Thanks for giving good suggestions!
Unfortunately, changing cost functions and several different options of move (cmass, big, giant, and ginormous) also failed to align the partial anatomical 3d image. But, the alignment process was a success by changing another option. In the original script that I used before, I differentiated the whole brain 3d and the partial image by using '-anat' and '-epi' though both images are 3d. After I changed the parts to '-dset1 and -dset2', I think the alignment was successful, as you can see in the attached image. I do not know why the difference was made by changing the parts. Anyway, I would like to appreciate your help, and I hope my experience will be helpful to other AFNI users. I also attached the original and new script.
Have a good day.
############ original script ############
align_epi_anat.py \
-anat ${subj}_expvol+orig \
-epi ${subj}_expvo02l+orig \
-epi2anat \
-suffix _tse_aligned
############ new script ############
align_epi_anat.py \
-dset1 ${subj}_expvol+orig \
-dset2 ${subj}_expvol02_deobl+orig \
-dset2to1 \
-suffix _tse_aligned