Could you try doing "Coregister (Est & Res)" in FMRI part of SPM to your individual structurals first and then use the resliced image (with 'r' prefix) to build the head model?

Vladimir

On Wed, Sep 4, 2013 at 12:06 PM, dandil dNd <[log in to unmask]> wrote:

Dear Vladimir,

I have not measured electrod locations and I am using only individual's MRI images. SPM identifies fiducial automatically. It is noticeable that when I am using template structural images instead, using the same ERP data, registration results seem correct. I do not know what is wrong with me.

Thank you for your attention.

Bests,

Dandil