forming one datset from multiple data files

[David Ziegler]

Hi Jan-Mathijs,

Thanks for the detailed response!  I am very much aware of the dangers
of concatenating data over sessions and assuming the sensor space is the
same.  In my case, the "runs" were all acquired during a single session
(i.e., 6 runs, 7 min each, in a single 50 min session) in which head
position was pretty carefully monitored.

Your trick of manually defining allT4.grad to be the same as the
original data file works just fine.  I did originally try simply
specifying cfg.layout = NM306mag.lay (as well as other NM306***.lay
options), and these resulted in plots, but they were simply empty square
line grids with four boxes.  Not sure why this was the case, but as long
as your fix works, I am all set for the moment.

Thanks!
David

jan-mathijs schoffelen wrote:
> Dear David,
>
> The reason why the sensor info is explicitly removed by ft_appenddata
> is to ensure that people realize that combining multiple sessions may
> be problematic or even downright 'forbidden' for some subsequent steps
> in the analysis. Think of e.g. doing source analysis for a single
> subject in which several sessions are combined. Since the subject's
> position was slightly different during each recording sessions, there
> is in fact not a guarantee that during one of the sessions the subject
> would have sat facing backwards ;o). The leadfields computed in such a
> case (appending with in one of the sessions the subject facing
> backwards) will clearly be wrong for most of the data. Of course if
> you were able to somehow compensate for the differences in position,
> e.g. by applying the maxfilter, things may be different.
> Yet, indeed for visualizing the results, and if you are confident that
> there were no gross differences across the sessions with respect to
> the positioning of the subject, there is no objection against keeping
> the gradiometer info. Although I am a bit puzzled by the fact that you
> do not seem to be able to visualize the data as you have it (because I
> thought that provided you give the plotting function an appropriate
> layout-file, in your case something like NM306xxx.lay, I would assume
> that it just works even without sensor position info; for the layout
> files, have a look in fieldtrip/templates, or at the wiki), you could
> of course 'fool' fieldtrip by appending a grad-structure to your
> concatenated data: allT4.grad = dataT4_list1.grad;
>
> Hope this helps,
> Jan-Mathijs
>
>
> On Apr 14, 2010, at 9:19 PM, David Ziegler wrote:
>
>> Hi Fieldtrippers,
>>
>> I have a similar situation where I have 3 "runs" of trials that were
>> collected separately on a neuromag306 system.  I took Ingrid's advice
>> and
>> ran ft_appenddata on my preprocessed (e.g., trigger-based trial
>> selection,
>> artifact rejection, and preprocessing) data files to combine the three
>> datasets into a single file.  The function worked, but with the
>> warning that
>> the sensor info was not consistent across trials:
>>
>>>> cfg=[];
>>>> allT4 = ft_appenddata(cfg, dataT4_list11, dataT4_list8, dataT4_list9);
>> input dataset 1, 308 channels, 32 trials
>> input dataset 2, 308 channels, 32 trials
>> input dataset 3, 308 channels, 32 trials
>> Warning: sensor information does not seem to be consistent across the
>> input
>> arguments
>>> In ft_appenddata at 106
>> concatenating the trials over all datasets
>> removing sensor information from output
>> output dataset, 308 channels, 96 trials
>>
>> Is there a better way to concatenate several runs of similar trials such
>> that the sensor information is preserved?  I can generate an time-locked
>> average on the resulting concatenated data, but I am not able to plot it
>> using multiplot or topoplot, just by viewing individual single channels
>> (presumably due to the stripping of the sensory info).
>>
>> Thanks for any advice!
>> David
>>
>> ----------------------------------
>> The aim of this list is to facilitate the discussion between users of
>> the FieldTrip  toolbox, to share experiences and to discuss new ideas
>> for MEG and EEG analysis. See also
>> http://listserv.surfnet.nl/archives/fieldtrip.html and
>> http://www.ru.nl/neuroimaging/fieldtrip.
>>
>
> Dr. J.M. (Jan-Mathijs) Schoffelen
> Donders Institute for Brain, Cognition and Behaviour,
> Centre for Cognitive Neuroimaging,
> Radboud University Nijmegen, The Netherlands
> J.Schoffelen at donders.ru.nl
> Telephone: 0031-24-3668063
>
> ----------------------------------
> The aim of this list is to facilitate the discussion between users of
> the FieldTrip  toolbox, to share experiences and to discuss new ideas
> for MEG and EEG analysis. See also
> http://listserv.surfnet.nl/archives/fieldtrip.html and
> http://www.ru.nl/neuroimaging/fieldtrip.

--

David A. Ziegler

Department of Brain and Cognitive Sciences
Massachusetts Institute of Technology
43 Vassar St,  46-5121
Cambridge, MA  02139

Tel: 617-258-0765

Fax: 617-253-1504

daz at mit.edu <mailto:daz at mit.edu>






----------------------------------
The aim of this list is to facilitate the discussion between users of the FieldTrip  toolbox, to share experiences and to discuss new ideas for MEG and EEG analysis. See also http://listserv.surfnet.nl/archives/fieldtrip.html and http://www.ru.nl/neuroimaging/fieldtrip.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20100415/3ba66d53/attachment-0002.html>