some questions about VMHC

Submitted by chlai on Tue, 12/03/2013 - 10:35

Dear REST experts:

Recently I worked on the new function of VMHC on the REST software.

However, I had some questions which might need your kind help and explanations.

(1) Can the REST solve the issue of "micromovements." Power et al. (2012) have shown that movements from "frame" to "frame" (i.e., EPI image(n) to EPI image(n+1)) can lead to artifactual effects in resting state data. Did the authors take any measures to address this? Yan et al. (2013) in a recent paper recommend using the "Friston24" option in the covariation strategy as well as a group level covariation for mean framewise displacement. It is also unclear whether other sources of signal were removed (e.g., global signal, white matter signal, CSF signal).

(2) Are the VMHC results shown bilaterally independently??
The VMHC measure should be identical in those two regions, and indeed a unihemispheric mask should probably be used to constrain multiple comparisons. Is that possible? How to do it?

(3) It is irritating that the regions with significant differences are absolutely symmetric and have bilaterally identical volumes. Could there be an artifact behind this statistical finding?

Thank you for your kind help.

Best

Dr. Lai

Submitted by ZHANG_RESTadmin on Tue, 12/03/2013 - 14:38

 1. Since VMHC is another type of seed-based FC analysis, you can easily remove the effect caused by microhead movement by checking FD covariates in nuisance regressors in DPARSFA or Friston24 also.   When you do this, the nuisance correlation caused by head motion can be modeled out.
 
 2. To show result or to correct for multicomparisons, you will create unilateral hemisphere mask. To do this, simply load brain mask (this is bilateral one) into matlab by using rest_ReadNiftiImage, then set all voxels with x > 30 (only show left side result) to zero. After that, use rest_WriteNiftiImage to write this mask out into your disk.

 3. This is because your VMHC result is bilaterally identical or symmetric, nothing should be worried about. Just use your unilateral mask to see one-side result.

Submitted by chlai on Thu, 12/05/2013 - 15:03

Dear professor Zhang:

Thank you for your kind explanations and help.

It helps me a lot.

However, for the second point about the creation of unilateral brain mask by rest ReadNiftiImage and rest WriteNifti Image.

I tried to type in the above commands in the matlab command window. But it seems not work for any further steps. No GUI window appears. Can you be so kind to give me the instructions about how to perform these two commands? Can that be performed by using REST viewer? But it seems no this function in the REST viewer.

Look forward to receiving your kind reply.

Dr. Lai

Submitted by ZHANG_RESTadmin on Mon, 12/09/2013 - 17:46

 its not a GUI. you should create a new m file and write a script based on those functions i had mentioned before. 

 its like the script i wrote below but you need to revise it to fit your situation:

[data head] = rest_ReadNiftiImage('THE_FOLDER_OF_YOUR_REST_SOFTWARE\mask\BrainMask_05_61x73x61.img');
data(31:61,:,:) = 0;
head.fname = fullfile(pwd,'\','result.img');
rest_WriteNiftiImage(data, head);




Submitted by chlai on Wed, 12/11/2013 - 10:41

Dear professor Zhang:

I follow your suggestion to establish the script fit for my files.

However, I found several errors happened as the follows (I restart MATLAB and run "rest_Fix_Read_Write_Error" but still faced the same bugs):

??? Error using ==> rest_Fix_Read_Write_Error at 64
Meet error while fixing read write error. Please restart MATLAB, and run "rest_Fix_Read_Write_Error" before starting anything.

Error in ==> rest_WriteNiftiImage at 92
rest_Fix_Read_Write_Error; %YAN Chao-Gan, 100426.

Error in ==> test at 4
rest_WriteNiftiImage(data, head);

Can you be so kind to instruct how to debug this problem?

Thank you for your kind help.

Dr. Lai

Submitted by ZHANG_RESTadmin on Wed, 12/11/2013 - 15:52

The alternative way is only using SPM8 (make sure only spm8 path is in matlab and REST path were removed)

First, you need copy BrainMask_05_61x73x61.img/hdr files into your current folder.
Then, run this:

head = spm_vol('THE_CURRENT_FOLDER\BrainMask_05_61x73x61.img');
data = spm_read_vols(head);
data(31:61,:,:) = 0;
head.fname = fullfile(pwd,'\','result.img');
spm_write_vol(head, data;

I think the problem you met is you should update your spm8, and then setpath of REST (with only REST folder, not with subfolders),
then, setpath of SPM8 (both folder and subfolders), then try again.   If still not work, you may change to another PC with windows as operation system.
Or just change to another one's computer which can run REST, you will not meet the above read and write problem.