I notice that you said you wanted to process the data with AFNI, so I will suggest an AFNI-based processing solution, using “afni_proc.py” for your preprocessing (which in afni_proc.py includes alignment, motion correction, blurring, etc. as well as GLM regression modeling).
I am not sure how the slignment ended up as such here, because I don’t know about the processing tool you used, the options used, nor its internal details. You can see the afni_proc.py help here: https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/programs/afni_proc.py_sphx.html#ahelp-afni-proc-py
… which has examples to try, e.g., Ex. 9 or Ex. 11 for rest. If you have task here, you might want to check out the AFNI Bootcamp example here (the demo data is downloadable, so you can follow along with the tutorial):