[FieldTrip] wPLI values higher than normal coherence

[daniel.strahnen at uni-ulm.de]

Dear FieldTrip - Community,

 

I am using the wPLI routine from the FieldTrip toolbox to detect true
connectivity.

However, I found that the wPLI values that I obtain are even higher than the
coherence values and the maximum often goes up to 1.

Based on the theory behind the two measures I would rather expect the
opposite.

Has anybody stumbled upon the same phenomenon?



 

Here is the code I used, Output_data contains 7 channels (samples x
channels) and it is a continuous recording over 10 minutes which is
truncated into 4 second trials:

 

    data.label = {'Ref'; 'PFCd'; 'PFCu'; 'dHC'; 'vHCu'; 'vHCd'; 'MD'};

    data.fsample = 1000; 

    j = 1;

    for k = 1:20    

        data.time{k} = j:(k*30000);

        data.trial{k} = Output_data((j:(k*30000)),:)';

        j = k*30000 + 1;

    end

 

    cfg.length    =  4;              

    cfg.overlap   =  0.5;             

    [data] = ft_redefinetrial(cfg, data);

 

    %% Freqanalysis to obtain the cross spectral density

    cfg_freq = [];

    cfg_freq.method = 'mtmfft';

    cfg_freq.output = 'fourier';

    cfg_freq.keeptrials = 'yes'; 

    cfg_freq.taper = 'dpss';

    cfg_freq.tapsmofrq = 2;

    cfg_freq.foilim = [0 48];

    cfg_freq.pad = 'nextpow2';

    freq_data = ft_freqanalysis(cfg_freq, data);

    

    %% coh analysis

    cfg_conn = [];

    cfg_conn.method = 'coh';

    coh_str = ft_connectivityanalysis(cfg_conn, freq_data); 

    coh_str = ft_checkdata(coh_str, 'cmbrepresentation',
'full','datatype','freq'); 

    

    %% wPLI analysis

    cfg_conn = [];

    cfg_conn.method = 'wpli_debiased';

    wpli_str = ft_connectivityanalysis(cfg_conn, freq_data); 

    wpli_str = ft_checkdata(wpli_str, 'cmbrepresentation',
'full','datatype','freq');

 

 

Thank you very much and best regards

 

Daniel

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20191217/d4b821c0/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image003.png
Type: image/png
Size: 42736 bytes
Desc: not available
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20191217/d4b821c0/attachment.png>