Dear Martin,
What you did won't work because the elements in trialdef correspond to
event types and not to single events. It's not possible to use
spm_eeg_definetrial to assign two different trial labels in this
situation because there is only one event type in the file. What you
should do is either epoch first and reassign condition labels later as
shown in the manual (look a little bit further down) or you can do
something like:
----
labels = textread('condition_labels.txt', '%s');
D = spm_eeg_load('Mdspm8_faces_run1.mat');
S = [];
S.dataset = fullfile(D.path, D.fname);
S.pretrig = 400;
S.posttrig = 600;
S.trialdef(1).conditionlabel = 'whatever';
S.trialdef(1).eventtype = 'STATUS';
S.trialdef(1).eventvalue = 1;
S.reviewtrials = false;
S.save = false;
trl = spm_eeg_definetrial(S);
S = [];
S.D = D;
S.epochinfo.trl = trl;
S.epochinfo.conditionlabels = labels;
D = spm_eeg_epochs(S)
-----
Best,
Vladimir
On Tue, Nov 10, 2009 at 6:12 PM, Martin Luessi
<[log in to unmask]> wrote:
> Hi,
>
> I'm following the EEG preprocessing instructions for the "Multimodal
> Faces" data set in SPM8. Unfortunately copy-pasting the conditions
> labels as described under 34.4.5 does not work for me (it does only
> paste into once cell, not the whole column, I'm using Matlab R2007b
> under Linux). So I wrote a script to do the epoching and assign the
> right condition labels. However, I must be doing something wrong, as
> Matlab stays busy forever (epoching as described under 34.4.4 is very
> fast). I'm using the following script:
>
> -------------------------------------------------------------------------------
> labels = textread('condition_labels.txt', '%s');
>
> S = struct;
> S.D = spm_eeg_load('Mdspm8_faces_run1.mat')
>
> S.pretrig = 400;
> S.posttrig = 600;
> S.trialdef = struct
>
> for ii=1:length(labels)
> S.trialdef(ii,1).conditionlabel = labels{ii};
> S.trialdef(ii,1).eventtype = 'STATUS';
> S.trialdef(ii,1).eventvalue = '1';
> end
>
> S.reviewtrials = false
> S.save = true
>
> D = spm_eeg_epochs(S)
> --------------------------------------------------------------------------
>
> Any hints on what I'm doing wrong are greatly appreciated.
>
> Martin
>
> ______________________________________
> Martin Luessi, M.S.
> PhD Candidate
> Department of Electrical Engineering and Computer Science
> Northwestern University
> http://ivpl.eecs.northwestern.edu/people/mluessi
>
