Hi Alfonso,

I’m working with resting-state fMRI data and analyzing functional connectivity differences using CONN. I created a psychometric scale for the patient group based on pre-scan clinical interviews. Unfortunately, we don’t have equivalent data for the control group, so for the second-level covariate I included the actual scores for the patient group and entered zeros for the controls.

In the second-level analysis, I included “patient,” “control,” and “patient*PsychometricScale” as covariates and used a between-subjects contrast of [0 0 1] to evaluate the association between this psychometric scale and functional connectivity within the patient group. I then ran a seed-to-voxel analysis and found some significant results.

My question is mostly about how to properly interpret these findings. Specifically, am I right in thinking that this contrast is effectively asking: “What is the effect of the psychometric scale within the patient group?” And more specifically, is it valid to interpret the results as: “Within the patient group, as scores on the psychometric scale increase, we observe positive (or negative) correlations between the seed region and these brain areas.”?

Also, while I understand that this model focuses on within-group effects in the patient sample, I was wondering whether the results might still offer some indirect insight into group differences—particularly in how this psychometric dimension might contribute to the altered connectivity patterns seen in patients compared to controls. I’m not sure if that kind of interpretation would be pushing beyond the scope of this model, so any guidance on that would also be really appreciated.

Thanks a lot for any thoughts!

Best,

Nico