TDT: is it possible to do leave-one-subject-out cross validation?

1

hi ~
As a new user of the decoding toolbox (TDT), I notice that this toolbox is quite easy to use when i need to do a leave-one-run-out cross validation .
but what if i need to do a leave-one-subject-out cross validation ?

thanks a lot ~

2

Hi,

check out the templates folder, there is a template for between-subject analyses (decoding_template_between_subject.m).

Best,
Martin

3

Thank u ~~~

But if I have 5 subjects now, and each time, I use 4 subjects to train, 1 subject to test. (with only two conditions in each run)
How should I set the cfg.files.chunk and cfg.files.label?

Hoping for your reply~