3dresample forces the grids to match, but doesn't align the data. It might be useful to find out why @auto_tlrc failed for this dataset. @auto_tlrc should be used with template that is a similar modality to the dataset. If a different type of data is used the alignment to the template won't work well. Also the skull stripping part of the data can fail for some datasets if you are using skull stripping in the @auto_tlrc. You can try skull stripping separately with 3dSkullStrip and then try @auto_tlrc with -no_ss.
You can also align to templates using some align_epi_anat.py or 3dAllineate. For either, you may want to use lpa cost function. See the alignment across sessions page here for some ideas.
[
afni.nimh.nih.gov]
[
afni.nimh.nih.gov]
If all else fails, you can upload the data, and I can take a look.