Detailed file descriptions¶
This document present the specifications of the CAPS format version 1.0.0
.
Versions of Clinica, BIDS, and CAPS specifications¶
Clinica version | BIDS version supported | CAPS version supported |
---|---|---|
0.9.x | 1.7.0 | 1.0.0 |
0.8.x | 1.7.0 | no version |
The dataset_description.json file¶
Specifications¶
This file MUST be present at the root of a CAPS dataset and MUST contain the following minimal information:
{
"Name": "name identifier for the dataset",
"BIDSVersion": "1.7.0",
"CAPSVersion": "1.0.0",
"DatasetType": "derivative"
}
Name
: String identifier of the dataset. It can be the name of your study for example. By default Clinica generates a random UUID for this field. When running a pipeline which will create a new CAPS dataset, you can use the--caps-name
option to provide a name. If the CAPS dataset already exist, the existing name will be kept.BIDSVersion
: The version number of the BIDS specifications that the BIDS input dataset is using when this CAPS dataset was generated.CAPSVersion
: The version number of the CAPS specifications used for this dataset.DatasetType
: Either "raw" or "derivative". For a CAPS dataset this should always be "derivative" as it contains processed data.
In addition, the dataset_description.json
file MAY contain a Processing
key which is a list of objects describing the different processing pipelines that were run on this CAPS.
Here is an example for a CAPS dataset containing the outputs of two pipelines: t1-linear
and pet-linear
:
{
"Name": "e6719ef6-2411-4ad2-8abd-da1fd8fbdf32",
"BIDSVersion": "1.7.0",
"CAPSVersion": "1.0.0",
"DatasetType": "derivative",
"Processing": [
{
"Name": "t1-linear",
"Date": "2024-08-06T10:28:21.848950",
"Author": "ci",
"Machine": "ubuntu",
"InputPath": "/mnt/data_ci/T1Linear/in/bids"
},
{
"Name": "pet-linear",
"Date": "2024-08-06T10:36:27.403373",
"Author": "ci",
"Machine": "ubuntu",
"InputPath": "/mnt/data_ci/PETLinear/in/bids"
}
]
}
A Processing
is described with the following fields:
Name
: The name of the processing. For Clinica pipelines, this is the name of the pipeline.Date
: This date is in iso-format and indicates when the processing was run.Author
: This indicates the user name which triggered the processing.Machine
: This indicates the name of the machine on which the processing was run.InputPath
: This is the full path (on the machine on which the processing was run) to the input dataset of the processing.
Potential problems¶
The dataset_description.json
file for CAPS datasets was introduced in Clinica 0.9.0
.
This means that results obtained with prior versions of Clinica do not have this file automatically generated.
Clinica will interpret this as a <1.0.0
dataset and should error with a suggestion of a minimal dataset_description.json
that you should add to your dataset.
In this situation, create this new file with the suggested content and re-start the pipeline.
You might also see the following error message:
Impossible to write the 'dataset_description.json' file in <FOLDER> because it already exists and it contains incompatible metadata.
This means that you have version mismatch for the BIDS and/or CAPS specifications. That is, the versions indicated in the input (or output) dataset(s) does not match the versions currently used by Clinica. If this happens, it is recommended to re-run the conversion or the pipeline which initially generated the dataset with the current version of Clinica.
Subjects and Groups folders¶
In the following, brackets [
/]
will denote optional key/value pairs in the filename, while accolades {
/}
will indicate a list of compulsory values (e.g. hemi-{left|right}
means that the key hemi
only accepts left
or right
as values).
Finally:
participant_id
denotesparticipant-<participant_label>
;session_id
denotessession-<session_label>
;group_id
denotesgroup-<group_label>
;long_id
denoteslong-<long_label>
.
T1 MRI data¶
t1-linear
- Affine registration of T1w images to the MNI standard space¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ t1_linear/
├─ <source_file>_space-MNI152NLin2009cSym_res-1x1x1_affine.mat
├─ <source_file>_space-MNI152NLin2009cSym_res-1x1x1_T1w.nii.gz
└─ <source_file>_space-MNI152NLin2009cSym_desc-Crop_res-1x1x1_T1w.nii.gz
The desc-Crop
indicates images of size 169×208×179 after cropping to remove the background.
t1-volume
pipeline - Volume-based processing of T1-weighted MR images¶
Segmentation¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ t1/
└─ spm/
└─ segmentation/
├─ normalized_space/
│ ├─ <source_file>_target-Ixi549Space_transformation-{inverse|forward}_deformation.nii.gz
│ ├─ <source_file>_segm-<segm>_space-Ixi549Space_modulated-{on|off}_probability.nii.gz
│ └─ <source_file>_space-Ixi549Space_T1w.nii.gz
├─ native_space/
│ └─ <source_file>_segm-<segm>_probability.nii.gz
└─ dartel_input/
└─ <source_file>_segm-<segm>_dartelinput.nii.gz
- The
modulated-{on|off}
key indicates if modulation has been used in SPM to compensate for the effect of spatial normalization. - The possible values for the
segm-<segm>
key/value are:graymatter
,whitematter
,csf
,bone
,softtissue
, andbackground
. - The T1 image in
Ixi549Space
(reference space of the TPM) is obtained by applying the transformation obtained from the SPM segmentation routine to the T1 image in native space.
DARTEL¶
dataset_description.json
groups/
└─ <group_id>/
├─ <group_id>_subjects_visits_list.tsv
└─ t1/
├─ <group_id>_iteration-<index>_template.nii.gz
└─ <group_id>_template.nii.gz
- The final group template is
<group_id>_template.nii.gz
. - The
<group_id>_iteration-<index>_template.nii.gz
obtained at each iteration will only be used when obtaining flow fields for registering a new image into an existing template (SPM DARTEL existing templates procedure).
Note for SPM experts
The original name of <group_id>_iteration-<index>_template.nii.gz
is Template<index>.nii
.
DARTEL to MNI¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ t1/
└─ spm/
└─ dartel/
└─ <group_id>/
├─ <source_file>_target-<group_label>_transformation-forward_deformation.nii.gz
└─ <source_file>_segm-<segm>_space-Ixi549Space_modulated-{on|off}[_fwhm-<X>mm]_probability.nii.gz
Atlas statistics¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ t1/
└─ spm/
└─ dartel/
└─ <group_id>/
└─ atlas_statistics/
└─ <source_file>_space-<space>_map-graymatter_statistics.tsv
Statistics files (with _statistics.tsv
suffix) are detailed in appendix.
t1-freesurfer
- FreeSurfer-based processing of T1-weighted MR images¶
The outputs of the t1-freesurfer
pipeline are split into two sub-folders, the first one containing the FreeSurfer outputs and a second with additional outputs specific to Clinica.
FreeSurfer outputs:
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ t1/
└─ freesurfer_cross_sectional/
└─ <participant_id>_<session_id>/
├─ label/
├─ mri/
├─ scripts/
├─ stats/
└─ surf/
Clinica additional outputs:
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ t1/
└─ freesurfer_cross_sectional/
└─ regional_measures/
├─ <source_file>_parcellation-wm_volume.tsv
├─ <source_file>_segmentationVolumes.tsv
└─ <source_file>_hemi-{left|right}_parcellation-<parcellation>_thickness.tsv
For the file: *_hemi-{left|right}_parcellation-<parcellation>_thickness.tsv
, _thickness
is just an example for the cortical thickness, we also have other measurements defined in the table below:
Name | Suffix | Description |
---|---|---|
Cortical thickness | _thickness |
Cortical thickness between pial surface and white surface |
Cortical volume | _volume |
Volume of gray matter |
Cortical surface area | _area |
Cortical surface area |
Cortical mean curvature | _meancurv |
Mean curvature of cortical surface |
- The
hemi-{left|right}
key/value stands forleft
orright
hemisphere. - The possible values for the
parcellation-<parcellation>
key/value are:desikan
(Desikan-Killiany Atlas),destrieux
(Destrieux Atlas) andba
(Brodmann Area Maps). - The TSV files for Brodmann areas contain a selection of regions, see this link for the content of this selection: http://ftp.nmr.mgh.harvard.edu/fswiki/BrodmannAreaMaps).
The details of the white matter parcellation of FreeSurfer can be found here: https://surfer.nmr.mgh.harvard.edu/pub/articles/salat_2008.pdf.
Example - Content of the TSV files
Content of sub-CLNC01_ses-M000_T1w_segmentationVolumes.tsv
:
Measure:volume Left-Lateral-Ventricle Left-Inf-Lat-Vent ...
/path/to/freesurfer/segmentation/ 12345.6 12.334 ...
This file contains the volume of the different subcortical structures after segmentation.
Content of sub-CLNC01_ses-M000_T1w_parcellation-wm_volume.tsv
:
Measure:volume wm-lh-bankssts wm-lh-caudalanteriorcingulate ...
/path/to/freesurfer/wm parcellation/ 2474.6 1863.7 ...
Content of sub-CLNC01_ses-M000_hemi-left_parcellation-desikan_thickness.tsv
:
lh.aparc.thickness lh_bankssts_thickness lh_caudalanteriorcingulate_thickness …
/path/to/freesurfer/cortical thickness/parcellation 2.048 2.892 …
t1-freesurfer-longitudinal
– FreeSurfer-based longitudinal processing of T1-weighted MR images¶
FreeSurfer unbiased templates¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <long_id>/
└─ freesurfer_unbiased_template/
└─ <participant_id>_<long_id>/
├─ label/
├─ mri/
├─ scripts/
├─ stats/
└─ surf/
FreeSurfer longitudinal outputs¶
The outputs are split into two sub-folders, the first containing the FreeSurfer longitudinal outputs and a second with additional outputs specific to Clinica.
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ t1/
└─ <long_id>/
└─ freesurfer_longitudinal/
└─ <participant_id>_<session_id>.long.<participant_id>_<long_id>/
├─ label/
├─ mri/
├─ scripts/
├─ stats/
└─ surf/
Clinica additional outputs:
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ t1/
└─ <long_id>/
└─ freesurfer_longitudinal/
└─ regional_measures/
├─ <participant_id>_<session_id>_<long_id>_parcellation-wm_volume.tsv
├─ <participant_id>_<session_id>_<long_id>_segmentationVolumes.tsv
└─ <participant_id>_<session_id>_<long_id>_hemi-{left|right}_parcellation-<parcellation>_thickness.tsv
where each file is explained in the t1-freesurfer
subsection.
Note
The naming convention <subject_name>.long.<template_name>
is imposed by FreeSurfer.
Diffusion imaging data¶
dwi-preprocessing-*
- Preprocessing of raw diffusion weighted imaging (DWI) datasets¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ dwi/
└─ preprocessing/
├─ <source_file>_space-<space>_desc-preproc_dwi.bval
├─ <source_file>_space-<space>_desc-preproc_dwi.bvec
├─ <source_file>_space-<space>_desc-preproc_dwi.nii.gz
└─ <source_file>_space-<space>_brainmask.nii.gz
The resulting DWI file after preprocessing.
According to the subtype of pipeline run, <space>
can be:
A brain mask of the preprocessed file is provided.
dwi-dti
- DTI-based processing of corrected DWI datasets¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ dwi/
└─ dti_based_processing/
├─ native_space/
│ ├─ <source_file>_space-<space>_model-DTI_diffmodel.nii.gz
│ ├─ <source_file>_space-<space>_{FA|MD|AD|RD}.nii.gz
│ └─ <source_file>_space-<space>_DECFA.nii.gz
│─ normalized_space/
│ ├─ <source_file>_space-MNI152Lin_res-1x1x1_affine.mat
│ ├─ <source_file>_space-MNI152Lin_res-1x1x1_deformation.nii.gz
│ └─ <source_file>_space-MNI152Lin_res-1x1x1_{FA|MD|AD|RD}.nii.gz
└─ atlas_statistics/
└─ <source_file>_space-<space>_res-1x1x1_map-{FA|MD|AD|RD}_statistics.tsv
The DTI is saved under the *_model-DTI_diffmodel.nii.gz
filename.
The different maps based on the DTI are:
FA
: fractional anisotropy.MD
: mean diffusivity.AD
: axial diffusivity.RD
: radial diffusivity.DECFA
: directionally-encoded colour (DEC) FA.
Current atlases used for statistics are the 1mm version of JHUDTI81
, JHUTract0
and JHUTract25
(see Atlases page for further details).
Statistics files (with _statistics.tsv
suffix) are detailed in appendix.
Note
The naming convention for suffixes follows the BIDS derivative specifications except for the statistics files (specific files for our needs) and the _deformation.nii.gz
file (it is equivalent to the _warp.nii.gz
file in the BEP014 specifications).
dwi-connectome
- Computation of structural connectome from corrected DWI datasets¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ dwi/
└─ connectome_based_processing/
├─ <source_file>_space-{b0|T1w}_model-CSD_diffmodel.nii.gz
│─ <source_file>_space-{b0|T1w}_model-CSD_tractography.tck
└─ <source_file>_space-{b0|T1w}_model-CSD_parcellation-{desikan|destrieux}_connectivity.tsv
- The constrained spherical deconvolution (CSD) diffusion model is saved under the
*_model-CSD_diffmodel.nii.gz
filename. - The whole-brain tractography is saved under the
*_tractography.tck
filename. - The connectivity matrices are saved under
*_connectivity.tsv
filenames.
Current parcellations used for the computation of connectivity matrices are desikan
and desikan
(see Atlases page for further details).
PET imaging data¶
pet-volume
- Volume-based processing of PET images¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ pet/
└─ preprocessing/
└─ <group_id>/
├─ <source_file>_space-T1w_pet.nii.gz
├─ <source_file>_space-T1w[_pvc-rbv]_pet.nii.gz
├─ <source_file>_space-Ixi549Space[_pvc-rbv]_pet.nii.gz
├─ <source_file>_space-Ixi549Space[_pvc-rbv]_suvr-<suvr>_pet.nii.gz
├─ <source_file>_space-Ixi549Space_brainmask.nii.gz
└─ <source_file>_space-Ixi549Space[_pvc-rbv]_suvr-<suvr>_mask-brain_pet.nii.gz
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ pet/
└─ preprocessing/
└─ atlas_statistics/
└─ <source_file>_space-<space>[_pvc-rbv]_suvr-<suvr>_statistics.tsv
- The
_trc-<label>
key/value describes the radiotracer used for the PET acquisition (currently supported:18FFDG
and18FAV45
). - The
[_pvc-rbv]
label is optional, depending on whether your image has undergone partial volume correction (region-based voxel-wise (RBV) method) or not. - The possible values for the
suvr-<suvr>
key/value are:pons
for FDG-PET andcerebellumPons
for different types of amyloid PET.
Statistics files (with _statistics.tsv
suffix) are detailed in appendix.
pet-surface
- Surface-based processing of PET images¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ pet/
└─ surface/
├─ atlas_statistics/
│ └─ <source_file>_trc-<label>_pet_space-<space>_pvc-iy_suvr-<suvr>_statistics.tsv
├─ <source_file>_hemi-{left|right}_midcorticalsurface
└─ <source_file>_hemi-{left|right}_trc-<label>_pet_space-<space>_suvr-<suvr>_pvc-iy_hemi-{left|right}_fwhm-<label>_projection.mgh
- The
trc-<label>
key/value describes the radiotracer used for the PET acquisition (currently supported:18FFDG
and18FAV45
). - The
[_pvc-iy]
label describes the partial volume correction used in the algorithm for the projection (Iterative Yang). - The possible values for the
suvr-<suvr>
key/value are:pons
for FDG-PET andcerebellumPons
for different types of amyloid PET. - The
fwhm
represents the FWHM (in mm) of the Gaussian filter applied to the data mapped onto the FsAverage surface. The different values are 0 (no smoothing), 5, 10, 15, 20, and 25. - Files with the
_midcorticalsurface
suffix represent the surface at equal distance between the white matter/gray matter interface and the pial surface (one per hemisphere). - Files with the
projection
suffix are PET data that can be mapped onto meshes. If_space_fsaverage_
is in the name, it can be mapped either onto the white or pial surface of FsAverage. If_space_native_
is in the name, it can be mapped onto the white or pial surface of the subject’s surface ({l|r}h.white
,{l|r}h.pial
files fromt1-freesurfer
pipeline). - Files with the
statistics
suffix are text files that display average PET values on either_space-desikan
or_space-destrieux
atlases.
Example of this content can be found in appendix.
pet-surface-longitudinal
- Surface-based longitudinal processing of PET images¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ pet/
└─ <long_id>/
└─ surface_longitudinal/
├─ atlas_statistics/
│ └─ sub-<label>_ses-<lalbel>_long-<label>_trc-<label>_pet_space-<space>_pvc-iy_suvr-<suvr>_statistics.tsv
├─ sub-<label>_ses-<lalbel>_long-<label>_hemi-{left|right}_midcorticalsurface
└─ sub-<label>_ses-<lalbel>_long-<label>_trc-<label>_pet_space-<space>_suvr-<suvr>_pvc-iy_hemi-{left|right}_fwhm-<label>_projection.mgh
Explanations on the key/values can be found on the
pet-surface
section.
pet-linear
- Affine registration of PET images to the MNI standard space¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ pet_linear/
├─ <source_file>_space-T1w_rigid.mat
├─ <source_file>_space-T1w_pet.nii.gz
├─ <source_file>_space-MNI152NLin2009cSym_res-1x1x1_suvr-<suvr_label>_pet.nii.gz
└─ <source_file>_space-MNI152NLin2009cSym_desc-Crop_res-1x1x1_suvr-<suvr_label>_pet.nii.gz
The desc-Crop
indicates images of size 169×208×179 after cropping to remove the background.
Statistics¶
statistics-surface
- Surface-based mass-univariate analysis with SurfStat¶
Group comparison¶
dataset_description.json
groups/
└─ <group_id>/
└─ statistics/
├─ participants.tsv
└─ surfstat_group_comparison/
├─ <group_id>_<group_1>-lt-<group_2>_measure-<measure>_fwhm-<label>_correctedPValue.jpg
├─ <group_id>_<group_2>-lt-<group_1>_measure-<measure>_fwhm-<label>_correctedPValue.jpg
├─ <group_id>_<group_1>-lt-<group_2>_measure-<measure>_fwhm-<label>_correctedPValue.mat
├─ <group_id>_<group_2>-lt-<group_1>_measure-<measure>_fwhm-<label>_correctedPValue.mat
├─ <group_id>_participants.tsv
├─ <group_id>_output.log
└─ <group_id>_glm.json
In the case above, _correctedPValue
indicates that these are maps of corrected p-values.
Other types of maps include, but are not limited to:
Name | Suffix | Description |
---|---|---|
Corrected p-value | _correctedPValue |
Corrected P-values for vertices and clusters level based on random field theory |
Uncorrected p-value | _uncorrectedPValue |
Uncorrected P-value for a generalized linear model |
T-statistics | _TStatistics |
T statistics for a generalized linear model |
FDR | _FDR |
Q-values for False Discovery Rate of resels |
The <group_1>-lt-<group_2>
means that the tested hypothesis is: the measurement of <group_1>
is lower than (lt
) that <group_2>
.
Value for measure
can be ct
(cortical thickness from t1-freesurfer
), fdg
(from pet-surface
) or user-defined maps.
The value for fwhm
corresponds to the size of the surface-based smoothing in mm and can be 5
, 10
, 15
or 20
.
The JPEG files are simple snapshots.
The *.mat
files can be read later by tools like PySurfer and Surfstat.
Example
groups/
└─ group-ADvsHC/
└─ statistics/
├─ participants.tsv
└─ surfstat_group_comparison/
├─ group-ADvsHC_AD-lt-HC_measure-ct_fwhm-20_correctedPValue.jpg
├─ group-ADvsHC_participants.tsv
├─ group-ADvsHC_output.log
└─ group-ADvsHC_glm.json
Group comparison between patients with Alzheimer’s Disease (group_1
= AD
) and healthy subjects (group_2
= HC
).
ADvsHC
defines the group_label
.
The group-ADvsHC_glm.json
contains the information for your generalized linear model, for example:
{
"DesignMatrix": "1 + age + sex + group",
"StringFormatTSV": "%s %f %f",
"Contrast": "group",
"ClusterThreshold": 0.001
}
This file describes the model that you want to create, you should include the factor and covariates in your generalized linear model as a column name in this TSV file.
For example, the linear model formula is: CorticalThickness = 1 + age + sex + group
, the contrasts (factors) group
, age
and sex
are the covariates.
Additional information is included in the log file.
The content of group-ADvsHC_participants.tsv
is:
participant_id session_id sex group age
sub-CLNC0001 ses-M000 Female CN 71.1
sub-CLNC0002 ses-M000 Male CN 81.3
sub-CLNC0003 ses-M000 Male CN 75.4
sub-CLNC0004 ses-M000 Female CN 73.9
sub-CLNC0005 ses-M000 Female AD 64.1
sub-CLNC0006 ses-M000 Male AD 80.1
sub-CLNC0007 ses-M000 Male AD 78.3
sub-CLNC0008 ses-M000 Female AD 73.2
(Note that to make the display clearer, the rows contain successive tabs, which should not happen in an actual TSV file.)
The <group_id>
key/value stands for the group_label
for your analysis.
It can be used to run different analyses for different subjects or different analyses for the same subjects.
The example image here maps statistically significant differences in cortical thickness between a group of patients with Alzheimer’s disease and a group of healthy controls (yellow: correction at the vertex level; blue: correction at the cluster level).
Correlation analysis¶
dataset_description.json
groups/
└─ <group_id>/
└─ statistics/
├─ participants.tsv
└─ surfstat_correlation_analysis/
├─ <group_id>_correlation-<label>_contrast-{negative|positive}_measure-<measure>_fwhm-<label>_correctedPValue.jpg
├─ <group_id>_correlation-<label>_contrast-{negative|positive}_measure-<measure>_fwhm-<label>_correctedPValue.mat
├─ <group_id>_output.log
├─ <group_id>_participants.tsv
└─ <group_id>_glm.json
- The
correlation-<label>
here describes the factor of the model which can be, for example,age
. - The
contrast-{negative|positive}
is the sign of the correlation you want to study, which can benegative
orpositive
. - All other key/value pairs are defined in the same way as in the previous section.
statistics-volume
- Volume-based mass-univariate analysis with SPM¶
dataset_description.json
groups/
└─ <group_id>/
├─ <group_id>_participants.tsv
└─ statistics_volume/
└─ group_comparison_measure-{graymatter|18FFDG|18FAV45|<custom_user>}/
├─ <group_id>_{RPV|mask}.nii
├─ <group_id>_covariate-<covariate>_measure-<label>_fwhm-<n>_regressionCoefficient.nii
├─ <group_id>_<group_1>-lt-<group_2>_measure-<label>_fwhm-<n>_{TStatistics|contrast}.nii
├─ <group_id>_<group_2>-lt-<group_1>_measure-<label>_fwhm-<n>_{TStatistics|contrast}.nii
├─ <group_id>_report-1.png
├─ <group_id>_report-2.png
└─ <group_id>_{mask|RPV|VarianceError}.nii
Suffixes are described in the table below:
Name | Suffix | Description |
---|---|---|
T-statistics | _TStatistics |
T statistics for a generalized linear model |
Resels per voxels | _RPV |
Image of the estimated resels per voxels (known as RPV.nii in SPM) |
Variance of the error | _VarianceError |
Image of the variance of the error (known as ResMS.nii in SPM) |
Weighted parameters | _contrast |
Image of the estimated weighted parameters (known as con_000X.nii in SPM) |
Regression coefficient | _regressionCoefficient |
Image of the estimated regression coefficient for beta_000X.nii in SPM) |
Report | _report-{1|2} |
SPM report containing FWE/FDR peak/cluster thresholds to report for subsequent corrections |
The <group_1>-lt-<group_2>
means that the tested hypothesis is: "the measurement of <group_1>
is lower than (lt
) that of <group_2>
".
The value for measure
can be graymatter
(output of t1-volume
), 18FFDG
or 18FAV45
(output of pet-volume
), or user-defined maps.
The value for fwhm
corresponds to the size of the volume-based smoothing in mm.
Corrected results are stored under the following hierarchy:
dataset_description.json
groups/
└─ <group_id>/
└─ statistics_volume/
└─ group_comparison_measure-<label>/
└─ <group_id>_<group_1>-lt-<group_2>_measure-<label>_fwhm-<n>_{FDRc|FDRp|FWEc|FWEc}
├─ <group_id>_<group_1>-lt-<group_2>_measure-<label>_fwhm-<n>_desc-{FDRc|FDRp|FWEc|FWEc}_axis-{x|y|z}_TStatistics.png
└─ <group_id>_<group_1>-lt-<group_2>_measure-<label>_fwhm-<n>_desc-{FDRc|FDRp|FWEc|FWEc}_GlassBrain.png
FWEp
(resp.FDRp
) corresponds to correction for multiple comparisons with family-wise error (FWE) (resp. false discovery rate [FDR]) correction at the peak (=voxel) level with a statistical threshold of P < 0.05.FWEc
(resp.FDRc
) corresponds to correction for multiple comparisons with family-wise error (FWE) (resp. false discovery rate [FDR]) correction.
A statistical threshold of P < 0.001 was first applied (height threshold). An extent threshold of P < 0.05 corrected for multiple comparisons was then applied at the cluster level.
Machine Learning¶
machinelearning-prepare-spatial-svm
- Prepare input data for spatially regularized SVM¶
dataset_description.json
subjects/
└─ <participant_id>/
└─ <session_id>/
└─ machine_learning/
└─ input_spatial_svm/
└─ <group_id>/
├─ <source_file_t1w>_segm-{graymatter|whitematter|csf}_space-Ixi549Space_modulated-on_spatialregularization.nii.gz
└─ <source_file_pet>_space-Ixi549Space[_pvc-rbv]_suvr-<suvr>_spatialregularization.nii.gz
dataset_description.json
groups/
└─ <group_id>/
└─ machine_learning/
└─ input_spatial_svm/
├─ <group_id>_space-Ixi549Space_gram.npy
└─ <group_id>_space-Ixi549Space_parameters.json
At the subject level, it contains SVM regularization of gray matter/white matter/CSF maps or PET data that accounts for the spatial and anatomical structure of neuroimaging data.
At the group level, it contains the Gram matrix with respect to gray matter/white matter/CSF maps needed for the SVM regularization and the information regarding the regularization. An example of JSON file is:
{
"MaxDeltaT": "0.0025",
"Alpha": "0.0025",
"Epsilon": "10E-6",
"BoundaryConditions": "TimeInvariant",
"SigmaLoc": "10",
"TimeStepMax": "0.07760115580830161",
"SpatialPrior": "Tissues (GM,WM,CSF)",
"RegularizationType": "Fisher",
"FWHM": "4"
}
Appendix - Content of a statistic file¶
<source_file>_space-<space>_map-<map>_statistics.tsv
Statistic file for a given atlas. The TSV file summarizes regional volumes or averages for a given parametric map. With the help of pandas (Python library), it can be easily parsed for machine learning purposes.
Possible values for _map-<map>
key/value are:
- For T1:
graymatter
(gray matter),whitematter
(white matter),csf
(CSF) andct
(cortical thickness) - For DWI:
FA
(fractional anisotropy),MD
(mean diffusivity, also called apparent diffusion coefficient),AD
(axial diffusivity),RD
(radial diffusivity),NDI
(neurite density index),ODI
(orientation dispersion index) andFWF
(free water fraction). - For PET:
18FFDG
(18F-Fluorodeoxyglucose),18FAV45
(18F-Florbetapir),18FAV1451
(18F-Flortaucipir),11CPIB
(11C-Pittsburgh Compound-B),18FFBB
(18F-Florbetaben) and18FFMM
(18F-Flutemetamol).
Example
Content of sub-CLNC01_ses-M000_T1w_space-Hammers_map-graymatter_statistics.tsv
:
index label_name mean_scalar
0.0 Background 0.0011357992189
1.0 Left Hippocampus 0.576250553131
2.0 Right Hippocampus 0.681780695915
3.0 Left Amygdala 0.577247679234
...
(Note that to make the display clearer, the rows contain successive tabs, which should not happen in an actual TSV file.)
Contact us !¶
- Check for past answers on Clinica Google Group
- Start a discussion on GitHub
- Report an issue on Github