TDT: problem identifying regressors

Summary of what happened:

I’m trying to edit the decoding_template to get my analysis to run, but I keep getting an error that I’m not sure I understand. I believe - but I’m not sure - that it isn’t able to correctly pool my regressors. My data is from an SPM analysis where we had 2 conditions (‘grammatical’ and ‘ungrammatical’) with several variants of each condition (several different types of ‘grammatical’ and an equal number of ‘ungrammatical’ types). We have one beta map for each type of each condition (28 ‘grammatical’ and 28 ‘ungrammatical’). The regressors are named things like “HGa” “HGb” “LGa” etc. (for the ‘grammatical’ ones) and “HNGa” “HNGb” “LNGa” etc. (for the ‘ungrammatical’ ones). I’m not exactly sure how to properly list this in the TDT decoding_template. I tried this:

labelname1 = ‘regexp:[HL]G[a-n]’
labelname2 = ‘regexp:[HL]NG[a-n]’

However when try to run the edited decoding_template (which I’ve renamed decoding_template_MVPA_subj001) I get the following error:

>> decoding_template_MVPA_subj001

labelname1 =


labelname2 =


getting betas from /Volumes/data1/Complex_seq_analysis/001/MVPA
Error using make_design_cv (line 230)
Empty decoding steps found in design in step(s) 1. Maybe you have only one chunk and want to do
cross-validation (which doesn't make sense).

Error in decoding_template_MVPA_subj001 (line 98) = make_design_cv(cfg);

Any ideas what this means? Any suggestions about how to get tdt to do a MVPA on this data?

Thanks for any thoughts!


Looking at this thread: TDT: Decoding having only one run, I believe that the issue is I have only one run in my design. However, we have separate betas for each individual stimulus, which represent one of two conditions (and these individual stimuli were presented in different random orders to each subject), so I should be able to use these. However, I don’t understand how to get TDT to do this. It seems to expect separate runs, which we don’t have.