Hi everyone,
My question has a few parts to it:
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? (this is
what is happening to me now; it only lets me enter an F contrast of 1 -1
or -1 1)
2. I am using the plot tool to look at a single regression of over 40
PETs from many subjects (one scan/subject) to assess the relationship of
activation at a certain voxel with a continuous variable, using the
aforementioned F contrast (1 -1 or -1 1--I get the same line either way).
a. Can I infer from seeing a high r value (~0.7) with an inverse
correlation relationship (r val is neg) that there indeed is more
hypometabolism in the area with increasing value of my continuous variable?
b. Even though I get the same line equation and r value with the
different F contrasts (1 -1 and -1 1) I get slightly different areas of
activation. It is not clear to me why exactly this happens.
c. Final question, what exactly (in the plot tool) is the real difference
between "predicted" and "adjusted"? What is the line being "adjusted" for
exactly?
Thank you so much for all your help!
Best Regards,
Aviva Olsavsky
|