Release SwE Version 2.2.0

[BryanGuillaume]

Changes made since the previous version (v2.1.1)

• enh: add GIfTI support
• enh: add CIfTI support
• enh: all the non-parametric WB p-values are now computed using a tolerance of 10^-8 for comparing real numbers
• enh: the maximum TFCE scores are now saved into SwE.mat
• enh: the WB procedure now makes voxel-wise inference on Z or X instead of T or Z (the main reason for this is to be consistent with cluster-wise inference which forms clusters based on Z or X)
• enh: improve the thresholding description in the result display
• enh: add a new function swe_compareVersions able to compare versions of the toolbox
• enh: add a (hidden) way to supply a WB resampling matrix instead of letting the toolbox generate one by itself
• enh: add the effective number of degrees of freedom Approximation I for the WB
• enh: replace the loop over planes by a loop over chunks of data
• enh: improve the way some X-scores are computed in order to prevent Inf values
• enh: add the option to constrain volume clusters to be within the ROI boundaries in CIfTI analyses
• enh: modify the naming convention for CIfTI, GIfTI and .mat outputs to better match PALM naming convention
• fix: the p-value thresholding is now inclusive for WB analyses
• fix: for WB analyses, the p-values displayed are systematically the WB p-values instead of the parametric p-values
• fix: fix a result display error observed for the classic parametric SwE
• fix: fix some typos
• fix: fix some bugs in the progress display and improve it
• fix: fix several bugs in the result display GUI
• fix: ensure that each CIfTI outputs contain an appropriate CIfTI XML file
• fix: replace the nan values in output files by zero values (this is to avoid issues when reading the CIfTI outputs in other software packages)
• fix: keep only one "NameMap" in the XML extension of the toolbox dscalar.nii outputs
• fix: fix bugs related to zeroing nan values
• fix: fix several bugs related to the naive estimation of the effective number of degrees of freedom
• fix: fix a bug in the detection of sub-design matrices
• fix: fix a bug where some ROI volume labels with multiple 'V_' would not be processed correctly

Link to relevant issues

• Issue #5
• Issue #41
• Issue #102
• Issue #107
• Issue #121
• Issue #122
• Issue #123
• Issue #124
• Issue #132
• issue #140
• issue #148
• issue #150
• issue #153
• issue #154
• issue #156

[BryanGuillaume]

All checked have passed. Thus I am ready to merge and create the new release. @nicholst, do you want to review something before I create this new release?

[nicholst]

Looks good! Go for it! I'll do PR.