Unrecognized function or variable 'desired_nfreq'.

[YaronLa]

I have been trying to follow the ROI tutorial, but I have run into some trouble. Specifically, when I attempt to compute Coherence and cross spectrum, I receive the same error. I have attached the code along with the error messages for reference.

If anyone have any ideas or suggestions on how to resolve this issue, I would be extremely grateful for the help.

EEG = pop_dipfit_settings( EEG, 'hdmfile','path_to_file','path_to_file,'path_to_file','coordformat','MNI','coord_transform',[0 0 0 0 0 0 1 1 1] ,'chansel',[1:61] );
EEG = pop_multifit(EEG, [1:55] ,'threshold',100,'plotopt',{'normlen','on'});
EEG = pop_roi_activity(EEG, 'leadfield',EEG.dipfit.sourcemodel,'model','LCMV','modelparams',{0.05},'atlas','LORETA-Talairach-BAs','nPCA',3, 'fooof', 'off');
EEG = pop_roi_connect(EEG, 'methods', { 'CS', 'COH'}, 'snippet', 'on', 'snip_length', 60, 'fcsave_format', 'mean_snips');

#Error message from roi_connect
Unrecognized function or variable 'desired_nfreq'.

Error in data2cs_event (line 114)
        if desired_nfreq ~= 0

Error in data2spwctrgc (line 108)
    CS = data2cs_event(data(:, :)',ndat, ndat - (floor(ndat/2)), ndat , [], CSpara);

Error in roi_connect (line 121)
        conn_mult = data2spwctrgc(source_roi_data, EEG.pnts, g.morder, 0, g.naccu, [], tmpMethods2);

Error in pop_roi_connect (line 219)
        EEG = roi_connect(EEG, 'morder', g.morder, 'naccu', g.naccu, 'methods', g.methods,'freqresolution',g.freqresolution); % compute connectivity over one snippet

[nguyen-td]

Thanks, I was able to reconstruct the error and will work on it today.

[arnodelorme]

Should be fixed in eb43bef

[YaronLa]

Thank you for the help. I can now calculate Cross-spectrum as well! However, There are still several connectivities that are giving error codes. I hope you can help.
When calculating Coherence:

EEG = pop_roi_connect(EEG, 'methods', {'COH'}, 'snippet', 'on', 'snip_length', 60, 'fcsave_format', 'mean_snips', 'freqresolution', 0);

Index in position 3 exceeds array bounds. Index must not exceed 101.

Error in data2spwctrgc (line 116)
            pow = real(diag(CS(:,:,ifreq)));

Error in roi_connect (line 121)
        conn_mult = data2spwctrgc(source_roi_data, EEG.pnts, g.morder, 0, g.naccu, [], tmpMethods2);

Error in pop_roi_connect (line 219)
        EEG = roi_connect(EEG, 'morder', g.morder, 'naccu', g.naccu, 'methods', g.methods,'freqresolution',g.freqresolution); % compute connectivity over one snippet

Error in eeg_eval (line 136)
            TMPEEG = feval(func, TMPEEG, g.params{:});

Error in pop_roi_connect (line 183)
            [ TMPEEG, com ] = eeg_eval( 'pop_roi_connect', TMPEEG, 'params', optionsSubj );

When calculating GC, PDC and DTF I recieve following error:

EEG = pop_roi_connect(EEG, 'methods', {'GC'}, 'snippet', 'on', 'snip_length', 60, 'fcsave_format', 'mean_snips', 'freqresolution', 0);

Error using schur
Input to SCHUR must not contain NaN or Inf.

Error in lyapslv (line 11)
[U,T] = schur(A);

Error in dlyap (line 64)
X = lyapslv('D',A,[],-Q);

Error in varma2iss (line 44)
O=dlyap(A,K*R*K');

Error in data2sctrgcmim (line 220)
          [eA2, eC2, eK2, eV2, eVy2] = varma2iss(reshape(A, nsubsetvars, []), [], SIG, eye(nsubsetvars));

Error in roi_connect (line 98)
        conn_mult = data2sctrgcmim(source_roi_data, EEG.pnts, g.morder, 0, g.naccu, [], inds, tmpMethods1, [], 'freqresolution', g.freqresolution);

Error in pop_roi_connect (line 219)
        EEG = roi_connect(EEG, 'morder', g.morder, 'naccu', g.naccu, 'methods', g.methods,'freqresolution',g.freqresolution); % compute connectivity over one snippet

Error in eeg_eval (line 136)
            TMPEEG = feval(func, TMPEEG, g.params{:});

Error in pop_roi_connect (line 183)
            [ TMPEEG, com ] = eeg_eval( 'pop_roi_connect', TMPEEG, 'params', optionsSubj );

There should be no such input as far as I am aware.
If you have any ideas or suggestions on how to resolve this issue, I would be extremely grateful for the help.

[CPernet]

@arnodelorme is there still a problm with the freq vector

Error in data2spwctrgc (line 116)
            pow = real(diag(CS(:,:,ifreq)));

it seems it reaches beyond the size of the cross frequency sprectrum

[nguyen-td]

The first problem should be fixed with #36 / d820d24 but needs to be accepted first

[YaronLa]

Thank you very much. Now both Cross-Spectrum and Coherence is working! Now I only have the 'schur' issue when calculating GC, PDC and DTF

[stefanhaufe]

Hi, if I get a minimal working example (ideally including data), I can look into that. It does not sounds like a DTF or PDC issue but more like a GC issue, as varma2iss is not used by DTF/PDC.

I suspect a problem related to singularity or ill-conditioning of the data or cross-spectral matrices, especially in combintion with the "conditional" version of GC, or when calculating these multivariate measures on spaces that are too high-dim compared to the rank of the data.

[nguyen-td]

We have some very basic pipelines in the test_pipes folder, e.g., pipeline_connectivity.m . Here, we load some sample EEGLAB data and run the pipeline. I replaced

roiconnect/test_pipes/pipeline_connectivity.m

Lines 38 to 43 in eb43bef

	measures = { 'CS' 'COH' 'DTF' 'wPLI' 'PDC' 'MIM' 'MIC' 'GC' };
	measures = { 'CS' 'COH' 'wPLI' 'PDC' 'MIM' 'MIC' 'GC' };
	measures = { 'CS' 'COH' 'MIM' 'GC' };
	measures = { 'CS' 'COH' 'MIM' };
	measures = { 'TRGC' 'MIM' };
	measures = { 'MIM' };

with only measures = { 'DTF' 'PDC' 'GC' }; and did not receive any errors after calling pop_roi_connect(), so it might have something to do with your data.

If possible, could you please confirm this by running the pipeline yourself and tell us if it works or not?

[arnodelorme]

OK, I have merged the request. Feel free to reopen if the bug persist.