Two-Sample t-test for data sets with different ranges

, ,
1

I have two data sets:

  1. First-level t-statistic scores for a stimulation protocol
  2. Intensity data for tracer dyes (from the Allen Brain Institute database)

I would like to see in how far the stimulus-evoked response differs from the projection atlas. I am thinking of using a two-sample unpaired t-statistic test ( https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FEAT/UserGuide#Unpaired_Two-Group_Difference_.28Two-Sample_Unpaired_T-Test.29 ), but the problem is that the t-statistics are much higher in magnitude (roughly -5 to +10) than the intensity scores (hard bounded between 0 and 1).

Would it make sense to just divide all maps by their standard deviation and then just use FLAME as linked above, or is perhaps another approach better suited for the task?