Question bout using svm classifier to decode EEG

Nov '7 - Hello, I want to use the svm classifier in brainstorm to distinguish two groups of 16-channel EEG signals, I want to know whether the decoding process is performed at each time point? That's to say, is the signal at each time point taken out for classifier training and testing?

Have you read the decoding tutorial?


Yes, I have read the tutorial, but I don’t know how the accuracy at time t is calculated.

I guessed the time courses were processed point by point.

I found that the data is not what you used in the tutorial.
The pics results showed in the tutorial is almost perfect. But I can not reproduce the pics even although I carefully follow each steps in your nice tutorials.

Where to download the data used in the tutorial?
In addition, where to download all the subjects' data? I think it is an interesting topic.

@pantazis @MartinC

In your tutorial, you used the data from subject 04. But the sample you provided for us to downloading is subject 06. May I ask you to provide the subject 04 if possible Thanks.

The file available on the Download page of the Brainstorm website contains the file subj04NN_sess01-0_tsss.fif, which is the one used in the tutorial.

Maybe you refer to an older version of this tutorial, which was using a file named mem6-0_tsss_mc.fif.
In that case, simply download the new dataset from the website.

I just realized that the dataset had not been updated properly on the website.
Now I understand better the problem you reported...
I fixed the download link, and will try to fix the FTP download soon.