AFNI version info (afni -ver): AFNI_25.0.04 'Severus Alexander'
Good afternoon,
I am writing a gen_group_command.py for a 3dMEMA and looking at various options from the afni 3dMEMA page. Can anyone point me to additional reading related to why/how to choose some of these options? Specifically, I'm wondering whether I should use -model_outliers and how to decide whether to set my -covariates_model center and slope to "same" or "different."
Could you share more details about your group-level analysis? Specifically, how many individuals are in each group, and what quantitative variables are included?
We are looking at two groups: deaf vs hearing and we have one behavioral score that we'd like to include in our analysis.
In one analysis, we want to be able to look at the interaction, i.e., if our score relates to activation more for one group than the other, in another analysis, we would ignore group and simply see if any area correlates with the behavioral score.
As of now, we have about 10 participants per group.
Maybe -model_outliers is not warranted with such a small group?
In one analysis, we want to be able to look at the interaction, i.e., if our score relates to activation more for one group than the other.
If this interaction effect is the main focus of your analysis, you should allow different slopes for the two groups. In this case, centering wouldn’t be necessary.
In another analysis, we would ignore group and simply see if any area correlates with the behavioral score.
Since the primary focus here is the overall association, centering wouldn’t matter in this case either.
Maybe -model_outliers is not warranted with such a small group?
Given the small sample size, it may actually be beneficial to include the -model_outliers option to account for potential outliers.
Thank you, Gang. I also want to add 3dClustSim to my 3dMEMA results, and am having a hard time figuring out how to do that. I've read this Taylor et al. (2018) page, and my undertanding was -3dClustSim needs to be run separately, after running 3dMEMA. However, I was also told I should be able to add it to my 3dMEMA gen_group_command.py script.
Can you please clarify how/when to add 3dClustSim when using 3dMEMA? Below is my current script. I've tried placing -3dClustSim in several locations (under -options, right after 3dMEMA), and I continue to get an error about difficulty parsing arguments.
To perform multiple testing adjustment with 3dClustSim, you need to write a separate script. See its usage guide for details. Additional discussions on result reporting can be found in this blog post and this video recording.
Indeed, Gang got here first, but I was going to mention that the -options .. option will add options for the given -command .. program, and 3dMEMA does not have any Clustsim-related option directly.
--pt
The
National Institute of Mental Health (NIMH) is part of the National Institutes of
Health (NIH), a component of the U.S. Department of Health and Human
Services.