I am new to RSA and trying to conduct this analyis using TDT.
Before running decoding_template_similarity, I think I have to create RDM or RSM.
To do that, should I use pattern_similarity.m?
If it’s correct, how can I create RDM(RSM)?
In my case, I’d like to compare behavioral RDM and neural RDM.
I think therefore I would assign the matrix of behavioral data and neural data(fMRI) in x or y.
I am sorry if this question is irrelevant but I do whatever it takes to achieve this analysis.
I would appreciate if you could give me an advice.
Since TDT was created in a way to automate all processing as much as possible, it provides you with two options:
Return RDMs (which you get with decoding_template_similarity)
or
Compute RDMs, run similarity analyses on them, and return results
(which you get with returning the decoding output rsa_beta and setting everything up according to make_design_rsa.
I coded the latter option about 6 years ago thinking it could serve as an all-encompassing version uniting all variants of RSA, pattern similarity analysis, and pattern component modeling. I noticed later that it just makes everything complicated since most people just want to correlate their RDMs.
Let me know if you want to run the latter of the two options, and I’ll see if and how we can get it to run more easily than what is the current default.
I am deeply sorry for my too late reply because I did not find the reply you have already given…, so sorry.
Thank you so much for your helpful answer but I’d like to make sure your comment again.
If my understanding is correct, I can create neural RDM by using decoding_templates_similarity or rsa_beta and make_desing_rsa?
Actually, I’ve just got the res_cfg.mat (the neural RDM?) by the former code (sorry, I haven’t chosen the latter codes).
But I’m still not so sure how I can compare this neural RDM and the behavioral RDM that I created as mat file.
I really appreciate if you could give me another comment.