Dear experts,
We perform resting state preprocessing in volume until nuisance regression step included. After this, we perform surface analysis to calculate FC. Now, we are going to start task-fMRI experiments and we would like to know what is the recommended moment to project from volume to surface in a task-fMRI pipeline.
On the other hand, is it possible to unlink the first level analysis step from nuisance regression in order to calculate FC with task-fMRI data?
Thank you very much in advance.
Best regards,
Marina
Hi, Marina-
There examples for these cases in the afni_proc.py help file.
For example, for surface analysis with task-based FMRI:
https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/programs/afni_proc.py_sphx.html#s03-ap-surface-class-demo-basic-surface-analysis
Resting state surface analysis could be done similarly. A modern example of resting state analysis with surface projection is provided as part of the afni_proc.py multi-echo resting state FMRI demo, which can be downloaded and unpacked with:
@Install_APMULTI_Demo1_rest
The full AP command looks like this (see the script for what the variables are and how they can be populated:
afni_proc.py \
-subj_id ${subj} \
-blocks despike tshift align volreg mask combine surf blur scale \
regress \
-radial_correlate_blocks tcat volreg \
-copy_anat ${anat_cp} \
-anat_has_skull no \
-anat_follower anat_w_skull anat ${anat_skull} \
-anat_follower_ROI aaseg anat ${roi_all_2009} \
-anat_follower_ROI aeseg epi ${roi_all_2009} \
-anat_follower_ROI FSvent epi ${roi_FSvent} \
-anat_follower_ROI FSWe epi ${roi_FSWe} \
-anat_follower_erode FSvent FSWe \
-surf_anat ${surf_anat} \
-surf_spec ${surf_specs} \
-blip_forward_dset "${epi_forward}" \
-blip_reverse_dset "${epi_reverse}" \
-dsets_me_run ${dsets_epi_me} \
-echo_times ${me_times} \
-combine_method OC_tedort \
-combine_tedort_reject_midk no \
-tcat_remove_first_trs ${nt_rm} \
-tshift_interp -wsinc9 \
-align_opts_aea -cost lpc+ZZ -giant_move -check_flip \
-volreg_align_to MIN_OUTLIER \
-volreg_align_e2a \
-volreg_warp_final_interp wsinc5 \
-volreg_compute_tsnr yes \
-blur_size ${blur_size} \
-mask_epi_anat yes \
-regress_motion_per_run \
-regress_make_corr_vols aeseg FSvent \
-regress_censor_motion ${cen_motion} \
-regress_censor_outliers ${cen_outliers} \
-regress_apply_mot_types demean deriv \
-html_review_style pythonic
This assumes you ran FreeSurfer previously to get the mesh. You could use FreeSurfer ROIs for your surface analysis. If you don’t have multi-echo FMRI data, you could remove those parts—namely, the combine block, and the “-echo_times …” and “-combine_* …” options. The “-blip_* …” options are because the demo also has reveres phase encoded EPIs, for B0 distortion correction. If you don’t have that kind of data, you could also leave out those options.
–pt