Hi Aviva,
> Hi everyone,
> I want to see the linear relationship between a variable and
> activation at a voxel using single regression.
> 1. Is there any scenario under which in a single regression analysis
> the program should not allow me to even enter a 1 or -1 t contrast?
> Only lets me enter F contrasts now. (says my t contrast is invalid
> when I try to enter 1 or -1)
>
I don't know .. it should not be ??
> 2. Does the F contrast (1 -1) give me the result I want—that is, the
> regression line for my variable and activation?
>
Yes a F contrast will gives you the result you want but not 1 -1 .. if
you type that you set the constant regressor (the intercept if you
prefer) to -1 .. For a t test you type 1 (positive correlation) or -1
(negative correlation) thus for a F test (two-sided) just type 1 (or -1
it is the same) and the map will show voxels with either positive or
negative correlation
> 3. Can I infer hypometabolism if the line has negative slope?
>
Well, you can infer that when your variable increases the metabolism
decreases
Best
Cyril
|