fMRIPrep Tutorial #6: Group Analysis


Overview

Our group-level analysis requires that we preprocess all of the individual subjects and run 1st-level analyses for each of them. To do that, we will put our preprocessing and 1st-level analysis code in a for-loop.

Modifying the fmriprep.sh Script

Previously, we used the fmriprep.sh script that had a hard-coded subject number. To put this code into a for-loop, we will need to include this line of code just below the first line:

subject=$1

And modify the “subj” line to the following:

subj=$subject

You can make the changes yourself if you like, and save the file as fmriprep_Scripted.sh; otherwise, you can download the script for here. Also make sure to download the script doDecon.sh to your Flanker directory.

Running the for-loop

To create our for-loop, we will simply combine all of the code we ran in the previous chapters, making sure that they are in the correct order. Assuming that you already ran the script “make_Timings.sh”, you should already have a file in your Flanker directory called subjList.txt, which contains a list of all the subjects in your study.

The combined code will look like this:

for i in `cat subjList.txt`; do
  cd code;
  bash fmriprep_Scripted.sh $i;
  cd ../derivatives/fmriprep/${i}/func
  for run in 1 2; do
    3dmerge -1blur_fwhm 4.0 -doall -prefix r${run}_blur.nii \
        ${i}_task-flanker_run-${run}_space-MNI152NLin2009cAsym_res-2_desc-preproc_bold.nii.gz;
    done
  for run in 1 2; do
    3dTstat -prefix rm.mean_r${run}.nii r${run}_blur.nii;
    3dcalc -a r${run}_blur.nii -b rm.mean_r${run}.nii \
       -c ${i}_task-flanker_run-${run}_space-MNI152NLin2009cAsym_res-2_desc-brain_mask.nii.gz                            \
       -expr 'c * min(200, a/b*100)*step(a)*step(b)'       \
       -prefix r${run}_scale.nii;
    done
    rm rm*;
    3dmask_tool -inputs *mask.nii.gz -union -prefix full_mask.nii
  mkdir stimuli;
  cp ../../../../${i}/func/*.1D stimuli;
  cp ../../../../doDecon.sh .;
  for reg in trans_x trans_y trans_z rot_x rot_y rot_z; do
    for run in 1 2; do
     awk -v col=$reg 'NR==1{for(i=1;i<=NF;i++){if($i==col){c=i;break}} print $c} NR>1{print $c}' ${i}_task-flanker_run-${run}_desc-confounds_regressors.tsv > ${reg}_run${run}_all_tmp.txt;
     sed '1d' ${reg}_run${run}_all_tmp.txt > ${reg}_run${run}_noHead_tmp.txt
     sed '1!d' ${reg}_run${run}_all_tmp.txt > ${reg}_run${run}_Head_tmp.txt
    done
  done
  NT=`3dinfo -nt r1_scale.nii`
  if [ -f zeros_tmp.txt ]; then rm zeros_tmp.txt; fi
  for ((j=0; j<$NT; j++)); do echo 0 >> zeros_tmp.txt; done
  for reg in trans_x trans_y trans_z rot_x rot_y rot_z; do
    for run in 1 2; do
      if [ $run -eq 1 ]; then
        cat ${reg}_run${run}_noHead_tmp.txt > ${reg}_run${run}.txt
      else
        cat zeros_tmp.txt ${reg}_run${run}_noHead_tmp.txt > ${reg}_run${run}.txt
      fi
    done
  done
  rm *tmp*
  tcsh doDecon.sh $i
  cd ../../../..
done

Each block of this code is discussed in the previous chapters; the only difference is that we now use variables for input where appropriate. You can either copy and paste this code into the Terminal while in the Flanker directory, or download this script from Andy’s GitHub page (again, making sure to run it from the Flanker directory). This will take a long time to run - maybe a few days, depending on the speed of your computer.

Running the Group Analysis

When the script has finished, do the same quality assurance checks that we covered in the previous chapters. If everything looks OK, then we are ready to do a group-level analysis using the same approach as in the AFNI tutorial. The code will look like this:

#!/bin/bash -xef

dirA=derivatives/fmriprep

# specify and possibly create results directory
results_dir=test.results
if [ ! -d $results_dir ]; then mkdir $results_dir; fi

# ------------------------- process the data -------------------------

3dttest++ -prefix $results_dir/Flanker-Inc-Con                     \
          -setA Inc-Con                                            \
             01 "$dirA/sub-01/func/stats.sub-01+tlrc[7]" \
             02 "$dirA/sub-02/func/stats.sub-02+tlrc[7]" \
             03 "$dirA/sub-03/func/stats.sub-03+tlrc[7]" \
             04 "$dirA/sub-04/func/stats.sub-04+tlrc[7]" \
             05 "$dirA/sub-05/func/stats.sub-05+tlrc[7]" \
             06 "$dirA/sub-06/func/stats.sub-06+tlrc[7]" \
             07 "$dirA/sub-07/func/stats.sub-07+tlrc[7]" \
             08 "$dirA/sub-08/func/stats.sub-08+tlrc[7]" \
             09 "$dirA/sub-09/func/stats.sub-09+tlrc[7]" \
             10 "$dirA/sub-10/func/stats.sub-10+tlrc[7]" \
             11 "$dirA/sub-11/func/stats.sub-11+tlrc[7]" \
             12 "$dirA/sub-12/func/stats.sub-12+tlrc[7]" \
             13 "$dirA/sub-13/func/stats.sub-13+tlrc[7]" \
             14 "$dirA/sub-14/func/stats.sub-14+tlrc[7]" \
             15 "$dirA/sub-15/func/stats.sub-15+tlrc[7]" \
             16 "$dirA/sub-16/func/stats.sub-16+tlrc[7]" \
             17 "$dirA/sub-17/func/stats.sub-17+tlrc[7]" \
             18 "$dirA/sub-18/func/stats.sub-18+tlrc[7]" \
             19 "$dirA/sub-19/func/stats.sub-19+tlrc[7]" \
             20 "$dirA/sub-20/func/stats.sub-20+tlrc[7]" \
             21 "$dirA/sub-21/func/stats.sub-21+tlrc[7]" \
             22 "$dirA/sub-22/func/stats.sub-22+tlrc[7]" \
             23 "$dirA/sub-23/func/stats.sub-23+tlrc[7]" \
             24 "$dirA/sub-24/func/stats.sub-24+tlrc[7]" \
             25 "$dirA/sub-25/func/stats.sub-25+tlrc[7]" \
             26 "$dirA/sub-26/func/stats.sub-26+tlrc[7]"

As explained in the AFNI group analysis chapter, this code will run a one-sample t-test on the Inc-Con beta weights from each subject’s stats file. The output will be placed in the folder test.results, and you can examine them in the AFNI viewer:

../../_images/06_GroupResults_fMRIPrep.png

Compared to the results from AFNI, the cluster appears to be slightly larger, and with a higher peak:

../../_images/06_GroupResults_AFNI.png

As you can see, the basic pattern of the results did not change, but the results did appear to get slightly stronger.

Next Steps

Now that we have seen how fMRIPrep can improve our results, we can take a closer look at some of the advanced options, which we will cover in the next chapter.

Video

For a video demonstration of scripting and group analysis, click here.