Hi, Francois,
I met a problem when I did the compute source:
How can I solve it?
Best,
As indicated in the error message, please make sure that the channel count is the same when you estimate the data or noise covariance statistics and when you run the source estimation process above. From the error message, it may be that you wish to estimate sources after you have marked channels as bad, and which channels were not marked as such when you estimated the noise/data covariance matrices.
I got the noise covariance matrix from the empty room recording and the data covariance matrix from the raw recording after pre-processing. Then I computed the source using the data covariance, noise covariance and head model. This error message did not happen on other subjects. I don't know how to fix it.
Simply make sure the channels marked as bad for both recordings are the same.
I found the problem is that the data covariance matrix calculated from raw recording has a zero line, like below:
Besides, I checked the raw data. There also is a zero line in it:
I think there may be a bad channel. Do you have some suggestions for fixing it?
Thank you for your replay. I have solved it.