Hi,
I've looked through the archive, but couldn't quite find an answer to this:
If you have 2 groups in TBSS and are using some behavioral measure to
correlate with FA, so that your EVs are
group1 group2 behgrp1 behgrp2
then how can you interpret the contrasts 0 0 1 -1 and 0 0 -1 1?
Specifically, you can reach significance in the first case if the
correlations are both positive but greater for group 1, if the correlation
for group 1 is positive and group 2 negative, or if both slopes are negative
for each group but steeper for group 2. Clearly these scenarii say different
things about your data, so how to tell them apart?
The most obvious might be to look at 0 0 1 0, 0 0 -1 0, 0 0 0 1, and 0 0 0
-1 to evaluate whether each correlation within each group is positive or
negative, but I feel that's subject to loopholes if some of these are
sub-threshold...
Another way might be to create a mask and then run fslmeants --showall --m
mask to extract the individual FA values and then plot them against the
behavioral data.
I feel like I'm missing something though... it should be easier, right?
Thanks,
Cherif
|