Skip to content

MICA-MNI/MICA-7t

BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

127 Commits
functions
functions
 
 
README.md
README.md
 
 

Repository files navigation

Precision NeuroImaging 7T

Scripts for sorting, organizing and processing the 7T database

0 Running the protocol with Aaron

# 1. log in to Aaron, source and activate conda environments 
source /export02/local/conda/etc/profile.d/conda.sh

# 2. conda envt for Day 1 and 2 
conda activate py38env

# 3. conda envt for Day 3 and 4
conda activate py382env

# 4. go to 7T dir
cd /data/mica3/7T_task_fMRI/from_micaopen/micaopen/7T_task_fMRI

# 5. open the GUI 
python run_tasks.py

# 6. if running Day 3 or 4, open another terminal and repeat the 1-4 steps then

python test_vlc.py

1 . Transfering the data (C

The files from the 7t scan are in /data/dicom/PNC001_Day1_?????. First, find and claim data using find_mri and find_mri -claim script. Then copy 7T data to our folder /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/dicoms.

SUBID=PNC001
ses1=01
ses=ses-${ses1}
find_mri ${SUBID}
find_mri -claim ${dicoms_directory}
mkdir /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}
mkdir /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/beh
mkdir /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/dicoms
mkdir /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/dicoms_sorted
cp -r ${dicoms_directory_returned_from_previous_command} /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/dicoms

2. Sorting the dicoms

This step is to sort the dicoms to /data_/mica3/BIDS_PNI/sorted using the dcmSort script.

dcmSort /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/dicoms /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/dicoms_sorted

3. From sorted dicoms to BIDS

Once the dicoms are sorted we can run the 7t2bids to transform all the dicoms into NIFTIS and rename and organize the files accoding to BIDS.

7t2bids -in /data_/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/dicoms_sorted -id ${SUBID} -bids /data_/mica3/BIDS_PNI/rawdata -ses ${ses1}

4. Copy behavior data

This step is to copy the behavior data from cognitive tasks.

cp -r /data/mica3/7T_task_fMRI/7T_task_fMRI/logs/sub-${SUBID}/$ses/beh/* /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/beh/

Please note that for Day3 we use different folder for now, and the rs-fMRI data is also named as different name (ses-03_2). The script should be therefore replace with:

cp -r /data/mica3/7T_task_fMRI/7T_task_fMRI_NE/logs/sub-${SUBID}/$ses/beh/* /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/beh/
cp -r /data/mica3/7T_task_fMRI/7T_task_fMRI_NE/logs/sub-${SUBID}/${ses}2/beh/* /data/mica3/BIDS_PNI/sorted/sub-${SUBID}_${ses}/beh/

Processing 7T with micapipe

You can run any module of the pipeline locally (-mica), on the mica.q (-qsub) or all.q (-qall). But you should always use one of these flags.

  1. Set singularity environment and directories
#!/bin/bash
# micapipe v0.2.0 "Northern Flicker"

sub=PNC001
ses=01

# Variables
bids=/data/mica3/BIDS_PNI/rawdata
out=/data/mica3/BIDS_PNI/derivatives
tmp=/data/mica2/temporaryNetworkProcessing
fs_lic=/data_/mica1/01_programs/freesurfer-7.3.2/license.txt

# run this container
micapipe_img=/data_/mica1/01_programs/micapipe-v0.2.0/micapipe_v0.2.3.sif

micapipe first stage modules: structural processing

  1. First run the structural processing with the flag -uni for MP2RAGE 7T data
#There are two uni images with 0.5mm and 0.7mm, for our purposes, we only process 0.5mm
# call singularity
singularity run --writable-tmpfs --containall \
	-B ${bids}:/bids \
	-B ${out}:/out \
	-B ${tmp}:/tmp \
	-B ${fs_lic}:/opt/licence.txt \
	${micapipe_img} \
	-bids /bids -out /out -fs_licence /opt/licence.txt -threads 6 -sub ${sub} -ses ${ses} \
	-proc_structural -uni -T1wStr acq-uni_0p5-T1map,acq-inv1_0p5-T1map,acq-inv2_0p5-T1map

Surface processing

  1. Here we run a denoising algorithm on the t1nativepro to enhance contrast in grey/white matter to facilitate the surface generation.

This step might be incorporated into the pipeline in the future but is still work on progress...

# cd to micapipe subject directory
id1=sub-PNC022/ses-01/
id=sub-PNC022_ses-01

Nifti=${id1}/anat/${id/\//_}_space-nativepro_T1w.nii.gz
outStr=${id/\//_}_space-nativepro_T1w_nlm
outdir=${id1}/anat

bash /host/yeatman/local_raid/rcruces/git_here/MRI_analytic_tools/Freesurfer_preprocessing/denoiseN4 $Nifti $outStr $outdir 15
  1. Once the denoised data is ready, run the surface processing module with the -fastsurfer and -T1 flags
# Variables
bids=/data/mica3/BIDS_PNI/rawdata
out=/data/mica3/BIDS_PNI/derivatives
tmp=/data/mica2/temporaryNetworkProcessing
fs_lic=/data_/mica1/01_programs/freesurfer-7.3.2/license.txt

# run this container
micapipe_img=/data_/mica1/01_programs/micapipe-v0.2.0/micapipe_v0.2.3.sif

#make sure to mount this, otherwise, it won't work
t1nlm=${out}/micapipe_v0.2.0/sub-${sub}/ses-${ses}/anat/sub-${sub}_ses-${ses}_space-nativepro_T1w_nlm.nii.gz

# call singularity
singularity run --writable-tmpfs --containall \
	-B ${bids}:/bids \
	-B ${out}:/out \
	-B ${tmp}:/tmp \
	-B ${fs_lic}:/opt/licence.txt \
    -B ${t1nlm}:/opt/T1.nii.gz \
	${micapipe_img} \
	-bids /bids -out /out -fs_licence /opt/licence.txt -threads 6 -sub ${sub} -ses ${ses} \
	-proc_surf -T1 /opt/T1.nii.gz

Run CNN

c/o Donna Note: CNN generated masks should be applied to Fastsurfer before manual QC

To apply the mask:

# 1. Generate the new binary mask from the CNN inference

mask_inference=/host/percy/local_raid/donna/7T_NNunet/new/nnUNet_results/Dataset500_Segmentation/nnUNetTrainer__nnUNetPlans__3d_fullres/inference/PNC_122.nii.gz
fsdir=/data/mica3/BIDS_PNI/derivatives/fastsurfer/sub-PNC022_ses-01

#2. Erase the mask and the norm
rm ${fsdir}/mri/mask.mgz ${fsdir}/mri/norm.mgz

#3. Replace the mask
mri_convert $mask_inference ${fsdir}/mri/mask.mgz

#4.  Multiply the orig_nu.mgz with the inference_mask
mrconvert ${fsdir}/mri/orig_nu.mgz ${fsdir}/mri/orig_nu.nii.gz
fslmaths $mask_inference -mul ${fsdir}/mri/orig_nu.nii.gz ${fsdir}/mri/norm.nii.gz

#5. Convert norm.nii.gz to mgz
mrconvert ${fsdir}/mri/norm.nii.gz ${fsdir}/mri/norm.mgz

#6.  Remove files previouslly created by the first run of recon-surf
rm ${fsdir}/mri/wm.mgz ${fsdir}/mri/aparc.DKTatlas+aseg.orig.mgz ${fsdir}/mri/orig_nu.nii.gz

#7. re-run fastsurfer
sub=PNC022
ses=01
/data/mica1/01_programs/MICA-7t/functions/post-qc_fastsurfer.sh -sub ${sub} -ses ${ses} \
         -out /data_/mica3/BIDS_PNI/derivatives/fastsurfer

Fastsurfer QC

The main outputs of fastsurfer deep volumetric segmentation are found under the mri/ directory: aparc.DKTatlas+aseg.deep.mgz, mask.mgz, and orig.mgz. The equivalent of freesurfer's brainmask.mgz now is called norm.mgz.

Warning!! Please make sure your eraser and brush values when editing are set to zero, otherwise, it will create issues on the subsequent steps.

  1. The edits should be perfom on the mask.mgz file. However, maybe it's easier to correct over the file called norm.mgz. Once the edits are perform you can replace mask.mgz with the binarized version of the corrected norm.mgz.

  2. Run the next script after you are done with the edits. It will create new surfaces based on on the edits and generate a file named qc_done.txt under the subject's directory e.g. fastsurfer/sub-PNA002_ses-01.

sub=PNA002
ses=01
/data/mica1/01_programs/MICA-7t/functions/post-qc_fastsurfer.sh -sub ${sub} -ses ${ses} \
         -out /data_/mica3/BIDS_PNI/derivatives/fastsurfer
post-qc_fastsurfer.sh details

post-qc_fastsurfer.sh will do the next steps:

# Convert from mgz to nifti
mri_convert norm.mgz norm.nii.gz

# --- IF YOU EDIT THE norm.mgz
# Binarize the mask edits
fslmaths norm.nii.gz -thr 1 -uthr 1 -binv mask_edited.nii.gz

# Generate the new mask from the norm.nii.gz edited
fslmaths norm.nii.gz -mul mask_edited.nii.gz -bin mask.nii.gz

# Generate the new norm multiplying the mask
fslmaths norm.nii.gz -mul mask.nii.gz norm.nii.gz

# Replace mask
rm mask.mgz norm.mgz norm.mgz~
mri_convert mask.nii.gz mask.mgz
mri_convert norm.nii.gz norm.mgz

# remove nifitis
rm mask_edited.nii.gz mask.nii.gz norm.nii.gz

# remove files previouslly created by the first run of recon-surf
rm wm.mgz aparc.DKTatlas+aseg.orig.mgz

Run the command recon-surf.sh using a singularity container to generate the new surfaces:

# Subject id
sub=sub-PNA002
ses=ses-01

# output directory
SUBJECTS_DIR=/data_/mica3/BIDS_PNI/derivatives/fastsurfer

# path to singularity image
fastsurfer_img=/data_/mica1/01_programs/fastsurfer/fastsurfer-cpu-v2.0.0.sif

# freesurfer licence
fs_licence=/data_/mica1/01_programs/freesurfer-7.3.2/

# Number of threads for parallel processing
threads=15

# Remove this variable from `env` because it could lead to an error withing the container
unset TMPDIR

# Run only the surface recontruction with spectral spherical projection (fastsurfer default algorithm instead of freesurfer)
singularity exec --nv -B ${SUBJECTS_DIR}/${sub}_${ses}:/data \
                      -B "${SUBJECTS_DIR}":/output \
                      -B "${fs_licence}":/fs \
                       ${fastsurfer_img} \
                       /fastsurfer/recon_surf/recon-surf.sh \
                      --fs_license /fs/license.txt \
                      --t1 /data/mri/orig.mgz \
                      --sid ${sub}_${ses} --sd /output --no_fs_T1 \
                      --parallel --threads ${threads}
                      
# Change the outputs permission, in case that someone else has to work on them
chmod aug+wr -R ${SUBJECTS_DIR}/${sub}_${ses}

touch ${SUBJECTS_DIR}/${sub}_${ses}/qc_done.txt

micapipe second stage modules

One shot processing after reconsurf


bids=/data/mica3/BIDS_PNI/rawdata
out=/data/mica3/BIDS_PNI/derivatives
tmp=/data/mica2/temporaryNetworkProcessing
fs_lic=/data_/mica1/01_programs/freesurfer-7.3.2/license.txt
fsdir=/data/mica3/BIDS_PNI/derivatives/fastsurfer/${sub}_${ses}

# run this container
micapipe_img=/data_/mica1/01_programs/micapipe-v0.2.0/micapipe_v0.2.3.sif

# call singularity
singularity run --writable-tmpfs --containall \
	-B ${bids}:/bids \
	-B ${out}:/out \
	-B ${tmp}:/tmp \
	-B ${fsdir}:${fsdir} \
	-B ${fs_lic}:/opt/licence.txt \
	 ${micapipe_img} -bids /bids -out /out \
	-sub ${sub} -ses ${ses} -proc_surf -surf_dir ${fsdir} -fs_licence /opt/licence.txt -threads 10 \
        -post_structural \
	-proc_dwi -dwi_rpe /bids/${sub}/${ses}/dwi/${sub}_${ses}_acq-b0_dir-PA_epi.nii.gz -regSynth \
	-GD -proc_func \
	-mainScanStr task-rest_run-2_echo-1_bold,task-rest_run-2_echo-2_bold,task-rest_run-2_echo-3_bold \
	-func_pe /bids/${sub}/${ses}/fmap/${sub}_${ses}_acq-fmri_dir-AP_epi.nii.gz \
	-func_rpe /bids/${sub}/${ses}/fmap/${sub}_${ses}_acq-fmri_dir-PA_epi.nii.gz \
	-MPC -mpc_acq T1map -regSynth \
	-microstructural_img /bids/${sub}/${ses}/anat/${sub}_${ses}_acq-T1_0p5-T1map.nii.gz \
	-microstructural_reg FALSE \
	-SC -tracts 40M

cleanup - Change the module name and subject name accordingly

micapipe_img=/data_/mica1/01_programs/micapipe-v0.2.0/micapipe_v0.2.3.sif
bids=/data/mica3/BIDS_PNI/rawdata/
out=/data/mica3/BIDS_PNI/derivatives
fs_lic=/data_/mica1/01_programs/freesurfer-7.3.2/license.txt
tmp=/data/mica2/temporaryNetworkProcessing
sub=sub-PNC009
ses=ses-04
echo "cleaning ${idBIDS} directory"
micapipe_cleanup -sub "${sub}" \
        -ses "${ses}" \
        -bids '/data/mica3/BIDS_PNI/rawdata' \
        -out '/data/mica3/BIDS_PNI/derivatives' \
        -post_structural

Processing times

Module Cores 7T-PNI 3T-MICs CPU
proc_struct 15 122 ± 16 48 ± 10 yes
proc_surf 15 188 ± 36 961 ± 205 yes
post_struct 15 303 ± 41 75 ± 13 yes
proc_func 15 94 ± 8 103 ± 7 yes
proc_dwi 15 ? 184 ± 11 yes
SC 15 ? 918 ± 299 yes
MPC 10 14 ± 3 8 ± 2 no
GD 10 96 ± 21 171 ± 25 yes
proc_flair 10 - 2 ± 0 yes
Total - 818 ± 125 ± -

Processing times diferences between versions

micapipe v0.1.4 v0.2.0 Difference
proc_struct 88 ± 17 48 ± 10 faster
proc_surf 961 ± 205 ~120 faster
post_struct 125 ± 14 75 ± 13 faster
proc_func 101 ± 8 103 ± 7 similar
proc_dwi 246 ± 37 184 ± 11 faster
SC 906 ± 427 918 ± 299 similar
MPC 7 ± 1 8 ± 2 similar
GD 159 ± 21 171 ± 25 slower
Total 2593 ± 730 1627 ± 367 966 ± 363

Derivatives size

Directory size
freesurfer ~830
micapipe/anat ~820M
micapipe/dwi 13G
micapipe/func 24G
micapipe/maps
micapipe/surf
micapipe/logs 31M
micapipe/xfm 2.6G
micapipe/QC 46M
micapipe/ ~10-40G

Rawdata size

Directory size
anat 495M
dwi 1.2G
fmap 15M
func 7.7G
Total 9.4G

7T MRI acquisition protocol

Session Acquisition BIDS dir BIDS name
01/02/03 "*anat-T1w_acq-mprage_07mm_UP" anat acq-mprage_T1w
01/02/03 "*anat-T1w_acq-mp2rage_05mm_UP_INV1" anat acq-inv1_T1map
01/02/03 "*anat-T1w_acq-mp2rage_05mm_UP_INV2" anat acq-inv2_T1map
01/02/03 "*anat-T1w_acq-mp2rage_05mm_UP_T1_Images" anat acq-T1_T1map
01/02/03 "*anat-T1w_acq-mp2rage_05mm_UP_UNI_Images" anat acq-uni_T1map
01/02 "*dwi_acq_b0-dir_PA_SBRef" dwi acq-b0_dir-PA_sbref
01/02 "*dwi_acq_b0-dir_PA" dwi acq-b0_dir-PA_epi
01/02 "*dwi_acq_b2000_90d-dir_AP_SBRef" dwi acq-b2000_dir-AP_sbref
01/02 "*dwi_acq_b2000_90d-dir_AP" dwi acq-b2000_dir-AP_dwi
01/02 "*dwi_acq_b700_40d-dir_AP_SBRef" dwi acq-b700_dir-AP_sbref
01/02 "*dwi_acq_b700_40d-dir_AP" dwi acq-b700_dir-AP_dwi
01/02 "*dwi_acq_b300_10d-dir_AP_SBRef" dwi acq-b300_dir-AP_sbref
01/02 "*dwi_acq_b300_10d-dir_AP" dwi acq-b300_dir-AP_dwi
01/02/03 "*func-rsfmri_acq-mbep2d_ME_19mm" func task-rest_bold
01 "*func-epiencode_acq-mbep2d_ME_19mm" func task-epiencode_bold
01 "*func-epiretrieve_acq-mbep2d_ME_19mm" func task-epiretrieve_bold
01 "*func-pattersep1_acq-mbep2d_ME_19mm" func task-patternsep1_bold
01 "*func-patternsep2_acq-mbep2d_ME_19mm" func task-patternsep2_bold
01/02/03 "*fmap-fmri_acq-mbep2d_SE_19mm_dir-AP" fmap acq-fmri_dir-AP_epi
01/02/03 "*fmap-fmri_acq-mbep2d_SE_19mm_dir-PA" fmap acq-fmri_dir-PA_epi
01/02/03 "*fmap-b1_acq-tra_p2" fmap acq-b1tra_fieldmap
01/02/03 "*fmap-b1_acq-sag_p2" fmap acq-b1sag_fieldmap
02 "*func-semantic1_acq-mbep2d_ME_19mm" func task-semantic1_bold
02 "*func-semantic2_acq-mbep2d_ME_19mm" func task-semantic2_bold
02 "*func-spatial1_acq-mbep2d_ME_19mm" func task-spatial1_bold
02 "*func-spatial2_acq-mbep2d_ME_19mm" func task-spatial2_bold
02/03 "*anat-T2star_acq-me_gre_07mm" anat T2starw
02/03 "*T2Star_Images" anat T2starmap
03 "*func-movies1_acq-mbep2d_ME_19mm" func task-movies1_bold
03 "*func-movies2_acq-mbep2d_ME_19mm" func task-movies2_bold
03 "*func-movies3_acq-mbep2d_ME_19mm" func task-movies3_bold
03 "*func-movies4_acq-mbep2d_ME_19mm" func task-movies4_bold
03 "*anat-angio_acq-tof_03mm_inplane" anat angio
03 "*anat-angio_acq-tof_03mm_inplane_MIP_SAG" anat acq-sag_angio
03 "*anat-angio_acq-tof_03mm_inplane_MIP_COR" anat acq-cor_angio
03 "*anat-angio_acq-tof_03mm_inplane_MIP_TRA" anat acq-tra_angio
03 "*_MTON" anat mt-on_MTR
03 "*_MTOFF" anat mt-off_MTR
03 "*_T1W" anat acq-T1w_MTR

About

scripts for 7t sorting and organizing

Resources

Readme
Activity
Custom properties

Stars

3 stars

Watchers

2 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •  

Languages