Dear Nese
> We have a paired study design of the same subjects (scan 1-scan 2) and
> we would like to see if there is a correlation between initial tracer
> binding obtained in scan 1 and reduction (%reduction preferred over
> absolute reduction) in binding observed between scan 1 and scan 2.
>
> Is there a design to be used to investigate a positive or negative
> regional correlation between (Scan 1-Scan2) and Scan1 on a voxel
> by voxel basis ?.
This is difficult ,in the context of SPM, because you need to specify a
different set of explanatory variables for each voxel. I can only
suggest that you take global activity as an index of uptake in scan 1
and look for a difference x global interaction (modeling both the main
effects of global and scan (1 vs 2)). To do this in terms of % change
you could consider log transforming your data using imagecalc prior to
statistical analysis.
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|