Help with correlation with FBA metrics


#1

Dear MRTrixers,

I am trying to study the correlation between a clinical measurement and FBA metrics. I just wanted to make sure that my design and contrast matrices do what they’re supposed to do.

My design matrix looks like this:

Clinical measurement gender demeaned age
6.075 1 3.75
7.215 0 2.75
-0.875 1 10.75
0.83 0 9.75
3.43 0 -5.25
2.205 1 5.75
6.99 1 -5.25
-6.345 1 -8.25
2.79 0 -8.25
-6.855 0 5.75
-11.945 1 -2.25
-3.52 0 -9.25

and my contrast matrix looks like this:

1 0 0

I’m trying to study the correlation of the first column (the clinical measurement) with the FBA metrics (FD, FC and FDC), while correcting for gender and demeaned age. Am I doing this right??

Any input is appreciated. Thanks for the constant support!

Best,
Joe


#2

Hello again,

I’ve been trying to figure this out on my own, but I haven’t reached a solution. I noticed in the FBA documentation it says:

For correlation analysis, a column of 1’s will not be automatically included (as per FSL randomise).

Does this mean I have to add an extra column of 1’s to my design matrix? So an extra column of 1’s at the start and then a contrast matrix like this:

0 1 0 0

Is that what is meant by the comment in the documentation?

Any help is appreciated! Thanks.

Best,
Joe


#3

Hi Joe,
I had similar questions that I put to the MRTrix team in this thread:

Let me know if it helps otherwise I can go back over some of my design and contrast matrices and see if I cna be any more help.
Michael


#4

Hi Michael,

Thanks for the help. I did come across your post while searching the forum for a similar question to mine, but I didn’t find what I was looking for. However, upon further digging into the forum, I have come across a post with basically the same question as mine. But thank you anyway for offering your help, very kind of you :grinning:

Cheers,
Joe


#5

No worries Joe. I hadn’t seen that thread. You’re helping me now. Cheers!