Dear Paul,

I am sorry for bothering you again. I have pre-processed my data, but wanted to make sure that what I did makes sense. Please see below for how I specified the afni_proc.py -- I modified the pre-processing a bit compared to what was suggested in the NI paper: I am pre-processing the data separately for each run whilst regressing out the polynomial degrees (up the third order given the average run length of 380 à default: 1 + floor(380 / 150.0) = 3) to cover scanner drifts. As recommended, I don’t do band pass filtering. I have further modified the -regress_censor_motion to be 0.3 instead of 0.2 based on some resting state recommendations I found in the afni_proc.py help. Afterwards, I take the final output file (the residual files errts*) and concatenate and reshuffle the volumes relating to each video clip without demeaning them. The output of this step is then used to compute the ISC. Is this halfway reasonable? Do you think I should demean the volumes for each magic trick before reshuffling/concatenating?

# specify actual afni_proc.py
afni_proc.py -subj_id $subj.magictrickwatching_perRun \
-blocks despike tshift align tlrc volreg blur mask regress \
-copy_anat $fsindir/$fsanat \
-anat_follower_ROI aaseg anat $fsindir/aparc.a2009s+aseg_rank.nii \
-anat_follower_ROI aeseg epi $fsindir/aparc.a2009s+aseg_rank.nii \
-anat_follower_ROI FSvent epi $fsindir/$fsvent \
-anat_follower_ROI FSWMe epi $fsindir/$fswm \
-anat_follower_erode FSvent FSWMe \
-dsets $epi_dpattern \
-tcat_remove_first_trs 0 \
-tlrc_base /usr/share/afni/atlases/MNI152_T1_2009c+tlrc \
-tlrc_NL_warp \
-volreg_align_to MIN_OUTLIER \
-volreg_align_e2a \
-volreg_tlrc_warp \
-regress_ROI_PC FSvent 3 \
-regress_make_corr_vols aeseg FSvent \
-regress_anaticor_fast \
-regress_anaticor_label FSWMe \
-regress_censor_motion 0.3 \
-regress_censor_outliers 0.1 \
-regress_apply_mot_types demean deriv \
-regress_est_blur_epits \
-regress_est_blur_errts \
-regress_run_clustsim no \
-regress_polort 3

One more question arises: In addition to the movie clip watching, we also did a pre and post task resting state scan and I am interested in changes in the resting state functional connectivity due to the task. Can I preprocess that resting state data in the same way? Would I still include both runs (i.e. pre and post task) into the same afni_proc.py script or would I treat them independently?

Any help with these issues is highly appreciated!
Many thanks and best regards,
Stef