Montréal Paris Neurobanque (MPN) 7T MRI Data Processing Pipeline

Overview

This repository hosts scripts and tools for processing and managing high-resolution 7T MRI data as part of the MPN initiative. The aim is to facilitate open data sharing and streamline quality control (QC) and preprocessing using an integrated pipeline that connects LORIS, CBRAIN, and micapipe, following BIDS standards.

Seamlessly manages raw and BIDS-formatted data, facilitating initial QC annotation.	Connects to LORIS to run QC and preprocessing tools, extracting and feeding back QC metrics and initial derivatives.	Performs standardized preprocessing of MRI data and generates derivatives.

Repository Contents

File	Description
README	Detailed documentation on the project's goals, setup instructions, and usage guidelines.
LICENSE	Information on the repository's licensing terms for open-source distribution.
Dockerfile	Configuration to containerize the pipeline for reproducibility and easy deployment.
Functions	Directory with the functions.

Workflow

The data processing workflow begins by transferring raw MRI data, in both BIDS and MINC format, to the LORIS platform. Once uploaded, initial quality control (QC) annotations are performed on LORIS using both automated tools and human evaluations. The data is then linked to CBRAIN, where automated QC metrics are extracted for further analysis. Following this, the QC reports on LORIS are reviewed and classified as either "pass" or "fail." Once the data is approved, it is transferred back to LORIS for preprocessing with micapipe. micapipe then generates initial derivatives while applying additional QC measures to ensure the integrity of the data throughout the entire pipeline.

MRI transfering steps

Option 1. Raw DICOM to to NIfTI BIDS

1. Organizes raw DICOM into a temporary structurated directories
2. Transforms the sorted dicoms into BIDS
3. Run BIDS validator through deno

dcm2bids.py --dicoms_dir MPN00001_sorted/ --bids_dir /BIDS_MPN/rawdata --sub MPN00001 --ses v1

Option 2. Sorted DICOM to NIfTI BIDS

dcm2bids.py --dicoms_dir MPN00001_sorted/ --sorted_dir MPN00001_sorted/ --bids_dir /BIDS_MPN/rawdata --sub MPN00001 --ses v1

3. Integrated BIDS validation

deno run --allow-write -ERN jsr:@bids/validator {bids_dir} --ignoreWarnings --outfile {bids_dir}/bids_validator_output.txt

Running micapipe v0.2.3 with container

mpn_micapipe.sh <subject> <session> <path to singularity image>

Naming dictionary

Anatomical

N	7T Terra Siemens acquisition	BIDS	Directory
1	anat-T1w_acq_mprage_0.8mm_CSptx	T1w	anat
2	anat-T1w_acq-mp2rage_0.7mm_CSptx_INV1	inv-1_MP2RAGE	anat
3	anat-T1w_acq-mp2rage_0.7mm_CSptx_INV2	inv-2_MP2RAGE	anat
4	anat-T1w_acq-mp2rage_0.7mm_CSptx_T1_Images	T1map	anat
5	anat-T1w_acq-mp2rage_0.7mm_CSptx_UNI_Images	UNIT1	anat
6	anat-T1w_acq-mp2rage_0.7mm_CSptx_UNI-DEN	desc-denoised_UNIT1	anat
7	anat-flair_acq-0p7iso_UPAdia	FLAIR	anat
8	CLEAR-SWI_anat-T2star_acq-me_gre_0*7iso_ASPIRE	acq-SWI_T2starw	anat
9	Romeo_P_anat-T2star_acq-me_gre_0*7iso_ASPIRE	acq-romeo_T2starw	anat
10	Romeo_Mask_anat-T2star_acq-me_gre_0*7iso_ASPIRE	acq-romeo_desc-mask_T2starw	anat
11	Romeo_B0_anat-T2star_acq-me_gre_0*7iso_ASPIRE	acq-romeo_desc-unwrapped_T2starw	anat
12	Aspire_M_anat-T2star_acq-me_gre_0*7iso_ASPIRE	acq-aspire_part-mag_T2starw	anat
13	Aspire_P_anat-T2star_acq-me_gre_0*7iso_ASPIRE	acq-aspire_part-phase_T2starw	anat
14	EchoCombined_anat-T2star_acq-me_gre_0*7iso_ASPIRE	acq-aspire_desc-echoCombined_T2starw	anat
15	sensitivity_corrected_mag_anat-T2star_acq-me_gre_0*7iso_ASPIRE	acq-aspire_desc-echoCombinedSensitivityCorrected_T2starw	anat
16	T2star_anat-T2star_acq-me_gre_0*7iso_ASPIRE	T2starw	anat
17	anat-mtw_acq-MTON_07mm	acq-mtw_mt-on_MTR	anat
18	anat-mtw_acq-MTOFF_07mm	acq-mtw_mt-off_MTR	anat
19	anat-mtw_acq-T1w_07mm	acq-mtw_T1w	anat
20	anat-nm_acq-MTboost_sag_0.55mm	acq-neuromelaninMTw_T1w	anat
21	anat-angio_acq-tof_03mm_inplane	angio	anat
22	anat-angio_acq-tof_03mm_inplane_MIP_SAG	acq-sag_angio	anat
23	anat-angio_acq-tof_03mm_inplane_MIP_COR	acq-cor_angio	anat
24	anat-angio_acq-tof_03mm_inplane_MIP_TRA	acq-tra_angio	anat

The acquisitions acq-romeo_part-phase_T2starw, acq-aspire_part-mag_T2starw, and acq-aspire_part-phase_T2starw each have five echoes. The final string will include the identifier echo- followed by the echo number. For example: acq-aspire_echo-1_part-mag_T2starw.

Field maps

N	7T Terra Siemens acquisition	BIDS	Directory
1	fmap-b1_tra_p2	acq-[anat|sfam]_TB1TFL	fmap
2	fmap-b1_acq-sag_p2	acq-[anat|sfam]_TB1TFL	fmap
3	fmap-fmri_acq-mbep2d_SE_19mm_dir-AP	acq-fmri_dir-AP_epi	fmap
4	fmap-fmri_acq-mbep2d_SE_19mm_dir-PA	acq-fmri_dir-PA_epi	fmap

Functional

N	7T Terra Siemens acquisition	BIDS	Directory
1	func-cross_acq-ep2d_MJC_19mm	task-rest_bold	func
2	func-cloudy_acq-ep2d_MJC_19mm	task-cloudy_bold	func
3	func-present_acq-mbep2d_ME_19mm	task-present_bold	func

Each functional MRI acquisition includes three echoes and a phase. The final string will contain the identifier echo- followed by the echo number (e.g., task-rest_echo-1_bold). Additionally, the string part-phase will be included to identify the phase (e.g., task-rest_echo-1_part-phase_bold).

Naming convention | Diffusion weighted Images

N	7T Terra Siemens acquisition	BIDS	Directory
1	*dwi_acq_b0_PA	acq-b0_dir-PA_dwi	dwi
2	*dwi_acq_b0_PA_SBRef	acq-b0_dir-PA_sbref	dwi
3	*dwi_acq_multib_38dir_AP_acc9	acq-multib38_dir-AP_dwi	dwi
4	*dwi_acq_multib_38dir_AP_acc9_SBRef	acq-multib38_dir-AP_sbref	dwi
5	*dwi_acq_multib_70dir_AP_acc9	acq-multib70_dir-AP_dwi	dwi
6	*dwi_acq_multib_70dir_AP_acc9_SBRef	acq-multib70_dir-AP_sbref	dwi

The string part-phase will be included to identify the phase acquisitions (e.g., acq-multib38_dir-AP_part-phase_dwi).

Abbreviation Glossary

Abbreviation	Description
AP	Anterio-Posterior
PA	Postero-anterior
mtw	Magnetic transfer weighted
sfmap	Scaled flip angle map
tof	Time of flight
multib	Multi shell N directions
semphon	Semantic-phonetic
romeo	Rapid opensource minimum spanning tree algorithm
aspire	Combination of multi-channel phase data from multi-echo acquisitions

References

1. Eckstein K, Dymerska B, Bachrata B, Bogner W, Poljanc K, Trattnig S, Robinson SD. Computationally efficient combination of multi‐channel phase data from multi‐echo acquisitions (ASPIRE). Magnetic resonance in medicine. 2018 Jun;79(6):2996-3006. https://doi.org/10.1002/mrm.26963

2. Dymerska B, Eckstein K, Bachrata B, Siow B, Trattnig S, Shmueli K, Robinson SD. Phase unwrapping with a rapid opensource minimum spanning tree algorithm (ROMEO). Magnetic resonance in medicine. 2021 Apr;85(4):2294-308. https://doi.org/10.1002/mrm.28563

3. Sasaki M, Shibata E, Tohyama K, Takahashi J, Otsuka K, Tsuchiya K, Takahashi S, Ehara S, Terayama Y, Sakai A. Neuromelanin magnetic resonance imaging of locus ceruleus and substantia nigra in Parkinson's disease. Neuroreport. 2006 Jul 31;17(11):1215-8. https://doi.org/10.1097/01.wnr.0000227984.84927.a7

Requirements

Package	Version
python	3.8
dcm2niix	1.0.20240202
jq	1.6
bids_validator	2.0.0
deno	2.0.6

Runing singularity

# Define directories
bids=/BIDS_MPN/rawdata/
dicoms=/BIDS_MPN/dicoms

# Path to singularity image
img=<path_to_image>/dcm2bids_v0.1.2.sif

# Define subject and session 
sub=MPNphantom
ses=v1

# Call singularity
singularity run --writable-tmpfs --containall \
	-B ${bids}:/bids -B ${dicoms}:/dicoms \
    ${img} --sub $sub --ses $ses --dicoms_dir /dicoms --sorted_dir /dicoms --bids_dir /bids