Abnormal structure image segmentation in rats

[Ernest861]

I used BIDScoin to convert the T2 structure image it looks ok,

then I found an error during processing, I don't know what the error is, can any friend help me to check it?

The T2 info:
“{
"Modality": "MR",
"MagneticFieldStrength": 9.4,
"ImagingFrequency": 400.627,
"Manufacturer": "UIH",
"ManufacturersModelName": "uMR 9.4T",
"DeviceSerialNumber": "H00011",
"StationName": "UIHMR",
"BodyPartExamined": "RATBRAIN",
"PatientPosition": "HFS",
"SoftwareVersions": "R002",
"MRAcquisitionType": "2D",
"SeriesDescription": "t2_fse_tra_0.1x0.6mm",
"ProtocolName": "t2_fse_tra_0.1x0.6mm",
"ScanningSequence": "SE",
"SequenceVariant": "NONE",
"SequenceName": "fse",
"ImageType": [
"ORIGINAL",
"PRIMARY",
"M",
"IF"
],
"SeriesNumber": 602,
"AcquisitionTime": "14:32:12.286000",
"SliceThickness": 0.6,
"SpacingBetweenSlices": 0.96,
"NumberOfAverages": 5,
"EchoTime": 0.0462,
"RepetitionTime": 3,
"FlipAngle": 180,
"PhaseEncodingDirectionDisplayed": "A->P",
"EchoTrainLength": 13,
"AcquisitionMatrixPE": 290,
"ReconMatrixPE": 580,
"TotalReadoutTime": 342.066,
"PixelBandwidth": 220,
"PhaseEncodingAxis": "j",
"ImageOrientationPatientDICOM": [
1,
0,
0,
0,
1,
0
],
"InPlanePhaseEncodingDirectionDICOM": "COL",
"ConversionSoftware": "dcm2niix",
"ConversionSoftwareVersion": "v1.0.20220505"
}”

[gdevenyi]

Please provide the exact commands you used to run the pipeline as well as all the logs.

[Ernest861]

Here is the commands: Whether using --anat_autobox or --anat_inho_cor method=SyN,otsu_thresh=2,multiotsu=true, the image is such that the structural states are enlarged and incorrectly delineated.

	 Running RABIES - version: 0.5.0
240513-03:43:41,837 nipype.workflow INFO:
	 CLI INPUTS: 
-> rabies_stage = preprocess 
-> plugin = MultiProc 
-> local_threads = 6 
-> scale_min_memory = 1.0 
-> min_proc = 1 
-> figure_format = png 
-> verbose = 1 
-> bids_dir = /test_dataset 
-> output_dir = /test_dataset_outputs 
-> bold_only = False 
-> anat_autobox = True 
-> bold_autobox = False 
-> apply_despiking = False 
-> HMC_option = optim 
-> isotropic_HMC = False 
-> voxelwise_motion = False 
-> apply_slice_mc = False 
-> detect_dummy = False 
-> data_type = float32 
-> anat_inho_cor = {'method': 'SyN', 'otsu_thresh': '2', 'multiotsu': False} 
-> anat_robust_inho_cor = {'apply': True, 'masking': True, 'brain_extraction': True, 'template_registration': 'SyN'} 
-> bold_inho_cor = {'method': 'Rigid', 'otsu_thresh': '2', 'multiotsu': False} 
-> bold_robust_inho_cor = {'apply': False, 'masking': False, 'brain_extraction': False, 'template_registration': 'SyN'} 
-> commonspace_reg = {'masking': False, 'brain_extraction': False, 'template_registration': 'SyN', 'fast_commonspace': False} 
-> bold2anat_coreg = {'masking': False, 'brain_extraction': False, 'registration': 'SyN'} 
-> nativespace_resampling = inputs_defined 
-> commonspace_resampling = inputs_defined 
-> anatomical_resampling = 0.15x0.15x0.15 
-> apply_STC = False 
-> TR = auto 
-> tpattern = alt-z 
-> stc_axis = Y 
-> interp_method = fourier 
-> anat_template = /home/rabies/.local/share/rabies/DSURQE_40micron_average.nii.gz 
-> brain_mask = /home/rabies/.local/share/rabies/DSURQE_40micron_mask.nii.gz 
-> WM_mask = /home/rabies/.local/share/rabies/DSURQE_40micron_eroded_WM_mask.nii.gz 
-> CSF_mask = /home/rabies/.local/share/rabies/DSURQE_40micron_eroded_CSF_mask.nii.gz 
-> vascular_mask = /home/rabies/.local/share/rabies/vascular_mask.nii.gz 
-> labels = /home/rabies/.local/share/rabies/DSURQE_40micron_labels.nii.gz 

attached log
rabies_preprocess.log

Thank you!

[gdevenyi]

-> anat_template = /home/rabies/.local/share/rabies/DSURQE_40micron_average.nii.gz

This is a mouse template, you need to select the SIGMA rat template.

[Ernest861]

Taking a hint from the forums Using SIGMA instead of Dsurge I used atlas from sigm and got the same error as @KimMinhHuynh.

Traceback (most recent call last):
File "/home/rabies/miniforge/bin/rabies", line 7, in
exec(compile(f.read(), file, 'exec'))
File "/home/rabies/RABIES/scripts/rabies", line 3, in
execute_workflow()
File "/home/rabies/RABIES/rabies/run_main.py", line 44, in execute_workflow
workflow = preprocess(opts, log)
File "/home/rabies/RABIES/rabies/run_main.py", line 202, in preprocess
check_template_overlap(opts.anat_template, opts.CSF_mask)
File "/home/rabies/RABIES/rabies/run_main.py", line 336, in check_template_overlap
raise ValueError(
ValueError: The file /atlas_folder/SIGMA_Rat_Anatomical_Imaging/SIGMA_Rat_Anatomical_InVivo_Template/SIGMA_InVivo_CSF_mask.nii does not appear to overlap with provided template /atlas_folder/SIGMA_Rat_Anatomical_Imaging/SIGMA_Rat_Anatomical_InVivo_Template/SIGMA_InVivo_Brain_Template.nii.

the log file:
`c240516-09:19:31,688 nipype.workflow INFO:
Running RABIES - version: 0.5.0
240516-09:19:31,689 nipype.workflow INFO:
CLI INPUTS:
-> rabies_stage = preprocess
-> plugin = MultiProc
-> local_threads = 6
-> scale_min_memory = 1.0
-> min_proc = 1
-> figure_format = png
-> verbose = 1
-> bids_dir = /test_dataset
-> output_dir = /test_dataset_outputs
-> bold_only = False
-> anat_autobox = True
-> bold_autobox = False
-> apply_despiking = False
-> HMC_option = optim
-> isotropic_HMC = False
-> voxelwise_motion = False c
-> apply_slice_mc = False
-> detect_dummy = False
-> data_type = float32
-> anat_inho_cor = {'method': 'SyN', 'otsu_thresh': '2', 'multiotsu': False}
-> anat_robust_inho_cor = {'apply': True, 'masking': True, 'brain_extraction': True, 'template_registration': 'SyN'}
-> bold_inho_cor = {'method': 'Rigid', 'otsu_thresh': '2', 'multiotsu': False}
-> bold_robust_inho_cor = {'apply': False, 'masking': False, 'brain_extraction': False, 'template_registration': 'SyN'}
-> commonspace_reg = {'masking': False, 'brain_extraction': False, 'template_registration': 'SyN', 'fast_commonspace': False}
-> bold2anat_coreg = {'masking': False, 'brain_extraction': False, 'registration': 'SyN'}
-> nativespace_resampling = inputs_defined
-> commonspace_resampling = inputs_defined
-> anatomical_resampling = 0.15x0.15x0.15
-> apply_STC = False
-> TR = auto
-> tpattern = alt-z
-> stc_axis = Y
-> interp_method = fourier
-> anat_template = /atlas_folder/SIGMA_Rat_Anatomical_Imaging/SIGMA_Rat_Anatomical_InVivo_Template/SIGMA_InVivo_Brain_Template.nii
-> brain_mask = /atlas_folder/SIGMA_Rat_Anatomical_Imaging/SIGMA_Rat_Anatomical_InVivo_Template/SIGMA_InVivo_Brain_Mask.nii
-> WM_mask = /atlas_folder/SIGMA_Rat_Anatomical_Imaging/SIGMA_Rat_Anatomical_InVivo_Template/SIGMA_InVivo_WM_mask.nii
-> CSF_mask = /atlas_folder/SIGMA_Rat_Anatomical_Imaging/SIGMA_Rat_Anatomical_InVivo_Template/SIGMA_InVivo_CSF_mask.nii
-> vascular_mask = /atlas_folder/SIGMA_Rat_Anatomical_Imaging/SIGMA_Rat_Anatomical_InVivo_Template/SIGMA_InVivo_Brain_Mask.nii
-> labels = /atlas_folder/SIGMA_Rat_Brain_Atlases/SIGMA_Anatomical_Atlas/InVivo_Atlas/SIGMA_InVivo_Anatomical_Brain_Atlas.nii

240516-09:19:31,689 nipype.workflow INFO:
INPUT BIDS DATASET:
test_dataset/
.bidsignore
sub-2/
sub-2_scans.tsv
anat/
sub-2_task-t2fsetra01x06mm_acq-t2fsetra01x06mm_echo-1_T2w.nii.gz
sub-2_task-t2fsetra01x06mm_acq-t2fsetra01x06mm_echo-1_T2w.json
testjoson/
func/
sub-2_task-epibold2shtra015x06mm_bold.nii.gz
sub-2_task-epibold2shtra015x06mm_bold.json
sub-9/
sub-9_scans.tsv
anat/
sub-9_task-t2fsetra01x06mm_acq-t2fsetra01x06mm_echo-1_T2w.nii.gz
sub-9_task-t2fsetra01x06mm_acq-t2fsetra01x06mm_echo-1_T2w.json
func/
sub-9_task-epibold2shtra03x06mm_bold.nii.gz
sub-9_task-epibold2shtra03x06mm_bold.json
`

[gdevenyi]

If you collecting your scans on a bruker, we recommend brkraw for conversion to ensure proper orientation (provided the orientation details were set properly in paravision)

[Ernest861]

Thank you both very much for the information, I collect the image by united-imaging uMR9.4T preclinical system.
I convert them from dicom to nii by bidscoin.
I gaze they use the cdm2niix2bids to convert, but i am not sure the parameter.

[gdevenyi]

dcm2niix excepts the DICOMs to be standards compliant regarding orientation of data. I am not aware of any options which can override orientation information. If something is wrong on NIFTI output, here are the possibilities.

1. The DICOMs have the wrong orientation info due to the DICOM export being wrong
2. The information regarding the animal was set wrong on the scanner (quadraped, Supine/Prone/Head-First/Feet First), but the DICOM export is correct

I suggest you do a few things

1. Check your scanner console settings are correct for the animal you scan
2. Check that the DICOM metadata regarding the type of sample (animal, positioning/orienation) is correct with something such as dcmdump
3. Open up the NIFTIs in itksnap to check their orientation. Correctly specified data will have the anatomical A-P, S-I and L-R labels in the display be correct with respect to the animal scan.

[Ernest861]

Thanks for the advice, I'll do all the steps on this one.

The processing after manual reorientation is complete and it seems that the first short BOLD sequence is more correct and the second BOLD fails. I would like to confirm that the signal to noise ratio of the image is viewed temporal_features.png, is it possible to extract the values.

the first session

the second session

attached file
rabies_preprocess.log
rabies_analysis.log
preprocess_QC_report.zip

[Gab-D-G]

This is correct, session 1 looks like registration worked, session 2 failed. You can use the tSNR map for session 1, although it looks like there are only 18 frames in your session 1, I'm not sure that's intended.