Re: [Freesurfer] Contrast for continuous-by-continuous interactions

Hi Avi, if you want to compute an interaction between two continuous variables, you have to create a new regressor where you multiply them together. The contrast you have created would test whether the age slope is different than the IQ slope. This contrast does not make sense since the to variables have inconsistent scaling (so the slopes can't be compared).

On 7/30/2021 9:42 AM, Avi Gharehgazlou wrote:

External Email - Use Caution

Hi experts,

I am using the command line stream on Freesurfer for my analyses. Although I have successfully designed all my contrasts, I wanted confirmation from you regarding one contrast that I am unsure about. Will you please check if the below contrast appropriately investigates age-by-IQ interaction effects across my two groups (TD and NDDs)?

I am also copying here information regarding my regressors and the first few lines of my FSGD file for further detail.

*Contrast:* To examine age-by-IQ interaction effects across groups (TD and NDDs): agexIQinteraction_acrossgroups.mtx

0 0 0 0 0.5 0.5 0.5 0.5 0 0 0 0 -0.5 -0.5 -0.5 -0.5

My regressors:

·2 discrete factors: Diagnosis, sex

·4 groups/levels/classes: TD M, TD F, NDD M, NDD F

·3 continuous covariates: age, SA, IQ

**

·Number of classes: 4

·Number of variables: 3

**

·NregressorsDODS= Nclasses*(Nvariables+1)= 4*(3+1)= 16

·Regressor 1: ones for TDM subjects, 0 otherwise. Codes intercept for TDM

·Regressor 2: ones for TDF subjects, 0 otherwise. Codes intercept for TDF

·Regressor 3: ones for NDDM subjects, 0 otherwise. Codes intercept for NDDM

·Regressor 4: ones for NDDF subjects, 0 otherwise. Codes intercept for NDDF

·Regressor 5: age for TDM subjects, 0 otherwise. Codes for TDM

·Regressor 6: age for TDF subjects, 0 otherwise. Codes for TDF

·Regressor 7: age for NDDM subjects, 0 otherwise. Codes for NDDM

·Regressor 8: age for NDDF subjects, 0 otherwise. Codes for NDDF

·Regressor 9: SA for TDM subjects, 0 otherwise. Codes for TDM

·Regressor 10: SA for TDF subjects, 0 otherwise. Codes for TDF

·Regressor 11: SA for NDDM subjects, 0 otherwise. Codes for NDDM

·Regressor 12: SA for NDDF subjects, 0 otherwise. Codes for NDDF

·Regressor 13: IQ for TDM subjects, 0 otherwise. Codes for TDM

·Regressor 14: IQ for TDF subjects, 0 otherwise. Codes for TDF

·Regressor 15: IQ for NDDM subjects, 0 otherwise. Codes for NDDM

·Regressor 16: IQ for NDDF subjects, 0 otherwise. Codes for NDDF

My FSGD file:

GroupDescriptorFile 1

Title TD_vs_NDDs_sex_age_SA_IQ

Class TDM

Class TDF

Class NDDM

Class NDDF

Variables age SA IQ

Input 100092_01 TDM 0.060861784 1.747581397 0.386693428

…

I'd really appreciate your confirmation of my contrast,

Thanks so much,

Avi

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer