Problem of a noisy mastoid electrode for re-referencing EEG data

Discussions
,
1

Hi,

I have a question regarding re-referencing. In my study I first reference the EEG signal to one mastoid electrode and then re-referencing to the other mastoid electrode (in my case, TP10 electrode). In some subjects TP10 has a very noisy signal (despite doing all the filtering, etc.) which corrupts the data which make it impossible to get proper ERPs for further analysis.

What should I do in this case? Should I remove such person from the analysis? Is there any procedure in Brainstorm to solve such problem?

Best,

Artur

2

Are you trying to obtain two separate sets of results, one for each mastoid?
If you don't, you should not proceed in this way, as the the second re-referencing would cancel the first one. Use the menu a " New re-referencing montage (linked ref)" montage instead.

image

In some subjects TP10 has a very noisy signal (despite doing all the filtering, etc.) which corrupts the data which make it impossible to get proper ERPs for further analysis.
What should I do in this case? Should I remove such person from the analysis? Is there any procedure in Brainstorm to solve such problem?

Indeed, if the quality of the signal recorded at your reference electrode is poor, this propagate the noise to all the electrodes...
Maybe a solution could be to process all you data in average reference? This allows to remove different sets of bad channels in each subject, without having to change the referencing scheme.