[FieldTrip] ICA

Robert Oostenveld r.oostenveld at donders.ru.nl
Sat Feb 26 10:15:35 CET 2011


Hi Holly

The ft_databrowser function was initially written for identifying segments (artifacts, sleep spindles and spike-wave discharges) from continuous data. We wanted those segments to be marked relative to the original data file, i.e. the segments shoudl be marked by their begin- and end-sample.

Over time, we started using the function for more and more stuff, but there are still some "relicts" in the code that rely on the sample indexing. We are working on improving the function.

To keep you posted on the improvements, I have added you and Vladimir to the respective "bugs" in our bugzilla system. Through bugzilla you will receive an email when the issues are resolved so you know when to download an improved version that will work for you. 

best regards
Robert


On 24 Feb 2011, at 17:06, Vladimir Litvak wrote:

> Hi Holly,
> 
> I'm not sure why the databrowser needs the trl struct or whether we
> can add something to our exported FT struct to make it happy. I'm
> CCing to the Fieldtrip list to perhaps get some answers. I got exactly
> the same complaint from someone else last week.
> 
> 
> Best,
> 
> Vladimir
> 
> 
> On Thu, Feb 24, 2011 at 2:48 PM, Holly Rossiter
> <H.Rossiter at ion.ucl.ac.uk> wrote:
>> Hi Vladimir,
>> 
>> 
>> 
>> I am trying out using ICA on my data to see what difference it makes to the
>> eye blinks and whether it is a step worth doing or not. I am using a
>> fieldtrip script...
>> 
>> 
>> 
>> D= spm_eeg_load;
>> 
>> data = D.ftraw(0);
>> 
>> cfg = [];
>> 
>> cfg.channel = 'MEG';
>> 
>> ic_data = ft_componentanalysis(cfg,data);
>> 
>> cfg = [];
>> 
>> cfg.viewmode = 'component';
>> 
>> cfg.continuous = 'yes';
>> 
>> cfg.layout = 'CTF275.lay';
>> 
>> cfg.blocksize = 30;
>> 
>> cfg.channels = [1:10];
>> 
>> ft_databrowser(cfg,ic_data);
>> 
>> 
>> 
>> But it comes up with an error saying
>> 
>> 
>> 
>> Error using ==> ft_databrowser at 217
>> 
>> no trials to display
>> 
>> 
>> 
>> I think this may be because I don’t have a trl matrix due to using
>> spm_eeg_load. Do you know how I can create one of these using spm format
>> data? Or how else to write the script so that ICA will work?
>> 
>> 
>> 
>> Kind regards,
>> 
>> 
>> 
>> Holly
> 
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip





More information about the fieldtrip mailing list