Indeed, I posted yesterday some major changes to the way the clusters work:
https://github.com/brainstorm-tools/brainstorm3/commit/8403a86afff14f13af7ecadbc85e229becfc395a
Now the clusters are saved in the channel file, while previously they were not saved anywhere. This explains most of what you describe:
- You can't load the clusters when there is no data loaded
- When you modify the clusters and then close the display figures, you are asked if you want to keep these changes, exactly in the same way as the edition of the events.
- The Cluster tab is opened automatically when loading the data file
- You can create once your cluster in the original raw file, and then you don't have to recreate it all the time.
- You can call the process "Extract > Clusters time series" without having to load any cluster first, but they must be available for all the data file selected in the Process1 tab.
- A cluster with a given name can be defined with a different list of channels for different folders or subjects.
This is the first step to implementing new tools for clustering SEEG contacts by anatomical regions.
I was not expecting this Clusters feature to be used much by Brainstorm users, and decided to reuse it for supporting new features.
Sorry for the inconvenience. I hope you'll find a way to reproduce what you were doing before with this new behavior.
If you don't, or if you observe any weird behavior that you don't understand, please let me know. There are probably still some bugs to fix. Suggestions for new features are welcome too.