[FieldTrip] using ft_definetrial

Tue Jul 14 14:53:03 CEST 2015

Hi Nikola,

I have no experience with EEG data myself. But If I understand correctly, the relevant trigger information is in your .trg-file.
So, if you can define variable triggers as below, it should be relatively straigthforward:

triggers= [  0.000000         0 0
20.568359   2490491 1
23.787109   2879419 9
28.597656   3460687 1
33.056641   3999475 3
45.923828   5554243 1
...]
It seems that you have 118*1024 samples (segments) per second, is that correct? And your EEG-data is time-labeled in segments, not seconds?

Now, your trigger values of interest (1) are in column 3, not in 2, that is the crucial bit probably:

u1=find(ismember(triggers(:,3),[1]));

And then you can do

trl= [triggers(u1,2)-.2*1024 triggers(u1,2)+6*1024 repmat(-.2*1024,numel(u1),1)];

You might want to change 0.2 into e.g. 0.25, since otherwise you might end up with non-integer values.

To get rid of your last error, make sure that there are no negative sample numbers in your cfg.trl (since there’s no data befor sample nr 0).

Best,
Stan

From: fieldtrip-bounces at science.ru.nl [mailto:fieldtrip-bounces at science.ru.nl] On Behalf Of Nikola Valchev
Sent: dinsdag 14 juli 2015 14:36
To: 'FieldTrip discussion list'
Subject: Re: [FieldTrip] using ft_definetrial

Hi Pelt,
Thanks a lot for the help. Somehow I forgot that I can specify my own trl matrix manually...
(I’m sorry for the long email but I try to be as clear as possible so that it takes less time.)

However, there is still something wrong which I don’t understand. What I do is to use ft_read_event to read all the triggers and events from the cnt file. What I get is the event structure with the fileds
type
sample
value
offset
duration

I presume that these are read from the trg file (i.e. s14_CI_1.trg which is associated with the s14_CI_1.cnt).

Inside that file I have:
  0.000000         0 0
20.568359   2490491 1
23.787109   2879419 9
28.597656   3460687 1
33.056641   3999475 3
45.923828   5554243 1
...

First column should be the times, second the segment and the third the trigger. I use your commands to create the matrix:
u1=find(ismember(triggers(:,2),[1]));

trl= [triggers(u1,2)-.2*1024 triggers(u1,2)+6*1024 repmat(-.2*1024,numel(u1),1)];

And at the end I get a matrix:

  1.0e+003 *

   -0.2038    6.1450   -0.2048
   -0.2038    6.1450   -0.2048
   -0.2038    6.1450   -0.2048
   -0.2038    6.1450   -0.2048
...

Then I attempt to preprocess again:
cfg=[];
cfg.dataset='C:\eeg_anita\data\S14\s14_CI_1.cnt';
cfg.headerfile='C:\eeg_anita\data\S14\s14_CI_1.cnt';
cfg.trl=trl;
cfg_test=ft_preprocessing(cfg);

and I get an error:
??? Error using ==> ft_read_data at 206
cannot read data before the begin of the file

Error in ==> ft_preprocessing at 563
      dat = ft_read_data(cfg.datafile, 'header', hdr, 'begsample', begsample, 'endsample', endsample, 'chanindx', rawindx,
      'checkboundary', strcmp(cfg.continuous, 'no'), 'dataformat', cfg.dataformat)

Nikola

From: fieldtrip-bounces at science.ru.nl<mailto:fieldtrip-bounces at science.ru.nl> [mailto:fieldtrip-bounces at science.ru.nl] On Behalf Of Pelt, S. van (Stan)
Sent: 14 July 2015 09:25
To: FieldTrip discussion list
Subject: Re: [FieldTrip] using ft_definetrial

Hi Nikola,

You might also try a workaround:

% read the events from the data
event = ft_read_event('C:\eeg_anita\data\S14\s14_CI_1.cnt');

% search for "trigger" events
triggers=[];
for i=1:size(event,1)
     a=find(strcmp(event(i).type,'trigger')==1);
     if a>0
         triggers=[triggers; event(i).value event(i).sample];
     end
end

% make cfg.trl: I think this is where things go wrong in your script. You could check what’s in triggers(:,1). Otherwise, what’s below should work:
u1=find(ismember(triggers(:,1),[1 3 6 9])); % select your eventvalue(s) of interest

trl= [triggers(u1,2)-.2*fs triggers(u1,2)+6*fs repmat(-.2*fs,numel(u1),1)]; % 0.2 prestim, 6s post-stim, with fs being your sampling frequency

% next, do preprocessing, with cfg.trl=trl;

From: fieldtrip-bounces at science.ru.nl<mailto:fieldtrip-bounces at science.ru.nl> [mailto:fieldtrip-bounces at science.ru.nl] On Behalf Of Jörn M. Horschig
Sent: dinsdag 14 juli 2015 8:38
To: 'FieldTrip discussion list'
Subject: Re: [FieldTrip] using ft_definetrial

Hi Nikola,

seems to be a nasty problem. I’d start with using event = ft_read_events(cfg.dataset) and then check up the event-structure on how they are defined there. ft_trialfun_general is literally checking that event-structure, so defining them like they are specified in there might help.

Best,
Jörn

--

Jörn M. Horschig, PhD, Software Engineer
Artinis Medical Systems<http://www.artinis.com/>  |  +31 481 350 980

From: fieldtrip-bounces at science.ru.nl<mailto:fieldtrip-bounces at science.ru.nl> [mailto:fieldtrip-bounces at science.ru.nl] On Behalf Of Nikola Valchev
Sent: Monday, July 13, 2015 9:51 PM
To: 'FieldTrip discussion list'
Subject: Re: [FieldTrip] using ft_definetrial

Nope. That’s not it. No matter if I declare it as a double or string when I call the eventtype ‘trigger’ instead of ‘?’ they are not read.

Nikola

From: fieldtrip-bounces at science.ru.nl<mailto:fieldtrip-bounces at science.ru.nl> [mailto:fieldtrip-bounces at science.ru.nl] On Behalf Of Pelt, S. van (Stan)
Sent: 13 July 2015 19:30
To: FieldTrip discussion list
Subject: Re: [FieldTrip] using ft_definetrial

What about

cfg.trialdef.eventvalue     = 1;

Op 13 jul. 2015, om 18:36 heeft Eelke Spaak <eelke.spaak at donders.ru.nl<mailto:eelke.spaak at donders.ru.nl>> het volgende geschreven:

cfg.trialdef.eventvalue     = {'1'};

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20150714/38844240/attachment.html>