Dear Afni experts,
I wanted to use physio and B0 correction on my EPI scans for resting state analysis. I wanted to see if there are any implementations currently (or potentially in the future) for B0 correction within afni?

Now for the main quesiton: I wanted to verify the pipeline in terms of if you suggest motion correcton or slice timing first when using phsyio data/B0 correction.

Right now I was thinking the following:
1) Time lock physio and EPI sequence
2) RetroTS or some sort of program to get the regressors of pulse and respiration(any recommendations on which tool is the most robust?). I assume outputs of the tool would feed into the ricor block inputs. If I wanted to do slice based correction can you tell me the appropriate flag callout (is sb23/RICOR needed for slice based or just a generic pathway from the example in afni_proc.py tutorial?).

3) Afni_proc.py

afni_proc.py -subj_id $sub -script $scrpt_file1 \
-out_dir $out_dir \
-dsets $rest_file \
-copy_anat $anat_file \
-blocks despike ricor tshift volreg regress \
-tcat_remove_first_trs $NP_cat \
-ricor_regs_nfirst $NP_cat \
-ricor_regs sb23/RICOR/r*.slibase.1D \
-tshift_interp -heptic \
-tshift_opts_ts -tpattern seq+z \
-volreg_interp -heptic \
-volreg_base_ind 0 $NP_last \
-regress_censor_motion 0.5 \
-regress_censor_outliers 0.1 \
-regress_apply_mot_types demean deriv

*I plan to regress motion parameters at a later stage after anaticor/bandpass/smoothing with 3dDeconvolve.

4) Run B0 correction using FUGUE or Fieldmap unless AFNI has its own implementation that I am not aware of.

I have heard some people mention motion correction before slice timing correction (therefore despike, volreg, ricor, slice timing) due to the fact that motion has a larger effect on the data vs slice timing for resting state and so I wanted to hear your thoughts on the matter (I believe previous posts suggested slice timing before motion correction so just wanted to make sure this is still the approach you suggest). Also do you suggest running B0 corrections at the end or at another point in the pipeline where its effects can be optimized.


Finally, one of the sequences I might include for each subject (aside from conventional scans) is multi band multiecho resting state data, and if so do you see any changes in the order of steps for this data? Would the ricor block and tshift be able to handle such data or any pitfalls/caveauts I should be aware of ? Is there a way to write the slice order to a nifti header since it is very specific? Finally, would I just use the @slicetimingfile in order pass a text file with the slice order to afni_proc.py? I did not see the info written to the fileheader so I wanted to see if there is a way to verify the correction has taken place.

Thanks,
Ajay



Edited 1 time(s). Last edit at 10/15/2014 04:26PM by AjaySK.