3dDeconvolve AM2: How to specify -gltsym doing contrasts for ABI-sensitive regressor#1 rather than the default regressor#0?

AFNI Message Board
1

Hello, I'm using -stim_times_AM2 in 3dDeconvolve for each subject and i got two set of coef and tstat for each condition: #0 coef and tstat is standard activation map, and #1 coef and tstat make activation map of voxels whose BOLD response changes with changes in ABI. I also used -gltsym to get contrast between two conditions, for example, -gltsym 'SYM: +Angry_Approach -Angry_Withdraw', but it only generated contrast for regressor #0 by default. What should i do if I'd like to make contrasts for regressor #1 (ABI-sensitive) for each subject? Thank you for reading and I'd appreciate it if anyone gives advice!**
--------------------------------------
I tried -gltsym 'SYM: +Angry_Approach[1] -Angry_Withdraw[1]' to get contrast for regressor #1. Is it right?**

2

I tried -gltsym 'SYM: +Angry_Approach[1] -Angry_Withdraw[1]' to get contrast for regressor #1. Is it right?

Yes, I think that's correct.

Gang Chen

1 Like
3

Thank you for helping! yesterday I tried a few times and found -gltsym 'SYM: +Angry_Approach[1] -Angry_Withdraw[1]' would make contrasts for regressor #1, while -gltsym 'SYM: +Angry_Approach[0] -Angry_Withdraw[0]' made contrasts for regressor #0. However if I don't specify which set of statistics, -gltsym 'SYM: +Angry_Approach -Angry_Withdraw' will generate contrast based on both regressor by default, as same as the results of -gltsym 'SYM: +Angry_Approachp[0..1] -Angry_Withdraw[0..1]'. That's what I found. Anyway, my needs have been solved, thank you for your help again!

4

Yes, not including the index results in using the sum of all such components. Note that 3dDeconvolve outputs the numerical row for the contrast, i.e. the list of weights across all regressors. Search for "GLT matrix from", e.g.

------------------------------------------------------------
GLT matrix from 'SYM: vis -aud':
  0  0  0  0  0  0  0  0  0  0  0  0  1 -1  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 
 
------------------------------------------------------------
GLT matrix from 'SYM: 0.5*vis +0.5*aud':
  0  0  0  0  0  0  0  0  0  0  0  0 0.5 0.5  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
  • rick
1 Like
5

Thanks rick, I searched the forums for "GLT matrix from" and added '> detail.txt 'to the end of my 3dDeconvolve script to see what my GLT matrix is. So actually the GLT matrix corresponds to all the regressors that I use, including drift effect regressors, all the conditions of interest, and 6 head motion regressors. When I use AM2 regression, the GLT matrix part of conditions that i'm interested in will become twice the size of the original, where each condition corresponds to two adjacent regressors, i.e. #0 and #1. Thus I can specify in -gltsym which regressor I want to contrast with for a specific condition. I think I understand this part better. Thanks again for your help!

6

That's right. The GLT applies to the full regression matrix. It can also be more than one row if you ask for it. Compare column indices against X.xmat.1D, even using a text editor to review the header of, or consider something like:

1d_tool.py -infile X.xmat.1D -show_group_labels
  • rick