Known bugs

---

Main window and database

• 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
• Load simultaneously datasets with slightly different sample rates: does not work

Figures issues (recordings / results)

• Stat: t-values cannot be displayed on 3D MRI slices
• Figure menu "Set colormap max": does not work from the 3D figures with some versions of Matlab (eg. R2007a)

• Cloned figures do not behave well
  Example: Clone + scouts + moving in time : currents on cortex are not updated

• 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)

Computation issues

• 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)

Scientific issues

• Grand average computation does not weight the different subjects with the number of trials for each subject
• Rounding errors with time values: sometimes there are +/- 1 samples selected than expected