Known bugs

Main window and database

Multiple screens: does not work in all the configurations
Cause: Too many possible configuration possible (each version of Linux and Matlab handle this in a different way)
Workaround: work only with your main screen
Tree is not updated well and some nodes are missing
Cause: Unpredictible order of execution of Java/Matlab operations
Workaround: Press F5 to refresh
Colormap sliders for contrast / brightness do not work with versions of Matlab < 7.4
Workaround: Manipulate the colormap by clicking on the colorbar and moving up/down (brightness) and left/right (contrast)
Some user preferences are not saved in compiled versions
Tree menu: "File > XTerm" does not work with all the linux distributions
Tree menu: "Group conditions" does not update all the filename links (in results files, linkresults files, etc.)
Display time series t=[0,end] + Display time series t=[begin,end]: time bounds are not good for the second figure
Save figures as AVI movies: only available on Windows operating system
Popup menus and "Open file" dialog take a few seconds to appear on some XP systems

Stat: t-values cannot be displayed on 3D MRI slices
Colormap menus do not work from the 3D figures with some versions of Matlab (eg. R2007a)
Workaround: Use the Colormaps menu in the main Brainstorm window.
Display topography + display channels from the tree : crashes if there are bad channels
Channel selection on 2D Residuals : don't work properly for magnetometers only (indices issue)
MRI Viewer may not work well on MacOS
2D/3D topography of modelized data (Fsynth) are always display with EEG colormap, even it is MEG .
2D/3D topography of modelized data (Fsynth): does not work for multimodal reconstruction
Modelized data / Residuals: no sensors selection available
"Save image" may capture a blank image on Linux

Tree menu "Project on default anatomy" : does not work for kernel / absolute values
Tree menu "Project on default anatomy" : does not work for kernels if there are some bad channels
OS_MEG crashes Matlab when scalp in not very well aligned with sensors
Workaround: align your them well...
OS_MEG (overlapping spheres) fails when the scalp surface is not perfect (some garbage inside the head)
Workaround: do your scalp extraction with FreeSurfer
"Project on default anatomy" cause Out of memory errors when processing FULL results (=> proceed by time blocks)
Permutation tests do not work on "kernel only" results

Grand average computation does not weight the different subjects with the number of trials for each subject
New t-tests and old t-test do not give the same results, because the processing of bad channels is different:
- Old tests: reject a channel when it is bad at least for one file
- New tests: reject a channel only when it is bad in all the files