You can't test the correlation between the drug effect and plasma concentrations. You only have 1 value for the drug effect. I think what you are actually asking, are the two groups different, after accounting for plasma concentration. This is an ANCOVA and I've described different options below.
Using a single covariate means that you think the covariate has the same relationship in both groups.
Using two covariates means that you think the covariate has a different relationship in each group.
Now, when the covariate is 0 in one group, the covariate would be all 0s. Thus, it doesn't matter if you use one or two covariates. You can either let SPM form the interaction or choose not to. The interaction would model it as two columns.
placebo>drug: 1 -1 0 0 (assuming 2 covariates; this takes into account plasma concentrations by decreasing the error term)
placebo_conc>drug concentration: 0 0 1 -1 (this will be the same as negative concentration contrast since beta for placebo concentration is 0).
Hope this helps.
Best Regards, Donald McLaren
=================
D.G. McLaren, Ph.D.
Postdoctoral Research Fellow, GRECC, Bedford VA
Research Fellow, Department of Neurology, Massachusetts General Hospital and
Harvard Medical School
Office: (773) 406-2464
=====================
This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED
HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is
intended only for the use of the individual or entity named above. If the
reader of the e-mail is not the intended recipient or the employee or agent
responsible for delivering it to the intended recipient, you are hereby
notified that you are in possession of confidential and privileged
information. Any unauthorized use, disclosure, copying or the taking of any
action in reliance on the contents of this information is strictly
prohibited and may be unlawful. If you have received this e-mail
unintentionally, please immediately notify the sender via telephone at (773)
406-2464 or email.
On Fri, Nov 18, 2011 at 6:55 AM, Yoan Mihov
<[log in to unmask]> wrote:
Dear SPM community,
In a design with two independent samples (placebo vs. drug) and variable plasma drug concentrations within the second group, due to interindividual differences in drug metabolism, I'd like to run the following tests on the BOLD signal:
[placebo < drug] x plasma concentration, i.e., one-tailed test for a positive correlation between the drug-placebo contrast and plasma drug concentration
[placebo < weighted_drug], i.e., compare both groups while "correcting/accounting" for the variability of the plasma drug concentration (ANCOVA)
I'd be thankful for any concrete suggestions on the following points, and truly indebted for any reference to articles/manuals/tutorials on this specific issue.
1. Model specification (two-sample t-test):
The plasma drug concentration within the placebo group is, obviously, 0
for all individuals: should I use only one covariate vector, or two?
Should I define an Interaction?
2. SPM contrast manager
How should I define the above-mentioned contrasts?
Many thanks in advance,
Yoan
--
Yoan Mihov, Dipl.-Psych.
NEMO Research Group
Department of Psychiatry
University of Bonn
Sigmund-Freud-Str. 25
53105 Bonn
Germany
Tel.: +49 (0)228 287 11165
web: http://web.me.com/renehurlemann/Website/Team.html
