Hi all!
I have a dataset that I imported to Brainstorm a while ago. This dataset was already preprocessed: filtering, ICA, and thresholded (removing the most noisy epochs).
Now, I have realised that I have a lot of ocular artifacts in my recordings, and I would like to pose some issues before assessing this problem:
Thank you very much in advance!
1 Like
If start again from the continuous files, you could add another layer of cleaning.
However, for EEG we don't recommend using SSP, see the last paragraph of the introduction of the tutorial you're citing: https://neuroimage.usc.edu/brainstorm/Tutorials/ArtifactsSsp
What could make more sense would be to try to identify additional ICA components matching your artifacts from the existing decomposition, and/or removing more data segments.
(removing the most noisy epochs)
If you are starting from imported epochs, the only thing you can do is mark more epochs as bad and recompute your averages.
Is there any way to add, for each subject in the database, the EOG channels? Because I discarded the EOG channels before importing them...
No, I don't see how you could add back something you deleted...
Unless you simply marked these channels as bad?
Thank you for your answer!
I have realised that it doesnt make much sense to add another preprocessing step once removed the noisy epochs, thus not having continuous recordings.
Thus, I have thought that maybe I can import the signals before the epoch selection, what is just after an ICA process that I performed a time ago. Nevertheless, applying 2 ICA processes is not correct, is it?
If is not correct I would import the registers before my old ICA process, and perform only one ICA process in Brainstorm.
Finally, when importing the EOG channels I have realised that it do not distinguish between vEOG or hEOG, it only state EOG. Is there any way to indicate this characteristic in Brainstorm? or Brainstorm only works with "EOG" channels? (not supporting the vertical (vEOG) or horizontal (hEOG) differenciation). The same question holds for differenciating between left and right EOG channels.
Thank you again, and sorry for bothering you with my (probably a bit weird) questions!!
Thus, I have thought that maybe I can import the signals before the epoch selection, what is just after an ICA process that I performed a time ago. Nevertheless, applying 2 ICA processes is not correct, is it?
If you still have the Brainstorm database, you should be able to simply change the list of ICA components removed from the recordings. Double-click on the Link to raw file, then menu Artifacts > Select active projectors in the Record tab.
Then you'll need to delete the imported epochs and all the files derived from them, and import/compute everything again.
If is not correct I would import the registers before my old ICA process,
How old is "old"?
Finally, when importing the EOG channels I have realised that it do not distinguish between vEOG or hEOG, it only state EOG. Is there any way to indicate this characteristic in Brainstorm? or Brainstorm only works with "EOG" channels? (not supporting the vertical (vEOG) or horizontal (hEOG) differenciation). The same question holds for differenciating between left and right EOG channels.
I'm not sure I understand your questions.
Each channel has a name (e.g. 'vEOG') and a type (e.g. 'EOG').
In many processes, you can select data channels either by typing their names or their types.
Some processes only accept one channel in input (e.g. the processes that detect events on a single channel), in that case only the name can be used.
The introduction tutorials show how to manage a dataset with two EOG channels (VEOG, HEOG):
https://neuroimage.usc.edu/brainstorm/Tutorials/ArtifactsDetect#Detection:_Blinks
Thank you very much Francois
If you still have the Brainstorm database, you should be able to simply change the list of ICA components removed from the recordings. Double-click on the Link to raw file , then menu Artifacts > Select active projectors in the Record tab.
Then you'll need to delete the imported epochs and all the files derived from them, and import/compute everything again. How old is "old"?
I performed the ICA process a few years ago, using EEGLAB probably (I don't remember...). Thus I think it will be impossible to re-select the ICA components to discard. Thus, should I perform again the ICA process (taking the signals before the old ICA) rather than add another layer of ICA?
I'm not sure I understand your questions.
Each channel has a name (e.g. 'vEOG') and a type (e.g. 'EOG').
In many processes, you can select data channels either by typing their names or their types.
Some processes only accept one channel in input (e.g. the processes that detect events on a single channel), in that case only the name can be used.
The introduction tutorials show how to manage a dataset with two EOG channels (VEOG, HEOG):
https://neuroimage.usc.edu/brainstorm/Tutorials/ArtifactsDetect#Detection:_Blinks
Sorry, maybe I posed this question confusingly. I meant that if it is relevant the specific type of the EOG sensor (vEOG/hEOG/right/left). For example, in the link you share it use VEOG to detect the blink components. Is it possible to use HEOG? It is relevant to select a right or left EOG channel? Is there any way to combine them all to improve the detection?
Thank you again Francois!!
Thus I think it will be impossible to re-select the ICA components to discard. Thus, should I perform again the ICA process (taking the signals before the old ICA) rather than add another layer of ICA?
Re-using old file structures that you are not sure how to manipulate is risky.
If you want to update your pre-processing pipeline, it's probably safer to start over from continuous files.
In that case, compute the ICA decomposition with runica from Brainstorm.
For example, in the link you share it use VEOG to detect the blink components. Is it possible to use HEOG?
VEOG is best for detecting blinks.
HEOG is best for detecting saccades.
https://neuroimage.usc.edu/brainstorm/Tutorials/BadSegments#Saccade_SSP
It is relevant to select a right or left EOG channel?
EOG should be used as bipolar channels (eg. EOGL - EOGR).
Understood! Thank you very much Francois!!
)