help > Interpretation of Z-score sign according to the behavioral variable
Showing 1-2 of 2 posts
Display:
Results per page:
Mar 31, 2020  01:03 PM | anzianom
Interpretation of Z-score sign according to the behavioral variable
Hello, 

I am performing analyses of lesion masks with a single behavioral variable where higher positive scores mean worse performance. I don't have any voxel reaching the significance threshold but I see that the results files and a check of the z maps of my lesion voxels provide zscores both positive and negative. 

Based on the software statistics: 
Is any adjustement of my behavioral variable needed to account for the behavioral meaning (for ex. changes of the sign if higer scores mean worse performance)?

Based on the exemple you give on the tutorial page of the site I can assume that the software performs for each voxel a t-test between the mean of the beahvioral scores in the lesioned voxel vs the same in the spared voxel (lesioned - spared) and that I have to interprete the resulting zscore based on this.
Is this correct? 

If it is correct, in my case lesions in voxels with positive zscore should worsen the performance and the opposite for negative zscores.
Is this kind of reasoning correct or does the software use different statistical logic? 

I hope my doubts were clearly explained.

Thank you in advance for your help 

Marco Anziano
Apr 6, 2020  09:04 AM | Isabelle Faillenot - hospital Saint-Etienne
RE: Interpretation of Z-score sign according to the behavioral variable
Hi,
I copy the answer of chris roden on the MriCroN forum. May be it helps.

"As regards to positive and negative correlations, I assume you are using NiiStat not NPM or Stephen Wilson's VLSM. With NiiStat, a positive statistic means that brighter images are correlated with higher behavioral scores and a negative correlation means that brighter images are associated with lower behavioral scores. The convention is to draw our lesions as bright regions (lesioned voxels = 1) on a dark background (unlesioned voxels = 0). The inference depends on your behavior: on many standardized tests like WAB-AQ or MoCO lower scores mean poorer performance. On the other hand, for reaction time tasks, higher response times means poorer performance."