subsetting from a structure array inside a for loop

Hello everybody, I am trying to write up a code that opens a bunch of eeglab datasets (set format), and stores a part of it for later. Here's my attempt:

conditionList={
    'subj1',
    'subj2',
    'subj3',
    'subj4'
}

chan = 'Cz' % the channel to use

for i = 1:length(conditionList)

      condition = char(conditionList(i));

      EEG = pop_loadset('filename', [condition '.set'], 'filepath', '~/MATLAB/');

      data = squeeze(EEG(i).data( strcmpi(chan, {EEG.chanlocs.labels}), :, 1:120)); 
      % the subset of the data we want to look at
      % EEG.data structure is made of three vectors: channels, points, and trials.

end

The following error pops up: "Index exceeds matrix dimensions". I am not an expert at matlab, so any advice would be greatly appreciated!

Answer 1

[Walter Roberson]

You have a comment that

% EEG.data structure is made of three vectors: channels, points, and trials

which suggests that any one EEG(i).data is a scalar structure with three fields . But in the line

      data = squeeze(EEG(i).data( strcmpi(chan, {EEG.chanlocs.labels}), :, 1:120));

you are treating any one EEG(i).data as if it is a 3D array. Is it really true that you have a 3D array of structures, each with three fields? Or should you perhaps be going down one layer, such as

      data = squeeze(EEG(i).data.points( strcmpi(chan, {EEG.chanlocs.labels}), :, 1:120));

??

I am also concerned because you have EEG(i), indicating that EEG is a structure array, and you have {EEG.chanlocs.labels} . With EEG being a structure array, for {EEG.chanlocs.labels} to be valid, each chanlocs would have to be a scalar structure, and each labels would have to be a character vector, which would tend to suggest that each EEG(i) is only for one channel, which would argue against the possibility that EEG(i).data being a structure array with channels as a field.