Your design is not necessarily wrong, but your reasoning is not correct.
Since these are categorical variables, there are actually two ways that
you can set up the model: interaction and non-interaction. The
interaction model allows you to test the interaction between your
variables (eg, is the difference between A and B affected by race? or
does E2 slope differ between A and B?). If there is such a significant
effect (regardless of whether you are interested in it or not), you'd
need to use the interaction model which would have 6 classes (3 soft
factor times 2 race factors). You would then have 2*6 regressors for 12
total. If you wanted to test the effect of E2, you would then create a
contrast with 6 zeros followed by 6 ones (or 6 0.166666). Your design
(ie, coding each group as a 1 in a given column) is non-interaction
model. If that works for you, then it is fine. Sometimes people will run
the interaction model, test for effects, if there are no effects, then
use the non-interaction model. Typically, this works out well because it
is so hard to get effects in neuroimaging:). Your contrast is right for
the non-interaction model.
On 6/14/2025 11:08 PM, Muskan Khetan wrote:
External Email - Use Caution
Hi,
I have a question related to setting up the FSGD file for my analysis.
Specifically, I’m looking into the effects of hormone levels on the
brain in a single group (i.e., not comparing different classes). My
model includes both continuous and categorical covariates.
From what I’ve gathered, most examples on the FreeSurfer website focus
on group comparisons, which isn't the focus of my analysis. Therefore,
I’ve set up an FSGD file with only one class (see the attached
screenshot for reference). In this file:
*
Variables like |soft_A|, |soft_B|, |soft_C|, |child_raceMixed|,
and |child_raceWhite| are *categorical covariates*.
*
Since these are covariates and not grouping factors, I haven’t
created separate classes for them.
I believe this approach is appropriate, but I wanted to confirm
whether that understanding is correct.
Additionally, regarding the contrast file:
*
My contrast vector is |0 1 0 0 0 0 0 0 0 0 0|, where the number of
elements matches the number of variables in my model (including
the intercept).
*
As I’m specifically interested in the effect of |E2_mean|, I’ve
set its position to |1| and all others (including covariates and
intercept) to |0|.
Is this the correct way to set up the FSGD and contrast files for this
type of single-group analysis?
Thank you in advance for your guidance.
Best,
Muskan
_______________________________________________
Freesurfer mailing list [email protected]
To unsubscribe send an email [email protected]
https://mail.nmr.mgh.harvard.edu/mailman3/lists/[email protected]/
_______________________________________________
Freesurfer mailing list -- [email protected]
To unsubscribe send an email to [email protected]
https://mail.nmr.mgh.harvard.edu/mailman3/lists/[email protected]/
The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Mass General Brigham
Compliance HelpLine at https://www.massgeneralbrigham.org/complianceline
<https://www.massgeneralbrigham.org/complianceline> .
Please note that this e-mail is not secure (encrypted). If you do not wish to
continue communication over unencrypted e-mail, please notify the sender of
this message immediately. Continuing to send or respond to e-mail after
receiving this message means you understand and accept this risk and wish to
continue to communicate over unencrypted e-mail.
