New Features |
|---|
| Nonparametric Permutation Inference | Conventional parametric statistical testing of fMRI data is an efficient technique to test the significance of hypotheses as long as the assumptions are met. Recent critique on parametric cluster-extent multiple comparison testing (Eklund et al., 2016) suggest to complement parametric fMRI analyses with more robust nonparametric inference methods. This release introduces the "Randomise Plugin" that provides nonparametric permutation inference for multi-subject designs. Permutation testing uses random shuffles of observations to obtain the correct distribution of a test statistic under a null hypothesis providing inference with nominal false positive rates. The Randomise Plugin provides direct support for common multi-subject designs including single-group contrast testing and group comparisions. Other more complex designs can be specified using a flexible graphical user interface. Furthermore, the Randomise Plugin allows to apply the thresold free cluster enhancement (TFCE) method to suppress background noise; while TFCE does not provide statistical inference (p values) when used in isolation, it can be integrated in permutation testing further demonstrating the usefulness of nonparametric statistics. A disadvantage of permutation testing is that it is computationally intensive because thousands of permutations need to be evaluated; the Randomise Plugin uses parallelized code (multi-threading) to achieve reasonable computation times when running on multi-core CPUs. For more details consult the documentation of the plugin and the overview provided in topic "The Multiple Comparisons Problem" in the "Statistical Data Analysis" chapter of the User's Guide. |
| Open NIfTI Files | While BrainVoyager offers the "NIfTI-1 converter" plugin to import and export NIfTI files since many years, this version introduces a more integrated approach allowing to open and save NIfTI-1 files directly from the File menu. Besides convenience, the new routines also better handle conversion to and from BrainVoyagers internal representations so that manual transformation adjustments should not be necessary in most cases. This version supports direct opening of 3D anatomical (as VMR) and 4D functional (as FMR-STC) files. While 3D files are opened in native and normalized MNI or Talairach space, 4D files are currently only opened if they are in native space (i.e. if NIfTI file contains "Scanner" orientation matrix). For further details how NIfTi files are processed when loaded from the File menu, consult the topic "Processing NIfTI Files" in the "Coordinates and Transformations" chapter of the User's Guide. |
Enhancements |
|---|
| Maximum Probability Maps | Improved prob maps with MPM and non-overlapping POI generation. More details coming soon... |
| Bidirectional MNI Transformation | MNI transformation matrices are generated when running the standard template-based normalization procedure. The resulting matrices contain 12 parameters reflecting not only scales, rotations and translations but also shear operations. These transformations were not supported when applied in opposite direction, e.g. when going backwards from MNI space to native space, e.g. when transforming a segmented volume from MNI space back to native space. Forward and inverted 12-parameter (MNI) matrices are now supported for volumes as well as for meshes and VOIs. Furthermore, several dialogs (including "3D Volume Tools", "Mesh Transformation" and "Transformation Matrix") have been updated to properly reflect (decomposed) shear transformations. Information in the "Log" pane now also describe decomposition of 12-parameter matrices when running spatial transformation operations. |
| Scripting API | New scripting commands have been added supporting the creation of surface maps from available volume maps (CreateSurfaceMapFromVolumeMap, CreateSurfaceMapFromVolumeMapDepth) with parameters to control how volume map data is sampled from the vertices of the target mesh. Also new commands to smooth surface maps (SmoothMap, SmoothMapLags) allow to control this process with all options also available in the GUI. The Python script "test_create-smp.py" has been provided in the "PythonScripts" folder demonstrating the use of these new scripting commands including a description of all parameters. Furthermore, the "add_motion_regressors.py" Python script can be used to append motion regressors to standard single-run .SDM design matrices. |
Bug Fixes |
|---|
| Huge VTC Files | Huge VTC files beyond 8 GB can be created and used for whole-brain statistical analysis without problem. Accessing time courses of such files when linked to a VMR (by clicking on voxels or by using VOIs) leads to corrupted time course displays for voxels stored beyond 8 GB in the VTC file in previous versions. This issue has been fixed by using 64-bit integer values when accessing voxel time courses in VTC files. |
| Black Protocol Dialog on Mac | The "Protocol" dialog was drawn in black on macOS in BrainVoyager 20.4 making it difficult to use the dialog since text was not readable. This issue has been fixed. |