Dear Jakob,
1) spm_eeg_convert_arbitrary_data is an example script which you might
have to adjust to convert your data correctly. If it works correctly
it should create an SPM dataset so there is no need to convert again.
2) Not sure what the problem is with the reviewing tool but (a) make
sure that there are no NaNs in your data. (b) the reviewing tool might
also not like completely flat data (like all zeros).
3) If you want to use DCM on this kind of data you should set the
channel type to 'LFP' (as in the example script) so they should
appear in the 'Other' tab and not in 'EEG'.
4) The button you are using in DCM is for loading saved DCM files and
not for loading data. I suggest that you look carefully at the manual
chapter and follow the instructions there.
5) Note that unlike GC and SEM, DCM assumes a particular generative
model for the data, specifically neural mass model driven by coloured
noise. So if you want to compare DCM to the other methods the data
should be generated with this kind of model as done in DCM methods
papers. Random, band limited noise with shifting is unlikely to work.
Best,
Vladimir
On Mon, Aug 22, 2011 at 7:02 PM, Jakob Volhard <[log in to unmask]> wrote:
> Hi everyone,
>
> I have a little problem and maybe somebody can help me on that. I have to admit that I am a newbe with SPM and DCM.
>
> I am trying to make a comparison between DCM and other methods (Granger Causality and unified SEM), using fictional time-series data.
> This data consists of 5 trials with 1000 datapoints on 5 sources, so it is similar to EEG data (for example after performing an ICA). So I tried to analyse this fictional data just as EEG-Data. For this I used the script "spm_eeg_convert_arbitrary_data.m" which converts my time-series data. Then I used the convert function (this works), but in the upcoming graphic interface the first error messages occures in the EEG tab. When I click on one of the 4 buttons "increase/decrease contrast/width of the plotted time window" nothing happens and the following error message occures:
> ___________________________
> ??? Reference to non-existent field 'gpa'.
>
> Error in ==> spm_eeg_review_callbacks>updateDisp at 1100
> ud = get(D.PSD.handles.gpa,'userdata');
>
> Error in ==> spm_eeg_review_callbacks at 412
> updateDisp(D,3);
>
> ??? Error while evaluating uicontrol Callback
> _____________________________
>
>
> I also get an error message when I click on the different trials:
> _____________________________
> ??? Error using ==> set
> Values must be monotonically increasing
>
> Error in ==> spm_DisplayTimeSeries at 347
> set(ud.v.handles.gpa,...
>
> Error in ==> spm_eeg_review_callbacks>updateDisp at 994
> ud = spm_DisplayTimeSeries(D.data.y(:,:,trN(1))',options);
>
> Error in ==> spm_eeg_review_callbacks at 723
> updateDisp(D,1)
>
> ??? Error while evaluating uicontrol Callback
> ______________________________
>
> What do those errors mean?
>
> When I ignore all this and open the DCM menu and load the dataset, the message "Reference to non-existent field 'DCM'." comes up, so I cannot perform a DCM on the data.
>
> Does anybody have an idea what the problem may be?
> Or does anybody know whether it is possible to perform a 'pure' DCM analysis without a head-model or ICA?
>
> Thanks for your help!
> Jakob
>
|