Regress out covariates

Submitted by nabila11 on

Hi experts, 

I'm quite new to REST for the  analysis of  my resting state data. I want to do functional connectivity analysis but i have problems when i want to regress out motion parameters and ROI signales from my data. 

I don't understand the way to arrange my covariates  (rp files and RoiSignals) in a specific folder ; so i can use them in the Regress Out Covariates step. Can you help me with that please ? 

Thank you, 

Best regards, 

Nabila 

YAN Chao-Gan

Thu, 01/07/2016 - 02:08

Permalink

Hi Nabila,

You can use the pipeline DPARSF (http://rfmri.org/dparsf) to extract and regress out the covariate automatically.

Best,

Chao-Gan

Hi,

Thank your for your fast reply.

i have already extract them, i have my rp files (resulting from spm analysis) and i have extract the ROI signales from brain masks, CSF and WM for each of my subject. But what i'm asking (and i'm sorry if i misspoke), is how do i have to arrange these files (2 per subject, 15 subjects) for using them in the covariables section of  the functional Connectivity GUI ? Because when i open it, it allow me to select only one file.

I don't know if it is possible or i have to do it subject per subject ?

Thank you,

Nabila

YAN Chao-Gan

Fri, 01/08/2016 - 01:10

Permalink

Hi,

You have to merge them into one text file for each subject.

And then setup as follows:

Process multiple subjects with different covaribles in batch mode. Give the Covariables file as: 
Covariables_List:
X:\Structural_Functional\Reprocess\Sub3Cov.txt
X:\Structural_Functional\Reprocess\Sub2Cov.txt
X:\Structural_Functional\Reprocess\Sub1Cov.txt
 
Best,
Chao-Gan

nabila11

Mon, 01/11/2016 - 16:38

Permalink

Hi,

So, i merged my covariates (rp and ROISignals) into one text file for each subject.

I have now all my 16 Cov.txt for each subject. But if i understood well, for obtaining the Covariables file (Covariables_List), i have to write a text file where i enter the paths of the Cov.txt or i have just to enter the directory where there are all the 16 texte files ?

Thank you very much for your help,

Best,

Nabila

Hi,

Yes thank you Yan. But which one of the 2 possibilities do i have to choose ? to put all the Cov files in one directory like this :

   

and then enter the directory here, like in the 2nd picture ?

 

An other question is with functional connectivity analysis, when i tried to run an analysis, i have this error displayed (below), is this already happened to you ? do you know what's the problem ?

Exception occured.    (MATLAB:textread:FileNotFound)
    Error using textread (line 166)
File not found.
    166#line,        textread,    in "C:\Program Files\MATLAB\R2013a\toolbox\matlab\iofun\textread.m"
    1393#line,        (parfor body),    in "C:\Users\E308-8\Documents\programs\REST_V1.8_130615\fc_gui.m"
    1370#line,        btnComputeFC_Callback,    in "C:\Users\E308-8\Documents\programs\REST_V1.8_130615\fc_gui.m"
    96#line,        gui_mainfcn,    in "C:\Program Files\MATLAB\R2013a\toolbox\matlab\guide\gui_mainfcn.m"
    39#line,        fc_gui,    in "C:\Users\E308-8\Documents\programs\REST_V1.8_130615\fc_gui.m"

Thank's a lot again,

Nabila

YAN Chao-Gan

Wed, 01/13/2016 - 09:14

Permalink

You have to input the configuration text file (the text file has a list of Covs, starting with the line "Covariables_List:") instead the directory in the second window.

Best,

Chao-Gan

Thank you Chao-Gan,

I finally managed to do it. So i done my functional connectivity analysis (voxel-wise and ROI-wise), but i have to check with you some points (and I'm sorry if these questions may seem a bit ridiculous):
- the iput files for these analyzes are  the Functional Images Normalized Smoothed Detrended Filtred files and not those which are "Covremoved", right ?
-So I got my zFCMap for my two analyzes. For the ROI-wise ones, when i have to extract ROI signals after (as you show us in your ppt), should i  to chose in this step these zFCMap that i obtained with my ROI-wise analysis or should i chose the Covremoved files ?
I don't understand the exact utility of this analysis step ? I mean why i have to extract my ROI signals after my ROI-wise analysis are already done ?

Best,

Nabila

 

 

1. Use the covremoved one, as you want the nuisance signals have been regressed out.

2. "I mean why i have to extract my ROI signals after my ROI-wise analysis are already done". Why do you want to do ROI-wise analysis after voxe-wise analysis? It depends on your hypothesis.

Best

Chao-Gan

Forums
Discuss (controversial) issues