Coherence N x N frequency bands

Requests
1

Hello,
i am calculating a coherence analysis (MScohere NxN) with 30 subjects (2 groups, 2 conditions, each conditions consists out of three runs). Each run has 30 preprocessed epochs with 4 seconds. I calculated Mscohere for each epoch. Is there an option to get the averaged coherence values over the different frequency bands (alpha etc.)? At the moment I can only extract data for every 0,5 Hz. For example, when calculating PLV a clustering in frequency bands is possible.

Thank you very much!

image

image
image293×521 33.8 KB

2

Hi @christinakehm,

Indeed, there is process for this Frequency > Group in time or frequency bands.

  1. Select Group by frequency bands
  2. Then input the bands to compute in each row as:
    bandName / freq1, freq2 / function

Note: By default Brainstorm uses the cannonical frequency bands (delta, theta, ...). If you change these band definitions, they can be reset with: iconCondition > Reset options (more info)

Best,
Raymundo

3

Thank you very much!

Best regards,

Christina

4

Hey,
averaging the data by frequency bands was successful, thanks again! To determine group effects I planned to export the data to SPSS. Is there in Option to get the coherence data values in a row? If I export the Data to Excel I get Matrices (similar to a classic correlation matrix), but instead I would need the data in a wide-format (e. g. F4xF4 = x, F4xF3 = y, F4x.....). Is there an option in brainstorm to get the data arranged like this?

Thank you very much in advance.

Greetings

Christina

5

Is there in Option to get the coherence data values in a row?

The simplest option is maybe to script this.
For example:

ConnectFile = '/.../.../timefreq_connectn_....mat';
ConnectMat = in_bst_timefreq(ConnectFile);
R = bst_memory('GetConnectMatrix', ConnectMat);

for iFreq = 1:size(R,4)
    fid = fopen(sprintf('test-freq%d.txt', iFreq), 'wt');
    for iRow = 1:length(ConnectMat.RefRowNames)
        for iCol = 1:length(ConnectMat.RowNames)
            fprintf(fid, '%sx%s\t%f\n', ConnectMat.RefRowNames{iRow}, ConnectMat.RowNames{iCol}, R(iRow, iCol, 1, iFreq));
        end
    end
    fclose(fid);
end
6

Thank you very much!