Hi Shruti,

On 2 April 2016 at 18:23, Shruti Narasimham <[log in to unmask]> wrote:
Hello,

I have 2 rs-fMRI data from 2 healthy controls and 2 patients. I would like to use a GLM to model the effect of realignment parameters and WM,CSF signals.

Effect on the time-series, or effect between groups? For the former, it may not make much sense, as the motion parameters are derived from the imaging data itself. And for sure, the WM signal is highly correlated with the WM, and same goes for CSF. For the latter (between subjects), with only 4 participants, it won't be possible to draw any conclusion.

In any case, please see more below:
 
I input EV1 as my "healthy controls" and EV2 as "patients". 
  1. How and where do I use my realignment paramters (6) and WM,CSF signals (2) for each subject in this GLM model? 
These go in the 1st level, as additional EVs. For the realignment parameters, there is an option to add them automatically. For the WM/CSF, you need to add the respective EVs.
 
  1. How do I obtain the residual time series of this GLM model (raw data minus the 8 nusiance variables).
After running the 1st level with the above EVs, you'd take the file with the residuals (I think it's called res4d.nii.gz).

All the best,

Anderson

 
Sorry if these questions are basic, but I am very new to FSL and therefore need some help and advice. 


Thank you.

Regards,
Shruti

--
Shruti Narasimham
PhD Candidate
Trinity College Dublin
Ireland