分配给结构数组时在Matlab中“索引超过矩阵维数”

Question

I am trying to assign values to an array contained within a struct in Matlab. I am perplexed by the "Index exceeds matrix dimensions" error, because as far as I can tell, I am following the example of the official documentation and doing something that has worked for me in the past. I hope this is enough relevant context:

for m = 1:tally
    matlabbatch{2}.spm.stats.fmri_spec.sess(m).scans = files{m};
    matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(1).name = 'button_press';
    matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(1).onset = events.bp.onsets{m};
    matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(1).duration = 0;
    matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(1).tmod = 0;
    matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(1).pmod = struct('name', {}, 'param', {}, 'poly', {});

    sess_onsets = events.conditions.sessions(tally).onsets{1};
    for n = 1:numel(sess_onsets)
        len = numel(sess_onsets{n});
        cond_list = events.conditions.names{tally};
        cond_onsets = sess_onsets{n};
        matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(n+1).name = cond_list{n};
        matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(n+1).onset = cond_onsets;
        matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(n+1).duration = repmat([4], len, 1);
        matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(n+1).tmod = 0;
        matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(n+1).pmod = struct('name', {}, 'param', {}, 'poly', {});
    end

    matlabbatch{2}.spm.stats.fmri_spec.sess(m).multi = {''};
    matlabbatch{2}.spm.stats.fmri_spec.sess(m).regress = struct('name', {}, 'val', {});
    matlabbatch{2}.spm.stats.fmri_spec.sess(m).multi_reg = rp{k};
    matlabbatch{2}.spm.stats.fmri_spec.sess(m).hpf = 128;
end

The error comes on this line:

matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(n+1).name = cond_list{n};

When I step to this line in the debugger, it is this expression

matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(n+1)

that gives the "index exceeds matrix dimensions error." It does this even when n is 1. Abstractly, based on my experience with other languages, that makes sense. Cond is length 1 and I'm trying to assign to the second slot. But it doesn't make sense to me based on my experience with and (obviously flawed) knowledge of arrays in Matlab structs. In the past I've just made a struct array longer by announcing I'm going to with the new subscript.

Here are Matlab's instructions for how to assign to a struct . The only difference I can see is that they don't use the (1) subscript the first time they assign to the array, and I tried that and it made no difference.

This code works fine:

>> a.b(1).name = 'apple';
>> a.b(2).name = 'banana'
a = b: [1x2 struct]

I am struggling to understand the difference. Any assistance you can provide is greatly appreciated, and please let me know if I need to give more information. Thanks!

Answer 1

Thanks very much to all who tried to help me. I am feeling like a doof, and somewhat guilty for not communicating accurate information to people who were trying to help me. This question should have been titled "I don't know how to use a debugger." I thought it was failing on the first execution because when I stepped to that line and pasted in

matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(n+1) 

I got an error. But I now understand that the reason for that error is that pasting in that expression does not do the work of creating the next array element -- I needed

matlabbatch{2}.spm.stats.fmri_spec.sess(m).cond(n+1).name = foo

to make that element exist. If I had clicked "continue" I would have realized that the real failure was in cond_list{n} when n = 5. This was just a conceptual failure in my code, and not anything anyone else would likely have recognized. Thanks very much for your time, and I'm sorry for leading you astray with my claims.