My research work is on EEG Source Localization. I installed Brainstorm and trying to compute EEG Sources – Source analysis. I acquired an Epileptic patient EEG data from hospital. Processed this data for computing sources. Data acquisition is done using 10-20 system 21 channel data. I used ICMB 152 – Generic 10-20 19 template (for electrode positions). By using this default electrode cap only 19 channels from the patient EEG data are updated and the remaining two channels not updated. A1 and A2 are not updated. With this 19 channel data, I computed noise covariance matrix and later source analysis.
While calculating the noise covariance matrix 21 channel raw EEG data is used by the software, even though in the channel file only 19 channels are been updated. In the covariance matrix, A1 and A2 channels are appearing as empty line as shown in the figure given below.
Whether these empty lines in the above noise covariance matrix influence the source computation(Calculating source)?
How can I delete the two lines?? (I removed from the channel file but the noise covariance matrix is calculated from the raw signal -21 channel signal)
This action only updates the location of the sensors. If you want to have the positions for A1 and A2 you can get the electrode locations from the cap ICBM152 > Generic > ASA 10-20 94
The noise covariance will always have the size of the total number of sensors, regardless their type. The noise covariance matrix, it is computed only for channels of type EEGMEG, ECOG, SEEG and NIRS. And it is set to zero for all the other types.
In this case, your sensors A1 and A2 have been probably identified as EEG REF, so they are not included in the computation of the noise covariance matrix.
If you want to use A1 and A2 as EEG electrodes instead of reference electrodes depends on the analyses that you want to perform. To do so, you would need to change their channel type, check this link about editing the Channel file:
This does not affect source estimation, as only EEG sensors are used, and as stated above, A1 and A2 are not EEG by EEG REF type.
There is no need to remove them from the covariance matrix. This covariance matrix has a size that matches the total number on sensors. When source estimation is performed using EEG sensors, only the subregion of the covariance matrix related to EEG is used.
It is not recommended to delete manually channels from the channel file, as by doing so, it will not match the recordings, potentially generating problems with dimension agreement.
Thank you so much for the clarification. It is very useful.
In the data that i have A1 and A2 are reference signals, can you please help to compute covariance matrix withoutg A1 and A2(which appear as empty lines currently)
Is there an specific reason to remove those empty lines?
Brainstorm will store the covariance matrix in the same size as total number of sensors, regardless their type that is why the rows and columns for A1 and A2 are included. On the further processing those lines will be ignored as they do not correspond to EEG sensors.
The way to remove those empty lines once for all is to remove the A1 and A2 sensors for the data. However, this is not trivial as you need to manually modify the structures of the data, see this post:
