BrainVoyager 24.1 |
|---|
| This minor release provides important improvements and bug fixes. |
New Features |
|---|
Enhancements |
|---|
| BBR for Across-Session Alignment | The BBR tool has a new input option to provide an initial spatial transformation (TRF) file. In case that a source and target volume are too far off for successful BBR, the TRF file can bring an anatomical mesh (reconstructed from a corresponding anatomical VMR) closer to the target volume. One use case is for BBR-based FMR-VMR coregistration when the anatomy was recorded in a different scanning session than the functional data. To get a useful TRF file, one can use the 'Mesh Spatial Transformation' dialog that has been improved for this purpose. You can change a spatial transformation (e.g., moving the cortex mesh up or down, or rotating it around an axis) and see the mesh contour continouly updated inside the VMR View window when pressing the 'Apply' button. For this workflow it is best to arrange the VMR View and 3D Viewer next to each other before entering the 'Mesh Spatial Transformation' dialog. Note that one needs not do the tricky alignment (that's for BBR) but bring the mesh close enough to the location and orientation of the target volume. For details, see topic 'Across-Session BBR' in the 'Coregistration' chapter of the User's Guide. |
| Postprocessing of DNN Segmentation | After running the deep-learning neural network segmentation tool, the probabilistic tissue volume maps are converted into discrete regions-of-interest (VOIs) for further processing. This VMPs to VOIs conversion is the first step in the 'DNN Segmentation Postprocessing' dialog that appears after DNN-based segmentation. While DNN Segmentation leads to excellent results, one might want to influence the conversion at the WM / GM and GM / CSF boundaries beyond the implemented "winner-takes-all" approach. Previous releases allowed to "boost" WM over GM and GM over CSF so that tissue boundaries could be moved to one side. This release offers a second new approach to shift boundaries using absolute thresholds for WM and GM, which is easier to use and often leads to even better results. For more details, see topic 'Postprocessing' in the 'Deep Neural Network Based Segmentation' chapter. |
| Atlases Support | This release adds the Juelich cytoarchitectonic atlas. The atlas is available via the 'Open Juelich MNI Atlas' item in the 'Atlases' menu (or as part of the 'Files' menu if chosen in the 'Settings' dialog). It is now also possible to provide a set of VOIs (that may be, for example, converted from a volume map and further edited) as input to the atlas-based cluster identification tool. This new variant is available via the 'Atlas-Based Identification of Regions-of-Interest' item in the 'Atlases' menu below the 'Atlas-Based Identificaton of Map Clusters' item (which was introduced in the 24.0 release under the name 'Identify Map Clusters Using Atlas Parcels'). For more informaiton about the new atlas tools, consult the 'Atlases' chapter of the User's Guide. |
| Combine TRFs Including MNI Transformation | The 'Spatial Transf' tab allows to combine multiple .TRF files by loading a first one and by appending subsequently loaded ones while having turned on the 'Append after current' option in the 'Combination of current and loaded transformation' field. Iinternally, combining multiple matrices corresponds to multiplication of the current and the newly added affine spatial transformation matrix. One can then save the combined matrix as a new .TRF file to disk and use it to apply a chain of spatial transformations with only one data resampling step. The combination tool also multiplied a MNI TRF matrix in the same way without the special non-linear axes scaling information stored in a MNI .TRF file. Since this version, the scaling informatoin is used appropriately and the saved comvbined TRF matrix is marked as a TRF leading to MNI space. For this to work, the only limitation is that the MNI TRF is the last added transformation, which is the case for typical normalization workflows. |
| Event-Related Averarging Maps | It is now possible to edit plot related parameters (colors, line thickness, font sizes) using the new 'Edit EMP Plot Information' dialog that can be invoked using the added 'Edit Plot Params' button in the 'Event-Related Averaging Map' dialog. The edited plot parameters are directly stored in the .EMP file and can, thus, be changed without re-calculating the event-related averaging map. The provided parameters can adjust nearly all parameters of the (voxel and VOI level) event-related averaging line plot, and key parameters of event-related bar plots. For more details of event-related averaging maps, please consult topic 'Event-Related Averaging Maps' in the 'Event-Related Averaging' chapter of the User's Guide. |
| Framewise Displacement | Framewise displacement caulculates the difference of all 6 motion estimates with respect to the previous time point (derivative) and summarizes those differences in a single 'framewise displacement' (FD) value per time point. A FD time course is now shown when running motion correction (below the plots of the 6 standard parameters). Next to the standard "_3DMC.sdm" motion parameter confound design matrix, a "_3DMC-FD.sdm" file with the FD time course is now saved that can also be used as a confound predictor when running GLMs. An FD plot, mean FD value and the "_3DMC-FD.sdm" file can also be calculated using the 'Framewise Displacement' dialog that can be invoked by clicking the 'Framewise Displacement From Motion Params ' item in the 'Options' menu. |
| Time Course Movie with Framewise Displacement | The 'Time Course Movie' dialog shows slices of functional documents (VTC, FMR) over time, which is a useful feature to check image quality and head motion. The dialog now shows now a framewise displacement time course above the usual time line indicator in case that the "_3DMC-FD.sdm' file can be found for the opened VTC / FMR document. Note that one should open the file prior to motion correction (or preprocessing including motion correction) so that the file will be found, and to ensure that the FD time course matches the visible motion in the document window. You can 'scrub' the knob of the timeline to go backward / forward around sections with high FD peaks to see how the slices move at those moments. It might be good to preload the entire time course data from disk by clicking the 'Preload All' button once, which wil make such inspections much faster. |
| Spatial Transformation of VMPs | VMP files can now be transformed using the user interface provided in the 'Spatial Transformation' field in the 'Space' tab of the 'Volume Maps' dialog (it is also still possible to use the 'Export VMP' button in the 'Spatial Transf' tab of the '3D Volume Tools' dialog but this is deprecated and might be removed in a future release). For fast computation, VMP spatial transformations can now run on the GPU using Metal on macOS (Lanczos and Sinc interpolation) and OpenCL on Windows and Linux (Sinc interpolation). |
| Python Support | The document method 'compute_run_glm_save_residuals()' has been added to the 'PyDoc' API, which can be used to run a GLM to obtain and save the residual voxel time courses in a output VTC file. Note that this GLM variant will NOT correct for serial correlations to not affect the timing information of the data. Any subsequent GLM on the residualized VTC will, of course, use serial correlation correction (if not turned off). The method accepts one optional string parameter, which can be used to set the trailing substring of the saved VTC file (default: "_residuals"). For an example, inspect the 'glm_residuals.py' script in the 'PythonScripts' folder. The document method 'get_func_path()' has been added to easily obtain the BIDS 'func' folder corresponding to the participant 'anat' folder in which the current document (e.g. anatomical VMR) might be located. If the document is not located in a BIDS 'anat' folder, the current location of the hosting VMR document will be returnd. |
| Setting Map Spread Range | When clicking on a overlaid volume map, a spreading algorithm is applied to include supra-threshold voxels in the neighborhood of the clicked voxel up to a certain distance. While the spreading can be changed in the '3D Volume Tools' panel, it is now possible to quickly change the maximum spread using the 'VMR View' context menu, i.e. by right (or Option) clicking inside one of the 3 orthographic slice views. A simple 'Spread Range' dialog can then be invoked by clicking the 'Set Map Spread on Click' item in the context menu. Setting the 'Maximum Spread' value in the dialog to 1, for example, will stop spreading, i.e. only the time course of a single supra-threshold voxel will be shown on click. |
| GIFTI X_Form | When loading GIFTI mesh files, the file header often contains a spatial transformation 'x_form' matrix. The x_form matrix is usually declared as being a Talairach normalization matrix. In case the mesh is already in normalized space, the matrix ia an identity matrix (no transformation) but it may also specify an actual transformation. The xform matrix was ignored by BrainVoyager in previous releases to keep a mesh in the same space as a corresponding anatomical volume. Since this release it is possible to apply the x_form transformation matrix by turning on the 'Apply x_form transformation matrix' option in the 'Settings' dialog inside the 'GIFTI Import' field of the '3D Viewer' tab. The x_form matrix will then be applied (if non-ID) for all subseequently imported GIFTI files until turned off. |
| Exporting CSV Files | Data tables are now exported as CSV (comma separated value) files instead of TXT files for better viewing support (e.g. in QuickLook on macOS) and easy import in Microsoft Excel and Apple Numbers software. |
| Toggling VOI Visibility | to show or hide VOIs in the 'Volumes-Of-Interest' dialog required first selecting one or more VOIs, followed by a click on 'Show VOIs' to show them in the associated volume and a click on 'Hide VOIs' to hide and deselect all VOIs. The dialog now has a new 'Show immediately' option allowing to toggle visibility of a VOI using CTRL-click (Windows, Linux) / CMD-click (macOS) if enabled (default). To get the old behavior, simply turn off the new option. |
Bug Fixes |
|---|
| Time Course Plot and MSEC PRTs | When using millisecond protocols, time courses must start from 0 (first data point) instead of 1 in case of volume protocols. While correctly done in the 'Protocol' dialog and the 'Singe-Study GLM' dialog, it was not correct for 'ROI Time Course Plot' windows. Time courses, thus, appeared to be shifted one TR to the left erroneously. This issue has been fixed. |
| Python Warnings at First Evaluation | When running Python code from the Python IDE or BV Notebooks the first time in a session, several warning messages are printed in the respective output panes. These internal warnings can be safely ignored and it is recommended to re-evaluate the Python code to get a clean output without the warning messages. All subsequent code evaluations do not show these warning messages. |