They are not identical, for sure. In general, I discourage people from modeling categorical variables (eg, Group1/2 below) as continuous variables because you make assumptions that might not be reasonable (eg, females will have twice the thickness as males). I would use the first method.
On 5/2/2024 8:20 PM, Lydia Chung wrote:
External Email - Use Caution
Hi FS support team,
I am going to be using mri_glmfit to test 1) main effects (which regions show differences in cortical thickness between group 1 and group 2) and 2) interaction effects (does the relationship between IV and DV depend on moderator). Depending on the model, the IV is sometimes a binary categorical variable and sometimes a continuous variable. The moderator is always continuous. So, interaction effects are either continuous x continuous OR categorical x continuous.
One thing I'm trying to understand is whether the two examples below are basically two different approaches that answer the SAME question: Do people in Group 1 differ from Group 2 on cortical thickness? Version 1 is the one provided by FS and Version 2 is an analog of another setup I have been provided by colleagues. Are these models answering the same or different questions? The freesurfer link below also shows an example of how to do an interaction (Group x Age) using Version 1 setup; for Version 2 setup of an interaction, I know I would multiply the IV and Moderator before this step so that I would have an additional "interaction variable" column to add as one of the 'Variables' listed in the fsgd code. So, I'm also curious if the two different methods of testing an interaction (in addition to the first question about the main effect) will get you the identical answer OR if there is something conceptually different? Do the nuances of this setup have to do with the difference between using DODS or DOSS?
_ _ _Main effect of Group on cortical thickness Version 1 (copied from FS example. link here https://secure-web.cisco.com/1cgzre8FgLw0c17s9QNhmU_14vQ714XlLV20U4K14P96Afbs-6Heo84PLySiwVhDEZMhxP0C6FgIow7J7CguJV6osIK9fC2z1hN85kjx1hYSlGn-ewBZLibOXP73prq6qzehGuWIRDsDCEiRBlSxahnAXEQvrZFCHQKMOQR2X4kI3qYW5Lz5tLkGPv3ihsbW73z7AgR__zjM9Q0tck9Z_lN10U8oIYt6XTvyXshAZinK2vbV5cFWo8PxMMpXWi37uJxfQNvoOksCQwKocaxlN5ontLF542jDxjQ-MXsnM1VfM2onrGG6H9M-rle6KQe7JYj9siXKlYK3v1WBNbA874g/https%3A%2F%2Fsurfer.nmr.mgh.harvard.edu%2Ffswiki%2FFsgdf2G2V)_ GroupDescriptorFile 1 Title OSGM Class Group1 Class Group2 Variables Age Weight Input subject1 Group1 30 100 Input subject2 Group2 40 120
Contrasts: 1 -1 0 0 0 0 (to test main effect of group; this feels like an anova approach?)
_Main effect of Group on cortical thickness Version 2_ GroupDescriptorFile 1 Title OSGM Class Subjects Variables Group Age Weight Input subject1 Subjects 0.5 30 100 Input subject2 Subjects -.5 40 120
contrasts: 0 1 0 0 (to test main effect of group; this feels like a linear regression approach?)
Thank you in advance for your help!
Lydia
Lydia Wu-Chung, M.A. Doctoral candidate BMED Lab Department of Psychological Sciences Rice University 6500 Main St - MS201 Houston, TX 77030 Lab Phone: 713-348-8126 Email: lydia.wu@rice.edu mailto:lw50@rice.edu
_ _
_ _
_Confidentiality Text:_
The information contained in this e-mail message may be privileged, confidential, and/or protected from disclosure. This e-mail message may contain protected health information (PHI); dissemination of PHI should comply with applicable federal and state laws. If you are not the intended recipient, or an authorized representative of the intended recipient, any further review, disclosure, use, dissemination, distribution, or copying of this message or any attachment (or the information contained therein) is strictly prohibited. If you think that you have received this e-mail message in error, please notify the sender by return e-mail and delete all references to it and its contents from your systems.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer