| Table of Contents |
|---|
Individual files
CSPM copies the T1 ref file from the project preprocess folder, renames it to the study name, and adjusts it for FSL (reslices so as to remove nifti rotations).
The results of the "run_first_all" are the other files in this folder. These include 1) the affine normalization to the template
- *_to_std_sub.mat: affine normalization parameters.
- *_to_std_sub.nii.gz: normalized T1 (I assume); this should be the file passed to FIRST for creating the segmentations.
and 2) the segmentations.
- *.bvars: raw output of the segmentation for each structure.
- *.vtk: surface in vtk format (a standard surface rendering file format).
- *.bvars: raw output of the segmentation for each structure.
- *_all_fast_firstseg.nii.gz: nifti file with final segmentations of all structures in native (T1 ref) space.
- *_all_fast_origieg.nii.gz: ?nifti file with first-pass segmentations of all structures in native (T1 ref) space (maybe).
THe "CON747.logs" contains the log files for the steps run by run_first_all. Errors are recorded here.
Set
Set folder
Set-level files are in subfolders of LinuxShare\FSL_FIRST using the standard set naming convention:
I D [set id] N [total number of studies] G [number of groups]
For example, the set ID 209 with 45 studies in 3 groups is 
.
Set-level components: structures, designs, variables and covariates, quality checking for normalization and segmentation
In order of usage:
slicesdir_T1toMNI: check normalization
Inside this folder, open the html file "index.html" in a browser to see the result of the normalization. The brain outlines in red should be in roughly the right place.
slicesdir_segmentations: check normalization
Inside this folder, open the html file "index.html" in a browser to see the result of the segmentations. This is a critical quality check step.
varfiles: variables and covariates
The most useful file is "allvariablestable.csv," which you will use to create designs.
designs: contains subfolders with designs (manually created)
Each design for this set is placed in a subfolder of "designs." The name should reflect the model. Examples include "2sample," "groupAgeSex", "groupRM" (RM = repeated measures). The design files themselves are always named "design" for ease of scripting the next steps.
Structures: from L_Acc to R_Thal
These contain structure-specific files.
Structures within a set
Structure content: scaling subfolders an concatenated bvars
The structure folder contains the concatenated bvars which is simply all studies' bvars files for that structure concatenated. This file is what gets processed in the vertex analysis.
The four scaling options for the FIRST vertex analysis are separated into four subfolders:
Scaling content: design subfolders and example scripts
Example scripts are created, which may be useful for manual processing (not needed for CSPM processing).
Each scale subfolder contains design subfolders. These are taken from the set designs.
Design content: design files, displacement files, mask, and results-by-correction method subfolders
The design files are copied from the "[set folder]\designs\..." subfolders. These include the contrast files "design.con" and "design.fts," as well as the design matrix used by the statistical analysis (randomise).
The displacements for each subject are stored in the "..._vertx*.nii.gz" file, which has one volume for each subject. The scale is displacement towards the mean surface in voxel units. The mean surface is represented by the "..._vertex*_mask.nii.gz" file.
The results for each correction method are stored in subfolders (clustere, clusterm, tfce, vox).
