Hello TDT users,
I have some questions about TDT toolbox function!
In my project, it’s a block design,and I have a imbalance of data in each run. Such as in_group stimulus have 4 blocks,out_group stimulus have 4 blocks,shape stimulus have 8 blocks.
So, my questions are:
1.If I want to test whether participants can classify the in_group stimulus and shape stimulus, or out_group stimulus and shape stimulus. In this case,I used make a decoding design matrix to conduct leave-one-chunk-out cross-validation.To avoid imbalance data ,I set cfg.design.unbalanced_data = ‘ok’ in my script, and the output I set the ‘AUC_minus_chance’.I don’t know whether this process can correctly handle my problem of data imbalance.
2.And I think make_design_boot_cv is more suitable for my data, so I tried a few things. I set n_boot=10;balance_test=1; used the default design; scale.method = ‘min0max1’;output='accuracy_minus_chance’and ‘AUC_minus_chance’.I also don’t know Whether I am dealing with this correctly.
Any feedback would be great, Thanks!