AFNI Class Notes 10/11/17:

afni_proc.py

 Example:
set subj = CAJe_110
set top_dir   = /mnt/nfs2/users/alans/CAJe/${subj}BRIKS
afni_proc.py -subj_id $subj \
        -script proc.$subj -scr_overwrite   \
        -blocks despike tshift align tlrc volreg blur mask scale \
    -despike_opts_3dDes -ignore 2 \
        -copy_anat $top_dir/${subj}+orig \
        -dsets $top_dir/${subj}MS+orig.HEAD \
    -tcat_remove_first_trs 2 \
    -tlrc_base MNI_avg152T1+tlrc \
        -volreg_align_to MIN_OUTLIER \
        -volreg_align_e2a  \
        -volreg_tlrc_warp   \
    -blur_in_automask \
        -blur_to_fwhm -blur_size 6.0

Running afni_proc.py produces: proc.CAJe_110

This can be run to create the needed files. Primarily:

File
Description
Blocks
Notes
pb00.CAJe_110.r01.tcat+orig
 combined or cut original functional scan
 tcat  auto: controlled above by
-tcat_remove_first_trs 2
pb01.CAJe_110.r01.despike+orig 3dDespiked file
despike
 option: helps bad data muchly
pb02.CAJe_110.r01.tshift+orig   3dTshift  tshift  option: suggested for event based
pb03.CAJe_110.r01.volreg+tlrc   Fulling aligned file
 align, tlrc, volreg
 required: This is a key step using align_epi_anat.py
 pb04.CAJe_110.r01.blur+tlrc  blurred file
 blur  option: since we used -blur_in_automask -blur_size 4.0 this is blur "to"
 pb05.CAJe_110.r01.scale+tlrc  scaled  scale  option: preps for 3dREMLfit
 CAJe_110_ns+tlrc  anatomical underlay
 tlrc  option: we are in tlrc not mni due to : -tlrc_base TT_N27+tlrc
 outcount_rall.1D  3dToutcount file  outcount  auto: data quality signal. vital to observe to assess signal
 dfile_rall.1D  volreg residuals
 volreg  This is the useful in regression later and data quality now.
 out.mask_ae_overlap.txt  outcome of alignments
 align  Gives info of anat v. epi: %(A \ B) gives the epi outside or anat and if this is really high that is bad news.
 out.mask_ae_dice.txt  outcome of alignments
 align  quick measure of overlap less detailed but can be of use.

plot your timeseries quality files with 1dplot outcount_rall.1D, etc.

You can also get an estimate of movement with SumMove