Error in 'signed_decision_values'

dear TDT-experts,
Thank you for your pretty toolbox !

I have an error when using the ‘signed_decision_values’ instead of ‘accuracy_minus_chance’. (I have only 2 runs and 2 conditions).
I am lost…

Reference to non-existent field ‘Label’.
Error in transres_signed_decision_values (line 51)
prev = sort(decoding_out(1).model.Label);
Error in decoding_transform_results (line 260)
output = feval(fhandle,decoding_out,chancelevel,cfg,data);
Error in decoding_generate_output (line 35)
output = decoding_transform_results(curr_output,decoding_out,chancelevel,cfg,data);
Error in decoding (line 568)
results = decoding_generate_output(cfg,results,decoding_out,i_decoding,curr_decoding,current_data);

clear cfg
cfg = decoding_defaults = ‘correlation_classifier’
cfg.decoding.method = ‘classification’;
cfg.results.output = ‘signed_decision_values’;

cfg.results.dir = ‘blabla/06’;
beta_loc = ‘blabla/stat_ind’;
cfg.files.mask = ‘blabla/stat_ind/mask.nii’;
regressor_names = design_from_spm(beta_loc);
labelname1 = ‘ALLO_sessbin 1’
labelname2 = 'CTR_sess
bin 1’

cfg = decoding_describe_data(cfg,{labelname1 labelname2},[1 -1], regressor_names, beta_loc); = make_design_cv(cfg);
results = decoding(cfg);

Hi Isabelle,

Thanks for noticing this error. It looks like we set up the function only for libsvm without telling anyone (tss tss). The main part is checking that the label identity doesn’t change between iterations, otherwise things get messed up easily. But that should be fine in most cases. For that reason, please send me an email, I’ll provide you with an updated version that should do what you want, and if you give me the ok, I’d upload it for everyone to use.