI space does not have a specific output dimension. MNI space simply encodes the origin, rotation, and brain size in mm. 

When you normalize an image, you estimate the parameters to get your image to match a template, then you write out the normalized image. The estimate and write step are separate. In the write step, you can specify the bounding box and voxel size. Thus, the output is whatever you have specified in the write step.

Best Regards, 
Donald McLaren, PhD

On Mon, Sep 28, 2015 at 3:08 PM, Priya Aggarwal <[log in to unmask]> wrote:

I am trying to do normalization of T2-weighted images using SPM8. I have used EPI.nii template file. After normalization it produces images of size 79*95*68.

Although as far as i know, MNI template normalised images should be of size 91*109*91. Can someone help does MNI templates sizes differ?  Wfu_pickatlas mask is also MNI registered and having size 91*109*91.

So why EPI.nii MNI template give different size?

Many thanks in advance,

With regards,