Generally speaking I don't think it is a good idea to do a subtraction of t-maps. That isn't a very statistically driven approach. You cannot say definitively A > B via the substation of t-values, it becomes a very qualitative description of 'it had a different t-value which was larger) but you cannot be sure that difference is above chance level.
Depending on what you are doing it is very possible to do direct statistical comparisons of brain-behaviour relationships. For example if you have run a regression, you can directly contrast those regression effects (basically a 1 -1 contrast for the regression of behavioral measure A and B).
Alternately you can move to another approach like PLS which does a very good job at brain-behavior relationships and can identify patterns of voxels which show different sets of relationships between your behavioral scores (e.g. same relationships, opposite, A only, B only) in a model-free approach. Randy McIntosh's PLS package works quite nicely (https://www.rotman-baycrest.on.ca/index.php?section=84) and is not too hard to use.
Colin Hawco, PhD
Neuranalysis Consulting
Neuroimaging analysis and consultation
www.neuranalysis.com
[log in to unmask]
-----Original Message-----
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On Behalf Of Tamara Sussman
Sent: December-08-15 10:13 AM
To: [log in to unmask]
Subject: [SPM] comparing correlations
Good morning, SPM experts!
I am trying to find the best way in SPM to compare two brain-behavior correlations.
Is it possible and acceptable to subtract one z-map from another (or one t-map from another), and then look at the p values for the clusters in that map?
Thanks so much!
Tamara Sussman
|