Fixelcfestats: one-sample ttest using random sign-flipping

Dear MRtrix3 experts,
I would like to use fixelcfestats to perform a one-sample ttest. I have searched for the related topics but still have two concerns:

  1. Creating the correct contrast matrix and design matrix file
  2. Using the right options
    I used single column of ones in the design matrix and single 1 in the first and -1 in the second row of contrast matrix for negative and positive effect. Fixelcfestats can recognize them and run without errors , but I do not know whether it is correct or not.
    For the options, I only used the option -errors ise, is the exchange_whole option also necessary ? In the thread , @rsmith referred to the -exchange_whole option when sign-flipping was used. I tried to read the Winkler et al.,2014 paper about the restricted exchangeability, but still do not know how to create the exchange_whole file for one-sample ttest. Could you give some suggestions ?

Thanks in advance !!

Best wishes,
Kitty

Hi Kitty,

Creating the correct contrast matrix and design matrix file

This can only be verified with adequate knowledge of the experimental paradigm and scientific hypothesis. The task here is essentially to re-express that content in numerical matrix form, which can’t possibly be achieved without accurate knowledge of such.

For the options, I only used the option -errors ise, is the exchange_whole option also necessary ?
… still do not know how to create the exchange_whole file for one-sample ttest.

These are two completely different aspects of the experimental model. I can’t explain how to create exchangeability blocks for a one-sample t-test because the two concepts are unrelated. And I can’t really hope to explain those concepts here with greater accuracy than that presented in the originating article. But given it sounds like you only have a single column in your design matrix, I don’t see how restricted exchangeability could have any applicability in your circumstance.

Cheers
Rob