I did the procedure to create a template from scratch by following the steps explaining in this page : http://surfer.nmr.mgh.harvard.edu/fswiki/SurfaceRegAndTemplates

The objective was to create a template to use it in the command recon-all –qcache –target my template.

So, after creating mytemp2.tif I did mris_register for the subjects used for the template (80 subjects) and then I used make_average_subject --subjects <80 subjects> --out mytemplate. Finally I ran recon-all -qcache –target mytemplate in order to visualize my analyzes in QDEC. This procedure seems correct to you because when I just visually compared your template (fsaverage) with mine it looks like very different, as if mine was more smoothed (cf ppt file) ? What do you think ?