[FieldTrip] Problem plotting independent components

Eelke Spaak eelke.spaak at donders.ru.nl
Thu Dec 8 07:58:13 CET 2011


Dear Dave,

The output of an ICA should actually not contain complex numbers. It
can happen though, and this is indicative of a problem. You should
probably have a look at this FAQ:
http://fieldtrip.fcdonders.nl/faq/why_does_my_ica_output_contain_complex_numbers

Best,
Eelke


2011/12/7 Dave Deriso <dderiso at ucsd.edu>:
> Hi Johanna,
>
> Thanks for taking a look. I could not find a method inside
> ft_componentanalysis that extracts real values (although using real() could
> also accomplish this), however the warning seems to indicate that the
> complex values are ignored. Do you think this is still an issue? The error
> with ft_databrowser indicates a lack of "unique sample points" from some
> kind of grid, which is perhaps a bigger problem. Any thoughts? Thanks again!
>
> Cheers,
> Dave
>
> On Wed, Dec 7, 2011 at 3:59 AM, Johanna Zumer <johanna.zumer at donders.ru.nl>
> wrote:
>>
>> Hi Dave,
>> Perhaps this error is due to the complex data that ft_databrowser cannot
>> plot?  If you re-calculate the components and get out real-only values, do
>> you still get this error?
>>
>> Best,
>> Johanna
>>
>>
>> On 7 December 2011 02:07, Dave Deriso <dderiso at ucsd.edu> wrote:
>>>
>>> I should also note that when I tried the ft_databrowser function it was
>>> unable to plot the component topographies because there weren't enough grid
>>> points. Have you encountered this before? Thanks again for all of your great
>>> help!!
>>>
>>> My code:
>>>
>>> cfg = [];
>>> cfg.viewmode = 'component'
>>> ft_databrowser(cfg, comp);
>>>
>>>
>>> Output:
>>>
>>> creating layout from cfg.grad
>>> creating layout for neuromag306alt system
>>> the input is component data with 306 components and 306 original channels
>>> detected   0 visual artifacts
>>> redrawing with viewmode component
>>> fetching data... done
>>> fetching artifacts... done
>>> preprocessing data... done
>>> plotting artifacts...
>>> plotting events...
>>> plotting data...
>>>
>>> Warning: Imaginary parts of complex X and/or Y arguments ignored
>>> > In ft_plot_vector at 191
>>>   In ft_databrowser>redraw_cb at 1309
>>>   In ft_databrowser at 508
>>> Warning: Imaginary parts of complex X and/or Y arguments ignored
>>> > In ft_plot_vector at 191
>>>   In ft_databrowser>redraw_cb at 1309
>>>   In ft_databrowser at 508
>>>                                           ****** these errors repeat
>>> about 10 times, then:
>>> plotting component topographies...
>>> Error using griddata (line 79)
>>> Not enough unique sample points specified.
>>>
>>> Error in ft_plot_topo (line 153)
>>> [Xi,Yi,Zi] = griddata(chanX', chanY, dat, xi', yi, interpmethod); %
>>> interpolate the topographic data
>>>
>>> Error in ft_databrowser>redraw_cb (line 1377)
>>>       ft_plot_topo(chanx, chany, chanz, 'mask', laychan.mask,
>>> 'interplim', 'mask', 'outline', laychan.outline, 'tag', 'topography',
>>>       ...
>>>
>>> Error in ft_databrowser (line 508)
>>> redraw_cb(h);
>>>
>>>
>>>
>>>
>>>
>>> Cheers,
>>> Dave
>>>
>>>
>>> On Tue, Dec 6, 2011 at 11:17 AM, Dave Deriso <dderiso at ucsd.edu> wrote:
>>>>
>>>> Dear JM and Jörn,
>>>>
>>>> Thank you so much for your helpful suggestions. Matlab could not
>>>> find/open layout file: NM306.lay with both ft_topoplotTFR and ft_topoplotIC.
>>>> I have also tried each of the following layouts without success:
>>>>
>>>> cfg.layout    =3D data.grad;
>>>> cfg.layout    =3D 'neuromag306all.lay';
>>>> cfg.layout    =3D 'neuromag306cmb.lay';
>>>> cfg.layout    =3D 'neuromag306mag.lay';
>>>> cfg.layout    =3D 'neuromag306planar.lay';
>>>> (and no layout cfg, as prescribed by the bottom
>>>> of http://fieldtrip.fcdonders.nl/reference/ft_topoplotic)
>>>>
>>>>
>>>> As Jörn mentioned, the IC and sensor labels are different and,
>>>> therefore, doing a string comparison will return an error. The question is
>>>> then, how can I project the ICs back into sensor space and plot them
>>>> topographically using the pre-existing sensor layouts? If this accomplished
>>>> by the ft_topoplotIC function, shouldn't it expect the IC labels to be
>>>> different from the layout labels?
>>>>
>>>> Thanks again for all of your help!
>>>>
>>>> Cheers,
>>>> Dave Deriso
>>>>
>>>> --
>>>> UCSD Institute for Neural Computation
>>>> UCSD Department of Neurosurgery
>>>>
>>>>
>>>>
>>>>
>>>> On Tue, Dec 6, 2011 at 3:55 AM, "Jörn M. Horschig"
>>>> <jm.horschig at donders.ru.nl> wrote:
>>>>>
>>>>> Dear Dave,
>>>>>
>>>>> the labels do not match indeed, because ICs do not correspond to single
>>>>> channel (obviously). I am not exactly sure how ft_topoplotIC is built, but
>>>>> you could try using ft_databrowser instead, with cfg.viewmode = 'component'
>>>>> Hope that at least circumvents your problem.
>>>>>
>>>>> Best,
>>>>> Jörn
>>>>>
>>>>>
>>>>> On 12/6/2011 2:45 AM, Dave Deriso wrote:
>>>>>
>>>>> Dear Fieldtrip Users,
>>>>>
>>>>> I am trying to implement the tutorial on using ICA to extract eyeblink
>>>>> artifacts
>>>>> (http://fieldtrip.fcdonders.nl/example/use_independent_component_analysis_ica_to_remove_eog_artifacts).
>>>>> Everything works, except that the labels from the IC data do not match the
>>>>> labels from the layout file, causing an error in the topoplotIC
>>>>> function. Your help in solving this is most appreciated.
>>>>>
>>>>>
>>>>> Layout labels (I am using a Neuromag 306 MEG):
>>>>> 'MEG0113'
>>>>> 'MEG0112'
>>>>> 'MEG0111'
>>>>> 'MEG0122'
>>>>> 'MEG0123'
>>>>>
>>>>> IC labels:
>>>>> 'runica001'
>>>>> 'runica002'
>>>>> 'runica003'
>>>>> 'runica004'
>>>>> 'runica005'
>>>>>
>>>>> Matlab Output:
>>>>> ft_topoplotIC(cfg, comp);
>>>>> creating layout from data.grad
>>>>> creating layout for neuromag306alt system
>>>>> Error using ft_topoplotTFR (line 659)
>>>>> labels in data and labels in layout do not match
>>>>> Error in ft_topoplotIC (line 122)
>>>>> ft_topoplotTFR(cfg, varargin{:});
>>>>>
>>>>>
>>>>>
>>>>> Cheers,
>>>>> Dave Deriso
>>>>>
>>>>> --
>>>>> UCSD Institute for Neural Computation
>>>>> UCSD Department of Neurosurgery
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> fieldtrip mailing list
>>>>> fieldtrip at donders.ru.nl
>>>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Jörn M. Horschig
>>>>> PhD Student
>>>>> Donders Institute for Brain, Cognition and Behaviour
>>>>> Centre for Cognitive Neuroimaging
>>>>> Radboud University Nijmegen
>>>>> Neuronal Oscillations Group
>>>>>
>>>>> P.O. Box 9101
>>>>> NL-6500 HB Nijmegen
>>>>> The Netherlands
>>>>>
>>>>> Contact:
>>>>> E-Mail: jm.horschig at donders.ru.nl
>>>>> Tel:    +31-(0)24-36-68493
>>>>> Web: http://www.ru.nl/donders
>>>>>
>>>>> Visiting address:
>>>>> Trigon, room 2.30
>>>>> Kapittelweg 29
>>>>> NL-6525 EN Nijmegen
>>>>> The Netherlands
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> fieldtrip mailing list
>>>>> fieldtrip at donders.ru.nl
>>>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>>>
>>>>
>>>
>>>
>>> _______________________________________________
>>> fieldtrip mailing list
>>> fieldtrip at donders.ru.nl
>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>
>>
>>
>> _______________________________________________
>> fieldtrip mailing list
>> fieldtrip at donders.ru.nl
>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
>
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip




More information about the fieldtrip mailing list