[Eeglablist] Fw: Error using DIPFIT

Makoto Miyakoshi mmiyakoshi at ucsd.edu
Tue May 5 08:20:11 PDT 2015 

Dear Claudio (cc Ramon),

Great job to underpin the cause of the problem. Now it sounds more familiar
to me. It seems possible that dipfit could fail with the
non-straightforward ica matrices. Please file the problem to eeglab
bugzilla to receive our further support.
http://sccn.ucsd.edu/wiki/EEGLAB_Bugs

> Thus, the question now is: Is there a way to remove components and still
be able to use the dipfit plugin?

Run dipfit first, then reject ICs.

Makoto

On Mon, Apr 27, 2015 at 5:52 AM, Claudio Georgii <
Claudio.Georgii at stud.sbg.ac.at> wrote:

> Hey,
>
> sry for my previous e-mail, the problem might not be related to the
> import-plugin you may use. More likely it seems the pop_subcomp function
> induces this error. If you remove components (like blink, saccades, etc.)
> you reduce the dimension of the data and this leads to this error message,
> triggered by some controlling mechanisms.
>
> Thus, the question now is: Is there a way to remove components and still
> be able to use the dipfit plugin?
>
> Thanks in advance!
>
> Best,
> Claudio
>
> 2015-04-26 11:33 GMT+02:00 Claudio Georgii <Claudio.Georgii at stud.sbg.ac.at
> >:
>
>> Hey,
>>
>> i investigated the issue further and can now share additional information
>> with you:
>>
>> The Data i load into EEGlab are .mat-Files created with the brain vision
>> analyser (BVA).
>> The error message only appears if i load the data using the
>> 'pop_bvaoi'-plugin and does not appear if i use the 'pop_biosig'-plugin. In
>> addition, this error occurs if i load the raw files (.EDF) using the
>> corresponding plugin.
>>
>> I hope this helps in understanding what goes wrong with this error
>> message.
>>
>> Best,
>> Claudio
>>
>> 2015-04-23 16:22 GMT+02:00 Claudio Georgii <
>> Claudio.Georgii at stud.sbg.ac.at>:
>>
>>> Dear Salim,
>>> Dear Makato,
>>>
>>> i run into the same problem after using the dipfit autofit or coarse
>>> grid option. I tried different Matlabversions (R2011a, R2013a, R2015a) as
>>> well as different EEGLAB versions (13_4_3b, 13_4_4b).
>>>
>>> Here is my error message in full:
>>>
>>> Error using ft_datatype_raw (line 85)
>>> inconsistent number of trials in raw data structure
>>>
>>> Error in ft_datatype_comp (line 105)
>>>       raw = ft_datatype_raw(raw, 'version', rawversion,
>>>       'hassampleinfo', hassampleinfo, 'hastrialinfo',
>>>       hastrialinfo);
>>>
>>> Error in ft_checkdata (line 223)
>>>   data = ft_datatype_comp(data, 'hassampleinfo',
>>>   hassampleinfo);
>>>
>>> Error in ft_dipolefitting (line 142)
>>> data = ft_checkdata(data, 'datatype', {'comp', 'timelock',
>>> 'freq'}, 'feedback', 'yes');
>>>
>>> Error in dipfit_gridsearch (line 106)
>>> source = ft_dipolefitting(cfg, comp);
>>>
>>> Error in pop_dipfit_gridsearch (line 135)
>>>   EEGOUT = dipfit_gridsearch(EEG, 'component', select,
>>>   'xgrid', xgrid, 'ygrid', ygrid, 'zgrid', zgrid,
>>>   options{:});
>>>
>>> Error in pop_multifit (line 133)
>>>         EEG = pop_dipfit_gridsearch( EEG, [1:ncomps], ...
>>>
>>> I as well tried combinations of different preprocessing steps, for
>>> example: with or without component rejection, average referencing, epoching
>>> before ICA etc.
>>>
>>> Nothing works im quite desperate now. Even a preprocessing script that
>>> worked before, does not work anymore.
>>>
>>> Here are the requested information of my EEG.data (after
>>> ICA_runica_extended):
>>>
>>> I have a 72 channels recording. I excluded 7 physiological channels
>>> (EOG, EKG, etc.) as well as 7 headelectrodes due to their noise level and 1
>>> marker Channel --> Current Number of Channel is 57.
>>>
>>> >> size(EEG.data)
>>>
>>> ans =
>>>
>>>     57   512    96
>>>
>>> >> size(EEG.icaact)
>>>
>>> ans =
>>>
>>>      0     0
>>>
>>> >> length(EEG.chanlocs)
>>>
>>> ans =
>>>
>>>     57
>>>
>>> Thanks in advance for your help!
>>>
>>> Best,
>>> Claudio
>>>
>>>
>>>
>>> 2015-03-27 18:52 GMT+01:00 Makoto Miyakoshi <mmiyakoshi at ucsd.edu>:
>>>
>>>> Dear Salim,
>>>>
>>>> I don't see anything strange.
>>>> Go back to the data before IC rejection and try dipfit. Does it work?
>>>> You repeat going back the version of the processed data and at some point
>>>> it should work. Then compare what's the difference between valid one and
>>>> invalid one. Something critical happened between them.
>>>>
>>>> Makoto
>>>>
>>>> On Fri, Mar 27, 2015 at 3:32 AM, Salim Al-wasity <
>>>> salim_alwasity at yahoo.com> wrote:
>>>>
>>>>> Hi
>>>>> This dataset has 38 channels and 30 ICs. After trying the mentioned
>>>>> commands above I got:
>>>>>
>>>>> size(EEG.data)
>>>>> ans =
>>>>>
>>>>>           38        2094          66
>>>>>
>>>>>
>>>>> size(EEG.icaact)
>>>>> ans =
>>>>>
>>>>>           30        2094          66
>>>>>
>>>>>
>>>>> length(EEG.chanlocs)
>>>>> ans =
>>>>>
>>>>>     38
>>>>>
>>>>> When I run Autofit I got:
>>>>>
>>>>> Error using ft_datatype_raw (line 88)
>>>>> inconsistent number of channels in trial 1
>>>>>
>>>>> Maximum number of iterations reached. Fitting failed
>>>>> Transforming electrode coordinates to match head model
>>>>> the input is component data with 30 components and 38 original channels
>>>>> the input is raw data with 38 channels and 66 trials
>>>>>
>>>>>
>>>>>
>>>>> Sincerely
>>>>> Salim
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>   On Thursday, 26 March 2015, 18:09, Makoto Miyakoshi <
>>>>> mmiyakoshi at ucsd.edu> wrote:
>>>>>
>>>>>
>>>>> Try
>>>>>
>>>>> size(EEG.data)
>>>>> size(EEG.icaact)
>>>>> length(EEG.chanlocs)
>>>>>
>>>>> What do you get?
>>>>>
>>>>> Makoto
>>>>>
>>>>> On Sat, Mar 21, 2015 at 4:29 AM, Salim Al-wasity <
>>>>> salim_alwasity at yahoo.com> wrote:
>>>>>
>>>>> Dear Mr. Makoto
>>>>> Thanks for your reply.
>>>>> I still get the same error message when I modify the 'dipole_fit). I
>>>>> am wondering why EEGLAB says
>>>>> (inconsistent number of trials in raw data structure)
>>>>> and (the input is component data with 32 components and 44 original
>>>>> channels
>>>>> the input is raw data with 44 channels and 1 trials)?
>>>>>
>>>>> Sincerely
>>>>> Salim
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>   On Thursday, 19 March 2015, 18:39, Makoto Miyakoshi <
>>>>> mmiyakoshi at ucsd.edu> wrote:
>>>>>
>>>>>
>>>>> Dear Salim,
>>>>>
>>>>> Edit 'dipole_fit' line 117 and 119 'maxiter'. These numbers could be
>>>>> x10 and it usually solves the problem.
>>>>>
>>>>> Makoto
>>>>>
>>>>> On Fri, Mar 6, 2015 at 6:44 AM, Salim Al-wasity <
>>>>> salim_alwasity at yahoo.com> wrote:
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> Dears
>>>>> I got the following error when I tried to estimate the Dipoles of a
>>>>> dataset after removing some components.
>>>>> I tried different versions of EEGLAB (v13.3.2 and v13.4.4)
>>>>>
>>>>> The error message is:
>>>>>
>>>>> Error using ft_datatype_raw (line 85)
>>>>> inconsistent number of trials in raw data structure
>>>>>
>>>>> Maximum number of iterations reached. Fitting failed
>>>>> Transforming electrode coordinates to match head model
>>>>> the input is component data with 32 components and 44 original channels
>>>>> the input is raw data with 44 channels and 1 trials
>>>>>
>>>>> Sincerely
>>>>> Salim
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
>>>>> To unsubscribe, send an empty email to
>>>>> eeglablist-unsubscribe at sccn.ucsd.edu
>>>>> For digest mode, send an email with the subject "set digest mime" to
>>>>> eeglablist-request at sccn.ucsd.edu
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Makoto Miyakoshi
>>>>> Swartz Center for Computational Neuroscience
>>>>> Institute for Neural Computation, University of California San Diego
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Makoto Miyakoshi
>>>>> Swartz Center for Computational Neuroscience
>>>>> Institute for Neural Computation, University of California San Diego
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Makoto Miyakoshi
>>>> Swartz Center for Computational Neuroscience
>>>> Institute for Neural Computation, University of California San Diego
>>>>
>>>> _______________________________________________
>>>> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
>>>> To unsubscribe, send an empty email to
>>>> eeglablist-unsubscribe at sccn.ucsd.edu
>>>> For digest mode, send an email with the subject "set digest mime" to
>>>> eeglablist-request at sccn.ucsd.edu
>>>>
>>>
>>>
>>
>


-- 
Makoto Miyakoshi
Swartz Center for Computational Neuroscience
Institute for Neural Computation, University of California San Diego
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20150505/f4cd0a90/attachment.html>

More information about the eeglablist mailing list