Hello,
I performed ICA analysis and other preprocessing steps (e.g., filtering) on my raw EEG data using EEGLab and imported it to Brainstorm for statistical analysis. When I visually checked the data, I noticed that ICA was not applied and thus the data included artifacts. I attach screenshots to compare the views of the same file after ICA in EEGLab and Brainstorm. The first image shows the data after ICA in EEGLab. The second image shows the view of the same data in Brainstorm. For comparison, the third image shows the raw data before ICA. Is there a way to import ICA components into Brainstorm? Or performing ICA analysis using Brainstorm is the only way to apply ICA to my data?
Sungjin
Hi Sungjin,
Please check the tutorials (and here) and make sure you select the ICA components you want to reject in Brainstorm.
I'll defer to others on whether you can import EEGlab results in Brainstorm with the ICA components.
Cheers,
Marc
You can probably force EEGLAB to write the .set file with the ICA cleaning applied.
Unless you do this, the ICA components would not be applied to your Brainstorm dataset.
It has been on our todo list (https://neuroimage.usc.edu/brainstorm/Next#Pre-processing) to import the ICA results from EEGLAB but this was never set as a high-priority task. We'd need help from collaborators with this, in case you are interested 
Note that when estimating brain sources, the mixing matrices from the ICA cleaning should be applied to the forward model. This is not be possible when importing clean EEGLAB .set (this is also something that should be developed). I don't think the impact of skipping this is very significant, but to be perfectly clean with the current limitations of the EEGLAB datasets in Brainstorm, this would be in favor of computing the ICA in Brainstorm (with all the limitations it has...)
Yes, there is room for improvement on our end. Contributions welcome!
