Proc.py gltsym question

AFNI Message Board
1

Hi AFNI users,

I have a question regarding proc.py gltsym. Below is the part of the proc.py script I am using which I intended to subtract brain activation during NF from F (F-NF).

-regress_stim_labels F NF \
-regress_basis 'BLOCK(4,1)' \
-regress_censor_motion 0.3 \
-regress_opts_3dD \
-gltsym 'SYM: +F -NF' \
-glt_label 1 F_vs_NF \

If I’d like to do the subtraction on the other way around which is subtracting brain activation from F from NF (NF-F), can I simply change the gltsym into ‘SYM +NF -F’ and glt_label into NF_vsF?

Thanks for reading this question,
Jun

2

Yes. To be accurate,

-gltsym ‘SYM: +NF -F’ \
-glt_label 1 NF_vs_F \

The label can be almost anything that’s useful to help you remember. It could be ‘NF-F’ or ‘NFminusF’, if you like one of those more.

The SYMbolic expression needs to be carefully formatted. In particular, you need the space between the ‘+NF’ and the ‘-F’.

3

Okay, I’ll carefully format SYMboic expression. Thank you so much for your help! Appreciate it!

4

Keep in mind that ‘SYM: NF -F’ will not give you any
more information. It is just the negative of the ‘F -NF’
contrast (betas and t-stats are negated).

  • rick
5

Hi Rick,

Thank you so much for your comment. Could you please let me know how should I edit the script in order to get NF -F?

Best,
Jun

6

Hi Jun,

If you have already run the full process, then
it is enough to use the existing results. Since
‘NF -F’ is the negative as ‘F -NF’, anywhere that
F_vs_NF is positive means F is greater, while the
negatives are where NF is greater. The NF_vs_F
contrast will just reverse that.

But if you wnat to run it, it is okay to put both
GLTs in the command (the second one gets label 2):

   -regress_stim_labels F NF   \
   -regress_basis 'BLOCK(4,1)' \
   -regress_censor_motion 0.3  \
   -regress_opts_3dD           \
      -gltsym 'SYM: +F -NF'    \
      -glt_label 1 F_vs_NF     \
      -gltsym 'SYM: NF -F'     \
      -glt_label 2 NF_vs_F     \
  • rick
7

Hi Rick,

Thanks again for the comment. Can I have one more question? If I’d like to see the activation of only F, not subtraction, is it correct to run the script that looks like this?

-regress_stim_labels F \
-regress_basis ‘BLOCK(4,1)’ \
-regress_censor_motion 0.3 \
-regress_opts_3dD \
-gltsym ‘SYM: F’ \
-glt_label 1 F \

And, if I want to see the activation of NF, should I run the script below separately?

-regress_stim_labels NF \
-regress_basis ‘BLOCK(4,1)’ \
-regress_censor_motion 0.3 \
-regress_opts_3dD \
-gltsym ‘SYM: NF’ \
-glt_label 1 NF \

Best,
Jun

8

The single beta results are output by default,
there is no need to specify them.

  • rick
9

Thanks a lot for the quick response. Could you please check if I have checked correctly? If I am going to look at activation for a single stim file, I don’t need to specify the below in the proc py script?

-regress_stim_labels NF \
-regress_basis ‘BLOCK(4,1)’ \
-regress_censor_motion 0.3 \
-regress_opts_3dD \
-gltsym ‘SYM: NF’ \
-glt_label 1 NF \

Best,
Jun