Thanks for the quick response.

I found that the problem was the input. If I use the EPI that it is coregistered to the T1 of Freesurfer instead of the EPI that it is also normalized to the MNI space, it works and I can project the surface (that I created with the command that I sent you) on fsaverage 5.

Could you confirm that the command has no errors?

mri_vol2surf --mov EPI.nii --reg register.dat --hemi [l/r]h --projfrac 0.5 --interp trilinear --trgsubject ico --icoorder 5 --o EPI_surface.mgh

Best wishes,
Marina