Avoiding noisy EEG sections during power calculation

Discussions
1

Hi

I am trying to select bad sections on EEG and not to include them during power calculation:

2
21920×1080 313 KB

Firstly, when I reject bad sections, it becomes impossible to calculate power.

Secondly, it is not possible to select several EEG segments.

Thanks

2

Hi!
I think the problem is that you're rejecting trials, instead of bad segments,

for power computation, rejecting bad or noisy sections must be done prior to import it in database (otherwise, you are rejecting the entire trial), I mean, you should mark bad segments in the continous raw file, and then import it in database as epochs, and then perform the PSD computation

3

Thanks, I got it. Is there a way to mark bad segment of continuous EEG data in Brainstorm?

4

The cleaning process is the same, with mouse you select bad segments and then with mouse right button option you mark them as BAD

5

Thanks, but there is an issue in doing that with continuous data as I mentioned in my 1st message. My data is continuous.

6

Sorry, I didn't understand the issue: if you have continous data you can easily select and reject bad segments with the mouse, In case you have epochs or trials and you don't want to reject the entire trial, you can always concatenate the files into one continous file and then manually reject bad segments

You may check the tutorials, they are useful and step guided
:slight_smile:

7

As I understood, selecting bad segments applies to epoched data but in case of continuous data like I showed in the 1st message it seems that selecting bad segments and removing them do not work.

8

Actually the data you showed in the 1st post is not continous, it is epoched data (trials), that's why bS rejects entire trial when you mark bad segments, you should concatenate your data before mark bad segments

9

@niko808080
If you want to exclude some bad segments from the PSD computation, you need to select them in the viewer and create an event that includes the tag "bad". This is described in the introduction tutorials:
https://neuroimage.usc.edu/brainstorm/Tutorials/BadSegments#Manual_inspection

What you are doing with the menu "Reject trial" is to mark the entire files as bad, as described later in the tutorials:
https://neuroimage.usc.edu/brainstorm/Tutorials/Epoching#Import_in_database

To process continuous recordings, the general recommendation is to use the menu "Review raw file" instead of "Import MEG/EEG":
https://neuroimage.usc.edu/brainstorm/Tutorials/ChannelFile#Review_vs_Import

If you are not familiar with these concepts, I recommend you start by following all the introduction tutorials using the example dataset, before processing your own data.
If you are working with EEG, read also the tutorial "EEG and Epilepsy":
https://neuroimage.usc.edu/brainstorm/Tutorials/Epilepsy

10

Thanks Francois. I used ''Review raw file'' and is able to select bad segments.