Skip to content

statistics-volume - Volume-based mass-univariate analysis with SPM

This pipeline performs statistical analysis (currently group comparison) on volume-based features using the general linear model (GLM) [Friston et al., 1994]. To that aim, the pipeline relies on the tools available in SPM.

Volume-based measurements are analyzed in the IXI549Space (from SPM12). Currently, this pipeline mainly handles gray matter maps obtained from T1-weighted MR images using the t1-volume pipeline and standardized uptake value ratio (SUVR) maps obtained from PET data using the pet-volume pipeline.

Dependencies

If you only installed the core of Clinica, this pipeline needs the installation of Matlab and SPM, or of SPM standalone, on your computer. You can find how to install these software packages on the third-party page.

Running the pipeline

The pipeline is divided into two sub-pipelines:

  • statistics-volume: performs group comparison but no statistical correction. It generates an SPM report necessary to perform statistical corrections using the second sub-pipeline.
  • statistics-volume-correction: performs family-wise error rate (FWE) or false discovery rate (FDR ) correction at the peak/cluster level. The user has to report the values from the SPM report generated by statistics-volume in order to run this pipeline.

statistics-volume pipeline

The pipeline can be run with the following command line:

clinica run statistics-volume [OPTIONS] CAPS_DIRECTORY GROUP_LABEL {t1-volume|pet-volume|custom-pipeline}
                              SUBJECT_VISITS_WITH_COVARIATES_TSV CONTRAST

where:

  • CAPS_DIRECTORY is the output folder containing the results of the t1-volume or pet-volume pipeline and the output of the present command, both in a CAPS hierarchy.
  • GROUP_LABEL defines the group name for the analysis.
  • The third positional argument defines the type of volume-based feature: type t1-volume to use gray matter maps, pet-volume to use PET data or custom-pipeline to use your own data in CAPS directory (see below for details).
  • SUBJECT_VISITS_WITH_COVARIATES_TSV is a TSV file containing a list of subjects with their sessions and all the covariates and factors of the model (the content of the file is explained in the Example subsection of the statistics-surface pipeline).
  • CONTRAST is a string defining the contrast matrix or the variable of interest for the GLM, e.g. group or age.

Pipeline options:

  • --group_label_dartel: Name of the DARTEL template that Clinica needs to use to grab the input files.
  • --full_width_at_half_maximum: Value of the full width at half maximum (FWHM) used when smoothing the input files. Default value is 8 mm.

Pipeline options if you use inputs from pet-volume pipeline:

  • --acq_label: Name of the label given to the PET acquisition, specifying the tracer used (trc-<acq_label>).
  • --suvr_reference_region: Reference region used to perform intensity normalization (i.e. dividing each voxel of the image by the average uptake in this region) resulting in a standardized uptake value ratio (SUVR) map. It can be cerebellumPons (used for amyloid tracers) or pons (used for FDG).
  • --use_pvc_data: Use PET data with partial value correction (by default, PET data with no PVC are used)

statistics-volume-correction pipeline

Once the statistics-volume sub-pipeline has finished, you need to open the SPM report (report1.png or report2.png file). This will look like as follows:

SPM report

You will need to report the following information in the statistics-volume-correction pipeline:

  • height_threshold: T value corresponding to an uncorrected p-value of 0.001
  • FWEp: height threshold (i.e. voxel-level (= peak) threshold)
  • FDRp: height threshold (i.e. voxel-level (= peak) threshold)
  • FWEc: extent threshold (i.e. cluster size threshold)
  • FDRc: extent threshold (i.e. cluster size threshold)

The pipeline can then be run with the following command line:

clinica run statistics-volume-correction [OPTIONS] CAPS_DIRECTORY T_MAP HEIGHT_THRESHOLD FWEP FDRP FWEC FDRC

where:

  • T_MAP is the name of the T statistic map used for the correction.

Optional parameters:

  • n_cuts: Number of cuts to display in the final figure

Outputs

Results are stored in the following folder of the CAPS hierarchy: groups/<group_id>/statistics_volume/group_comparison_measure-<measure-label>. The most important files are:

  • <group_id>_participants.tsv: copy of the subject_visits_with_covariates_tsv parameter file.
  • <group_id>_<grp_1>-lt-<grp_2>_measure-<msr>_fwhm-<fwhm>_TStatistics.nii: T statistics associated with the hypothesis group1 < group2.
  • <group_id>_mask.nii: voxels included in the analysis.
  • <group_id>_report.png: all the results of the two sample t-test generated by SPM. Contains information necessary to use the statistics-volume-correction sub-pipeline.

The <group_1>-lt-<group_2> means that the tested hypothesis is: "the measurement of <group_1> is lower than (lt) the measurement of <group_2>". The pipeline includes both contrasts so *<group_2>-lt-<group_1>* files are also saved.

The full list of output files from the statistics-volume-[correction] pipeline can be found in The ClinicA Processed Structure (CAPS) specifications.

Describing this pipeline in your paper

Example of paragraph:

These results have been obtained using the statistics-volume pipeline of Clinica [Routier et al., 2021]. This pipeline is a wrapper of the statistical analysis toolbox implemented in SPM. More precisely, a point-wise, voxel-to-voxel model was used to conduct a group comparison of whole brain voxels. The data were smoothed using a Gaussian kernel with a full width at half maximum (FWHM) set to <FWHM> mm. The general linear model was used to control for the effect of <covariate_1>, ... and <covariate_N>.

  • For FWEp: Statistics were corrected for multiple comparisons using the family-wise error (FWE) correction at the peak level with a statistical threshold of P < 0.05 FWE.
  • For FWEc: Statistics were corrected for multiple comparisons using the family-wise error (FWE) correction at the cluster level. A statistical threshold of P < <ClusterThreshold> was first applied (height threshold). An extent threshold of P < 0.05 corrected for multiple comparisons was then applied at the cluster level.
  • For FDRc: Statistics were corrected for multiple comparisons using the false discovery rate (FDR) correction at the cluster level. A statistical threshold of P < <ClusterThreshold> was first applied (height threshold). An extent threshold of P < 0.05 corrected for multiple comparisons was then applied at the cluster level.

Tip

Easily access the papers cited on this page on Zotero.

Support

(Advanced) Using volumes other than gray matter or PET SUVR maps

If you run the help command line clinica run statistics-volume -h, you will find the following options:

  • --custom_files CUSTOM_FILES: allows you to specify which file should be taken in the CAPS/subjects directory. For example, if you want to use the file <participant_id>_<session_id>_trc-18FAV45_pet_space-Ixi549Space_pet.nii.gz that is contained in CAPS/subjects/<participant_id>/ses-<session_id>/pet/preprocessing/<group_id>, you can use the argument *sub-*_ses-*_trc-18FAV45_pet_space-Ixi549Space_pet.nii.gz. If you want to specify the group, you can use <group_id>/sub-*_ses-*_trc-18FAV45_pet_space-Ixi549Space_pet.nii.gz.
  • --measure_label MEASURE_LABEL: specifies the name of the feature type. It will appear in the _measure-<MEASURE_LABEL> of the output files once the pipeline has run.