Dear Freesurfer experts,

I tested associations between a continuous predictor and thickness and area in qdec (and command line mri_glmfit, which gives the same results) and found a cluster being negatively associated with thickness orbitofrontal, but nothing for area orbitofrontal. After exporting mean thickness and area values from aparc.stats into Stata, consistent with my expectation I found a significant negative correlation with my predictor for thickness. However, I also found a significant positive association between birth weight and area.

To check what might be wrong I tested correlations between mean medialorbitofrontal thickness from rh.aparc.thickness and thickness in qdec and did the same for area. As expected, large average thickness values were positively associated with thickness in medialoribotfrontal and adjacent areas (see attached qdec screenshot thickness.png), but there were no associations using area (see attached qdec screenshot area.png). My expectatioin was that there should be positive area associations similar to those for thickness.

Hence my question: Is this expectation correct? If yes, why this discrepancy between thickness and area?

Thank you.

Wolff