[Eeglablist] Problems with Creating Study and running precomputing - preclustering

Fri Jul 1 16:11:40 PDT 2016

Dear Maya,

Sorry for slow response.

> And k-means algorithm worked when I chose the option k-means(no toolbox).

That's good to hear. Sometimes, these rarely-used options do not work well
since rarely checked or bug-reported.

The error you reported does not make sense to me. There must be a problem
is data preprocessing or setup. Make sure that you did NOT separate .set
files according to within-subject conditions, since there is known bug
related to that.

Other than that, I really can't say anything for sure unless I take a look
at data. If you haven't tried yet, you may want to follow EEGLAB STUDY
tutorial with the downloaded data to confirm our steps.

Sorry for just providing general solutions. Maybe you want to describe your
STUDY design. For example, what is the factorial design?

Makoto

On Fri, Jun 10, 2016 at 10:55 AM, Maya Pinacle <mayapinacle at gmail.com>
wrote:

> Hi,
>
> I realized that the preclustering parameters were kept under
> Study.etc.preclust rather than Study.preclust, which is a bit confusing
> really. And k-means algorithm worked when I chose the option k-means(no
> toolbox). So all my previous problems are solved.
>
> However even though I run precomputing and preclustering for all the
> options offered, in the pop_clustedit() GUI, except Plot Dipoles and Plot
> Spectra, non of the other options are working nor ploting a cluster or
> neither ploting a single component.
>
> For plotting the scalp maps and ERPs I am getting error ''EEGLAB error in
> function std_comppol() at line 62. Attempted to access r(2); index out of
> bounds because numel(r)=1.''
>
> And for plotting ERPimage, ERPSs and ITCs I get this error ''SUMSKIPNAN:
> DIM argument larger than the number of dimensions of x ( Error occured in
> function sumskipnan() at line 109)''.
>
> Obviously there is a matrix dimension problem at some point but I really
> cannot understand where it might be coming from.
>
> So I would really appreciate any help or comment on that please.
>
> Maya
>
>
>
>
>
>
>
>
>
> 2016-06-09 12:56 GMT+01:00 Maya Pinacle <mayapinacle at gmail.com>:
>
>> Hi Makoto,
>>
>>
>> thank you very much for the answer. I tried what you described in your
>> email. I didn't get the error message then. But the wierd thing is when I
>> run Build Preclustering Matrix from the Study > PCA Clustering Menu it runs
>> without no error, however at the end of the whole process there is no any
>> record of parameters in the Study.preclust structures (erpclusttimes,
>> specclustfreqs, erspclustfreqs, erspclusttimes), they are all empty is that
>> normal?
>>
>> And then when I try to run Cluster Components from the menu with K-means
>> and choosing 3 cluster for example, I get an error that is saying '' Too
>> many input arguments (Error occured in function pop_clust() at line 184)
>> which line calls k-means function actually.
>>
>> So I am stuck really and do you think is there any link between not
>> having any record of parameters in the anywhere under Study structure and
>> not being able to run Cluster Components? And do you think these all again
>> go down to the STUDY cannot read the data sets properly even though it
>> seems running the precomputing and preclustering without an error?
>>
>> For your first question, I have actually 5 condition, for each condition
>> I try to do seperate network analyisis to see within-condition network
>> using ICs belongs only to that condition. Do you think this might be a
>> problem as well? I mean I am not interested in with statistical comparison
>> between the conditions at this stage, the only thing that I am interested
>> in at the moment is to see whether is there a network specific to each
>> single condition.
>>
>> Thank you very much for any help indeed,
>>
>> Maya
>>
>>
>>
>>
>> 2016-06-09 3:09 GMT+01:00 Makoto Miyakoshi <mmiyakoshi at ucsd.edu>:
>>
>>> Dear Maya,
>>>
>>> Sorry for the trouble.
>>>
>>> > each collected for only one condition
>>>
>>> Do you mean there is no within-subject condition?
>>>
>>> > ''Warning: ICA ERSP or ITC data files computed with old version of
>>> EEGLAB for design 1 (and maybe other designs). These files may be corrupted
>>> and must be recomputed.''
>>>
>>> This message has been there for LONG time. It's a false alarm.
>>>
>>> > '' Error using eeglab (line 1006)
>>> Error: ()-indexing must appear last in an index expression.
>>> You are using the latest version of EEGLAB.
>>> Warning: STUDY moved since last saved, trying to load data files using
>>> relative path
>>> pop_loadset(): loading file C:\Users\tc025847\Desktop\Mitchell Replica
>>> Experiment Files\EEG only\DATA\3\Category_5_ICAremoved.set ...
>>> Reading float file 'C:\Users\tc025847\Desktop\Mitchell Replica
>>> Experiment Files\EEG only\DATA\3\Category_5_ICAremoved.fdt'... ''
>>>
>>> I don't know about this error.
>>> Maybe your folder structure is bad... my guess is that STUDY is trying
>>> to use relative path but failing. What I would try is that I would move all
>>> the .set and precompute files into one folder and create another STUDY
>>> there with exact same data and STUDY.design. In this way, you can skip
>>> precompute and immediately start from precluster.
>>>
>>> Makoto
>>>
>>>
>>> On Tue, Jun 7, 2016 at 3:45 AM, Maya Pinacle <mayapinacle at gmail.com>
>>> wrote:
>>>
>>>> Hi,
>>>>
>>>> I have just started using STUDY with EEGLAB, I have data from 10
>>>> subjects in different 10 folders, each collected for only one condition and
>>>> I would like to extract clusters of ICs of those datasets using STUDY. I
>>>> have already run ICA and Dipfit to extract ICs and corresponding dipoles.
>>>> When I create a Study with those 10 subject's data and save it to a folder
>>>> and again want to load it back to the EEGLAB, I always get this warning
>>>> message in a popup window which says
>>>>
>>>>
>>>> ''Warning: ICA ERSP or ITC data files computed with old version of
>>>> EEGLAB for design 1 (and maybe other designs). These files may be corrupted
>>>> and must be recomputed.''
>>>>
>>>>
>>>> and also in the command line there is a warning saying
>>>>
>>>> '' Error using eeglab (line 1006)
>>>> Error: ()-indexing must appear last in an index expression.
>>>> You are using the latest version of EEGLAB.
>>>> Warning: STUDY moved since last saved, trying to load data files using
>>>> relative path
>>>> pop_loadset(): loading file C:\Users\tc025847\Desktop\Mitchell Replica
>>>> Experiment Files\EEG only\DATA\3\Category_5_ICAremoved.set ...
>>>> Reading float file 'C:\Users\tc025847\Desktop\Mitchell Replica
>>>> Experiment Files\EEG only\DATA\3\Category_5_ICAremoved.fdt'... ''
>>>> ...
>>>>
>>>> This message continues for each single file that I have used to create
>>>> STUDY. However there is no change of datesets' location at all but STUDY
>>>> seems struggling to find the datasets. Eventually EEGLAB seems loading the
>>>> *.study file and I run preclustering on the data, however there is no
>>>> preclustering parameters are set either.
>>>>
>>>>
>>>> These are the code lines how I run precomputing and preclustering
>>>>
>>>> % precomputing
>>>>  [STUDY ALLEEG] = std_precomp(STUDY, ALLEEG,
>>>> 'components','erp','on','erpparams',{'rmbase' [-200 1]
>>>> },'scalp','on','spec','on','specparams',{'specmode' 'fft' 'logtrials'
>>>> 'off'},'erpim','on','erpimparams',{'nlines' 10 'smoothing'
>>>> 10},'ersp','on','erspparams',{'cycles' [3 0.5] 'freqs' [3 45] 'alpha' 0.01
>>>> 'padratio' 1},'itc','on');
>>>>
>>>>
>>>>  % preclustering
>>>>  [STUDY ALLEEG] = std_preclust(STUDY, ALLEEG, 1,{'spec' 'npca' 5 'norm'
>>>> 1 'weight' 1 'freqrange' [3 45] },{'erp' 'npca' 5 'norm' 1 'weight' 1
>>>> 'timewindow' [-500 1000] },{'scalp' 'npca' 5 'norm' 1 'weight' 1 'abso'
>>>> 1},{'dipoles' 'norm' 1 'weight' 10},{'ersp' 'npca' 5 'freqrange' [3 45]
>>>> 'timewindow' [-500 1000]  'norm' 1 'weight' 1},{'itc' 'npca' 5 'freqrange'
>>>> [3 45]  'timewindow' [-500 1000]  'norm' 1 'weight' 1},{'finaldim' 'npca'
>>>> 5});
>>>>
>>>>
>>>> These codes run smoothly however there is no parameters stored in
>>>> theStudy.cluster.precluster.preclusterparams and preclusterdata
>>>>
>>>>
>>>> And when I run all these processes on the sample data I got the same
>>>> error messages and nothing was stored in those Study structures at all.
>>>>
>>>>
>>>> Is there anyone encountered this problem before? Do you have any idea
>>>> what could be the problem? I would appreciate any help please.
>>>>
>>>> Thank you,
>>>>
>>>> Maya.
>>>>
>>>> _______________________________________________
>>>> 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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20160701/170a51f9/attachment.html>