[Eeglablist] comparing condition and group ERP's from a STUDY

Eduardo Vianna eduardo.vianna at ucla.edu
Wed Oct 22 19:01:03 PDT 2008


Dear Hamish,

I have been running into the same problem as yours.  I also noticed that the
number of subjects you have in each group is different (this was also my
case).  I have run a test with same number of subjects in both groups (for
tutorial data and my own data), and plotting ERP data works.  However,  when
a run my data with different number of subjects in each group, I get the
same problem you described.  Plotting ERSP and ITC should work fine with
different number of subjects per group, but plotting ERP and Spectra results
don¹t.

Eduardo Vianna, PhD
Center for Neurobiology of Stress
Division of Digestive Diseases
University of California, Los Angeles
10945 Le Conte Ave., Ste 2338C
Los Angeles, CA 90095


On 10/19/08 6:06 PM, "Hamish INNES-BROWN" <HINNES-BROWN at bionicear.org>
wrote:

> Hi Arno, thanks for your reply.
>  
> The weird thing is, I'm certain that all your points below are true.  Both
> conditions are present in all subjects, and the group really is a group (ie
> groups '1' and '2' have different subjects - children vs adults, and are not
> ways to sneak in extra conditions).
>  
> Just to be sure, I wrote a short script to save the EEG.setnames, EEG.group,
> EEG.condition, EEG.ntrials, pnts, srate, xmin and xmax for all the sets in the
> Study to a csv file, and all was correct.
>  
> % script for extracting the number of trials of each stimulus, for each
> % subject in a STUDY.
> %STUDY must be currently open
> % loop over sets in study.
>  
> nsets = size(STUDY.datasetinfo,2)
>  
> %make an empty matrix to fill up with the data
> %rows = sets
> output = zeros(nsets, 8);
>  
> for currentset = 1:nsets;
>     csetname = EEG(currentset).setname;
>     ID = str2num(csetname(2:4));
>     output(currentset,1) = ID;
>     output(currentset,2) = str2num(EEG(currentset).group) ;
>     output(currentset,3) = str2num(EEG(currentset).condition);
>     output(currentset,4) = EEG(currentset).trials ;
>     output(currentset,5) = EEG(currentset).pnts ;
>     output(currentset,6) = EEG(currentset).srate ;
>     output(currentset,7) = EEG(currentset).xmin ;
>     output(currentset,8) = EEG(currentset).xmax ;
> end
>  
> Results here if interested
> http://spreadsheets.google.com/ccc?key=pMWrciDbEoFRAgnrYtXy9MQ
> <http://spreadsheets.google.com/ccc?key=pMWrciDbEoFRAgnrYtXy9MQ&inv=hinnesbrow
> n at gmail.com&t=4926724069117154563&guest>
>  
> The other weird thing is that every other plotting button works perfectly
> well, I can see Spect, ERSP, and ITC plots for both conditions and groups,
> which makes me think the data must be OK, as I guess all these measures are
> being calculated from the same data..?
>  
>  
> And all the electrodes are interpolated - I ended up using v6.01 just to do
> the electrode interpolation, then switching back to 6.03b(oct2008) to pre-calc
> all the channel measures.  Selecting the individual setfiles and plotting
> ERP's also works fine, for both groups and condition types, so I dont think
> the data is missing or wrongly arranged.  Any other hints?
>  
> Many thanks!
>  
>  
> Hamish Innes-Brown
> Senior Research Officer
> 
> The Bionic Ear Institute
> Auditory Clinical Neuroscience Unit
> c/o St Vincents Hospital
> 6th Floor Daly Wing,
> 35 Victoria Pde, Fitzroy Vic 3065
> Tel: +61 3 9288 3523
> Fax: +61 3 6997 7518
> 
>  
> 
> 
> From: eeglablist-bounces at sccn.ucsd.edu
> [mailto:eeglablist-bounces at sccn.ucsd.edu] On Behalf Of arno delorme
> Sent: Saturday, 18 October 2008 9:58 PM
> To: Hamish INNES-BROWN
> Cc: eeglablist at sccn.ucsd.edu
> Subject: Re: [Eeglablist] comparing condition and group ERP's from a STUDY
> 
> Dear Hamish, 
> 
> some more information about groups and conditions. Optimally:
> 
> - All conditions are supposed to be present in all subjects. If not, the
> program should still work but might lead to the empty plots you are seeing
> them right now.
> - Groups are supposed to be for different subjects or different sessions
> (different ICA decomposition for sure). Using groups within the same subjects
> (to define a second condition) might work but not always. This is pushing the
> limits of the program.
> - You should interpolate missing electrodes. Once more, if you do not do that,
> the program should work but the average will take into account different
> subjects at each channel location.
> 
> We are working on these limitations. You will be able to select several
> conditions and/or groups.
> 
> Best,
> 
> Arno
> 
> On 14 oct. 08, at 07:35, Hamish INNES-BROWN wrote:
> 
>>  
>>  
>> 
>> Hi EEGLAB list. 
>>  
>> 
>> I'm currently trying to track down the source of a  problem i have plotting
>> condition and group ERP's from the STUDY channel  measures menu.
>>  
>> 
>> I have all the 'group' and 'condition' data loaded  into the EEG.group and
>> EEG.condition.  The correct group and condition  numbers pop up in the study
>> info window when all the subjects are loaded into  the study..  In EEG.group
>> and EEG.condition the values are strings (but  numerals).  I have groups '1'
>> and '2', and conditions '111' and  '122'.
>>  
>> 
>> The study looks like this
>>>> >> STUDY 
>> STUDY =  
>>            name:  'Study1'
>>     datasetinfo:  [1x58 struct]
>>         cluster: [1x1 struct]
>>          changrp: [1x63 struct]
>>            task:  ''
>>           notes: ''
>>        filename:  'STUDY_ALL.study'
>>        filepath:  'D:\_PROJECTS\_FB\_ANALYSIS\__EEGLAB\SET\08STUDY_ALL\'
>>         history: [1x3126  char]
>>         subject: {1x29 cell}
>>           group:  {'1'  '2'}
>>         session: []
>>       condition: {'111'   '122'}
>>          setind: [2x29  double]
>>             etc:  [1x1 struct]
>>        preclust: [1x1 struct]
>>           saved:  'yes'
>>  
>> 
>> Now, when plotting the channel measures, I select  "plot conditions on the
>> same panel", and get two plots as expected.  The  first (group '1') has two
>> lines, as expected.  One is labelled '111' and  the other '122'. All OK.
>> However underneath there is an empty plot,  labelled '2', but with no lines.
>>  
>> 
>> I thought this might be somehow a fault in my data  or the way the conditions
>> and groups are entered (maybe they should be  numbers, not strings?) so I
>> tried loading the example data (5subjects.zip),  and added a group variable.
>> Same problem however.
>> 
>>  
>> 
>> Anyone had any luck with this before?
>>  
>> 
>> Many thanks 
>> 
>> 
>> 
>>  
>> 
>> Hamish Innes-Brown
>> Senior Research Officer
>>  
>> 
>> The Bionic Ear Institute
>> Auditory Clinical Neuroscience  Unit
>> c/o St Vincents  Hospital
>> 6th  Floor Daly Wing,
>> 35  Victoria Pde, Fitzroy Vic 3065
>> Tel: +61 3 9288 3523
>> Fax: +61 3 6997 7518
>> _______________________________________________
>> 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
> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20081022/4a0d2a28/attachment.html>


More information about the eeglablist mailing list