Hi FSL experts,
I ran a higher level analysis in FEAT where I was interested in an interaction between a continuous variable and group. I got results where Group 1 slope > Group 2 slope in brain region A and Group 2 slope > Group 1 slope in brain region B.
My question arises because when I plot these interactions (by extracting cope values from the significant brain regions (A and B respectively)) against my continuous variable - I find that the slope for the first finding (Group 1 slope > Group 2 slope for brain area A) is more pronounced for Group 2. However, it is a negative correlation (whereas Group 1 is a positive correlation). My question is: is the reason why I am getting the result as Group 1 slope > Group 2 slope for brain region A in the feat report because Group 1 slope is positive and Group 2 slope is negative?
I guess ultimately I want to know if the absolute values of these slopes are taken or if we have a scenario as that described above where Group 1 will be greater than Group 2 because it is a positive compared to a negative slope.
Thanks,
Danielle
|